Abstract
Data from the Mexican Census reveal that between 2005 and 2015, nearly two million migrants returned voluntarily to Mexico from the United States. Currently, high rates of voluntary-return migration to Mexico continue at the same time that migration flows to the U.S. steadily decline. This return migration trend presents serious challenges for Mexico, a country that has long struggled to satisfy the health care demands of its population. However, little is known about return migrants’ health care needs. In this study, we examine the health risk profiles and healthcare utilization for Mexican return migrants and the non-migrant population. We examine how these outcomes are affected by both the migration and return migration experience of the returnee population, while paying close attention to age-group differences. We employ inverse probability weighting regression adjustment (IPWRA) and logistic regression analysis of a sample of 348,450 respondents from the 2014 National Survey of Demographic Dynamics (ENADID) to test for differences in health conditions between those Mexican return migrants and non-migrants. We then turn to the Survey of Migration at Mexico’s Northern Border (EMIF Norte, for its Spanish acronym) for the 2014–2017 period to further assess whether certain characteristics linked to aging and the migration experience influence the prevalence of chronic health conditions, and health insurance coverage among 17,258 returned migrants. Findings reveal that compared to non-migrants, returnees are more likely to be physically impaired. These poor health outcomes are influenced by the migration and return migration experience and vary by age group and duration of residence, the time that has elapsed since returning to Mexico. We do not find an association between return migration and mental or emotional distress. Policy implications are discussed in light of immigration reform and restrictions on eligibility for health insurance coverage for older adults in Mexico.
Keywords: Return migration, Aging, Mexico, Health, Chronic health conditions, Health insurance coverage
Introduction
In the last decade, Mexico has experienced unprecedented levels of return migration, a trend likely to continue in the foreseeable future (Tim Henderson 2018). It has also been argued that with the growing Mexican economy and changes in U.S. immigration policies, the composition of those who are choosing to return home has profoundly changed (Chort et al. 2015; Masferrer and Roberts 2012). Among immigrants to the U.S. from Mexico the number and characteristics of those who return to Mexico have fluctuated over time, with many increasingly returning home to reunite with family (Gonzalez-Barrera 2015). This return migration trend presents serious challenges for Mexico, a country that has long struggled to satisfy the healthcare demands of its population. For these reasons, this study examines potential differences in health and health insurance coverage between Mexican return migrant and non-migrant older populations. This question is of transnational importance for policymakers. If those choosing to return home are predominantly at a higher risk of poor health and disability, the burden of healthcare access and insurance coverage could significantly affect Mexican society.
Previous literature has examined the health status and access to medical care of Mexican immigrants in the United States. This research concluded that while lacking access to healthcare in the United States (Aguila and Zissimopoulos 2013; Martinez-Donate et al. 2017; Vargas Bustamante et al. 2012), Mexican immigrants tend to report better health outcomes than U.S.-born individuals despite their low socioeconomic status, constituting what is often referred to as the “Hispanic Health Paradox” (Crimmins et al. 2005; Markides and Eschbach 2005; Palloni and Arias 2004). Most research suggests that this Hispanic “health advantage” can be attributed to the selectivity of U.S. migrants from Mexico (Fuller Thomson et al. 2013; Riosmena et al. 2013; Rubalcava et al. 2008). According to this hypothesis, those who migrate do not constitute a random sample of the Mexican population, but rather a “healthier” sample of the population, as most of them had better pre-migration health outcomes than those who chose not to migrate.
Another strand of the literature posits that outward selectivity –rather than migration selectivity– explains this paradox. In this view, health differentials are mostly due to unhealthy migrants in the U.S. voluntarily choosing to return to Mexico because of their poor health condition –this inverse self-selecting issue is commonly known as “salmon bias” (Arenas et al. 2015; Markides and Eschbach 2005). Scholarly research on emigration selection, particularly on the “salmon bias” hypothesis, has increasingly looked at age-health related (Angel and Angel 2015) and healthcare (Martinez-Donate et al. 2017; Wassink 2016, 2018) push and pull factors for return migration. However, to date, we know little about the health and age-related health conditions of returnees and the extent to which these factors are associated with Mexican immigrants’ decisions to return.
The present study contributes to the literature by gaining further insight into the age-related healthcare needs of Mexican returning migrants from the U.S. by examining the prevalence of health conditions and the lack of health insurance coverage among this population. Our aim is to examine how the health risk profiles and healthcare utilization patterns for Mexican return migrants are affected by both the migration and return migration experience of the returnee population, while paying close attention to age-group differences. In a first step, we employ data from the 2014 National Survey of Demographic Dynamics (ENADID) and multivariate models to analyze differences in health conditions between Mexican return migrants and the non-migrant population. To account for self-selection that normally arises when looking at individuals who have chosen to migrate, we then employ an inverse probability weighted regression adjustment (IPWRA) approach to empirically estimate these differences. Finally, we perform supplementary analysis with repeated cross-sectional data from the Survey of Migration at the North Border of Mexico (EMIF Norte) for the 2014–2017 period to further explore age-group differences in how the conditions under which migrants migrate and return to their home country impact their likelihood of developing chronic health conditions and lacking access to healthcare. A better understanding of the health conditions of returnees, as well as their access to healthcare services, is critical to understanding how policy interventions can help reduce health disparities of this already vulnerable population.
Return Migration to Mexico and Health Selection
Previous research on Mexico confirms that poor health is associated with return migration to Mexico (Arenas et al. 2015; Ullmann et al. 2011; Wilson et al. 2014). While this reverse selection hypothesis implies that migrants returning home are in worse health than those who stay in the destination country, few studies have looked at how return migrants fare compared to the non-migrant population in their country of origin. Are they in worse health than non-migrants? Did they return to Mexico to seek health insurance coverage? More importantly, how did the migration and return migration experience affect their health? Despite collective agreement in the literature that Mexican return migrants tend to have worse health than those who never migrated, findings with regard to variations in the prevalence of health-related conditions among returnees paint a mixed picture.
According to a cross-sectional survey conducted by Familiar et al. (2011), Mexican return migrants have an increased risk of experiencing depressive symptoms when compared to the general population, excluding relatives of migrants (Familiar et al. 2011). Using the Mexican Health and Aging Study, Wong and Gonzalez-Gonzalez (2010) analyze the physical health outcomes of Mexican migrants who returned vis-a-vis those of the non-migrant population. The authors uncover that women returnees – particularly those at the lower end of the wealth distribution – are more likely to be disabled compared to the general population, indicating that not only do return migrants have worse health outcomes than those who stayed, but that there are notable gender differences in how the migration experience interacts with health (Wong and Gonzalez-Gonzalez 2010).
Ullmann et al. (2011) further examine health disparities between Mexican return migrants and the non-migrant population, finding evidence of poorer health outcomes for returnees when employing data from the Mexican Migration Project (Ullmann et al. 2011). While no significant differences are found with regard to self-rated health, diabetes, or hypertension, they find that return migrants have a higher prevalence of heart disease, emotional/psychiatric disorders, obesity, and smoking than non-migrants (Ullmann et al. 2011). Using the same data source and cox proportional hazards models, Wilson et al. (2014) further confirm that return migrants are at greater risk of developing chronic health conditions later in life when compared with the non-migrant population. Like Ullmann et al. (2011), they find that when compared to Mexicans who never migrated, return migrants have no significant higher risk of developing hypertension and diabetes. However, their findings show no effects for heart or lung issues, or for mental health. Furthermore, when looking at these health outcomes for those returnees that had been deported, they find that this population subgroup is at higher risk of experiencing both diabetes and poor mental health when compared to non-migrants (Wilson et al. 2014). Such detrimental health outcomes are thought to be the consequence of a lack of healthcare access combined with the challenging migration experience and the poor working conditions migrants face upon return (Wilson et al. 2014). Moreover, their results shed light on the link between the particularities of the migration trajectory, and how it affects the health conditions of the returnees.
