The Mexican Immigration Debate

Transcription

1

2 Michael B. Katz, Mark J. Stern, and Jamie J. Fader The Mexican Immigration Debate The View from History This article uses census microdata to address key issues in the Mexican immigration debate. First, we find striking parallels in the experiences of older and newer immigrant groups with substantial progress among second- and subsequent-generation immigrants from southern and eastern Europe and Mexican Americans. Second, we contradict a view of immigrant history that contends that early-twentieth-century immigrants from southern and eastern Europe found well-paying jobs in manufacturing that facilitated their ascent into the middle class. Both first and second generations remained predominantly working class until after World War II. Third, the erosion of the institutions that advanced earlier immigrant generations is harming the prospects of Mexican Americans. Fourth, the mobility experience of earlier immigrants and of Mexicans and Mexican Americans differed by gender, with a gender gap opening among Mexican Americans as women pioneered the path to white-collar and professional work. Fifth, public-sector and publicly funded employment has proved crucial to upward mobility, especially among women. The reliance on public employment, as contrasted to entrepreneurship, has been one factor setting the Mexican and African American experience apart from the economic history of most southern and eastern European groups as well as from the experiences of some other immigrant groups today. More immigrants to the United States now arrive from Mexico than from anywhere else. As half of the recent immigrants, Mexicans compose 28.8 percent of the foreign-born population (Bean and Stevens 2003). Unlike any other pair of nations, Mexico and the United States share a 2,000-mile border that separates countries whose living standards vary enormously. Mexico Social Science History 31:2 (Summer 2007) DOI / by Social Science History Association

3 158 Social Science History has become the United States second-largest trading partner (Massey et al. 2002; Zúñiga and Hernández-León 2005: xii). Proximity, transnational labor markets, and inequality keep the level of immigration high and hard to control and guarantee that it will be contentious. A growing scholarly literature investigates the impact of Mexican immigrants on the United States and the economic prospects of Mexican immigrants and their children. Lurking behind these concerns is an important historical comparison. Will Mexicans retrace the historical route that led immigrants from southern and eastern Europe who arrived in the early twentieth century to economic parity and success, or will they follow a different path that leaves them economically marginal, imperfectly assimilated, and heavily dependent on public aid? Are they, as some commentators would argue, of lower quality than earlier immigrants (Borjas 1985)? Will the Mexican second generation become white Americans or join the underclass in the nation s central cities (Waldinger and Bozorgmehr 1996; Smith and Edmonston 1997; Borjas 1999; Espenshade and Huber 1999; Rosenfeld and Tienda 1999; Rumbaut 1999; Jencks 2001a, 2001b; Lim 2001; Portes and Rumbaut 2001; Rumbaut and Portes 2001; Waldinger 2001; Waldinger and Lee 2001; Farley and Alba 2002; Perlmann 2005)? This article addresses some of the key issues in the Mexican immigration debate by comparing the economic assimilation of two cohorts of first- and second-generation immigrants with that of first- and subsequent-generation Mexicans. Like today s Mexican immigrants, displaced by the spread of capitalist agriculture and industrial development, southern and eastern Europeans came in huge numbers to the United States as target earners. They remitted massive amounts of money to family members left behind and often returned themselves. Two of these immigrant groups Italians and Poles show suggestive parallels to Mexicans today (Perlmann 2005: 2). Italian and Polish Immigrants: Paths to Economic Incorporation The analysis of Italians and Poles is based on the first-generation immigrants born in , who as young adults entered the United States at the peak of immigration, and the second generation born in , roughly the American-born children of the immigrant cohort. These early-twentiethcentury immigrants, who differed in cultural origins from the native-white

4 The Mexican Immigration Debate 159 American population and arrived from countries that had less developed educational systems than the United States, compose a more useful standard against which to measure the success of late-twentieth-century Mexican immigrants than do Europeans who arrived during immigration s midtwentieth-century nadir. The twentieth-century occupational history of ethnic cohorts points to the continued anchor of the first-generation Poles and Italians in the working class and the advantages and mobility of the second generation. First, consider the division of men into three broad categories: white-collar, bluecollar, and other (mainly agriculture and service work). Polish and Italian immigrants remained primarily blue-collar (Hutchinson 1956). Despite some movement over time into white-collar work, their mobility remained modest. In 1950, 70 percent of Polish and 63 percent of Italian immigrants remained in blue-collar work. Even in 1970 only 35 percent of second-generation Poles and 30 percent of second-generation Italians held white-collar jobs. Still, this was roughly double the fraction of white-collar workers in their parents generation. Second-generation men not only moved up the occupational ladder but also enjoyed greater work stability. As of 1970 a large proportion of immigrant men worked less than 48 weeks a year. Second-generation men, by contrast, more often worked the whole year. Among Poles, for example, 58 percent of first-generation 40- to 49-year-old men worked 48 weeks in the year prior to the census, compared to 77 percent of the second generation. Women followed roughly parallel paths. Between 1910 and 1950 white-collar work increased modestly among first-generation Italian and Polish women. In the second generation, however, white-collar work became common among women from each ethnic group about two-fifths of Poles and Italians. In the working class, immigrant men improved their situations. The declining fraction of Polish and Italian men employed as ordinary laborers offers the best evidence of this limited mobility. When they moved into white-collar work, it was often as proprietors of small businesses. Few were professional, clerical, or sales workers. Second-generation Polish and Italian men also improved their situations in the working class, escaping jobs as laborers more often than their fathers but, like their fathers, frequently ending up as proprietors. Was proprietorship the route to prosperity for these immigrant groups and their children, among whom so many worked for themselves, as some scholars have argued (see Portes and Zhou 1999)? The answer is equivocal.

