Introduction

Throughout its history, the Mexico-United States border has developed cross-border relationships tied to numerous mobilities of the population across the boundary line for various purposes. The porosity of the border, the cross-border social networks, and the structural differences between both countries have allowed the occurrence of intense cross-border processes (^{Alegría, 1992}). These crossings from one border to the other, far from being ahistoric, have depended on economic shifts and the actions of the nation-states to exercise their sovereignty and regulate the economic markets (^{Alegría, 1992}).

This work will examine the recent transformation of cross-border interactions. The objective is to study the changes in working and giving birth in the United States of the Mexican border population during the first decade of the 21^st century, as well as to examine the profiles of those who participated in these processes. Despite the continuity of strong structural differences between the neighboring countries, the conditions under which their practices are produced have changed in the last decade.

Taking this into consideration, the two cross-border process are studied through the conceptual framework of the transborder interactions (^{Bustamante, 1989}). Both crossing to work as well as to give birth are two phenomena that reflect the recurrence and the internationality of interactive processes between the border populations. They can be defined as cross-border interactions at the micro-dimensional level in a structural context of power asymmetry (^{Bustamante, 1989}). In particular, these are indicators of the population strategies to take advantage of the socioeconomic opportunities in a context of strong inequality between both countries (^{Ojeda, 1994}; ^{Escala and Vega, 2005}; ^{Cueva and Vásquez, 2009}).

Due to the North American economic opening and integration propitiated by the North American Free Trade Agreement of 1994 (NAFTA), transnational processes grew in Mexico as a whole, at the same time cross-border interactions did not fall (^{Alegría, 2002}; ^{Tuirán and Ávila, 2002}). It seems, however, that these interactions did decline in the past decade, when important recent processes in political, economic, and public safety terms took place in both neighboring countries.

First, the control of the southern U.S. border was reinforced, was surveillance in the U.S. interior part of the national security policy implemented after the terrorist attacks of 2001, which brought not only changes in undocumented immigration but also in the expansion of the anti-immigrant climate (^{Ochoa and Sánchez, 2011}). In consequence, the deportation or expulsion of Mexicans from the interior of the United States increased, with Mexican border cities playing a crucial role receiving these migrants (^{Masferrer and Roberts, 2012}).

Second, starting with the U.S. financial crisis that began in 2007, the economic situation both in the south of the United States as well as in the north of Mexico worsened. Unemployment along the southern U.S. border increased drastically, particularly in California and Arizona, border states with large immigrant populations (^{Mendoza, 2012}). Similarly, in the north of Mexico, unemployment increased due to the drop in U.S. market demand, as important sectors of the economy such as the export industry and services oriented toward U.S. consumption were affected (^{Calderón et al., 2010}).

Third, Mexico's northern border experienced changes in terms of public safety. Although the area had experienced violence since the beginnings of the past decade, the situation worsened as a result of the world economic crisis and the war between the government and the drug cartels (^{De la O and Flores, 2012}). In consequence, displacements increased from the states of Chihuahua, Nuevo León, and Tamaulipas toward other Mexican states and the United States (^{Hernández, 2014}; ^{Arceo-Gómez, 2012}).

Various studies have shown the recent change in cross-border interactions in this area of Mexico. Using a quantitative methodology, ^{Orraca (2015)} shows that the number of cross-border workers fell between 2000 and 2010, but that those who continued crossing had higher educational levels and salaries. ^{Vargas (2015)} also finds that, in this period, youngsters in households with cross-border workers dropped and that those born in the United States rose, whose births occurred at the beginning of the 1990s, when tougher U.S. actions to block border crossings began. ^{Velasco and Contreras (2014)}, using a qualitative methodology, argue that the new context of border control has propitiated greater uncertainty and differentiated opportunities for cross-border interactions, where only a smaller sector of society can take advantage of these opportunities to strengthen their socioeconomic position.

Based on these findings, this study states that the two transborder links have evolved in opposite ways: While cross-border labor mobility fell, childbirths in the United States rose. The study also supports the idea that these changes are associated with a greater socioeconomic selectivity of the population participating in these interactions. In addition, it is assumed that childbirths in the United States increased as a means for the upper-middle class to obtain double citizenship for their children and gain this way eventual access to the U.S. labor market.