More recently, Mudrazija et al. (2016) employ 2001–2012 longitudinal data from the Mexican Health and Aging Study to explore differences in health-related outcomes and the household composition of older return migrants in comparison to those who have never migrated. Their results indicate that there is a relationship between the household composition of returnees and their health status, suggesting that return migrants are at higher risk of developing health issues due to the effect that the migration experience has on their family support network (Mudrazija et al. 2016). Loría (2017) further investigates whether these negative health outcomes can be attributed to the post-return process or the U.S. experience by using data from the Mexican Migration Project. The findings suggest that negative health outcomes result from the migration experience abroad rather than from the post-return process, indicating migrants could be returning home to seek healthcare in Mexico (Loría 2017).
On the other hand, recent scholarship reveals that return migrants face significant challenges when looking for medical care in Mexico. According to Aguila and Zissimopoulos (2013), Mexican migrants have lower levels of healthcare access and social security benefits than the non-migrant population, leaving them more vulnerable to poverty in old age (Aguila and Zissimopoulos 2013). Martinez-Donate and colleagues (Martinez-Donate et al. 2014) reach a similar conclusion, uncovering a lack of access to healthcare that affects Mexican migrants living in both the U. S and Mexico (Martinez-Donate et al. 2014). Although this gap is believed to narrow over time, as Seguro Popular continues to incorporate previously uninsured informal sector workers and return migrants reintegrate into the Mexican labor market, recent analysis reveals large and persistent differences in health insurance coverage rates and access to care between return migrants and non-migrants (Wassink 2016, 2018).
Ultimately, the preponderance of evidence supports the notion that there are specific factors associated with adverse health outcomes of returnees: predominantly, migrants’ age, social and family networks, and migration trajectories. Furthermore, disparities in the prevalence of health conditions among returnees reflect the heterogeneity of the return migrant population, suggesting they are not equally vulnerable, and therefore have different healthcare needs. Exploring this heterogeneity with a particular focus on age-effects is of substantial importance, especially given a general consensus in the literature that Mexican return migrants lack health insurance coverage and access to adequate medical care (Aguila and Zissimopoulos 2013; Martinez-Donate et al. 2017; Wassink 2018).
Data and Methods
To identify age-group differences in the heath of return migrants and examine how the migration and return migration experiences impact the specific healthcare needs of returnees, this study employs data from two sources. First, to analyze the differences in health and access to health insurance between Mexican return migrants and non-migrants, we use data from the 2014 National Survey of Demographic Dynamics (ENADID). The ENADID is a nationally representative survey carried out by Mexico’s National Institute of Statistics, Geography, and Informatics (INEGI). These data are particularly relevant for our study, given the large representative sample of both return migrants and non-migrants in Mexico that allows us to compare outcomes between these two sub-populations (N = 345,348). It contains socio-economic and demographic information on Mexican households including place of birth, the number of members within the household who immigrated to the U.S. over the past five years, and whether they lived in the U.S. either a year or five years prior to being surveyed. We define return migrants as those with Mexican citizenship who reported having lived in the U.S. at any point in time within the past five years. Furthermore, the ENADID asks individuals to self-report if they experience any physical limitations or mental distress, letting us explore variations in their health conditions. Individuals also report whether or not they have health insurance coverage in Mexico and through which institutions in the event that they do.
To assess the prevalence of physical disability or mental distress, we constructed two dichotomous variables. These outcome variables are based on the ENADID’s self-reported prevalence of physical and mental or emotional limitations. Respondents were coded as physically limited if they reported experiencing difficulty when performing any of the following: walking, arm mobility, hearing, eyesight, bathing, and eating. In terms of mental or emotional distress, respondents were asked if they had any emotional or mental problems that interfered with their daily activities such as autism, depression, bipolarity or schizophrenia. Possible answers to these questions ranged from 1 to 4, where 1 meant completely disabled, and 4, no difficulty at all. We coded those individuals that responded with 1 or 2 in each of the physical or mental and emotional conditions as functionally limited or distressed, respectively. Approximately 20% of the sample was coded as physically limited while only 2% of the sample reported being mentally or emotionally distressed (N = 67,653, N = 6,784; respectively).
The variable of interest in our model is return migration. Since the ENADID asks respondents whether or not they resided in the U.S. one or five years prior to the survey,1 we were able to construct two predictor measures of return migration as in Wassink (2018). Our first measure refers to one-year return migrants, those individuals who reported living in the U.S. only a year prior to being surveyed and that returned to Mexico at any point in time within the past year (Wassink 2018). The second measure refers to five-year return migrants or those who reported returning to Mexico at any point in time within the past five years. When coding the binary variable for five-year return migrants, those who returned within the past year were excluded to form their own separate category. Although these two measures do not capture exactly how long the return migrant has resided in Mexico, these two measures serve as a proxy to assess whether return migrants’ health varies in response to the time the individual has been residing in Mexico. Our models then control for gender, age groups, educational attainment, marital status, socioeconomic status, employment status characteristics, and whether the respondent self-identified as indigenous, as well as population size and geographic variables for the locality where they reside.
Table 1 presents the summary statistics of our ENADID 2014 sample by measure of return migration. As expected, there are significant differences between returnees and the non-migrant population. Unlike the general population, return migrants are mostly male (74%), and less likely to be from indigenous descent (.07% vs .04%). On average, while returnees tend to have elementary or secondary education, fewer have high school degrees. In addition, there are fewer return migrants in higher socio-economic groups. In terms of marital status, we also observe that while most returnees are married, when compared to non-migrants, a higher proportion of them are single or separated (26% vs 15% in both cases). It can also be seen that migrants return disproportionately to rural areas, with a considerable number of them residing in localities with less than 2,500 inhabitants. Furthermore, there are less unemployed return migrants than non-migrants, and in line with the literature on labor market outcomes and return migration, a larger share of them are self-employed (Hagan and Wassink 2016). Finally, when looking at differences between return migrants who returned within a year prior to the survey versus those who returned within the five years prior to being surveyed, we observe that although marginal, there are some differences between both groups, particularly in terms of labor market outcomes and the size of the locality in which they reside.
Table 1.
Summary characteristics by return migrant group (ENADID, 2014)
| Variable | Non-migrants | One-year return migrants | Five-year return migrants |
|---|---|---|---|
|
| |||
| Socio-demographic characteristics | |||
| Female (%) | 52.0% | 26.2% | 25.5% |
| Age group (%) | |||
| 15–24 years | 10.54% | 20.54% | 9.88% |
| 25–34 years | 35.11% | 34.82% | 35.13% |
| 35–44 years | 29.89% | 19.64% | 30.57% |
| 45–54 years | 14.86% | 13.39% | 14.96% |
| 55–64 years | 6.82% | 5.36% | 6.92% |
| ≥ 65 years | 2.77% | 6.25% | 2.54% |
| Indigenous (%) | 0.07% | 0.02% | 0.04% |
| Educational attainment | |||
| Some elementary education (%) | 3.5% | 0.8% | 0.3% |
| Completed elementary education (%) | 36.4% | 32.0% | 32.6% |
| Secondary education (%) | 27.8% | 37.6% | 38.8% |
| High school or more (%) | 32.3% | 29.6% | 28.5% |
| Marital status | |||
| Separated (%) | 15.1% | 26.1% | 26.4% |
| Divorced (%) | 6.5% | 10.1% | 12.4% |
| Widowed (%) | 4.8% | 4.2% | 2.2% |
| Married (%) | 39.6% | 34.5% | 40.7% |
| Single (%) | 34.0% | 25.2% | 18.3% |
| SES | |||
| Low (%) | 20.1% | 27.0% | 22.5% |
| Mid-low (%) | 52.3% | 47.6% | 53.8% |
| Mid-high (%) | 20.0% | 18.3% | 18.6% |
| High (%) | 7.7% | 7.1% | 5.0% |
| Work status | |||
| Employed (%) | 40.0% | 48.7% | 51.3% |
| Self-Employed (%) | 13.8% | 15.1% | 20.7% |
| Unemployed (%) | 46.2% | 36.1% | 28.0% |
| Residence | |||
| ≥100,000 (%) | 43.4% | 27.8% | 31.4% |
| 15,000–99,999 (%) | 15.6% | 12.7% | 16.1% |
| 2,500–14,999 (%) | 16.1% | 12.8% | 18.7% |
| ≤ 2,500 (%) | 25.0% | 46.8% | 33.8% |
| Physical and mental limitations | |||
| Physical limitations | 21.1% | 17.5% | 22.7% |
| Mental or emotional distress | 2.1% | 0.8% | 1.9% |
| N | 313,910 | 126 | 1,750 |
The EMIF-Norte
Despite its strengths, a drawback of the ENADID is that it does not differentiate migrants who returned voluntarily to Mexico from those who were deported. Furthermore, the ENADID does not contain detailed information on the respondents’ migration experience and their access to healthcare services while abroad. Thus, to further assess how the migration and return migration experiences impact the age-related health and healthcare utilization patterns of Mexican return migrants, we use data from the Survey of Migration at the North Border of Mexico (EMIF Norte) for the 2014–2017 period.