5 160 Social Science History Earnings of the self-employed varied more than earnings in the private or public sectors. After all, the self-employed included professionals, proprietors of relatively large businesses, petty entrepreneurs, and pushcart vendors. Among both first- and second-generation men and women, the median earnings of the self-employed did not rank highest. Instead, public-sector employees earned the most. Indeed, the economic benefits of public-sector employment stand out as an important and largely neglected theme in the history of these groups. First-generation Poles and Italians developed distinctive industrial profiles. A large fraction of Poles worked in heavy manufacturing. Italians more often were found in mining, construction, transport, and personal services. First-generation Polish and Italian women, like men from the same backgrounds, remained primarily in the manual working class, concentrated in manufacturing and, early in the century, service (see table 1). The ability of many immigrant women to leave household service signaled modest mobility comparable to the exit of first-generation men from work as ordinary laborers. By the second generation, however, a substantial minority had entered clerical work or found jobs in retail trade, although the most common job among second-generation Polish women remained operatives. Retail trade or clerical work, however, increased from almost nothing among the immigrant generation to around one-quarter of the second generation.1 Thus, through their work experience, the occupational structure of Polish and Italian women began to bifurcate: a substantial fraction of the second generation moved into white-collar work, although as a whole they retained their anchor in the working class. Even more than occupation, earnings reveal the economic advantage of the second over the first generation. In the second generation, earnings for each group grew remarkably. Two principal sets of statistics are useful for examining earnings. One is the median earnings; the other is the distribution of earnings into quintiles. Unfortunately, because the census started to report earnings only in 1940, it is not possible to track earnings growth in the first generation. The statistics here are for male earnings only. Unless otherwise noted, earnings are given in 1990 dollars.2 The distribution of ethnic groups by earnings percentiles highlights even more vividly the advantage of the second generation (see table 2). Among the second generation, the share of Italians and Poles in the lowest quintile dropped dramatically, while the share in the top quintile jumped. Secondgeneration ethnics earned higher incomes from the start of their working

6 The Mexican Immigration Debate 161 Table 1 Occupational distribution, by ethnicity and cohort, for women over the age of 18 in the labor force Polish women First generation born Private household service Laborers Operatives Clerical Retail Second generation born Private household service Laborers Operatives Clerical Retail Italian women First generation born Private household service Laborers Operatives Clerical Retail Second generation born Private household service Laborers Operatives Clerical Retail Source: Ruggles et al lives, and their earnings increased as they aged, with the largest increment registered among men in their thirties and forties that is, given their birth between 1905 and 1914 in the immediate postwar years.3 Early in the twentieth century, the links between educational attainment and earnings remained far looser than they became later. Thus, in the first generation, earnings differences did not reflect educational attainment. For the second generation, however, earnings did increase with length of schooling, and higher education began to pay off more consistently. Few

7 162 Social Science History Table 2 Distribution of total personal income, by cohort and ethnicity, for men (percentage in income percentiles) Under First generation born Age Poles Italians Second generation born Age Poles Italians Age Poles Italians Source: Ruggles et al first-generation ethnics had a high school or postsecondary education, and a college education was rare for anyone in the first generation. (It was still relatively rare for native-born whites as well.) When immigrants did attend college, it was more often men than women who did so. College education, however, roughly doubled between generations. More than years of schooling, literacy highlights the differences in educational attainment separating early-twentieth-century immigrants from native whites. Immigrants arrived from countries that had less developed public education systems than the United States, where the common school system had produced nearly universal literacy among native whites. In Italy and Poland, educational underdevelopment left huge numbers illiterate. Among first-generation immigrants born between 1875 and 1884, for example, 36 percent of Italian men and 29 percent of Polish men could not read and write. Except among native whites, illiteracy was much higher for women: 43 percent of Italian and 40 percent of Polish female immigrants could not read or write. Thus immigrants arrived with an educational disadvantage not accurately reflected in the number of years of schooling. The ability to speak English a skill examined intensively by current-day social scientists concerned with immigrant adaptation also helped determine immigrant assimilation and earning capacity. Immigrants capacity to speak English increased notably with time in the United States. Between 1910 and 1920 the proportion of the first generation reporting English-speaking