In the first section of this study, the theoretical-empirical background of cross-border work and giving birth in the United States of the Mexican border population are synthesized; the second section details how the change in the migratory and socioeconomic context could have influenced cross-border links; the third section describes the methodology used in the statistical analysis; and the fourth section exhibits the results of the study. Finally, in the conclusions, the most important findings of this paper are discussed.

Theoretical-empirical background of selected cross-border links

Working and giving birth in the United States have been interpreted as two indicators of the existence of cross-border links between the populations on both sides of the border (^{Alegría, 1992}; ^{Ojeda, 1994}; ^{López, 1994}), due to the border location of both origins and destinations of the crossings of people involved (^{Alegría, 1992}). These phenomena are part of a border social dynamic that tends to take advantage of opportunities arising out of the asymmetry between the two neighboring countries (^{Ojeda, 1994}; ^{López, 1994}; ^{Escala and Vega, 2005}). They are considered to be two forms of the "transborder way of life", which while apparently disconnected, are both part of the border social system.

The crossing to the United States for work purposes by border residents of Mexico has been well studied (^{Acuña, 1983}; ^{Arámburo, 1987}; ^{Herzog, 1991}; ^{Alegría, 1992}, ²⁰⁰²; ^{Estrella, 1994}; ^{Escala and Vega, 2005}; ^{Coubès, 2008}; ^{Chávez, 2011}; ^{Orraca, 2015}). It is not the only cross-border interaction in labor terms, but it is emblematic of the relations that are generated from the asymmetries of the labor markets of both countries. The research about crossings in order to give birth in the United States has been less numerous (^{Guendelman and Jasis, 1992}; ^{Ojeda, 1994}; ^{López, 1994}; ^{Cueva and Vásquez, 2009}). The studies of both issues have allowed the understanding of the mechanisms that explain these types of crossings at a structural and micro or meso levels.

From a structural point of view, it has been argued that the cross-border processes take place due to the deep economic differences between the involved countries (^{Acuña, 1983}; ^{Bustamante, 1989}; ^{Alegría, 1992}). Thus, it is argued that the crossing of the commuters¹ is due to the wage differences between the pairs of border cities more than to factors that merely have to do with labor supply and demand (^{Alegría, 2002}). Also, from this perspective it has been argued that the crossings to give birth in the United States occur because of the binational differences in the structure of socioeconomic opportunities for children-for example, in the labor markets or educational services (^{Ojeda, 1994}; ^{López, 1994}; ^{Vargas, 2015}).

An essential aspect of the structural perspective for the study of cross-border interactions is the control that nation-states exercise over their territories to slow the volume or define the characteristics of the immigrants or foreigners who cross their borders (^{Alegría, 1992:42}). Alegría sees the border as a "selective barrier" for cross-border relationships, where countries only allow in what is not considered negative for their development. Also, ^{Kearny (2008)} considers the dividing line as a "classifying filter" that differentiates between populations in accordance with their socioeconomic classes and that transforms class relationships between those who cross the border and those who cannot. In this sense, border security and migration policies may limit the intensity of cross-border relations and the profiles of those who may perform them.

Thus, the last regularization of undocumented workers in the United States, the Immigration Control and Reform Act of 1986 (IRCA) (^{Congress of the United States of America, 1986}), better known as the Simpson-Rodino Act, delineated various characteristics of cross-border workers. Once they had obtained their residency card, various migrants decided to live on the Mexican side of the border and go to work as commuters. This law was created mainly for agricultural workers and, as a result, had gender and generational biases (^{Estrella, 1994}). Thus, in the past decade, studies emphasized that cross-border workers increasingly were men of older ages (^{Escala and Vega, 2005}; ^{Coubès, 2008}; ^{Orraca, 2015}).

Similarly, the selective nature of the commuters was documented. The majority had papers to cross legally (although not all to work), had more education than the average worker where they lived (although they took jobs they were overqualified for), and they used a car as a transportation mean (^{Alegría, 1992}; ^{Escala and Vega, 2005}; ^{Chávez, 2011}).