The EMIF Norte is carried out by Mexico’s College of the Northern Border (COLEF, for its Spanish acronym) with the goal of recording both authorized and unauthorized flows of Mexican-born migrants to and from the U.S. The survey is administered only to Mexican nationals, through a multistage sampling design that first selects transit points and border-crossing cities. It then employs probabilistic sampling techniques at the selected place and time to identify respondents. We combined the 2014 through 2017 waves of data of the EMIF-Norte and merge the Mexico City Airport questionnaire with the questionnaire at northern border-crossings to increase our sample size and capture a more representative sample of returnees. Furthermore, we restricted our analysis to those respondents who reported returning to Mexico for reasons other than to visit family. Overall, our sample consists of 17,258 returnees. Summary statistics are available in Table 5.
Table 5.
Summary characteristics for control variables (EMIF-Norte, 2014–2017)
| Variable | Mean | Min | Max |
|---|---|---|---|
|
| |||
| Socio-demographic characteristics | |||
| Female (%) | 47.33 | 0 | 1 |
| Age | 41.76 | 15 | 98 |
| Age group (%) | |||
| 15–24 years | 19.94 | 0 | 1 |
| 25–34 years | 21.93 | 0 | 1 |
| 35–44 years | 15.35 | 0 | 1 |
| 45–54 years | 13.89 | 0 | 1 |
| 55–64 years | 16.24 | 0 | 1 |
| ≥ 65 years | 12.65 | 0 | 1 |
| Indigenous | 3.42 | 0 | 1 |
| Years of schooling | 3.46 | 0 | 18 |
| Educational attainment (%) | |||
| None | 3.27 | 0 | 1 |
| Some elementary education | 80.87 | 0 | 1 |
| Completed elementary education | 13.87 | 0 | 1 |
| Secondary education | 0.03 | 0 | 1 |
| High school or more | 1.96 | 0 | 1 |
| Marital status (%) | |||
| Separated | 3.05 | 0 | 1 |
| Divorced | 9.56 | 0 | 1 |
| Widowed | 9.07 | 0 | 1 |
| Married | 42.87 | 0 | 1 |
| Partnered | 5.98 | 0 | 1 |
| Single | 29.47 | 0 | 1 |
| Migration experience | |||
| Legal migrant (first migration) (%) | 6.33 | 0 | 1 |
| Family Networks (if has family in the U.S.) (%) | 84.71 | 0 | 1 |
| Number of times crossed | 1.81 | 1 | 33 |
| English (if speaks English) (%) | 50.21 | 0 | 1 |
| Year first crossed | 2000 | 1995 | 2017 |
| Return migration experience | |||
| Deported (%) | 9.65 | 0 | 1 |
| Current migratory status (%) | |||
| Unauthorized | 17.88 | 0 | 1 |
| US citizen | 3.87 | 0 | 1 |
| US resident | 9.54 | 0 | 1 |
| Temporary visitor | 68.71 | 0 | 1 |
| Health status | |||
| Health coverage (%) | |||
| Access to health insurance in the U.S. | 32.63 | 0 | 1 |
| Access to health insurance in the Mexico | 69.8 | 0 | 1 |
| Health conditions (%) | |||
| Hypertension | 14.76 | 0 | 1 |
| High cholesterol | 3.85 | 0 | 1 |
| Diabetes | 12.66 | 0 | 1 |
| N | 17,258 | – | – |
Although less commonly used due to its sampling design, the EMIF-Norte has important strengths for the analysis of differences in age-related health conditions and access to healthcare among the return migrant population (Rendall et al. 2011). First of all, the EMIF-Norte is specifically designed to measure flows of Mexican migration and return migration, as it is intended to sample individuals at airports, bus depots and train stations to maximize its coverage of both authorized and unauthorized migration flows. Second, it reports yearly data on return migrants’ immigration status, whether the respondent returned voluntarily or not, and migrants’ previous migration histories (Rendall et al. 2011). In addition, it is the only survey in Mexico to ask returnees if health concerns motivated their decision to return. Finally, the survey asks respondents to self-report their perceived health status, whether or not they have access to health insurance in Mexico, and if any doctor or medically-trained individual has diagnosed them with any of the following chronic health conditions: hypertension, diabetes, and high cholesterol. These questions allow us to further analyze variations in health outcomes and health insurance coverage among returnees according to their return migration experiences.
It is important to note, however, that while providing a clearer picture of the health conditions affecting return migrants and their lack of access to healthcare, the inference derived from the EMIF-Norte cannot be generalized to the overall Mexican return migrant population. While migrants crossing through specific points along the Mexico-U.S. border are well represented, those who employed unauthorized or less popular border crossings are under sampled.
Empirical Strategy
The empirical analysis presented in this study is conducted in three stages. First, we begin by employing the ENADID 2014 and logistic regression models to estimate differences in the prevalence of physical limitations and mental distress among return migrants and non-migrants. Four models were estimated, one for recent return migrants and the other for five-year return migrants for both physical limitations and mental or emotional distress. Since empirical evidence indicates that there is selectivity (and reverse selectivity) regarding the migrant population (Arenas et al. 2015; Ullmann et al. 2011; Wilson et al. 2014), we then employ an inverse probability weighted regression adjustment (IPWRA) approach to test for health disparities between return migrants and non-migrants. Finally, we turn to the EMIF-Norte for the 2014–2017 period to conduct a more detailed age-group effects analysis on how the migration (whether they migrated legally or not) and return migration (whether they returned voluntarily or not) experiences impact the health and healthcare utilization patterns of U.S. immigrants returning to Mexico. For the second part of the analysis, logistic regression models that include age-group effects were conducted for the following five binary outcomes: access to health insurance coverage in Mexico, access to health insurance coverage in the United States), and diagnosis of hypertension, high cholesterol and diabetes, respectively.
Inverse Probability Weighted Regression-Adjustment
In observational studies, it is often the case that treatment was not randomly assigned to individuals, but rather that certain factors that influence the likelihood of receiving a specific treatment are related to the outcome of the treatment. Similarly, in the context of this study, it is most likely that return migrants do not constitute a random sample from the Mexican population, but that there are certain defining characteristics that pushed these individuals to migrate or to return to their country of origin. Furthermore, we can think of these characteristics as being related to the outcomes of interest being analyzed in our study. For example, poorer and less educated individuals are more likely to migrate and experience health issues, and less likely to have access to healthcare. Thus, when estimating the effect of being a return migrant on our binary health outcomes, it is probable that we will encounter bias due to the self-selectivity of the return migrant population.