8 The Mexican Immigration Debate 163 ability to the census taker more than doubled from 40 percent to 96 percent of Poles and from 44 percent to 90 percent of Italians.4 There is no way to know what level of proficiency these numbers signify. Nonetheless, they do reflect an enormous rise in the ability of immigrants to at least get along in English. Among the second generation, the ability to speak English, as is to be expected, was virtually universal. Important as they were, higher earnings and steady work were not the only metrics by which immigrants and their children measured financial success. A home of one s own property ownership also constituted a crucial goal a buffer against the inevitable earnings decline in old age and the only realistic means of accumulating capital (Golab 1977: 70). Judged from this vantage point, their American experience proved encouraging. Homeownership remained closely tied to age. Between the ages of and 60 plus, homeownership among male household heads more than doubled in the immigrant generation, reaching 68 percent for Poles and 75 percent for Italians. Thus, despite their anchor in the manual working class and modest earnings, an extraordinary number of first-generation Polish and Italian immigrants bought homes. Homeownership rates did not change much in the second generation, although fewer men in their thirties acquired property, probably the result of a delay imposed by depression and war. However, the homes of second-generation immigrants were worth much more than those of their parents, again underlining economic mobility between generations. With homeownership rather than earnings as the metric, immigrants emerge as more successful. Over time even working-class immigrants were able to purchase security and increase their wealth. However, another measure of economic standing poverty qualifies the image of modest first-generation success.5 Despite the financial progress of the immigrant generation, poverty among its households remained high. In 1950, 25 percent of working-age first-generation Poles and 24 percent of working-age first-generation Italians had incomes below the poverty line. Among the second generation, poverty rates dropped, although they were still high for younger families, 16 percent for Poles and 17 percent for Italians. By 1970, however, these poverty rates had plummeted to 10 percent for Poles and 9 percent for Italians.6 Nonetheless, the midcentury coexistence of extensive poverty with increased income and widespread homeownership underscores the emergence of economic differentiation in ethnic groups as a hallmark of immigrant history.

9 164 Social Science History Explaining Success The relative economic success of immigrants and their children after World War II did not result from either the presence of manufacturing jobs or the end of southern and eastern European immigration with the restrictive quotas passed in the 1920s. Rather, it reflected the impact of legislation and institutions, notably labor laws and unionization. Rising wages did not depend on a labor shortage induced by a shortage of unskilled or semiskilled workers willing to work for low pay. Instead, labor demand pulled African Americans from the South, Puerto Ricans, and eventually Mexicans into the nation s industrial economy. Thus the cessation of immigration did not exert its economic impact through an induced labor shortage. Rather, its economic consequences worked more indirectly through its influence on the social organization of ethnic communities, which by the 1920s and 1930s were dominated by the second generation. In 1930, 90 percent of Italians and Poles in their 40s were immigrants, while a majority of those in their 20s had been born in the United States. These second-generation ethnics constituted the backbone of the New Deal coalition and the constituency for the rise of organized labor (Cohen 1990). Membership in labor unions rocketed from under onetenth of the workforce in 1929 to about a quarter by the end of World War II. In mining, manufacturing, and transportation, which accounted for most of this increase, unionization had become the rule rather than the exception. Industries at the forefront of unionization employed a large fraction of immigrants and their children. In 1940, 30 percent of all foreign-born workers, compared to 25 percent of all blue-collar workers, were employed in the 12 industries leading the unionization wave. Among immigrants in 1939, 37 percent of Poles and 43 percent of Italians worked in these industries, which also employed 29 percent of second-generation Polish workers and 33 percent of second-generation Italian workers. Immigrants and their children had worked in these industries for a long time. But it took unionization to prod employers into offering decent wages and improved stability. Unionization in turn rested on New Deal legislation that legitimated the bargaining power of workers.7 In the 1930s the Wagner Act provided an administrative framework for unionization. The Fair Labor Standards Act set a minimum wage and regulated hours of work. The Federal Housing Authority guaranteed mortgages and facilitated the explosive growth of affordable housing in the suburbs. Social Security opened a path to an old age lived outside of poverty, while the

10 The Mexican Immigration Debate 165 rapid increase in unemployment compensation after World War II and the enactment of disability insurance in 1956 shielded this generation of workers from the two greatest risks to their financial well-being. After the war, unions, helped by an expanding economy, won large pay increases and health and retirement benefits. The GI Bill and expansion of higher education vastly increased opportunities for advanced education. With high rates of service in the armed forces, second-generation Italians and Poles were well placed to take advantage of the opportunities opened up by the GI Bill. Like other second-generation ethnic whites, 78 percent of second-generation Italian and Polish men born between 1918 and 1927 were veterans, compared to 74 percent of other whites and only 55 percent of African Americans.8 As a result, the percentage of college graduates jumped among the World War II cohort of second-generation ethnics and continued to climb among the cohort born in , most of whom had also served in the armed services. By contrast, after a small rise in the World War II cohort, the proportion of college graduates among African Americans dropped to its prewar level. As a result, by 1960 the fairly small gap in college graduation rates between white secondgeneration ethnics and African Americans had widened into a chasm, while between second-generation Italians and Poles and nonethnic native whites there was virtually no difference. Unionization, public social insurance, the GI Bill, and the expansion of higher education all brought unprecedented prosperity to the entire white working class, including the children and grandchildren of immigrants. As we shall see, Mexican immigrants arriving after the mid-1960s found a far less favorable institutional environment. Mexican Immigrant Paths to Economic Incorporation When Congress finally dismantled national-origin quotas in 1965, massive legal immigration resumed. Most of the new immigrants arrived from places deeply underrepresented among earlier newcomers. Prominent especially were Asia and Latin America, notably Mexico. Mexicans, of course, are not a new immigrant group. Until the midnineteenth century and the 1848 Treaty of Guadalupe Hidalgo, which ended the Mexican-American War, much of what is now the American Southwest was part of Mexico. And Mexicans have moved easily and frequently back