The porosity of the border has been also crucial for women of certain social groups to give birth in the United States. Only border women who have crossing documents have been able to plan their children's delivery in this country (^{Guendelman and Jasis, 1992}; ^{Cueva and Vásquez, 2009}). Also, ^{Guendelman and Jasis (1992)} showed that the majority of the Tijuana women who crossed to give birth in the United States belonged to the middle class and made use of private medical services; that is, this practice did not involve a cost to the U.S. treasury. Later, ^{Cueva and Vásquez (2009)} confirmed this pattern at the border of Matamoros with Brownsville.

From the micro and meso-social perspectives, the cross-border processes have been interpreted as part of the life strategies of individuals or families in a binational context of socioeconomic inequalities and cultural differences. Thus, cross-border labor mobility has been addressed as an option that allows workers to maximize their available socioeconomic resources at the border through better-paid jobs than they can obtain in Mexico yet maintain a residence at a lower cost than in the United States (^{Arámburo, 1987}).

It is not just the individual who decides to cross to the United States to work but also the family (^{Estrella, 1994}). This decision may correspond to a family strategy for cross-border life. Thus, considering a familial division of labor, more heads of household are found in these jobs (^{Coubès, 2008}). Even when the cross-border workers are doing jobs on the U.S. side they are overqualified for (^{Alegría, 1992}), they earn more than double what workers make on the Mexican side (^{Alegría, 2002}). The buying power of the cross-border workers and their permanent mobility between one side and the other allows them to purchase goods and services that are cheaper in one locality or the other and to be linked to border families and cultures (^{Escala and Vega, 2005}).

Giving birth in the United States has also been studied as a couple's or individual's strategy (^{Ojeda, 1994}) to give their children U.S. citizenship. This motivation considers the possibility that double citizenship will allow children to access the U.S. labor market, buffer their economic risks in the medium term, and even obtain legal emigration status for themselves and the family (^{Ojeda, 1994}; ^{Cueva and Vásquez, 2009}). In particular, highly educated women have expressed motivation for giving birth in the United States in the belief it can pay off in the medium term (^{Cueva and Vásquez, 2009}). To make these expectations a reality, cross-border networks are also necessary, as relatives and friends in the south of the United States provide information about the health system as well as lodging (^{Guendelman and Jasis, 1992}; ^{López, 1994}; ^{Cueva and Vásquez, 2009}).

Finally, from the family life strategies approach, having children and working in the United States may correspond to a single family strategy in two stages: have children on the other side so that, once they are adults, they can be commuters- and not necessarily international migrants (^{Ojeda, 1994}; ^{Escala and Vega, 2005:154}). Both practices are examples of the use of border links as a strategy of social mobility of the families who live on the border (^{Estrella, 1994}).

Impact of the recent changes on cross-border links

This section will show in greater detail how the new migratory and socioeconomic border contexts could have transformed the structural factors that facilitate cross-border interactions.

In terms of economic change, the wage gaps between both countries continue to exist, in spite of the impact of the crisis of 2008 on the Mexico-U.S. border labor markets. In contrast to the crises of the 1980s and that of 1994, which occurred on the Mexican side and whose devaluations caused an increase in cross-border workers (^{Alegría, 2002}; ^{Coubès, 2003}), the crisis of 2008 originated in the United States, but also affected the border economies because of their large interconnection with that country (^{Samaniego, 2009}). However, at the same time that rising unemployment in border cities brought about an incentive to work on the other side, the increase in unemployment in the south of the United States reduced the demand for the jobs these kinds of workers do.

The economic contraction also could have had an impact on access to U.S. health services for border populations between 2000 and 2010. The possibility of paying hospitals to give birth in the United States perhaps fell, as well as the access to health insurance in that country. Nevertheless, the crisis effect could have been different depending on economic status, with the lower middle class population being affected the most.

Migratory controls and the anti-immigrant climate in the United States could have had a negative impact on the employment of cross-border migrants, as both probably affected the demand for undocumented labor due to the restrictions placed on employers as well as the supply of undocumented workers because of their fear of being detained. Standing out for its effects on the economic participation of undocumented workers was the Legal Arizona Workers Act of 2007 (HB 2779), which prohibited employers from hiring foreigners without authorization to work and sanctioned employers who hired such foreigners (^{Lofstrom, Bohn, and Raphael, 2011}). In addition, Latino immigrant workers experienced labor instability and falling wages following the crisis of 2008 (^{Kochhar, Espinoza, and Hinze-Pifer, 2010}).