A common solution to adjust for this potential bias or confounding is to employ an inverse probability weighted regression-adjustment (IPWRA) approach. As implied by its name, IPWRA is able to remove confounding by using weights to construct a “pseudo-population” in which the treatment is independent of the measured confounders. These weights result from calculating the individual’s propensity to being treated given a set of covariates or their propensity score. Thus, the method combines propensity score matching (PSM) and regression adjustment (RA) by giving more weight to those individuals least likely to be in the observed group, and then uses linear regression models to predict the average treatment effect on the treated (ATET) or the average treatment outcomes for each treated individual. This method is commonly referred to as a “doubly-robust” approach as it will be consistent as long as either the PSM model or the RA model are correctly specified (Abadie and Spiess 2016; Austin and Stuart 2015). All analyses were done in Stata 14.1 using the “teffects” command (STATA 2013).
Results
Regression Analysis of Physical Limitations and Mental or Emotional Distress
Table 2 presents logistic regressions that examine the association between each measure of return migration and the prevalence of physical functional limitations and mental distress along with the standard errors of these estimates. As expected, the prevalence of both physical limitations and mental distress increases significantly with age. However, after controlling for age, results indicate there are significant differences in the effect of return migration on the risk of presenting physical limitations among migrants who returned to Mexico within the past year versus those who returned to Mexico at some point in time within the past five years. While we find no significant effects for recent return migrants (see column 1 of Table 2), migrants who have been residing in Mexico more than a year are 16% more likely to be physically impaired than non-migrants. On the other hand, similar to Wilson et al. (2014), we find no evidence of return migrants being at higher risk of mental or emotional distress (Wilson et al. 2014). The estimated difference in the return migration effect on physical limitations between these groups suggests that health could be deteriorating once migrants are back in Mexico as a result of the difficulties experienced when re-adapting and poor working conditions in Mexico.
Table 2.
Logistic regression estimates of the association between return migration and health (ENADID 2014)
| Physical limitation | Mental or emotional distress | |||
|---|---|---|---|---|
|
|
|
|||
| One-year | Five-year | One-year | Five-year | |
|
| ||||
| Return migrant | 0.95 | 1.16* | 0.5 | 1.1 |
| (0.23) | (0.08) | (0.53) | (0.23) | |
| Female | 0.97 *** | 0.97 ** | 1.17 *** | 1.16 *** |
| (0.01) | (0.01) | (0.03) | (0.03) | |
| Age group (ref: 15–24 years) | ||||
| 25–34 years | 1.33 *** | 1.33 *** | 1.70*** | 1.70*** |
| (0.03) | (0.03) | (0.09) | (0.09) | |
| 35–44 years | 2.25 *** | 2.25 *** | 2.50 *** | 2.47 *** |
| (0.06) | (0.06) | (0.13) | (0.13) | |
| 45–54 years | 5.00 *** | 4.99 *** | 3.27 *** | 3.24 *** |
| (0.18) | (0.18) | (0.16) | (0.16) | |
| 55–64 years | 7.67 *** | 7.67 *** | 3.34 *** | 3.34 *** |
| (0.34) | (0.34) | (0.21) | (0.20) | |
| ≥ 65 years | 17.36 *** | 17.38 *** | 5.69 *** | 5.68 *** |
| (0.92) | (0.92) | (0.30) | (0.29) | |
| Indigenous | 0.91 *** | 0.91 ** | 0.72 *** | 0.72 *** |
| (0.02) | (0.03) | (0.06) | (0.06) | |
| Educational attainment (ref: elementary) | ||||
| Some elementary education | 3.91 *** | 3.91 *** | 6.95 *** | 6.94 *** |
| (0.63) | (0.63) | (1.51) | (1.50) | |
| Secondary education | 0.75 *** | 0.75 *** | 0.63 *** | 0.63 *** |
| (0.01) | (0.01) | (0.02) | (0.02) | |
| High school or more | 0.56 *** | 0.56 *** | 0.34 *** | 0.34 *** |
| (0.01) | (0.01) | (0.02) | (0.02) | |
| Marital status (ref: single) | ||||
| Separated | 0.92 ** | 0.92 ** | 0.43 *** | 0.43 *** |
| (0.03) | (0.02) | (0.03) | (0.03) | |
| Divorced | 1.23 *** | 1.23 *** | 0.84 ** | 0.84 ** |
| (0.03) | (0.02) | (0.05) | (0.05) | |
| Widowed | 1.40 *** | 1.39 | 0.85 * | 0.85 * |
| (0.04) | (0.04) | (0.06) | (0.06) | |
| Married | 0.88 *** | 0.88 *** | 0.43 *** | 0.43 *** |
| 0.02 | (0.02) | (0.02) | (0.02) | |
| SES (ref: low) | ||||
| Mid-low | 1.10 *** | 1.09 *** | 1.21 * | 1.20 * |
| (0.02) | (0.02) | (0.09) | (0.09) | |
| Mid-high | 1.05 | 1.04 | 1.43 ** | 1.42 *** |
| (0.03) | (0.03) | (0.12) | (0.12) | |
| High | 0.79 *** | 0.79 *** | 1.27* | 1.27* |
| (0.03) | (0.03) | (0.14) | (0.14) | |
| Locality size (ref: ≤ 2,500) | ||||
| ≥ 100,000 | 1.31 *** | 1.31 *** | 1.66 *** | 1.66 *** |
| (0.02) | (0.02) | (0.10) | (0.10) | |
| 15,000 – 99,999 | 1.10 *** | 1.10 *** | 1.31 *** | 1.31 *** |
| (0.03) | (0.03) | (0.09) | (0.09) | |
| 2,500 – 14,999 | 1.00 | 1.00 | 1.15 * | 1.16 * |
| (0.02) | (0.02) | (0.07) | (0.07) | |
| Pseudo R-squared | 0.17 | 0.17 | 0.08 | 0.08 |
| N | 232,285 | 233,829 | 232,285 | 233,829 |
Results are odds ratios
p < 0.05
p < 0.01
p < 0.001.
Robust standard errors in parentheses to adjust for household level clustering. All models included state fixed effects
IPWRA Results of Health and Health Insurance Coverage
For IPWRA to produce reliable ATET estimates, the common support assumption must hold. This assumption stipulates that matches are only possible when the propensity scores of the treated group overlap with those of the non-treated in an “area of common support,” given that individuals who fall outside of this range must be excluded from the analysis due to the lack of comparison units (Abadie and Spiess 2016). Figure 1 displays the distribution of the estimated propensity scores for both return migrants who have been residing in Mexico a year or less, those who returned to Mexico within the past five years, and Mexican non-migrants or the comparison group. The bar measures the frequency of individuals within each propensity score range and each graph indicates that common support is achieved for both groups (recent return migrants and five-year return migrants) with very few observations falling off support. Thus, it is possible to proceed with IPWRA estimation.
Fig. 1.
Common support analysis for recent returned migrants and five-year return migrants (ENADID, 2014)
Table 3 presents IPWRA estimates along with the associated bootstrapped standard errors of each return migrant measure on the outcomes previously analyzed. Once again, we find no significant effects of return migration status on the prevalence of mental distress. However, in terms of physical limitations, we find that while there is no effect for one-year returnees, the average treatment effect on the treated (ATET) reveals that migrants who returned to Mexico within the past five years have a probability increase by three percentage points of being physically impaired compared with non-migrants.
Table 3.
IPWRA estimates of the association between return migration and health (ENADID 2014)
| Physical limitation | Mental or emotional distress | |
|---|---|---|
|
| ||
| One-year | −0.21% | −0.62% |
| (0.03) | (0.01) | |
| Five-year | 3.17% ** | 0.32% |
| (0.01) | (0.00) | |
Results are ATET estimates
p < 0.05
p < 0.01
p < 0.001.