11 166 Social Science History and forth across the border throughout the nation s history. Federally legislated immigration quotas in the 1920s did not restrict legal immigration from Mexico, although the legislation did create the border patrol, but in the Great Depression of the 1930s the government, worried about jobs for Americans and the burden of welfare dependency, repatriated more than 400,000 Mexicans (Alba and Nee 2003: 329n28).9 In 1943, responding to the labor needs of American agriculture, the federal government initiated the bracero program, which allowed Mexicans to enter the United States as guest workers. This program ended in 1964, after Congress refused to reauthorize it. Mexican immigration accelerated once more in the 1960s as Mexico s economic miracle began to unravel. Rapid population growth and declining economic fortunes in Mexico coincided with the imposition of new U.S. limits on Mexican migration in 1976, when Western Hemisphere nations became subject to the same 20,000 annual limit already faced by nations in the Eastern Hemisphere (Donato 1994). The result promoted an explosion of undocumented immigration. Especially in California, this undocumented immigration met the needs of growers for large quantities of cheap labor, and American border authorities did little to impede the flow. The substitution of undocumented workers for the older bracero system served both sides reasonably well (Peters 1998). Two-thirds of Mexican migrants in these years were men, most of them married. The vast majority emigrated from midsize towns and cities. Their average age was 21 to 23; they had received only five years of schooling, but they had worked for many years. Only one-quarter followed agricultural occupations in Mexico, another quarter were unskilled workers, about one-sixth worked in service jobs, and one-tenth were skilled. They had strong links to the United States: two-fifths had a parent and onesixth had a sibling with U.S. experience; in the towns from which they came, one-third of the townspeople had spent time in the United States (Massey et al. 2002).10 The world of Mexicans had become increasingly transnational (Smith forthcoming). After 1986 the stable and circular migratory system between Mexico and the United States broke down. The Immigration Reform and Control Act of 1986 (IRCA), which tried to limit immigration, instead inadvertently shifted the balance even more toward undocumented workers and discouraged immigrants from returning to Mexico. On the Mexican side, economic collapse intensified the pressure to emigrate and to remain in the United States (Bean forthcoming: 6; see also Lozano-Ascencio et al. 1999; Marcelli

12 The Mexican Immigration Debate 167 and Cornelius 2001; Massey et al. 2002). In 1982, as oil prices imploded, the Mexican government floated the currency, which promptly lost half its value. Inflation soared, the GDP fell, real wages plummeted 21 percent, unemployment increased, and a chasm opened between rich and poor (Massey et al. 2002). At the same time in the United States mounting hysteria over illegal immigration reinforced by domestic economic problems helped pass the IRCA, the most visible symbol of the government s attempt to crack down on undocumented migrants. The IRCA provided amnesty for undocumented immigrants, implemented sanctions for employers of unauthorized workers, and strengthened border controls (Donato et al. 1992; Durand et al. 2005: 11 12; Zúñiga and Hernández-León 2005: xvi). The war on undocumented Mexicans, however, did not reflect the realities of American labor market demand, which remained strong. Much of the need for cheap unskilled labor shifted to new immigrant destinations, where population growth, economic restructuring, and demand for low-wage, low-skilled workers (as in the meatpacking and poultry industries) drew Mexican immigrants. (The share of first-time Mexican immigrants working in agriculture already had fallen before the shift to the new destinations accelerated [Donato 1994].) Between 1990 and 2000 the share of Mexican immigrants going to nongateway states such as Iowa and Nebraska jumped from 13 percent to 35 percent (Bean and Stevens 2003; Durand et al. 2005: 13). Most of these new destinations consisted of smaller towns and cities, where for the most part Mexican immigration has become a family affair as immigrants have used the amnesty provision of the IRCA to sponsor the immigration of family members (Zúñiga and Hernández-León 2005: xxvii). Aside from sponsored family members, 80 to 85 percent of Mexican immigrants now are undocumented, more of them are dependents (i.e., children who do not work), they tend to stay longer, and many more become citizens (Gonzales Baker et al. 1998; Passel 2005). American immigration policy perversely has helped transform a seasonal, circular flow back and forth across the border into a permanent settlement and encouraged additional migration from Mexico to the United States (Massey et al. 2002: 136). The source for this analysis of Mexicans in the United States the 1 percent samples of the U.S. federal censuses created by the Integrated Public Use Microdata Series (IPUMS) project at the University of Minnesota poses two problems. The first is how the large number of undocumented immigrants informed estimates put the number at 3.8 million in