However, parents' desire to obtain double citizenship for their children might have increased with the new migratory situation. On the one hand, cross-border social networks have lasted for generations. Even after the terrorist attacks, more than half of the citizens or residents of the United States who crossed the Mexican border did so to visit relatives and friends (^{SANDAG and CALTRANS, 2010}). On the other hand, the economic crisis could have evidenced the higher value citizenship has to improve the access into the U.S. labor market.

It should be noted about border crossings that, after the terrorist attacks, surveillance of the border was reinforced and border wait times grew (^{SANDAG and CALTRANS, 2010}). This also must have made the daily travels of the cross-border workers to their places of work more difficult than the planned crossings to give birth. Those workers born in the United States or with work permits could have migrated to that country or acquired the SENTRI card (Secure Electronic Network for Travelers Rapid Inspection) (^{IVTS, 2015}).² Finally, societal violence could also have affected the permanent migration of some commuters to the United States. In this sense, ^{Orraca (2015)} documents that at the same time that Mexican migration to the United States decreased, the immigration of Mexicans with high educational levels took place to U.S. counties adjoining Mexico's northern border, with the migration of cross-border workers probably being part of the explanation.

In synthesis, the new economic, migratory, and social context of the last decade appears to have increased barriers for the cross-border workers. In contrast, children born in the United States may have risen, except those of lower class women, as the border controls did not affect legal border crossings nor cross-border social networks. Finally, as a result of more control for undocumented labor mobility in addition to the economic collapse, greater socioeconomic selectivity was to be expected in the cross-border links registered during this period.

Methodology

The basic sources of information for this study were the 10 percent samples of the Mexican population censuses of 2000 (^{INEGI, 2001}) and 2010 (^{Inegi, 2011}).³ The microdata of the two most recent population censuses allowed estimating the prevalence of the selected indicators of cross-border links and the socioeconomic characteristics of the populations involved in these processes. Only the population of the adjacent municipalities to the United States were included. The sample numbers are presented in each table.

With this information, a variable was created for each type of cross-border link: employment in the United States and the birth, in that country, of the northern border residents in 2000 and 2010. Cross-border workers were defined as the employed workforce aged 15 and up, whose workplace was in the United States, and acquired the value of 1 if they met this condition and 0 when they did not.

Births in the United States as a result of border crossings were identified through the place of birth of the population aged 0-4 years and the mother's country of residence five years before the census, with the objective of distinguishing between births that occurred specifically as a result of a border crossing and births to migrant mothers who had returned to Mexico.⁴ The population born in the United States as a result of border crossings was defined as 1 if the child was born in the United States and the mother lived in Mexico five years before the census, and 0 if the child was born in Mexico. A third category also was created of children born in the United States but whose mother lived in that country five years before the census, with the goal of showing the small numbers of children of mothers who returned to Mexico with respect to the total of the children born in the United States.

The identification of children's mothers was done using various methods depending on the census year due to differences in the availability of information. In 2010 it was possible to directly merge the mother's information with their children's. The mother could not be identified at home only in 2.5 percent of the cases; these children were excluded from the study. In contrast, in 2000 it was not asked who the parents of the children were. Therefore, when the study dealt with the children of the head of household, the mother was taken as the head in cases where there was a female head of household, or as the wife when the head of household was a male; the residence information five years before the census of any of these women was assigned to the child. For children who were not sons or daughters of a head of household, recent migratory information of the reproductive aged women at home was used. If any of these women was a recent immigrant from the United States, this characteristic was attributed to the infant. The indirect estimation to determine the number of children born in the United States to border resident mothers was validated with information from the 2010 census (^{Inegi, 2011}) and was quite close to the direct estimation.⁵ Thus, the place of residence of mothers who gave birth in the United States was determined utilizing the direct method in 2010 and the indirect method in 2000.

