Will the wall work? Patterns and driving forces of cross-border labor migration between Mexico and the USA

The border between Mexico and the USA is of paramount importance to the analysis of cross-border migration. Approximately 6 million trucks, 2.3 million passengers in buses, 141 million passenger cars and 42.2 million pedestrians were registered in almost fifty legal border crossings in 2016. Similarly, hundreds of thousands of unregistered crossings of refugees and migrant workers are reported annually.Footnote 1 In spite of all statements on an already existing, to-be-built, or under construction border wall between the USA and Mexico, this border seems highly permeable. Cross-border labor migration between Mexico and the USA amounts to at least half a million Mexican nationals yearly (temporary visa, mainly H1B, H2A, long-term emigrants, non-regular resident).Footnote 2 In addition to the size of flows, the number of Mexicans living in the USA is significant. Taking only the number of Mexico-born persons working in 2015 in the USA (some 7.7 million), this accounts for about 3% of the total number of employees in the USA (around 256 million) and for about 15% of Mexico’s employed population (around 52 million).Footnote 3

While acknowledging the importance of migration between Mexico and the USA, a crucial question to raise relates to the main influencing factors. Three key drivers could be distinguished: (restrictive or permissive) migration policy and regime, demographic and labor market developments, and transnational social spaces. Since the 2007 financial crisis, there was a (temporary) decrease of cross-border migration and of remittances from the USA to Mexico. Policies of the federal US-government and of several states could suggest that the significance of Mexico-USA migration will decline. One could argue that border control is more and more restrictive, and that due to economic growth and demographic changes in Mexico, the pressure for emigration to the USA is slowing down.

The aim of this paper is to examine the development of Mexico-USA migration (volume and types, remittances) considering the impact of economic, political and social driving forces. I will argue that all of the following three factors influence migration and remittances dynamics: the regulative framing of the migration regime, the demographic and labor market conditions, and the institutionalized transnational social spaces. Since the quality of available data does not allow for detailed regression analysis (Salas, Loría-Díaz de Guzmán, & Díaz, 2016, p. 186), the guiding assumption will be evaluated according to the procedure of stepwise exclusion of singular explanations.

Firstly, some remarks on social science labor market theory, on migration theory and on transnationalism research (Findings from labor market, migration and transnationalism research section) as well as on the development of the USA-Mexico migration regime (Development of the bi-national migration regime Mexico-USA section) will be outlined. Secondly, empirical findings from aggregate data and estimates of Mexican-US-American labor migration dynamics since the 1950s will be discussed taking into account the migration regime and the labor market (Aggregate data and estimates of migration 1940s to 2015 section). Based on the analysis of individual cross-border mobility and remittances dynamics, the factor of transnational social spaces will also be analysed (Individual cross border mobility and remittances section). Then conclusions will be drawn (Conclusions section).

Findings from labor market, migration and transnationalism research

Classic theories on international migration focus on the economic, demographic, political and socio-cultural conditions, forms and consequences of cross border mobility (Castles, de Haas, & Miller, 2014, chapter 2). The focus is on emigration and immigration, as well as on seasonal or circular migration. Important findings of this research are: (1) migrants normally are not from the poorest social strati (because migration itself requires a minimum of resources), (2) migration is embedded in social networks, (3) independently of simple push-pull-factors, migration chains follow historically grown relations between different countries and regions. During the twentieth century, the dominant focus to explain labor mobility was a rational-choice model of local and national push-pull factors. Existing transnational social relations and institutionalized orders as political regimes were not part of the explanatory model (Massey et al., 1994; for a broader approach of intervening actor groups in migration and the interrelation between formal authorization and social recognition of migrants see Ambrosini, 2018, pp. 16f, 21f). From the 1950s onwards, a social science labor market theory developed especially in the USA and Western Europe. In explicit critique of the dominant economic and individualistic-rationalistic model a main argument was that labor mobility – whether in the national or international context – is not only determined by push-pull-mechanisms of aggregated scarcity, but also by institutionalized social orders of internal and external, primary and secondary markets (Doeringer & Piore, 1971; Kerr, 1954).

Since the 1980s, the links between social science labor market theory and migration theories strengthened, especially in an international and comparative perspective. It could then be explained why labor force scarcity in the secondary labor market segment of highly industrialized countries – under given demographic conditions – must not necessarily lead to higher wages, but can also be answered by the (temporary) recruitment of labor migrants from other countries (cf. Papademetriou & Martin, 1991; Piore, 1979). The theory of the “new economics of labor migration” argued that migration decisions are primarily taken in frameworks of families and households (Stark, 1984, 1991). Decisions on cross-border labor migration therefore serve less individual benefit maximization than risk diversification of household income: “the migration of a family member, as a means of diversifying the family’s income portfolio, could reduce the overall risk associated with the generation of that income” (Stark, 1984, p. 207). More recent research also included the analysis of cross-border intra-organizational mobility like that of expatriates and inpatriates (e.g. Adick, Gandlgruber, Maletzky, & Pries, 2015).