13 168 Social Science History 2002 and 6.0 million in 2005 affects the data (Lowell and Suro 2002; Passel 2005). Although a precise answer is impossible, many of the undocumented were included in the census, and there is no reason to believe that the census presents an unrepresentative portrait of the Mexican population. Undocumented immigrants, however, add a complication to the comparison between early-twentieth-century newcomers and current Mexicans because their status poses a barrier to mobility not confronted by Italians, Poles, and others who arrived earlier. In fact, this barrier makes the achievements of Mexicans, discussed below, even more impressive. The second problem posed by the data concerns identifying the second generation. First, because individuals and families were part of a two-way migration, it is possible that someone identified as a second-generation Mexican American spent a substantial portion of his or her life in Mexico at the same time that someone identified as a Mexican immigrant had parents who spent much of their lives in the United States. For example, of individuals born in Mexico enumerated in 1970, 8 percent had a U.S.-born mother, and 9 percent had a U.S.-born father. Second, after 1970 the U.S. census ceased to ask for parental birthplace. Before 1980 the decennial census included a question on parents birthplaces that made the identification of the second generation straightforward. Beginning in 1980, the census dropped this question in favor of inquiries about a person s ancestry. The inclusion of questions about Hispanic origins only partially remedied this problem. As a result, for the last three censuses we have been able to differentiate only between Mexicans born in Mexico and individuals of Mexican ethnicity born in the United States. To compensate for this problem, we have used the Current Population Survey, which added questions on parents birthplaces in 1994 to compare the experience of U.S.-born Mexicans whose parents were born in the United States with the experience of those whose parents were foreign-born. Although the frequency of border crossing makes this an imperfect variable, the analysis provides a standard against which to compare the second-generation Italian and Polish cohorts with U.S.-born Mexicans in the late twentieth century. (For an alternative method of compensating for the absence of parental birthplace, see Perlmann 2005: ) The individuals identified as Mexican and born in the United States comprise the second and subsequent generations. According to the Current Population Survey, in 2000, among ethnic Mexicans born between 1945 and 1974, 58 percent were born in Mexico, 13 percent were U.S.-born with at

14 The Mexican Immigration Debate 169 Figure 1 Mexicans and Mexican Americans in blue-collar occupations, 1970 and 2000 Source: Ruggles et al least one foreign-born parent, and 29 percent were U.S.-born with two U.S.- born parents. In fact, there was little socioeconomic difference between the U.S.-born groups.11 Here we will refer to them mainly the children and grandchildren of immigrants as Mexican Americans. Most Mexican immigrant men remained in blue-collar work through midcentury (see figure 1). Although Mexican Americans moved into whitecollar work more often than immigrants did, a plurality still remained bluecollar. A large fraction of each group also worked in agriculture or as service workers. Women, however, moved much more decisively into white-collar jobs. Although employment patterns varied with local labor markets, more than 8 of 10 immigrant men worked at blue-collar jobs in There had been essentially no change in this number over time. The change that did occur was the 50 percent fall in the number in agricultural work between 1970 and Nonetheless, in 1990, 91 percent of California s farmworkers were immigrants, primarily from Mexico (Cornelius 1998). In 2000 the modal occupational category remained operative, and 35 percent worked in agriculture or some form of service job; during the last three decades of the twentieth century, the share in white-collar work hovered around 15 percent. Blue-collar workers still composed the majority of Mexican American workers in A large fraction, however, had reached white-collar work: 41 percent of the oldest cohort and 42 percent of the youngest cohort. In their

15 170 Social Science History occupations, Mexicans thus closely paralleled Italians and Poles who arrived in the early twentieth century. In their occupations, Mexican American men differed from U.S.-born white men, who were much more likely to work in professional or technical jobs or as managers, officials, and proprietors and less than half as likely to be laborers. At first this occupational advantage appears to reinforce the fear that second- and subsequent-generation Mexicans face a dim economic future that will leave them further behind the grandchildren and greatgrandchildren of earlier immigrants. In an important way, however, this inference misleads. A valid comparison of intergenerational mobility would compare individuals with similar social origins. But this is not how the data are usually presented. Most Mexican American men grew up in not very well-off working-class families, while many more of the U.S.-born whites had fathers with white-collar jobs and higher incomes. A study that controls for family class origins should be a priority for research on comparative, intergenerational mobility. For Mexican women, however, the story is quite different. In their occupations, Mexican American women increasingly resembled U.S.-born white women. As with African Americans, whose work history we have traced elsewhere, the trajectory of work resulted in the emergence of a gender gap among Mexican Americans (Katz et al. 2005a). While the majority of immigrant women were concentrated in blue-collar work, two-thirds to threequarters of Mexican American women were found in white-collar occupations. Among Mexican American women, employment in agriculture, apparel, and private household service plummeted, while it rose in medicine, hospitals, general merchandise, federal public service, banking and credit, insurance, and general business services. More than Mexican American men, Mexican American women were moving into the expanding sectors of the service economy. In New York City, Robert C. Smith (2002) found that between 1980 and 1990, 19 percent of Mexican American men and 30 percent of Mexican American women were upwardly mobile. In Los Angeles, Dowell Myers and Cynthia Cranford (1998) found Mexican American women moving into white-collar work more than Mexican American men. The modal occupational category of first-generation Mexican women was operative; for Mexican American women it was clerical. The shift was dramatic. In 1970, 37 percent of Mexican immigrant women born between 1945 and 1954 were operatives. A minuscule 2 per