The statistical analysis was done with the software Stata/SE and consisted of three parts. First, the 2000-2010 percentage change of cross-border link indicators was estimated for the total and the most-populated municipalities of the northern border, where 89 percent of the border population lives. Then the frequencies of cross-border workers and of births in the United States were calculated for five-year age groups both in 2000 and 2010. Finally, some demographic and socioeconomic characteristics were explored⁶ to find changes in the profiles of the populations participating in both processes.

Similarly, various types of statistical tests of differences were undertaken; the differences in characteristics between the groups with or without cross-border ties were determined for each year using chi squared and F tests, according to whether they were categorical or continuous variables, respectively. All of the variables showed significant differences within each year.

In addition, to analyze the change over time in the sociodemographic profiles of the populations with cross-border links and those without, different methods were used depending on whether a continuous or discrete variable was utilized. For the continuous variables, models of linear regression were performed with interactions of these variables over time. For the categorical variables, log-linear models were used, which are suitable for testing differences in three-way contingency tables (^{Agresti, 2007}). For these last models, the aggregate data of the 2000 and 2010 census samples were combined and interactions over time with all the categories of independent variables were tested. Also, a comparison was done of the saturated models with the models without the interaction between the cross-border link over time and the evaluated characteristic, through log-likelihood ratio tests. For simplicity, Tables 2 and 4 only show whether the interactions were or were not statistically significant at the 95 percent confidence level.⁷

The levels and the characteristics of the commuters between2000 and 2010

Between 2000 and 2010, the cross-border workers in the northern border region fell both in absolute numbers (from 87 000 to 75 000) and in relative ones, as their proportion of the economically active population dropped from 3.9 to 2.8 percent (Table 1). The only exceptions to this generalized decline were the municipalities of Ciudad Acuña and Piedras Negras, where the share of these workers among the employed labor force remained stable and had a slight increase, respectively.

Table 1 Employed Labor Force age 15 and above that Works in the United States. Total and Selected Cities. Mexico's Northern Border, 2000 and 2010 (weighted values) 

Source: Own calculations based on microdata from population censuses (^{INEGI, 2001}, ²⁰¹¹).

Similarly, the higher intensity of crossborder employment continued in the west, with respect to the east, as other studies have found (^{Alegría, 2002}; ^{Coubès, 2008}). In the west, the workers cross to work in the large urban areas of Southern California (San Diego and Los Angeles) and its agricultural areas (Imperial Valley) and those of Arizona (Yuma and Maricopa Valley). However, in the west is also where the drop in cross-border workers was the greatest; the drop was about 50 percent in some municipalities such as Mexicali and San Luis Río Colorado.

This decline in cross-border workers was registered in all age groups, although it was more pronounced after age 35; in 2010, the percentage of cross-border workers aged 35-39 fell to less than half of what it was in 2000 (Graph 1). In 2000, 15 years after the Simpson-Rodino Act (IRCA, 1986), all the age groups that were over 20 years old in 1986 (beginning with age 35 in 2000) had employment frequencies in the United States above five percent. Nevertheless, in 2010, the regularization of 1986 was a distant event; workers under age 44 were too young to have obtained papers back then. For the age groups between ages 30 and 44, obtaining documents to work in the United States, unless they had obtained citizenship by birth, was more difficult. Therefore, in these age groups the rates of cross-border employment plummeted.

Source: Own calculations based on the microdata of population censuses (^{INEGI, 2001}, ²⁰¹¹).

Graph 1 Percentage of the Employed Workforce that Was Working in the United States. Mexico's Northern Border, 2000 and 2010 

In 2010, the profile of the cross-border workers, compared with border workers in Mexico, continued to be predominantly masculine, older, and comprised of heads of household, although in 2010 these differences were less marked than in 2000 (Table 2).

The characteristics that were more reinforced during this period were tied to the relationship with the United States: the proportions of those born in the United States and of those who returned recently from that country were greater among cross-border workers than among those who work in Mexico; in contrast, there was a declining trend of cross-border workers born in non-border states and, most notably, who were recent internal migrants to the border. This shows that crossing to work in the United States is a border strategy that national migrants are using less, probably because of their lower insertion into networks that would allow their entry into the labor market in the United States, and their lower access to migratory documentation (citizenship, particularly).