Besides labor market and migration theories, formal legal regulations and policies are systematically included in concepts of migration regimes that can be understood as the historically grown migration-related principles, norms, rules and decision-making procedures.Footnote 4 It includes the specific individual and collective value orientations and standards, laws and regulations, the structures of collective actors as well as the practical policies and procedures relating to the control of migration as immigration and emigration. Moreover the (including or excluding) treatment of migrants living in a given country, e.g. in the form of assimilation or integration strategies, which in particular concern legal access to labor markets, should be considered. The dominant focus is on nationally and bi-nationally negotiated migration regimes. Migrant organisations are an important part of binational migration regimes (Pries & Sezgin, 2012, p. 20ff).

Since the 1990s, migration studies argued that not only the local and national economic and political conditions in countries of origin and of arrival are relevant, but that the cross-border mobility and communication of people, money and other resources can create more or less dense and stable transnational social spaces. These represent a genuine social-institutional factor that influences migration dynamics and is relatively independent of economic opportunities and political-legal frameworks. For the Mexican-USA-migration Massey, Alarcon, Durand, and Gonzalez (1987) show that cross-border social networks improve potential migrants’ knowledge about border control, smugglers, and labor opportunities in the USA. This can lower the risks and costs of migration, and the stock as well as historical tradition of migration increases opportunities for social capital and networks (Massey & Espinosa, 1997; for the European context see Ambrosini, 2018, p. 7; Faist, Fauser, & Reisenauer, 2013; Helbling & Leblang, 2019, pp. 253, 260f; Jacobson & Goodwin-White, 2018). In this context, transnational migration or transnationality of migration can be understood as a subtype of cross-border international migration based on the multi-directionality of mobility and the durability of strong cross-border social entanglements (Guarnizo, 2003; Khagram & Levitt, 2007).

Whenever there are dense and permanent institutionalized exchanges between local and regional levels in different countries that develop independently (or relatively independently) from the relevant national migration regimes, transnational migration regimes can consolidate. For the USA-Mexican case, Besserer (2002) revealed permanent and dense economic, political, cultural and social interaction structures between municipalities in rural areas of Mexico and district-related communities in the USA. These include established structures of transnational political decision making and power (for example, with regard to organizing village festivals or the taking over of local administration tasks). As shown in many studies (e.g. Besserer, 2002, 2016; Gil Martínez de Escobar, 2006), transnational migration regimes also comprise cultural, economic, religious-ritual and educational aspects.

Taking into account this background of conceptual approaches to the labor market, migration and transnationalism, some empirical evidence concerning cross-border labor mobility between Mexico and the USA can be presented as state of the art of research. Firstly, initial migration often leads to cumulative migration and building of social networks that reduce the risks and costs of migration (e.g., by changing expectations in the regions of origin and new migration-related demand structures in the arrival regions; Massey et al., 1987; Massey & Espinosa, 1997). Secondly, given long-term cross-border migration the social spaces in the regions of origin and arrival are increasingly intertwined with one another through complex processes. This could create new, transnational social spaces, e.g. by the withdrawal and return of qualified people (brain circulation), the reimbursement of funds, changes in economic expectations, new political and gender-specific entitlements, wage competition and labor market displacement (Aguilera & Massey, 2003; Basch, Glick-Schiller, & Blanc, 1997; Bryceson & Vuorela, 2002; Goldring, 1997; Goldring, 2001; Kearney & Nagengast, 1989; Palloni, Massey, Ceballos, Espinosa, & Spittel, 2001). Thirdly, international migration processes can be controlled and managed only to a limited extent, given that they follow collective logics in local, national and transnational social spaces. Border control is limited in social practice; measures of strengthening it often have counter-intended effects (Ambrosini, 2018, p. 11f). Increased negative control or closure of borders lead to higher fares for smugglers and life-risks for irregular migrants. It could fuel the unintended “caging” effect: legal and irregular resident migrant workers, who would move in and out in accordance with the employment opportunities under conditions of free mobility, remain in the countries of arrival.Footnote 5 Fourthly, in the face of growing environmental challenges, regional violent conflicts, growing local poverty, the boundaries between (voluntary) labor migration and (forced) refugee migration are increasingly fading. Also due to new transportation and communication technologies, the local, national, transnational and global level of migration processes are increasingly interwoven (Castles et al., 2014, p. 39f; Portes, 1996; UNHCR, 2007, 2011).