16 The Mexican Immigration Debate 171 cent worked in the skilled crafts and as laborers. Mexican American women, by contrast, were half as likely to work as operatives, half as likely to work in household service, and more than twice as likely to be in professional or technical jobs, teaching, or managerial work. By the end of the century, only one in three was in blue-collar work. In fact, by 2000 the occupational profile of Mexican American women did not look very different from that of native-white women. (These national trends mask the limited opportunities for immigrant women in rural areas, such as Nebraska, where they have few alternatives to work in meatpacking [Gouveia et al. 2005: 39].) The convergence between native-white and Mexican American women marked a huge shift since With the Mexican Americans, as with other ethnic groups and African Americans, women led the way into white-collar and professional work. (The proportion of African American women in white-collar work increased from 2 percent in 1940 to 65 percent in 2000, a leap much greater than that among African American men and one that brought them closer to white women.)12 In one way Mexicans differed from earlier immigrants. Poles and Italians were two or three times more likely than Mexicans to work for themselves.13 Thus Mexicans were not often able to start business careers by working for coethnics who provided crucial on-the-job training a first solid rung on the ladder to entrepreneurship (Raijman and Tienda 2000). Where Mexicans did own businesses, they provided crucial employment for coethnics recruited primarily through referrals from other employees (Cornelius 1998). This employment of coethnics did not always result in higher wages. Wages depended on the size of ethnic markets and the concentration of ethnic selfemployment (Spener and Bean 1999). For Mexicans, work in the public sector, not available to Italians and Poles earlier in the century, substituted for self-employment. Public-sector work played a small role in the experience of the immigrant generation but a substantial and generally increasing one among Mexican Americans. Indeed, public-sector employment rose particularly steeply between 1970 and 1980, reflecting the expansion of government social spending in those years. As with men, many more Mexican American than immigrant women found employment in the public sector. At the end of the twentieth century, the public sector directly employed about one-quarter to one-third of them. A rough aggregation of employment in the state-related sector (health and medicine, education, social service and nonprofits, other public service) accentuates the

17 172 Social Science History Figure 2 Real personal earnings of Mexicans and Mexican Americans, by age, 2000 Source: Ruggles et al Note: Real personal earnings are adjusted for the consumer price index and are reported in constant 1990 dollars. importance of publicly funded employment for Mexican American women. In 2000, 47 percent of Mexican American women and 25 percent of Mexican American men worked in this sector. As with African Americans, the expansion of public employment and publicly funded employment in private agencies opened opportunities for Mexican Americans, especially women. And it left them particularly vulnerable when the public sector faced cuts. Public-sector work in fact paid relatively well and offered the most certain route to modest prosperity. In both 1990 and 2000 the median earnings of men and women of Mexican background were higher in the public sector than in the private sector and close to the earnings of the small number of self-employed. Public-sector earnings also varied in a narrower band. Public employment did not hold out the prospect of wealth, but it did promise reasonable pay: a trade-off of comfort and security for the possibility realized by relatively few of wealth. In earnings as well as in occupation, Mexican American women, but not men, began to converge with native-white women. If we correct for educational attainment, the convergence is especially striking. In 2000 Mexican American women in their early forties earned $26,525, compared to $27,872 among non-hispanic white women. With Mexican American men earning only about one-half to two-thirds of family income, the contributions of married women proved crucial. The other important

18 The Mexican Immigration Debate 173 Figure 3 Educational attainment of Mexicans and Mexican Americans, 2000 Source: Ruggles et al point about earnings among immigrants is that they increased with years resident in the United States (see figure 2). Economic mobility, although not dramatic, was solid and unmistakable. Late-twentieth-century Mexican immigration coincided with tightening links between human capital and economic success a development that handicapped the many Mexican immigrants who arrived with minimal schooling.14 Mexican immigrants attended school for significantly fewer years than non-hispanic Americans (see figure 3). In 2000, 61 percent of Mexican men and 54 percent of Mexican women had eight years or less of schooling. This situation, however, does not prove that the educational gap between Mexicans and the U.S.-born is greater today than it was between immigrants and natives in the early twentieth century, as critics of Mexican immigration argue (see, e.g., Borjas 1999). Early in the twentieth century, as the discussion of Italians and Poles showed, an immense difference in literacy and common school education separated natives and newcomers. Educational attainment rose sharply among Mexican Americans: 41 percent of men and 50 percent of women had completed at least one year of college. Educational attainment kept rising among younger cohorts. The educational level of younger immigrants born between 1965 and 1974 was notably higher than that of those born 10 years earlier: only about one-third had eight years or less of schooling, reflecting educational changes in Mexico.