Another profile that was accentuated in 2010, compared with 2000, is the greater educational level of the cross-border workers compared with the population that works in Mexico. In 2010, the educational gap between Mexicans who worked in the United States and those who worked in Mexico widened compared with 2000; for example, Mexicans with high school and above (12 or more years of schooling) represented 48.8 percent of the Mexican workers in the United States but only 35.3 percent of those who worked in Mexico, while in 2000 these population segments were 32 and 26 percent, respectively.

Also, the sectors of economic activity in the United States changed between 2000 and 2010. In this period, among the commuters there was a reduction in the primary sector participation from 19 to 10 percent. This drop coincided with a reduction in laborers from 10 to 5 percent, which could have occurred because of the sharp drop in cross-border workers in Mexicali and San Luis Río Colorado, where cross-border agricultural workers have traditionally come from (^{Acuña, 1983}; ^{Estrella, 1994}). On the other hand, there was also a slight but significant greater increase of the commuters in the tertiary sector compared with that of the workers employed in Mexico. Finally, employment in manufacturing industry contracted, but this occurred on both sides of the border, and the greater participation of the cross-border workers in construction work remained stable, in comparison with the border workers employed on the Mexican side.

Despite these changes in the activity sectors and occupational positions, the difference in wages between U.S. and Mexico borders continued to be as important in 2010 as in 2000: wages were three times greater for cross-border workers in comparison with border workers employed in Mexico (see the average in Table 2).

Regarding the drop in commuters due to migration (^{Orraca, 2015}), this hypothesis was difficult to link to the results obtained. Those with a higher level of education would have had greater economic resources to migrate, but the proportion of less-educated cross-border workers fell. In addition, those who were older had greater access to legal documents to migrate, but the reduction in the number of commuters of intermediate ages was greater.

The changes in the intensity of the cross-border workers and their profiles between 2000 and 2010 were related to the transformation of the border to a less porous one. The results indicate that the lower cross-border mobility affected the less-educated workers and those in the agricultural sector; in other words, cross-border labor mobility was less a border practice (to which migrants could integrate into over time) and more a social practice that involves people with higher educational levels and more binational ties (by birth or previous migration).

The levels and profiles of the children born in the United States

Table 3 shows the percentage of children born in the United States as a result of a border crossing, in 2000 and 2010 and the percentage change. The trend of such births was on the rise. In 2000, 5.9 percent of the children aged 0-4 years were born in the United States to mothers who were residents of Mexico, while in 2010 this population grew to 9 percent; that is, during the last decade, this type of births rose 51 percent.

Table 3 Population Aged 0-4 Born in the United States to Mothers Residing in Mexico Five Years before the Census. Total and Selected Cities. Mexico's Northern Border, 2000 and 2010 (weighted values) 

Source: Own calculations based on the microdata of population censuses (^{INEGI, 2001}, ²⁰¹¹).

The increase of children born in the United States (Table 3) stands out in Ciudad Acuña, Mexicali, Piedras Negras, and Tijuana. Similarly, the higher proportion of children born in the United States who live in San Luis Río Colorado is noteworthy. It is also the municipality with the highest percentage of cross-border workers (Table 1), which reveals the importance of family border life strategies in this municipality. The only municipality where the percentage of children dropped was Nuevo Laredo, which could be due to the emigration of residents who had cross-border networks because of the violence the city experienced.

Graph 2 illustrates the percentage of children born in the United States as a result of a border crossing as the difference between the total of children born in this country and those whose mothers had recently emigrated from the United States, for each age group. This figure shows that although return migration contributed to the recent increase in children born in the United States who live in the northern border of Mexico, the majority of the children have mothers who were already residents. Between 2000 and 2010, the proportion of children aged 0-4 born in the United States augmented from 6.7 to 10.5 percent, while those born to mothers who came from the United States increased from 0.75 to 1.51 percent.

Source: Own calculations based on the microdata of population censuses (^{INEGI, 2001}, ²⁰¹¹).