Robust standard errors are shown in parenthesis
Overall, the ATET estimates confirm the findings from our previous logistic regression analysis on the effect of return migration on health –returnees’ health, in terms of physical or functional limitations, seems to deteriorate after returning from the U.S. to Mexico, suggesting that reintegration into the home country is a strenuous process for returnees and that Mexico’s poor working conditions situate return migrants at a higher risk of being physically impaired. While it is important to note that migrant cohort effects could also explain these differences, we believe that this is unlikely. When looking at returnees’ self-reported health by year of return, we do not observe a trend in poor health outcomes for earlier cohorts.
Given that previous scholarly work indicates that return migrants are less likely to have health insurance coverage in Mexico (Aguila and Zissimopoulos 2013; Martinez-Donate et al. 2017; Wassink 2018) this “health” gap between return migrants and non-migrants is likely to widen as returnees struggle to reintegrate economically and socially while lacking access to health insurance. This places older returnees at greater risk; they are not only more likely to be physically disabled, but they also experience more difficulty reintegrating themselves into the formal labor market, which would grant them access to health insurance coverage under Mexico’s public healthcare system (OECD 2016). Not surprisingly, our data show that the greatest proportion of return migrants (28%) that benefit from Seguro Popular –Mexico’s publicly subsidized health insurance program–are those belonging to the oldest age group (65 years and over).
Next, we turn to the EMIF-Norte to analyze how the migration and return-migration experiences affect the health and healthcare utilization patterns of Mexican-born U.S. immigrants who have returned to their home country, while placing special attention to age-group differences.
Does the Migration Experience Impact Returnees’ Health and Healthcare Utilization Patterns?
Return migrants surveyed in the EMIF-Norte rarely cite health issues as their primary reason for moving back to Mexico; less than 1 % of respondents reported returning home due to health concerns. However, when looking more closely, we observe that a large proportion of return migrants experience chronic health issues (20%) and barriers to health insurance coverage in both the U.S.2 (74%) and Mexico (33%).
Figure 2 displays health outcomes by age group. As anticipated, aging affects health outcomes, including both the prevalence of chronic health conditions and the likelihood of self-reported poorer health. Thus, we employ an age-effects approach to analyze how certain characteristics that are associated with the migration and return migration experiences moderate the effect that aging has on the returnee’s access to health insurance coverage and health outcomes. Theoretically, we know that both ‘illegality’ and deportation have numerous detrimental impacts on an individual’s health, economic stability, and access to care (Gibney 2013; Hagan et al. 2010; Lee 2013). While it is known that the migration status of returnees is significantly related to their ability to access health insurance while in the U.S. (Angel and Angel 2015), little is known about the effects of deportation once they have returned to their home country.
Fig. 2.
Chronic health conditions and self-reported health by age category (EMIF-Norte, 2014–2017)
Table 4 shows logistic regression estimates of the association between the return migration experience and health insurance coverage in the U.S. and Mexico. To capture differences in the return migration experience, the variable of interest in all models in Table 4 is whether the migrant returned to Mexico because of deportation or if, on the contrary, he/she chose to return voluntarily. Demographic controls plus migration characteristics are included in the models. Model 1 regresses whether the individual had health coverage or not on deportation status. Model 2 employs the same regression and includes an interaction term between age-group and deportation.
Table 4.
Logistic regression estimates of the association between the return migration experience, aging, and health insurance coverage (EMIF-Norte, 2014–2017)
| Mexico health insurance coverage | US Health insurance coverage | |||
|---|---|---|---|---|
|
|
|
|||
| Model 1 | Model 2 | Model 1 | Model 2 | |
|
| ||||
| Deported | 0.59 * | 0.57 | 0.25 *** | 0.06 *** |
| (0.12) | (0.18) | (0.06) | (0.03) | |
| Female | 1.78 *** | 1.75 *** | 0.70 * | 0.70 * |
| (0.27) | (0.27) | (0.10) | (0.10) | |
| Age group (ref: 15–24 years) | ||||
| 25–34 years | 0.63 * | 0.59 * | 0.66 * | 0.62 ** |
| (0.13) | (0.16) | (0.11) | (0.11) | |
| 35–44 years | 0.70 | 0.71 | 1.10 | 1.04 |
| (0.16) | (0.17) | (0.19) | (0.18) | |
| 45–54 years | 0.85 | 0.88 | 1.16 | 1.12 |
| (0.32) | (0.35) | (0.23) | (0.22) | |
| 55–64 years | 1.70 * | 1.73* | 1.20 | 1.17 |
| (0.38) | (0.41) | (0.17) | (0.17) | |
| ≥ 65 years | 2.39 *** | 2.42 *** | 0.72 * | 0.69 * |
| (0.50) | (0.54) | (0.10) | (0.10) | |
| Interaction analysis | ||||
| Deported × 25–34 years | – | 1.53 | – | 6.61 *** |
| – | (0.41) | – | (2.01) | |
| Deported × 35–44 years | – | 0.87 | – | 7.18 * |
| – | (0.18) | – | (5.66) | |
| Deported × 45–54 years | – | 0.47 ** | – | 6.49 * |
| – | (0.11) | – | (5.42) | |
| Deported × 55–64 years | – | 0.20 *** | – | 2.21 |
| – | (0.05) | – | (1.57) | |
| Deported x ≥ 65 years | – | 0.00 | – | 0.00 |
| – | (.) | – | (.) | |
| English (if speaks English) | 1.69 ** | 1.71 *** | 6.48 *** | 6.39 *** |
| (0.24) | (0.25) | (1.30) | (1.27) | |
| Family Networks (if has family in the U.S.) | 1.65 ** | 1.64 ** | 2.03 *** | 2.02 *** |
| (0.29) | (0.28) | (0.25) | (0.26) | |
| Educational attainment (ref: none) | ||||
| Some elementary education | 1.80 *** | 1.79 *** | 0.67 | 0.66 |
| (0.14) | (0.13) | (0.19) | (0.19) | |
| Elementary education | 1.00 | 0.99 | 0.52 *** | 0.52 *** |
| (0.07) | (0.07) | (0.06) | (0.06) | |
| Secondary education or more | 0.54 * | 0.53 * | 1.12 | 1.11 |
| (0.15) | (0.16) | (0.40) | (0.40) | |
| Marital status (ref: single) | ||||
| Partnered | 0.71 | 0.71 | 0.97 | 0.96 |
| (0.17) | (0.17) | (0.20) | (0.21) | |
| Separated | 0.78 | 0.78 | 1.02 | 1 |
| (0.20) | (0.19) | (0.24) | (0.24) | |
| Divorced | 0.50 * | 0.51 * | 0.36 * | 0.36 * |
| (0.15) | (0.15) | (0.18) | (0.18) | |
| Widowed | 0.92 | 0.92 | 0.75 | 0.74 |
| (0.06) | (0.15) | (0.21) | (0.20) | |
| Married | 0.85 | 0.85 | 0.71 | 0.71 |
| (0.20) | (0.20) | (0.17) | (0.17) | |
| Pseudo R-squared | 0.14 | 0.20 | 0.10 | 0.14 |
| N | 16,173 | 16,173 | 13,594 | 13,594 |
Results are odds ratios
p < 0.05
p < 0.01
p < 0.001.
Robust standard errors in parentheses clustered at the survey-year level
Findings from column 1, Table 4, indicate that migrants who were forcibly returned to Mexico are 41% less likely to have insurance coverage than those who choose to return voluntarily. When adding the interaction term between deportation and age-group (Model 2) we observe that the coefficient for deportation is no longer significant. However, results from the interaction analysis reveal that deported returnees belonging to the older-age groups are more at risk of lacking health insurance coverage. As shown in column 2, Table 4, deported return migrants belonging to the 45 to 54 age group and those belonging to the 55 to 64 age group are 53% and 80% less likely to have health insurance coverage in Mexico than younger deportees, respectively. We had no observations in the 65-year-old and over group. The significant interaction effect suggests that deportation contributes to older return migrants’ high levels of un-insurance, which may in part be the consequence of the economic and physical distress associated with forced removals combined with unsuccessful job searches (Table 5).