The review of conceptual and empirical findings suggests that all three drivers (migration regime, labor market, and transnational social spaces) should be taken into account (for similar arguments related to migration in Europe see Ambrosini, 2018; Baumann, Lorenz, & Rosenow, 2011; de Haas et al., 2018). Therefore, the development of the Mexico-USA migration regime will be addressed first in the following section.

Development of the bi-national migration regime Mexico-USA

Scholars (Delgado Wise & Márquez Covarrubias, 2009, p. 48; Durand, 2005, p. 48) divide the development of the Mexico-USA migration regime into five periods. The first phase dates up to the 1930s and it is characterized by a restrictive US-American immigration policy towards Mexico; country in which migration was not a policy issue. Starting from 1821, the year in which Mexico gained formal independence from Spanish colonial power, until 1848 almost all of today’s southern states of the USA belonged to the United States of Mexico. After the Mexican-American War of 1846–48, the USA began to introduce sharper border controls (Griswold del Castillo, 1990). However, due to past events, the immense extent and geographical conditions along the border, a flourishing informal trade, and mechanisms for irregular border crossings and professional smuggling services soon emerged and established (Andreas, 2009, p. 30). Migration policy was not explicitly issued, but put into practice e.g. in the deportation of Mexican workers from the USA after the Great Depression of 1929.

A second phase opened with the Bracero program (1942 to 1964) aiming at substituting (male) US-American workers (who left for military services during World War II) by mainly low-skilled Mexicans to work in the USA for a limited time. The agreement was continued after World War II, also following pressure from employers, who continued to need cheap labor for harvesting or construction work (Trigueros, 2009). This bilateral program represents a turning point in the migration policy for both countries. Mexico succeeded in negotiating a number of treaties to protect workers (aiming at non-discriminatory management, adequate working conditions and fair wages). The program allowed about five million Mexican migrants to work legally and with temporary employment contracts in the USA. Most of them came from the states of Jalisco, Michoacán, Guanajuato and Zacatecas (Durand & Massey, 2003), while Texas, California, New Mexico, Arizona and Arkansas were among the preferred destinations for agricultural workers in the 1950s (Trigueros, 2009, p. 64). During the Bracero program – similarly to the European “guest workers” programs – migration was dominated by male workers from rural regions, that migrated in a circular way with fixed-term employment contracts (Torre & Giorguli, 2015, p. 14).

At the same time, the number of Mexican job seekers and of available jobs in the USA exceeded the number of places offered by the Bracero program. This led to an increase in unregistered labor mobility. It is calculated that, during the 22-year term of the Bracero Agreement, a total of a further five million working migrants were employed under irregular conditions in the US additionally to the regular workers employed in this program (Durand & Massey, 2003). The negotiations on the renewal of the agreement were accompanied by heated public debates in both countries. In the USA trade unions, political associations and a part of the press spoke out against the Mexican immigrants.Footnote 6 Small businesses and farmers criticised the negative impact of the (poor) working conditions of immigrants on local workers (in the sense of increased pressure to downgrade) and stated to have a critical disadvantage over their larger competitors, as the Bracero program provided these cheap Mexican workers. The Kennedy government did not extend the program in 1964.

A third phase dates from 1964 to the Immigration Reform and Control Act (IRCA) 1986, characterised by a ‘policy of non-migration policy’. After the end of the Bracero program, the number of irregular border crossings increased substantially. Trigueros (2009, p. 67) states that “the Immigration and Nationality Act of 1965 for the first time limited the number of permits for Mexican immigrants to 66,000 per year.” Despite these constraints, the need for cheap labor was particularly high in the agricultural sector. Between 1970 and 1980, the number of Mexican irregular immigrants rose fivefold: “net undocumented immigrants into the United States increased from an estimated 23,000 annually by 1970 to 112,000 annually by 1980” (Vernez & Ronfeldt, 1991, p. 1189). Likewise, based on the demographic development (the population grew from 41 million in 1965 to 79 million in 1985; cf. Ruiz Nápoles & Ordaz Díaz, 2011, pp., 92–95), the number of job seekers in Mexico rose sharply after the “Mexican economic miracle” (1940–1970).

Although the formal-legal framework and the “policy of non-policy” did not provide for an intensification of cross-border occupational mobility, migratory movements, especially irregular migratory flows, continued to increase. During this period, the number of women and migrants from the central and northern states as well as from cities increased in importance. During this period, activities of Mexican workers with temporary contracts in the USA shifted from the primary to secondary and tertiary sectors (Trigueros, 2009, p. 67). Also, not only governmental agencies of both states were relevant as collective actors, but also networks and organizations for the defense of migrant interests, like the Chicano civil rights movement or trade union activists such as César Chávez.