19 174 Social Science History Figure 4 Real personal earnings of Mexicans and Mexican Americans, by educational attainment, 2000 Source: Ruggles et al However, the educational experience of Mexican American men and women went in opposite directions: the proportion of men with at least one year of college dropped, while among women it rose to half highlighting stagnant circumstances for men but not for women.15 As with African Americans, the history of education revealed the emergent gender gap that had opened among Mexicans. (Black women began to graduate from college with greater frequency than black men in the 1950s, and their lead increased gradually until the 1980s, when it rose sharply.) Unsurprisingly, Mexicans earnings rose with educational attainment (see figure 4). For Mexican American men, earnings increased from $22,757 for those with 12 years of schooling to $47,250 for college graduates. Although women earned less, college graduates earned substantially more than those with a high school education. Homeownership also reflected modest economic mobility. Like the Italians and Poles earlier in the century, most Mexican immigrant families managed to buy a home. Among Mexicans, homeownership went up with age, as it did among all other groups. By the age of 50 59, 66 percent of Mexican immigrant male household heads born between 1945 and 1954 owned their own homes, as did 77 percent of Mexican American men from the same

20 The Mexican Immigration Debate 175 Table 3 Income distribution, by gender, for Mexicans and Mexican Americans 1970 Personal income (percentiles) Under Mexicans Men Women Mexican Men Americans Women Mexicans Men Women Mexican Men Americans Women Source: Ruggles et al cohort. The value of homes, however, did not increase much with the age of owners or between generations. Still, over time the great majority of Mexican families acquired modest assets and the security of a home. Also like earlier immigrants, most Mexican immigrants eventually learned to speak English. Among Mexicans who had been in the United States for less than 5 years in 2000, 41 percent of men and 46 percent of women spoke only Spanish. Among Mexicans who had lived in the United States more than 20 years, only 7 percent of men and 10 percent of women remained monolingual. Virtually all Mexican Americans spoke English, and many were bilingual. The pessimists fear of huge numbers of Spanishspeaking children and grandchildren of Mexican immigrants Balkanizing America linguistically and displacing the English language clearly represents an imagined disaster, not a realistic future.16 Late in the twentieth century, the relative economic position of Mexican immigrants and Mexican Americans, like that of African American men and most of the working class, remained stagnant or began to slip. As Annette Bernhardt and her colleagues show in their book Divergent Paths (2001), the prospects of young, white, male workers began to deteriorate sharply in the late 1960s and the 1970s. It would be odd indeed if Mexican Americans proved an exception. The distribution of earnings by percentiles highlights this slippage among Mexicans and Mexican Americans (see table 3).

21 176 Social Science History Table 4 Poverty rate, by year of birth, for Mexicans and Mexican Americans, Percentage of individuals in poverty Census year Year of birth Mexicans Mexican Americans Source: Ruggles et al Between 1970 and 2000 the share of Mexican and Mexican American men in the lowest earnings quintile increased, while their share among the top two quintiles declined. Although Mexican American men improved on the income of immigrants, the relative positions of both groups slipped. Women both immigrants and Mexican American were overrepresented in the two lowest quintiles and were badly underrepresented in the top two. Thus it is not surprising that a large fraction lived in poverty (see table 4). However, poverty declined with age and was lower among Mexican Americans than among immigrants. Between 1980 and 2000, for example, among immigrant adults born between 1945 and 1954, poverty fell from 23 percent to 19 percent. In the same years it declined among Mexican Americans born a decade later from 24 percent to 15 percent. These numbers put the poverty rate of Mexicans above that of native whites but below that of African Americans. At first, poverty and the prospect of blocked mobility appear to distinguish the economic history of Mexicans and their U.S.-born descendants from early-twentieth-century immigrants, and even researchers who find parallels between early- and late-twentieth-century immigrants worry that economic assimilation will progress much more slowly among today s newcomers. Mexicans, they predict, could take a generation longer than earlier immigrants to reduce the gaps in occupation and income that distance them from native-born Americans (Bean and Stevens 2003: 112; Perlmann 2005: 96, 116). In part, the evidence already presented qualifies this concern by