Not surprisingly, results from columns 3 and 4 from Table 4 show that deportees are less likely than those who returned voluntarily to have had health insurance while residing in the U.S. Furthermore, when adding the interaction terms, we find that young and middle-aged deportees (25–54 years) are more likely to have had coverage in the U.S. than those unauthorized migrants in the15 to 24-year group. Nonetheless, the effects are smaller and non-significant for returnees in the older-age group (55–64 years), suggesting that although health insurance coverage for unauthorized migrants is correlated with aging, individuals 55–64 years are still more vulnerable to remain uninsured while in the U.S.
Interestingly, we also find that among return migrants, women are almost 30% less likely to have had health insurance coverage in the U.S. and nearly twice as likely to have health insurance coverage once they are back in Mexico. These results are robust across all models.
Next, we look at the association between the initial migration experience and the prevalence of chronic health conditions among return migrants. We do so by performing logistic regression estimates of the association between legal migration and whether respondents’ have been diagnosed with hypertension, high-cholesterol, or diabetes. In addition to all control variables, we further control for respondents’ immigration status at the time they were surveyed, to account for those returnees who might have migrated without documentation but managed to obtain legal status while abroad. These results, which are presented in Table 6, indicate that having migrated legally to the U.S. is associated with a lower prevalence of both hypertension and diabetes, when compared to those migrants who were initially undocumented. Moreover, when we add interaction terms of legal migration and age-group (Model 2), we find that aging remains strongly associated with a higher likelihood of experiencing hypertension and that young returnees that migrated legally are significantly less likely to experience diabetes (91%). We do not find any significant results for high cholesterol.
Table 6.
Logistic regression estimates of the association between the migration experience, aging and chronic health conditions (EMIF-Norte, 2014–2017)
| Hypertension | High cholesterol | Diabetes | ||||
|---|---|---|---|---|---|---|
|
|
|
|
||||
| Model 1 | Model 2 | Model 1 | Model 2 | Model 1 | Model 2 | |
|
| ||||||
| Legal migrant (first migration) | 0.45 *** | 0.40 *** | 0.94 | 1.17 | 0.56 *** | 0.85 |
| (0.07) | (0.06) | (0.43) | (0.68) | (0.07) | (0.92) | |
| Female | 1.06 | 1.05 | 0.75 *** | 0.75 *** | 0.97 | 0.97 |
| (0.06) | (0.06) | (0.06) | (0.05) | (0.04) | (0.04) | |
| Age group (ref: 15–24 years) | ||||||
| 25–34 years | 1.05 | 1.06 | 1.83 | 1.78 | 1.37 | 1.44 |
| (0.04) | (0.05) | (0.82) | (0.83) | (0.36) | (0.39) | |
| 35–44 years | 2.59 *** | 2.59 *** | 6.17 *** | 6.10 *** | 3.18 *** | 3.27 *** |
| (0.62) | (0.62) | (2.57) | (2.61) | (0.73) | (0.81) | |
| 45–54 years | 4.48 *** | 4.26 *** | 11.70 *** | 11.17 *** | 7.93 *** | 7.75 *** |
| (1.48) | (1.36) | (6.27) | (6.16) | (2.09) | (2.08) | |
| 55–64 years | 6.34 *** | 6.18 *** | 15.33 *** | 14.92 *** | 13.75 *** | 13.74 *** |
| (3.02) | (2.90) | (7.77) | (7.46) | (5.08) | (5.04) | |
| ≥ 65 years | 12.79 *** | 12.55 *** | 20.44 *** | 19.51 *** | 23.37 *** | 23.52 *** |
| (5.45) | (5.34) | (8.38) | (8.30) | (9.13) | (9.41) | |
| Interaction analysis (first migration) | ||||||
| Legal migrant × 25–34 years | – | – | – | 0.81 | – | 0.19 *** |
| – | – | – | (0.27) | – | (0.06) | |
| Legal migrant × 35–44 years | – | 0.42 | – | 0.59 | – | 0.41 |
| – | (0.25) | – | (0.37) | – | (0.45) | |
| Legal migrant × 45–54 years | – | 0.63 | – | 0.97 | – | 1.11 |
| – | (0.51) | – | (0.50) | – | (1.34) | |
| Legal migrant × 55–64 years | – | 2.18 ** | – | 0.62 | – | 0.74 |
| – | (0.49) | – | (0.38) | – | (0.94) | |
| Legal migrant x ≥ 65 years | – | 1.47 * | – | 0.00 | – | 0.59 |
| – | (0.25) | – | (.) | – | (0.70) | |
| English (if speaks English) | 1.24 | 1.24 | 1.27 * | 1.27 * | 0.92 | 0.92 |
| (0.28) | (0.28) | (0.14) | (0.14) | (0.13) | (0.13) | |
| Family Networks (if has family in the U.S.) | 2.73 *** | 2.73 *** | 4.50 *** | 4.47 *** | 3.04 ** | 3.03 ** |
| (0.60) | (0.61) | (0.63) | (0.60) | (1.06) | (1.08) | |
| Educational attainment (ref: none) | ||||||
| Some elementary education | 0.65 *** | 0.65 *** | 0.72 | 0.73 | 0.73 | 0.73 |
| (0.04) | (0.04) | (0.18) | (0.18) | (0.16) | (0.16) | |
| Elementary education | 0.62 ** | 0.62 ** | 0.62 | 0.63 | 0.62 * | 0.62 * |
| (0.11) | (0.11) | (0.25) | (0.26) | (0.13) | (0.13) | |
| Secondary education or more | 0.47 | 0.47 | 0.87 | 0.86 | 0.76 | 0.75 |
| (0.42) | (0.42) | (0.55) | (0.54) | (0.50) | (0.50) | |
| Marital status (ref: partnered) | ||||||
| Separated | 1.01 | 1.02 | 1.20 | 1.20 | 1.21 | 1.22 |
| (0.13) | (0.13) | (0.42) | (0.43) | (0.52) | (0.52) | |
| Divorced | 0.71 | 0.71 | 0.84 | 0.84 | 0.75 | 0.75 |
| (0.25) | (0.25) | (0.30) | (0.31) | (0.28) | (0.28) | |
| Widowed | 1.01 | 1.01 | 0.93 | 0.93 | 1.53 | 1.55 |
| (0.19) | (0.19) | (0.19) | (0.19) | (0.45) | (0.44) | |
| Married | 0.67 | 0.67 | 0.73 | 0.73 | 1.22 | 1.22 |
| (0.17) | (0.17) | (0.15) | (0.15) | (0.32) | (0.31) | |
| Single | 0.96 | 0.95 | 0.92 | 0.91 | 1.01 | 1.01 |
| (0.28) | (0.27) | (0.16) | (0.16) | (0.18) | (0.17) | |
| Current migratory status (ref: unauthorized) | ||||||
| US citizen | 3.87 ** | 3.89 ** | 3.22 * | 3.24 * | 1.87 | 1.88 |
| (1.74) | (1.76) | (1.55) | (1.59) | (0.82) | (0.82) | |
| US resident | 4.42 *** | 4.45 *** | 3.04 ** | 3.07 ** | 2.45 *** | 2.46 *** |
| (0.53) | (0.54) | (1.28) | (1.31) | (0.30) | (0.27) | |
| Temporary visitor | 1.93 *** | 1.97 *** | 1.77 * | 1.79 * | 1.72 * | 1.74 * |
| (0.33) | (0.35) | (0.41) | (0.42) | (0.34) | (0.35) | |
| Pseudo R-squared | 0.18 | 0.18 | 0.12 | 0.13 | 0.23 | 0.23 |
| N | 13,563 | 13,488 | 13,079 | 13,005 | 13,280 | 13,280 |
Results are odds ratios
p < 0.05
p < 0.01
p < 0.001.