A fourth phase began with the IRCA of 1986 that aimed at controlling the situation of irregular residents by means of broad legalization and at curbing the influx of irregular migrants (in particular through the expansion of border controls, the issuance of more temporary work visas and the threat of penalties for companies employing irregular workers; cf. Trigueros, 2009, p. 68). However, these objectives were only partially achieved. IRCA strengthened transnational social networks (including of irregular immigrants), as about three million people were given permanent residence permits (Torre & Giorguli, 2015, p. 15). The proposed sanctions against irregular migrant workers were limited and differently imposed (Durand, Massey, & Parrado, 1999, p. 522). The law improved conditions for legalized labor migrants and, worsened the employment situation for irregular workers. It fostered a caging effect, and family reunion became the quantitatively most important part of legal immigration from Mexico to the USA (Alarcón, 2011; Durand et al., 1999; Giorguli Saucedo, García Guerrero, & Masferrer, 2016).

This fourth phase extends to the turn of the century and can be characterized by tightened border controls and an expansion of legal labor migration opportunities. The Immigration Act (IMMACT90) of 1990 enabled immigration of skilled and specialized workers, and the number of annual work visas rose from about 54,000 to a total of 140,000 at the end of the 1990s with an increased share of H-1A and H-1B visas for qualified workers and academics (cf. Trigueros, 2009, p. 71). Massey, Pren, and Durand (2009) referred to this period as the “era of the contradiction” because “on the one hand the checks for the transit of goods, services, capital, information and certain categories of persons (professional, highly qualified technicians, employees of transnational companies) had been liberalized and new barriers and obstacles to the mobility of workers between the two countries had been established” (Massey, et al., 2009, p. 102).

Government agencies of the USA hoped to mitigate the problem of (stocks of) irregular migrants in the USA and (flows of) irregular migrants to the USA. In fact, tightened border controls were introduced. This reduced the circularity of irregular migration in terms of frequent changes between the countries, due to increasing risks and costs of irregular border crossings. Nevertheless, the volume of irregular migration was not curbed to the expected extent. The border between the two countries was too long and too difficult to control, and the number of jobseekers in Mexico continued to grow strongly (by the end of the millennium only around 12 million of the potential of almost 45 million persons were employed in formal, social insurance-based jobs). A differentiated service industry developed for irregular border crossings, which included a wide range from simply crossing the Rio Grande (as “mojado”) to a complete trip from the place of origin in Mexico, to the destination in the USA including a workplace.

The fifth and last phase of the development of the USA-Mexico migration regime starts with the new century, marked by the lack of a common migration agenda and increasing tensions concerning migration policies between the two countries. Further, organised violence generated more forced migration from Mexico. In the year 2000 in Mexico, for the first time in history, a government not appointed by the old state party PRI took over official business. One of the hopes associated with the new liberal-conservative party PAN (Partido de Acción Nacional) government was to negotiate more liberal migration provisions. However, the presidential election of George W. Bush in November 2000 and the terrorist attacks of September 2001 led to a wave of securitization in spite of liberalization. This holds for the two Bush term periods (2000–2008) and also for the subsequent government of Barack Obama (2008–2016).

No substantial bilateral agreements were signed; only in March 2005, the Presidents of Canada, Mexico and the United States signed the joint program “Security and Prosperity Partnership of North America (SPP)” as a regional dialogue initiative, which included far-reaching objectives e.g. the introduction of a common North American border Pass (Independent Task Force, Council on Foreign Relations [ITF], 2005, p. 8f). But this program was discontinued in August 2009, without effective joint mechanisms being agreed upon in migration questions.Footnote 7 Unilateral measures dominated. Massey et al. (2009) described this period as one of marginalisation, “in which Mexican migrants were forced to abandon their connections to their communities of origin because of the militarization of the border, which made it impossible to return, and in which they simultaneously feel themselves as strangers in a society that requires and demands them, but at the same time does not accept them, discriminates and oppresses them” (Massey et al., 2009, p. 102).

A crucial question that arises is to what extent the binational migration policy and regime actually determined or influenced the real cross-border migratory movements between Mexico and the USA (for a global and a European evaluation of this question see e.g. de Haas et al., 2018, p. 28f; Helbling & Leblang, 2019, p. 257f). Since there are no solid data available for the first phase, we will concentrate on the period since the 1940s.

Aggregate data and estimates of migration 1940s to 2015

Inhabitants with Mexican roots represent the largest immigrant group in the USA (Alarcón, 2011, p. 186). The intense economic, political, cultural and social exchange relations between the two countries go far back in time, before both nation states existed in today’s form. Indigenous communities on both sides of the border are still relying on a pre-colonial common history and ancestors (Castillo Ramírez, 2012). Names such as Los Angeles, San Francisco, and New Mexico point to the centuries-old colonial order that for more generations than today’s binational structure determined the life of today’s North America as “Nueva España” (Balderama & Rodríguez, 2006). So how did the volume of regular immigration from Mexico to the USA actually develop since the 1940s?

