After spouses depart: emotional wellbeing among nonmigrant Mexican mothers

Abstract

Nonmigrant family members play a central role in facilitating Mexico-U.S. migration by maintaining families, sustaining social relationships, and overseeing household economic organization in sending communities. This study investigates changes to the emotional wellbeing of nonmigrant mothers when their partners reside in the United States. We hypothesize that partner migration affects mothers' wellbeing through three pathways: directly via the toll of spousal separation, and indirectly via changes to the economic profile of the sending household and through changes to mothers' household responsibilities. We test these relationships using data on 2813 mothers aged 18-44 in 2002 and measured in three waves (2002, 2005, 2009) of the Mexican Family Life Survey. We employ a fixed-effect estimation strategy that improves causal attribution of women's wellbeing to spousal residential location. We find evidence of increases in some forms of distress-sadness, crying, difficulty sleeping-when spouses are in the United States but no meaningful increase in depressive symptomology. Though partner emigration shifts several aspects of women's lives in sending households, changes to household resources or time allocation do not account for the moderate shifts in emotional duress associated with spousal absence. Importantly, additional tests reveal that we would observe large and significant associations between spousal migration and mothers' emotional wellbeing using a less rigorous estimation strategy, raising caution about the interpretation of cross-sectional studies evaluating wellbeing in sending homes.

Keywords: Emotional health; Mexico; Migration; Transnational families.

Figure 1

Percentage point change in the probability of experiencing symptom when spouse lives in the United States, relative to previous or subsequent measurement of symptom when spouse coresides in Mexico, Mexican mothers surveyed in 2002, 2005, & 2009. Source: Mexican Family Life Survey waves 1, 2, & 3. Notes: Percent of total sample reporting symptom in 2002 in parentheses. Graph: estimate on spousal residence in the U.S. from fixed-effect linear probability models predicting each of 21 symptoms. Regressions adjusted for time-varying measures of household composition, household economic shocks, and period changes observed between surveys for all women.