Robust standard errors in parentheses clustered at the survey-year level
Our data also suggest having migrated legally to the U.S. is associated with a lower prevalence of both hypertension and diabetes, any type of legal current migratory status seems to be associated with a significantly higher likelihood of reporting hypertension and high cholesterol. Furthermore, in the case of those who are U.S. residents or possess temporary visitor visas, there is also a higher prevalence of diabetes. These results hint at the acculturation and health hypothesis which posits that as Hispanic immigrants integrate into American society, they also tend to become less healthy by acculturating their diet and certain health behaviors (see Lee et al. 2013, among others). However, our data are limited, and future studies should take into account duration of residence when looking at the association between the initial migration experience and health outcomes across different age-groups.
Conclusion
Our primary goal in this study is to contribute to the literature on return migration to Mexico by empirically assessing age-group differences in the extent to which return migrants are vulnerable in terms of health and a lack of health insurance coverage. Our findings suggest that return migrants are more vulnerable to experiencing functional limitations than non-migrants, supporting previous findings from Familiar et al. (2011). However, our findings do not show that return migrants have a higher prevalence of mental or emotional distress. Such results could be due to the ENADID’s measure of emotional and mental distress being different from those provided by the Mexican Migration Project given that they consist of different self-reported indicators (Ullmann et al. 2011; Wilson et al. 2014). A possible mechanism as to the lack of association between return migration and mental or emotional distress could be due to return migration from the U.S. being mostly voluntary (86%), primarily due to migrants wanting to reunite with family members (61%) (Gonzalez-Barrera 2015). Thus, the return process might be less stressful, but the conditions encountered once back home could still negatively affect the health of returnees.
Furthermore, our findings show that migrants who returned to Mexico within the past five years were shown to have a higher likelihood of being physically impaired than non-migrants, while there were no significant differences between those who recently returned (within the past year) and the general non-migrant population. Such findings should be taken with caution given the small sample size for recent return migrants (N = 126); however, these results further support the possibility that the post-return process could be more detrimental than the return migration process itself. Thus, the harmful effects of the reintegration process into the home country could matter more than the outward selectivity hypothesis, as opposed to findings from previous work by Loría (2017). These findings where further confirmed through our IPWRA analysis.
Our descriptive results indicate that not all returnees are equally vulnerable and that age-related factors considerably undermine health and wellbeing. For example, while late-life migrants are at risk of chronic illnesses, our findings indicate that both the migration and the return migration experiences adversely influence returnees’ access to healthcare throughout these stages, and hence, their health circumstances after returning. Unauthorized migrants are at higher risk of experiencing poorer health outcomes, as well as being uninsured both abroad and back home. This, in turn, suggests that migrants who are forced to return face greater challenges when trying to reintegrate themselves into the Mexican labor market. Furthermore, we also observe gender differences in the healthcare utilization pattern of returnees. Female return migrants are more likely to lack health insurance coverage while in the U.S. and are more likely to be insured once back home.
Previous findings along these lines illustrate that women migrants in the U.S. –particularly older adults– are less likely than non-Hispanic whites, African-American women, and Mexican-origin men, to report coverage, mainly as a result of their lack of participation in the U.S. labor force. (Angel et al. 2011; Angel et al. 2015; Wong et al. 2006; Wong and Gonzalez-Gonzalez 2010). Additionally, older migrant women who outlive their traditional “bread-winner” husbands become vulnerable to a loss of insurance coverage or a loss of full benefits in later life (Angel et al. 2011). However, more empirical work merits attention to have a better sense of why this is the case and whether women are more likely to return to Mexico to seek healthcare.
Taken together, our findings reveal a complex constellation of factors associated with return migration, aging, health, and coverage. Our results corroborate Wilson and colleagues findings that return migrants are at greater risk of poor health and disability when compared with the non-migrant population (Wilson et al. 2014). Furthermore, the fact that these detrimental outcomes are found only for those who returned to Mexico within the past five years further suggests that it is mostly the challenging conditions that migrants face back home which affect their health status. On the other hand, we also find that the characteristics associated with the migrant’s immigration history and trajectory –specifically legal migration and deportation–impact both migrants’ health and lack of health insurance coverage, and that these effects vary across different age groups (Mudrazija et al. 2016; Wilson et al. 2014).
These findings have important implications for policymakers. On the one hand, deportation puts returnees at greater risk due to the strenuous return migration experience and the barriers they face when seeking healthcare back home. On the other hand, even when return is voluntary, and health is not linked to the decision to return, the process of reintegrating both economically and socially into the home-country appears to have potentially detrimental effects on return migrant’s health. Older individuals face a dual disadvantage in terms of increasing poor health with fewer resources to draw on. This, combined with the fact that there is a significant gap in healthcare insurance coverage between return migrants and non-migrants, highlights the urgent need for the Mexican government to address the healthcare necessities of this already vulnerable population.
Our data also inform public policies aimed at older adults and people with disability. The EMIF-Norte data identified an important gap between need and awareness of Mexico’s three major programs aimed at addressing return migrants’ health: Ventanilla de Salud, Vete sano, Regresa Sano, and Asistencia a Repatriados Enfermos. These programs provide immigrant families with free health screenings, referrals to community health services, and access to healthcare at different stages of the migration process. Despite their importance, only 2 % of the total number of respondents in our sample reported being aware of these programs. More importantly, less than 3 % of respondents that reported returning due to health problems were aware of these resources, indicating that these programs are not serving the population they were intended to aid. Besides improving awareness of such programs, support and follow-up should be given to return migrants who have already resettled in Mexico.
In summary, our results point to the need for more comprehensive policies that support migrants’ health and healthcare needs with a focus on older adults and the post-return process. As the U.S. continues to tighten its immigration policies, it is likely that Mexico will see an increase in return migration due to the lack of safety net provisions in the U.S. and the need for family healthcare support, which will place a greater burden on Mexico’s healthcare system.
Funding
Jacqueline Angel has received research support for this study from the UT-Austin LBJ School of Public Affairs Policy Research Institute International Program. Ana Canedo received a travel grant from the NIH/National Institute on Aging #AG029767 to present an earlier version of the paper.
This research was supported by grant, P2CHD042849, Population Research Center, awarded to the Population Research Center at The University of Texas at Austin by the Eunice Kennedy Shriver National Institute of Child Health and Human Development. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.
Footnotes
The ENADID asks respondents that reported having migrated to the U.S. the following two questions: Five years ago, in August 2009, in what state of the Mexican republic or country did you live? A year ago, in August 2013, in what state of the Mexican republic or country did you live?
Health insurance coverage in the U.S. is defined as having access to any of the following: private health insurance, employer provided health insurance, Medicaid, Medicare, Medical Assistance, TRICARE or any other health insurance plan.
Compliance with Ethical Standards
Conflict of Interest The authors declare that they have no conflict of interest.