A first approximation of the migratory movements from Mexico to the USA is the number of immigrants per country of origin registered annually, as listed in the Yearbooks of immigration statistics (Fig. 1). There is a slight increase in documented immigration in the post-war period, which however reduces by the end of the 1980s to less than one hundred thousand annual immigrations (each of which is subject to different visa types). The enormous increase from 1989 to 1991 can be explained by the fact that the IRCA legislation allowed for the legalization of Mexicans, who resided irregularly in the USA (cf. Statistical Yearbook, 1989, p. 18). From 2000 to the year 2014, legal immigration moved to an average level of 160,000. The figure only represents registered new entries of Mexicans into the USA (including temporarily limited visa and permanent visa, the latter mainly for family reunification; see Trigueros, 2009, p. 62).

Fig. 1
figure 1

Annual registered immigrants from Mexico into the USA 1940–2014

The number of registered immigrants from Mexico to the USA (1949–2015) is slightly positively correlated with unemployment rate in the USA (Pearson = 0,12) and negatively correlated with unemployment rate in Mexico (Pearson = − 0,46).Footnote 8 The opposite correlation were to be expected from classic labor market and migration theory (the lower unemployment in country of arrival, the higher immigration, and: the lower unemployment in country of origin, the lower emigration). We could assume that official registered migration only reflects a part of the overall dynamics. These have their own momentum – beyond macro-economic indicators, and across the timeframes of the migration regime. There is only a strong effect of IRCA 1986 leading to a substantial increase of immigration (by regularization) during the years to follow.Footnote 9 The yearly regular immigration from Mexico to the USA oscillated between some 50 and 100 thousand until IRCA and 100 and 200 thousand thereafter. This correlates neither with the four periods of migration regime identified earlier nor with the macro data of labor market dynamics between both countries. An enormous expansion of the working population in both countries goes hand in hand with a comparatively constant registered labor migration from Mexico to the USA.

Flows of immigration and emigration are reflected in the stock of residents with Mexican citizenship at birth living in the USA. According to estimates, the number of Mexico-born and naturalized persons, Mexicans with permanent residence permit, certain registered non-immigrants (such as students, refugees) and irregular residents in the USA, raised constantly from less than one million in 1970 to almost 12 million immigrants in 2015 (Migration Polity Institute [MPI]), 2016). Counting all residents in the USA with regular residence status and Mexican roots (including the born-citizenship of at least one parent), the American Community Survey estimates even 31 million people in 2010.Footnote 10 Since around 2006, the number of Mexico-born immigrants stagnated at just under 12 million.

Recalling that the level of regular immigration from Mexico to the USA was below 100,000 per year until the 1980s, and then on average around 150,000 (Fig. 1), the strong decline of the Mexican-born population in the USA since the 2000s – in relative and absolute terms – is surprising. According to some authors (Giorguli Saucedo et al., 2016, pp. 17, 19; Maldonado, Morales, González González, Crow, & Schavon, 2016, pp. 102f, 106ff), the sustained annual inflows of Mexicans have not substantially increased the stock of that group in the USA since then, given that returns from the USA to Mexico increased and that regular immigration from Mexico to the USA declined. This is explained by declining employment opportunities since the crisis in 2007 and a more negative or sometimes even racist climate for Mexicans.

Other scholars state a recovery in both the immigration movements from Mexico to the USA and the stock of Mexico-born residents in the USA since 2014 (Canales & Meza, 2016, pp. 81, 86) and/or an increasing importance of migration between Mexico and the USA with a shift from irregular to regular trips (Verduzco, personal communication, 2017, p. 2). One aspect associated with these contradictory diagnosis in migratory flows and stocks is the calculation of the number of Mexicans irregularly crossing the border and living in the USA. While there are precise and long-term series for registered immigration to the USA, for obvious reasons the extent of irregular (labor) migration can only be estimated, and such estimates are hardly comparable for long periods. Calculations are usually based on the numbers of Mexicans in the USA without valid residence papers, on the numbers of deportees registered in Mexico, and on comparisons of data from American surveys that record irregular residents, with the statistics of registered immigrants.Footnote 11 As illustrated in Fig. 2, the stock of irregular Mexican immigrants living in the USA is estimated at almost half the level of that of regular immigrants.Footnote 12

Fig. 2
figure 2

Stock of Mexicans in the USA without valid residence papers

We can assume that, since the financial crisis in 2007, the stock of Mexicans residing without legal permit in the USA has decreased (for about one million). At the same time, the willingness to cross the border between Mexico and the USA irregularly has declined in recent years (Maldonado et al., 2016, p. 104). Due to limited data it is difficult to calculate the role played by intensified border controls (migration regime), limited employment opportunities (labor market) and increasing racism in the USA (social factor). Based on survey data, several scholars agree that the willingness of adult Mexicans to migrate to the USA for labor has declined overall since around 2005 (Salas et al., 2016, p. 187). Salas et al. (2016) try to move forward from a simple cost-benefit model described by Massey et al. (1994).