22 The Mexican Immigration Debate 177 Table 5 Occupational income scores, by length of residence in the United States, in metropolitan areas outside the South Gender Ethnicity by place of birth 0 5 years Years in the United States 6 10 years years years 21+ years Increase per year in U.S. Men Italian Polish Foreign-born Mexican Women Italian Polish Foreign-born Mexican Source: Ruggles et al Notes: For an explanation of occupational income score, see These scores are derived from a general linear model analysis that controls for age. highlighting the relatively slow economic progress of Italian and Polish immigrants. A set of regression analyses that estimate the increase in a number of measures of economic progress and assimilation per year in the United States points to the same conclusion (see table 5). The measures are occupational income score, Duncan occupational status score, percentage of workers in white-collar occupations, property ownership, and ability to speak English. The analyses compare immigrants who arrived at immigration s peaks: and The first immigration wave crested in 1907 and then receded abruptly with the beginning of World War I in The second peaked in 1991, when 1.8 million immigrants entered the United States. During the rest of the decade, economic recession combined with restrictive legislation to push immigration below 1 million entrants per year. The formal model includes main effects for ethnicity, years in the United States, age, and the interaction effect for ethnicity and years in the United States. It covers Italian, Polish, and Russian Jewish immigrants for the earlier period and Mexicans, other Latin Americans, and Koreans for the later one.17 Because European immigrants were almost totally absent from the South during the early twentieth century, this comparative analysis is necessarily restricted to nonsouthern metropolitan areas. The analysis of occupational status includes only labor force partici

23 178 Social Science History pants, and the regression for property ownership includes only household heads.18 The regressions underscore the parallels between early- and latetwentieth-century immigrant men. They also highlight the differences between women and men and the differences between Italian and Polish immigrant women early in the century and Mexican women today. For example, among men resident in the United States for 0 5 years, the occupational income score based on the relative ranking of median income in 1950 was and for Italian and Polish immigrants, respectively, and for Mexicans.19 All groups progressed with length of stay in the United States. For Italian, Polish, and Mexican men, the annual increase was In their initial rank, Italian women (19.03) outscored Polish (13.91) and Mexican (16.37) women. But Mexican women progressed faster: 0.23 per year, compared to 0.12 for Italian women and 0.14 for Polish women. The accelerated progress of Mexican women resulted from the expansion of occupational opportunity for all women in the late twentieth century (Katz et al. 2005b). Trends in white-collar occupations underline women s occupational advance. After 20 years in the United States, among most groups, more immigrant women than immigrant men worked in white-collar jobs: 23 percent of Mexican men, compared to 45 percent of Mexican women. Women moved into white-collar work faster than men as well. For Italians, the increase per year in the United States was 0.91 for men and 1.23 for women; for Poles, it was 0.61 for men and 1.15 for women. Again, however, the emergent gender gap among Mexicans stands out from the 0.44 increase for men, compared to the 1.23 increase for women. After 20 years in the United States, 35 percent of Italian immigrant men and 48 percent of Polish immigrant men owned property. The surprise is Mexican men s rapid increase in property ownership an annual rate of 2.04 percent, the highest of any of the groups. Among Mexican men, property ownership increased from 16 percent for those resident in the United States for 0 5 years to 61 percent for those resident for 21 years or more. Mexicans also arrived speaking English more often than earlier immigrants, and they retained their advantage.20 This discovery contradicts critics who inject Mexicans allegedly reluctant and slow acquisition of Englishlanguage skills into the politics of immigration (Woolard 1989; Citrin et al. 1990).

24 The Mexican Immigration Debate 179 Among immigrant men in the United States for 0 5 years, 23 percent of Italians and 15 percent of Poles spoke English, compared to 54 percent of Mexicans. All groups made rapid progress. For the most part, in immigrant groups, women s ability to speak English tracked men s. Thus, with earlytwentieth-century immigrants as the point of comparison, Mexican immigrant assimilation seems neither unusually tepid nor slow. Mexican men s advance was measured, close in tempo to Italians and Poles, but Mexican women s progress proved dramatic, distancing them from Italian and Polish women and from Mexican men as well. Discussion The Harvard economist George J. Borjas (1999: 26), one of the most prominent immigration scholars, asks: Are the immigrant waves arriving today as skilled as those that arrived twenty or thirty years ago?... this question has a long and less than honorable history in the immigration debate. The checkered history of this question, however, does not diminish its relevance for evaluating the economic impact of immigration or for thinking about the type of immigration policy that the United States should pursue. Researchers can find a convincing answer, according to Borjas (ibid.), by contrasting the skills and economic performance of the most recently arrived immigrants enumerated in the 1960 or 1970 census with the skills and economic performance of the most recently arrived immigrants enumerated in later surveys. Here is the problem. Immigrants entering the United States between 1955 and 1965 Borjas s reference group do not compose the most appropriate population for evaluating the prospects of more recent Mexican immigrants. Nationality-based quotas introduced in the 1920s, followed by the Great Depression and World War II, slowed immigration drastically. In 1905 more than 800,000 immigrants entered the United States; in 1955 the number was about 300,000. The 1905 arrivals represented more than 10 immigrants for every 1,000 residents; the 1955 newcomers represented fewer than 2 (Smith and Edmonston 1997: 32, fig. 2.1). Other than refugees, immigrants in the 1950s and 1960s arrived during years of strong economic growth, when their small numbers exerted minimal impact on the availability of jobs or on wages. Unlike the Mexicans who arrived later, mid-twentieth-century European