Graph 2 Percentage of the Population Born in the United States by Age Group. Total and Recent Immigrants. Mexico's Northern Border, 2000 and 2010 

In 2010, the percentage of children born in the United States living in Mexico's northern border region was much greater in the age groups 0-4 and 5-9 than in older age groups (Graph 2). This confirms the increase in the number of births in the United States in the last decade and not only in the previous five-year period. Also noteworthy is the decrease through time in the population born in the United States of various age groups. For example, children aged 0-4 born in the United States, who were 6.7 percent in 2000, 10 years later, when they were aged 10-14, only represented 5.2 percent. This pattern was consistent in the next age groups, although less abruptly at older ages, which may be associated with the combined effect of the emigration of this population segment and the immigration of children born in Mexico to the border northern region.

The profiles of the parents of the population aged 0-4 allows us to understand what type of population opts to give birth to their children in the United States (Table 4). Both in 2000 and 2010, the U.S. born children tended to live in households where one of the parents was also born in this country or worked there. It is worth noting that the birth of the parents in this country does not mean that they are non-Hispanic whites, as the birth of Mexican children in that nation has been a border family strategy that goes beyond the present generation (^{Ojeda, 1994}; ^{Cueva and Vásquez, 2009}). In contrast, U.S. born children tended to live less in households with parents who had recently moved from the interior of Mexico, compared to the children born in Mexico. This indicates that the parents having children in the United States have a longer period of residence at the border.

Table 4 Sociodemographic Characteristics of the Children Aged 0-4 to Mothers Residing in Mexico Five Years before the Census, According to Place of Birth. Northern Border of Mexico (weighted values) 

^a See Note 6 about how the information on parents and children was determined for 2000 and 2010.

^b Deflated income based on the national consumer price index in June 2002.

^c Television, refrigerator, washer, telephone, automobile, and computer.

Significant differences in percentages or averages in each census year (+ p < 0.05). Significant differences between the years 2000-2010 (^*** p < 0.001, ^* p < 0.01, ^* p < 0.05, NS = p > 0.5) (see methodology section).

Source: Own calculations based on the microdata of population censuses (^{INEGI, 2001}, ²⁰¹¹).

With regard to the socioeconomic profile by children's place of birth, as argued previously, the children born in the United States to border residents live in better social and economic conditions than those born in Mexico. The former have more educated mothers, and, on average, they have higher income, use private medical services and possess more durable goods.

The transborder characteristics of those born in the U.S. changed between 2000 and 2010. During this decade, the percentages of the children whose parents work in the United States fell, being the decline much greater among children born in United States. This characteristic dropped from 27 to 9.6 percent, while parental cross-border employment among those born in Mexico fell only from 3.8 to 1.8 percent. The number of children born in the United States with one of the parents' having been born in the United States remained stable. In addition, the percentage of parents who moved from the Mexican interior fell among the children who were born in Mexico, while this rate remained constant among those born in the United States.

In 2010, the phenomenon of giving birth in the United States was more selective in terms of human capital than in 2000, now mothers had more schooling. In general, there was an increase in educational levels at the border. Nevertheless, the advantage in mother's schooling of those born in the United States was greater than of those born in Mexico. The mothers who had gone to university rose among U.S. born children, compared with those born in Mexico.

Regarding the change in the use of medical services between 2000 and 2010, among U.S. born children, there was a greater increase in the users of the Mexican social security institutions⁸ and, above all, in the users of "other type" of services, compared with those born in Mexico. Among U.S. born children, the greater use of social security services could indicate that the practice of giving birth in the United States has increased among formal employees, and the greater use of "other" types of services could be tied to the demand for health services in the United States. The mothers of children born in the United States have greater formal education than before and possibly greater economic resources or knowledge about how to use the health services and health insurance plans in that country.⁹ In contrast, the use of medical services of social assistance-such as those provided by the Ministry of Health and Assistance- for impoverished populations increased among children born in Mexico, which is associated with the emergence of Seguro Popular (Popular Insurance) during this decade. These disparities exhibit the existing socioeconomic gap between parents with cross-border practices and those who do not perform them.

This large gap between children born in the United States and those born in Mexico is also observed in the differences of socioeconomic levels of households in 2000 and 2010. U.S. born children live in households with higher income and goods, profile that remained stable during the studied period. Although the differences between one year and the next were not very significant, the effect of the economic crisis in both populations is noticeable, since the mean deflated income dropped in both populations.