Salas et al. (2016, p. 188) explain the decision to seek an irregular border crossing from Mexico to the USA as the net profit that a person expects over a calculated period of time with the following influencing variables: “probability of not being deported again”, “probability of getting a job in the country of arrival”, “pay in the country of arrival”, “probability of getting a job in the country of origin”, “wage in the country of origin” and “total cost of migration, including physical and psychological costs due to crime”. Since there are not sufficient data available for the latter, the authors only consider three variables: (1) the proportion of economically active population ready to migrate, (2) the number of deported Mexicans who were not registered as delinquent in the USA, and (3) the ratio of the USA unemployment rate compared to that of Mexico. The variable that most closely identifies the (irregular) cross-border prospect of Mexicans (PEAC) is the number of non-criminalized deported Mexicans: the greater the risk of being deported, the lower the declared intention to pass the border irregularly. Calculations also show that the ratio of the unemployment rates of both countries have an independent influence, and not only an influence on the willingness to emigrate. “The latter empirical finding suggests that U.S. and Mexican labor market exhibit a high degree of integration, and that migration works as a communication channel” (Salas et al., 2016, p. 194).

While Salas et al. (2016) use the number of deported Mexicans as an independent variable to explain the intention of irregular border crossings, in other studies the number of deportations is treated as a proxy for the number of irregular border crossings themselves. The Pew Research Center (2017) reported an increase in arrests on the border Mexico-USA from about 331,000 in the fiscal year 2015 (1.10.2014–30.9.2015) to almost 409,000 in the fiscal year 2016 (Pew Research Center, 2017).Footnote 13 There is evidence that arrests and expulsions due to irregular border crossings increased since 2014, and this is mainly caused by intensified controls, rather than an increase in non-registered border crossings. The cost of irregular border crossing multiplied in recent years (from about 1000 to over 3000 US dollars per person).Footnote 14 At the same time, controls on irregular employment in the USA have become more extensive. Many employers shift from the old model of employing irregular residents to a policy of increased visa application.

To sum up, stagnation of cross-border mobility since the economic crisis in 2007, argued by many authors, can be explained only partly by worsened employment prospects in the USA, since total employment in the USA has been clearly expanding since 2010 (Bureau of Labor Statistics, 2017). Changes in the migration regime (extended regular visa options and, simultaneously, more restrictive border controls) also fuel the stagnation of Mexicans, who are residing in the USA irregularly (Pew Research Center, 2016) or who are considering an irregular border crossing. These results are consistent with international studies showing that (restrictive) state migration policies do have an impact, but are shaped by factors like labor market opportunities and the bilateral stock of migrants (de Haas et al., 2018; Helbling & Leblang, 2019). The method described by Salas et al. (2016) to measure the influence of economic and social variables on the emigration of Mexicans suggests a greater influence of general social context variables (such as the (in)security feeling in Mexico) than i.e. changes in unemployment rates in both countries.

Based on the analysis of individual labor mobility, the following section focuses on transnational social factors as explaining the dynamics of labor mobility. It also relates migration development to remittances flows. Related to the debates about the decrease or increase of recent cross-border mobility I will argue that remittances from the USA to Mexico do not show a long-term stagnation trend. The development of remittances suggests that transnational social spaces are a strong factor in explaining labor mobility between the two countries.

Individual cross border mobility and remittances

Despite methodological challenges, the longitudinal migration data of the Mexican Migration Project (MMP) is one of the best internationally accessible sources to analyze transnational migration dynamics between Mexico and the USA based on individual and household data.Footnote 15 Data allow analyzing the frequency of migration trips between both countries. Figure 3 shows, for all heads of households interviewed between 1987 and 2015 (n = 8.252), the number of border crossings from Mexico to the USA referred to by the interviewed persons (of at least one month’s duration). The figure has to be interpreted cautiously because the data are ‘right-censored’, therefore, only persons with a first trip until 2010 are included.