References
- Abadie A, & Spiess J (2016). Robust post-matching inference. In preparation. Retrieved from https://editorialexpress.com/cgi-bin/conference/download.cgi?db_name=EEAESEM2016&paper_id=3067. Accessed 12 Nov 2018. [Google Scholar]
- Aguila E, & Zissimopoulos J (2013). Retirement and health benefits for Mexican migrant workers returning from the United States. International Social Security Review, 66(2), 101–125. 10.1111/issr.12014. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Angel RJ, & Angel J (2015). Latinos in an aging world. Social, psychological, and economic perspectives. New York: Routledge. [Google Scholar]
- Angel JL, Montez JK, & Angel RJ (2011). A window of vulnerability: Health insurance coverage among women 55 to 64 years of age. Womens Health Issues, 21(1), 6–11. 10.1016/j.whi.2010.07.011. [DOI] [PubMed] [Google Scholar]
- Angel RJ, Angel JL, & Hill TD (2015). Longer lives, sicker lives? Increased longevity and extended disability among Mexican-origin elders. The Journals of Gerontology: Series B, 70(4), 639–649. 10.1093/geronb/gbu158. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Arenas E, Goldman N, Pebley AR, & Teruel G (2015). Return migration to Mexico: Does health matter? Demography., 52, 1853–1868. 10.1007/s13524-015-0429-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Austin PC, & Stuart EA (2015). Moving towards best practice when using inverse probability of treatment weighting (IPTW) using the propensity score to estimate causal treatment effects in observational studies. Statistics in Medicine, 34(28), 3661–3679. 10.1002/sim.6607. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Chort I, de la Rupelle M, Gonzalez-lozano H, Co CY, Gang IN, Yun M-S, … Masferrer CP (2015). Welcome home in a crisis : Effects of return migration on the non-. Review of Development Economics, 98(2), 58. 10.1257/aer.103.6.2121,2168. [DOI] [Google Scholar]
- Crimmins EM, Soldo BJ, Kim JK, & Alley DE (2005). Using anthropometric indicators for Mexicans in the United States and Mexico to understand the selection of migrants and the “Hispanic paradox”. Social Biology, 3(52), 164–177. 10.1080/19485565.2005.9989107. [DOI] [PubMed] [Google Scholar]
- Familiar I, Borges G, Orozco R, & Medina-Mora M-E (2011). Mexican migration experiences to the US and risk for anxiety and depressive symptoms. Journal of Affective Disorders, 130(1–2), 83–91. 10.1016/j.jad.2010.09.025. [DOI] [PubMed] [Google Scholar]
- Fuller Thomson E, Nuru-Jeter A, Richardson D, Raza F, & Minkler M (2013). The Hispanic paradox and older adults’ disabilities: Is there a healthy migrant effect? International Journal of Environmental Research and Public Health., 10, 1786–1814. 10.3390/ijerph10051786. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Gibney MJ (2013). Is deportation a form of forced migration? Refugee Survey Quarterly, 32(2), 116–129. Retrieved from. 10.1093/rsq/hdt003. [DOI] [Google Scholar]
- Gonzalez-Barrera A (2015). More Mexicans leaving than coming to the U.S. Pew Research Center, 202419, 32. Retrieved from http://www.pewhispanic.org/2015/11/19/more-mexicans-leaving-than-coming-to-the-u-s/. Accessed 26 Oct 2018. [Google Scholar]
- Hagan JM, & Wassink J (2016). New skills, new jobs: Return migration, skill transfers, and business formation in Mexico. Social Problems, 63(4), 513–533. 10.1093/socpro/spw021. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Hagan J, Castro B, & Rodriguez N (2010). The effects of U.S. deportation policies on immigrant families and communities: Cross-border perspectives. N.C. L. Rev. 1799, 88. Retrieved from http://scholarship.law.unc.edu/nclr/vol88/iss5/12. [Google Scholar]
- Lee AE (2013). “Illegality,“ health problems, and return migration: Cases from a migrant sending community in Puebla, Mexico. Regions and Cohesion, 8(3), 62–93. Retrieved from 10.3167/reco.2013.030104. [DOI] [Google Scholar]
- Lee S, O’Neill AH, Ihara ES, & Chae DH (2013). Change in self-reported health status among immigrants in the United States: Associations with measures of acculturation. PLoS One, 8, 10. 10.1371/journal.pone.0076494. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Loría MV (2017). Back in Mexico: The health of returned migrants and its relationship with health in the U.S. population Association of America Annual Meeting 2017. [Google Scholar]
- Markides KS, & Eschbach K (2005). Aging, migration, and mortality: Current status of research on the hispanic paradox, 60(Special_Issue_2). Retrieved from 10.1093/geronb/60.Special_Issue_2.S68. [DOI] [PubMed] [Google Scholar]
- Martinez-Donate AP, Zhang X, Rangel MG, Hovell M, Simon N-J, Mpa M/, … Martínez-Donate AP (2014). Healthcare access among circular and undocumented Mexican migrants: Results from a pilot survey on the Mexico-US border NIH Public Access. Int J Migr Bord Stud, 1(1), 57–108. 10.1504/IJMBS.2014.065069. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Martinez-Donate AP, Ejebe I, Zhang X, Guendelman S, Lê-Scherban F, Rangel G, Gonzalez-Fagoaga E, Hovell MF, & Amuedo-Dorantes C (2017). Access to health care among Mexican migrants and immigrants: A comparison across migration phases HHS Public Access. J Health Care Poor Underserved, 28(4), 1314–1326. 10.1353/hpu.2017.0116. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Masferrer C, & Roberts BR (2012). Going back home? Changing demography and geography of Mexican return migration. Population Research and Policy Review, 31(4), 465–496. 10.1007/s11113-012-9243-8. [DOI] [Google Scholar]
- Mudrazija S, López-Ortega M, Vega WA, Gutiérrez Robledo LM, & Sribney W (2016). Household composition and longitudinal health outcomes for older Mexican return migrants. Research on Aging, 38, 346–373. 10.1177/0164027515620241. [DOI] [PMC free article] [PubMed] [Google Scholar]
- OECD. (2016). OECD Reviews of health systems: Mexico 2016. Paris. Retrieved from 10.1787/9789264230491-en. [DOI]
- Palloni A, & Arias E (2004). Paradox lost : Explaining the Hispanic adult. Hispanic, 41(3), 385–415. [DOI] [PubMed] [Google Scholar]
- Rendall MS, Brownell P, & Kups S (2011). Declining return migration from the United States to Mexico in the Late-2000s recession : A Research Note, 1049–1058. 10.1007/s13524-011-0049-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Riosmena F, Wong R, & Palloni A (2013). Migration selection, protection, and acculturation in health: A binational perspective on older adults. Demography, 50(3), 1039–1064. 10.1007/s13524-012-0178-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Rubalcava LN, Teruel GM, Thomas D, & Goldman N (2008). The healthy migrant effect: New findings from the Mexican family life survey. American Journal of Public Health, 98(1), 78–84. 10.2105/AJPH.2006.098418. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Stata. (2013). STATA treatment effects reference manual. Stata Journal. [Google Scholar]
- Henderson Tim. (2018, September 20). Drop in Mexican-born immigrants attributed to hostility Here, opportunity there. Pew Research Center. [Google Scholar]
- Ullmann SH, Goldman N, & Massey DS (2011). Healthier before they migrate, less healthy when they return? The health of returned migrants in Mexico. Social Science and Medicine, 73, 421–428. 10.1016/j.socscimed.2011.05.037. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Vargas Bustamante A, Fang H, Garza J, Carter-Pokras O, Wallace SP, Rizzo JA, & Ortega AN (2012). Variations in healthcare access and utilization among Mexican immigrants: The role of documentation status. Journal of Immigrant and Minority Health, 14, 146–155. 10.1007/s10903-010-9406-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Wassink JT (2016). Implications of Mexican health care reform on the health coverage of nonmigrants and returning migrants. American Journal of Public Health, 106, 848–850. 10.2105/AJPH.2016.303094. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Wassink JT (2018). Uninsured migrants: Health insurance coverage and access to care among Mexican return migrants. Demographic Research, 38, 401–428. 10.4054/DemRes.2018.38.17. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Wilson FA, Stimpson JP, & Pagán JA (2014). Research article disparities in health outcomes of return. 10.1155/2014/468250. [DOI] [Google Scholar]
- Wong R, & Gonzalez-Gonzalez C (2010). Old-age disability and wealth among return Mexican migrants from the United States. Journal of Aging and Health, 22, 932–954. 10.1177/0898264310380742. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Wong R, Díaz JJ, & Higgins M (2006). Health care use among elderly Mexicans in the United States and Mexico: The role of health insurance. Research on Aging, 23(3), 393–408. 10.1177/0164027505285922. [DOI] [PMC free article] [PubMed] [Google Scholar]