Conclusions

The objective of this article was to study the changes in the levels of two cross-border links: employment and giving birth in the United States among the northern border residents of Mexico during the first decade of the 21^st century, as well as the profile of the populations involved in these processes. To achieve these goals, the census samples of the Mexican population (^{INEGI, 2001}, ²⁰¹¹) were analyzed by means of descriptive statistics and difference tests.

A finding of this study is that there effectively was a change in the intensity of both indicators of transborder links but in opposite directions. While the proportion of commuters fell, more cross-border births were recorded. Despite this apparent contradiction, we hypothesize that this could be due to the greater reinforcement of the border and migratory control in the United States together with the socioeconomic crisis in Mexico's northern border region.

In terms of commuters, the results show that the workers of intermediate ages are mainly those whose labor participation fell in the United States; it is possible that a lack of work documents was the main explanation for this reduction. Workers aged 44 and older could have benefited from the regularization of their migratory status through IRCA (^{Congress of the United States of America, 1986}) and not have been affected by the greater control of undocumented migration beginning in 2001. However, given that the number of commuters fell in all age groups, it is also possible that the crisis had a negative effect in the labor demand in the south of the United States, and the tougher border control favored the emigration of some cross-border workers to this country.

In addition, the child births in the United States of border residents increased between the periods 1995-2000 and 2005-2010. The analysis by age of U.S. born population exhibited that while an increase in this practice had been taking place since the 1980s, there was a higher increase in the 2000-2005 period, which also coincided with the greater blockade of the border and migratory control after 2001.

The practice of multiple citizenship has grown in other contexts of the globalized world as a response to the paradox of the opening of international markets to merchandise and capital together with the closing of borders to labor immigration (^{Mateos, 2015}). This phenomenon also applies to the Mexico-United States border and was reinforced by migratory control in recent years. The aspirations of the border societies to integrate with U.S. society could have grown, since a strong north-south inequality in economic and public safety terms continue to exist and undocumented migration to the United States has become more and more difficult. The smaller prevalence of cross-border workers confirms this greater difficulty of access to the U.S. labor market. In this context, citizenship "from birth" seems to be a very important route to assure this access.

This study also confirms that the populations engaged in cross-border links at the Mexico-U.S. border were more exposed to processes of socioeconomic selectivity from 2000 to 2010, but less exposed to transborder capital selectivity. Cross-border workers tend to have more high school or university studies, as well as to have been born in the United States and/or to have lived in that country. In the case of children born in the United States, they also increasingly tend to have more educated mothers with privileged access to health services, but fewer parents employed in the United States.

In this sense, it can be concluded that in light of the reinforcement of the border and migratory control, the border between Mexico and the United States has become more and more a "selective barrier" for the interactions that result from border crossings (^{Alegría, 1992}), now that only certain privileged populations in terms of human capital and U.S. citizenship can benefit from the asymmetries between both countries. The nature of the border interactions changed from 2000 to 2010, these became less dependent on the border location of the actors and cross-border social networks, and more on the position of the populations in the social structure and transnational links. Now the border sustains greater differences in terms of social stratification, as has been found in other studies (^{Velasco and Contreras, 2014}), where the population that has university studies is more privileged.

This article has some limitations relative to the sources of information. While the use of the population censuses allows analyzing border municipalities that cover the studied zone,¹⁰ the variables on employment status are less precise than those obtained by a longitudinal employment survey. A panel-type employment survey with representation at the municipal level could have allowed studying the labor conditions and place of work over time, and, therefore, tracking the reasons for the abandonment of cross-border work. Another disadvantage of this study was not being able to know whether the population had documents to work or live in the United States; having this information would have allowed us to examine the strategies of cross-border life of the workers, as well as of multiple citizenship.

The findings indicate that 2001 was a "breaking point" for cross-border interactions due to border reinforcement and migratory control, since their tendencies were modified in the long term. The cross-border interactions of the populations have depended on the porosity of the border and the cross-border networks in a context of adjacency of asymmetric economic structures and policies. In light of this new scenario, networks and contiguity remained, but the possibility of working or having children on the U.S. side depended less on them and more on the socioeconomic capital of the populations of Mexico's northern border; that is, these results reflect that in the last decade, the impact of border interactions on the population settled in Mexico's northern border region has changed. They passed from being a mechanism of social mobility to being a force that exacerbates social polarization.