Fig. 3
figure 3

Number of US trips per person according to the year of first labor migration

The figure illustrates the composition of the number of trips between Mexico and the USA according to the year of the first trip (grouped by persons who indicated one trip, two trips, three to five trips, six to ten trips, or eleven to twenty trips). The (four) periods of migration regime seem to have little influence on the number of trips by person. Only IRCA suggests having an effect given that the share of persons indicating 6–10 trips is decreasing substantially since 1986 (as year of the first trip; part of this finding probably has to be explained by the right-censored of data). The end of the Bracero program seems to have had little influence on the individual migration patterns.Footnote 16

Another survey of more than 600 transnational Mexican labor migrants from predominantly agrarian regions of the state of Puebla to the metropolitan area of New York revealed interesting results due to capturing retrospectively the total life and employment history of these interviewees. For instance, the survey allows measuring the patterns and reasons of shifting employment relations inside and between the two countries (Mexico and the USA) and inside and between economic sectors (Pries, 2004). The interviewees stated that they had switched to the USA for economic and employment reasons and from the USA to Mexico mainly for family reasons. Household considerations and other family members usually influenced all corresponding decisions. Opportunity structures for employment are transnationally aligned with family needs in both countries.

Concerning the specific occupational and sectoral positions during the labor trajectory (as a sequence of employment relations during the life course) some statistically significant patterns (not random with a probability of more than 99%) could be detected. Shifts of employment relations within Mexico, within the USA or between Mexico and the USA each follow certain specific patterns. Agricultural jobs are more frequent than (statistically) expected after a trip from the USA to Mexico. Vice versa, in case of a trip from Mexico to the USA an agricultural occupation was counted only half as often as expected (54 to 112). On the one hand this means that interviewees are shifting from agricultural employment in Mexico to predominantly industrial or service employment in the USA and, on their return to Mexico, above average, frequently engage in the agricultural sector. When a shift of employment occurs inside one of the two countries, either in Mexico or the USA, the new job – with a statistically significant probability – will not be in an agricultural occupation (only 102 inside-country job changes into agricultural activities, where 159 were expected). On the other hand, subsequent employment takes place significantly more often than expected inside the manual-industrial sector or inside areas of commerce, administration and hotels-restaurants, if previous employment occurred in the same country (Mexico or the USA) or in Mexico (that is, job shift is connected to a country shift from Mexico to the USA). When a previous employment took place in the USA and the following in Mexico, less than half of the expected number of job shifts went into the manual-industrial or the commercial/administrative/hotel-restaurant-sector.

Figure 4 displays the number of statistically significant employment shifts according to whether these occurred inside the same country or they were related to a shift between both countries, and to whether they occurred inside the same occupational sector or were combined with a shift between different sectors. The numbers indicate the standard residuals and could be read as the statistically significant correlation between changes of country and occupation. A value of zero would indicate no significant correlation between both, the possible shift of an occupation and of a country when changing from one employment to another. The circled arrows indicate employment shifts inside the same occupational sector. The linear arrows represent employment shifts that are combined with a shift between different occupations. The three ‘colors’ (white, grey and black) of the circled and linear arrows specify whether an employment shift was combined with no country shift (white arrows), with a country shift from Mexico to the USA (grey arrow) or with a country shift from the USA to Mexico (black arrow). Only statistically significant correlations (of more than 95% of probability of being non-random) are indicated.

Fig. 4
figure 4

Employment shifts according to change of country and occupation

As an example, for those interviewees that had been working in an occupation of the commercial, administrative, hotels-restaurants sector in Mexico there is a seven times higher propensity to work in the same occupational sector when they change employment and move from Mexico to the USA. Similar occupational sector stability stands for country shifts from Mexico to the USA in the case of manual-industrial work, technical and academic occupations. A very high sector stability is in place for all three kinds of country shifts (no shift, shift Mexico-USA, shift USA-Mexico) in the case of agricultural work and person related services. There are also significant (but with lower probabilities of 1.3 to 1.6) patterns for inter-sectoral change for all three kinds of country shifts.

The results suggest that labor market positioning and dynamics are embedded in three different units of reference: Mexico, the USA and transnational social spaces). Simple personal services, for example as a nanny or gardener, have a different social embedding in the USA than the same activities in Mexico. In the USA, this is usually a job paid on an hourly work basis not significantly different from other service work. In Mexico such activities are historically linked to a still existing paternalistic system of dependency and care relations. Relations between muchachas and patrones are less associated with modern wage-employment than with traditional paternalistic ownership and care. In a transnational social space, a child-care task can be conceived as a temporally limited moral family obligation being negotiated in the context of the larger family. Analysis based on person-related data suggest that labor mobility within and between the two countries is not fully explained by the migration regime and binational labor market mechanisms. Transnational labor mobility patterns and household logics also seem to influence and have their own inertia – independent of migration regimes and national labor markets.

Besides direct physical mobility, an important indicator of cross-border social relations between the USA and Mexico are remittances. This can be justified both within the framework of “new economic migration theory” and transnationalization research. The first suggests that decisions for migration are usually not taken by individual market players, but in the context of complex network structures, especially families and households. A crucial aim of migration is diversifying risks. Transnationalization research conceptualizes remittances not exclusively in an economic-rationalist way, but as an expression of strong transnational social ties: families themselves span transnationally; remittances then are an indicator of intensity of cross-border social relations and a factor in (re)structuring cross-border inequality.Footnote 17

Delgado Wise and Márquez Covarrubias (2009, p. 43) emphasize the NAFTA agreement of 1994 as a catalyst for the labor migration from Mexico to the USA, making remittances an important instrument for Mexico to obtain foreign exchange. Behind the more populous countries China and India (and the Philippines with an explicit labor migrants exporting strategy) Mexico is the country with the fourth largest volume of remittances (World Bank, 2016, p. 30). Based on data of the Bank of Mexico, Balderas (2009, p. 364) pointed out that remittances from the USA almost tripled from 1998 to 2004 (from 5.6 million to $15.4 million US dollars). They were on the same level with foreign exchange income from foreign direct investments, and higher than currency income from tourism.

Analyzing MMP data (71 communities in 13 Mexican states, over 4700 observations covering the period 1982 to 1999), Balderas (2009, p. 379f) finds that wage increases in the USA have a positive effect on the volume of remittances to Mexico. An increase in the hourly wages of the interviewed Mexican workers in the USA by one dollar leads to an increase of the average monthly remittance payment to Mexico by 16 US dollars. Mexican migrants with a bank account in the USA transfer significantly more money to Mexico (an average of 137 US dollars per month) than those without a bank account; those with children still living in Mexico transfer on average about 291 US dollars more each month than those without children in Mexico. Migrant workers without valid residence permits in the USA send about 87 US dollars more a month to Mexico than those with legal residency status.Footnote 18

As illustrated in Fig. 5, since the 1970s remittances from the USA to Mexico grew steadily up to the economic crisis of 2007/2008 (Banco de México, 1969, 2017). Some scholars assumed that due to restrictive border controls, a tense labor market situation in the USA and increasing racism towards Latinos and particularly Mexicans (especially since the 2015 presidential election campaign) transnational social ties would decrease. In fact, since 2009, the money flows have risen again and have already exceeded the pre-crisis level (see also Li Ng & Serrano, 2017, p. 134). Even if the actual volume of cross-border mobility and the stock of Mexican migrants in the USA would be stagnating during the last years (as discussed in Individual cross border mobility and remittances section), remittances flows indicate strong ongoing transnational social ties.

Fig. 5
figure 5

Revenues by Workers’ Remittances (quarterly amounts 1943–2017)

These transnational social areas play a significant role in labor mobility: “Macro-economic trends alone do not determine migration flows. A variety of other factors also influence the number of people moving from one place to another, and these can include everything from border controls in a receiving country to agricultural support policies in a sending country. An overarching factor is the operation of family networks that link migrants to relatives left behind. These networks can greatly facilitate new migration by providing access to housing, information about work opportunities and the comfort of familiar faces in a new land” (Passel & Suro, 2005, p. 11).


Previous research on the volume and patterns of labor migration between Mexico and the USA had suggested that these are influenced by labor market conditions, migration regime and transnational institutionalized social spaces. In light of the four commonly differentiated periods of the migration regime between Mexico and the USA since the 1940s, only for the beginning of the IRCA phase from 1986 onwards there was a clear impact on the registered migration dynamics between the two countries. Although IRCA aimed at reducing not registered migration, the volume of irregular border crossings from Mexico to the USA, and irregular stays of Mexican migrants in the USA rose for decades. According to Massey et al. (2009, p. 116), the likelihood of another irregular border crossing from Mexico to the USA continued to increase until 1998.Footnote 19

Contrary to predictions from classic migration theories, remittances from Mexican labor migrants did not tend to decline in the ‘maturing process’ of migration, but increased constantly, dropped with the crisis 2008 and then recuperated until 2017. This indicates that transnational social spaces are a genuine relevant factor. Three larger sets of explanatory factors have to be integrated: migration regimes, labor markets, and the institutionalized transnational social spaces.Footnote 20 Especially since the 2000’s, other societal aspects such as organized violence in Mexico have to be taken into account: “Now – and as an additional factor – we have rampant, international organized crime in Mexico that dramatically affects the undocumented migrant workers and should be considered a crucial and new psychological dissuasion factor” (Salas et al., 2016, pp. 194-196; see also Menjívar & Abrego, 2012). A question for further research is if in the future registered and irregular cross-border migration (and remittances) between the two countries will decrease (Giorguli Saucedo et al., 2016; Lowell, Villareal, & Passel, 2008, p. 8), stay at current levels or even increase (Canales & Meza, 2016, p. 85f). Strong cross-border ties and social spaces existing for hundreds of years and having stabilized during the last decades will probably maintain and flourish in the future – relatively independent of migration policies and regimes, and of labor market conjunctures.