Abstract
Migration from Mexico to the United States results in numerical imbalances between men and women in the communities left behind, but little is known about how these imbalanced sex ratios in Mexican communities affect family formation behavior. Using two waves of data from the Mexican Family Life Survey (MxFLS) and 2000 Mexican census data, we examine how the sex composition of the municipality population is associated with the likelihood that Mexican women and men marry. Consistent with demographic-opportunity theory, but not sociocultural theory, results from logistic regression analyses show that men’s probability of marrying is positively associated with the numerical availability of eligible women in the municipality population. Consistent with the argument that many Mexican migrant men, but not migrant women, remain eligible partners for those who remain in Mexico, the sex composition of the municipal population is significantly associated with men’s but not women’s likelihood of marrying.
Migration from Mexico to the United States is numerically dominated by men, and largely by young men (Donato et al., 2011; Hamilton, 2015; Raphael, 2013). This sex difference in Mexican migration to the United States contributes to an imbalance in the numbers of young men and women in many left-behind communities in Mexico (Parrado, 2004; Raphael, 2013). Social theory and empirical research on other countries suggests that sex ratio imbalances are related to a wide range of social phenomena, including marital and family behaviors (Dyson, 2012; Trent & South, 2011; Trent et al., 2015). Yet, we know little about the consequences of Mexico’s imbalanced sex ratios for women’s and men’s marital transitions.
Mexico presents a unique context for examining the impact of sex ratios on family behavior because Mexico’s sex ratio imbalances are driven largely by migration, and not by fertility (and a preference for sons) as is the case in other countries with noteworthy imbalanced sex ratios such as China and India (Agnihotri, 2000; Guillot, 2002; Guilmoto, 2008). Moreover, unlike many Asian countries, the concern in Mexico rests with female-dominated rather than male-dominated sex ratios. As in these other countries, however, sex ratio imbalances in Mexico are likely to help shape patterns of union formation. The relative availability of members of the opposite sex in Mexican communities serves as a source of potential marriage partners that presumably influences the probability of marrying.
In this paper, we use two waves of data from the Mexican Family Life Survey (MxFLS) merged with aggregated data from the 2000 Mexican census data to examine how the municipality sex ratio is associated with the likelihood that Mexican men and women marry. We estimate logistic regression models examining the relationship between the sex ratio of MxFLS respondents municipality of residence and their likelihood of marrying, controlling for several individual-level and community-level characteristics associated with the propensity to marry.
Background
Sex ratio imbalances in Mexican communities are fueled in large part by sex-selective out-migration. Migration of Mexicans to the United States is numerically dominated by young men (Donato, 2010; Donato et al., 2011; Raphael, 2013) and thus most of the U.S. Mexican-born population is composed of adult men (Donato et al., 2011). Male-biased sex ratios have been well documented for young adult Mexican nationals living in the U.S. For example, in 2007 the male-to-female sex ratio was 156 at ages 21 to 25 and 149 at ages 26 to 30 (Raphael, 2013).
Correspondingly, for Mexico as a whole, Mexican women in their twenties and thirties outnumber men of the same age (Raphael, 2013). This sex ratio imbalance has persisted for several decades. The national sex ratio for Mexico was 94.9 and 94.0 for ages 20 to 39 in 2000 and 2010, respectively, and 94.7 for the same age group in 2020 (U.S. Census International Database, 2022). Importantly, there is substantial variation in the sex ratio across regions and communities in Mexico (Raphael, 2013). Age and sex differentials in migration have resulted in a numerical imbalance of young men and women in the prime ages for marrying among many left-behind Mexican communities (Parrado, 2004; Raphael, 2013).
Theoretical Framework and Hypotheses
Perhaps the most common way sex ratios are thought to influence the propensity to marry is through demographic opportunity. Demographic-opportunity theory stresses the importance of the numerical availability of potential sexual and marital partners for the likelihood of engaging in various types of family-related behaviors including marriage. This theoretical perspective flows from work by social demographers on marriage markets (e.g., Lichter et al., 1992; Schoen & Kluegel, 1988) and by economists working in the New Household Economics (e.g., Angrist, 2002; Becker, 1981; Chiappori et al., 2002; Grossbard-Shechtman, 1993) and has been often utilized to understand opposite-sex union formation. Simply put, residing in a community containing an abundance of members of the opposite sex increases the likelihood of finding a suitable partner that eventuates in marriage. In contrast, when few potential mates are available, fewer people will marry.
Numerous studies across a wide range of geographic and social contexts provide support for demographic-opportunity theory, linking sex ratios specifically to marriage rates and individual transitions to marriage (Angrist, 2002; Fossett & Kiecolt, 1993; Lichter et al., 1992; McLaughlin et al., 1993). Early research examining the relationship between imbalanced sex ratios and the transition to marriage tended to focus on the U.S. These studies initially relied on aggregate data, and later on used individuals as the units of analysis, to examine the relationship between marriage and various measures of the sex ratio delimited by the local neighborhood or larger community. This research generally finds that women’s and men’s marriage rates and propensities are higher in geographic areas containing more eligible opposite-sex partners (e.g., Angrist, 2002; Fossett & Kiecolt, 1993; Guzzo, 2006; Lichter et al., 1991, 1992; Lloyd & South, 1996; McLaughlin et al., 1993; Schoen & Kluegel, 1988; South & Lloyd, 1992).
An alternative argument for how sex ratios affect marriage behavior, derived mainly from Guttentag and Secord (1983) and referred to as a sociocultural (Trent & South, 2011) or mating market (Schacht & Kramer, 2016; Schacht & Smith, 2017) approach, suggests that the effect of the sex ratio on marriage is gendered, with men reacting to a surplus of possible female partners by avoiding commitment to marriage. Guttentag and Secord (1983) stress the importance of dyadic power in interpersonal relationships between women and men and argue this affects how imbalanced sex ratios shape family behaviors. Guttentag and Secord contend that members of the sex that is in short supply are less dependent on their partners because a greater number of alternative relationships are available to them and they can thus more easily form relationships with other members of the opposite sex. This leaves members of the sex that is in relative oversupply in a dependent position with their opposite-sex partners because there are fewer members of the opposite sex with whom to form alternative partnerships.
Sociocultural theories suggest several responses to imbalanced sex ratios. In high-sex-ratio contexts, the shortage of women will result in men using their structural power to limit women’s roles and encourage their traditional roles as wives and mothers. In this context, women will be more likely to marry when they are relatively scarce, and men are plentiful (as also posited by demographic-opportunity theory). However, a different sex-role structure characterizes low-sex-ratio populations, as is found among many Mexican communities. In this context, according to sociocultural theory, the surplus of women will discourage men’s commitment to any one partner, and thus to marriage. In sharp contrast to demographic-opportunity theory, sociocultural theory argues that when men are exposed to a relative abundance of potential partners they will delay or refrain from marrying. Thus, under low sex ratio conditions, both men and women will be less likely to marry.
Tests of hypotheses derived from the sociocultural approach initially relied primarily on societal-level data. These hypotheses have received some support. South and Trent (1988) find that, in a sample of 117 countries circa 1980, high sex ratios are positively related to the percentage of women who are married, but that high sex ratios also limit women’s life options in other social roles. In India, a surplus of men increases the likelihood that women will marry (Trent et al., 2015), but also reduces women’s autonomy in having a say in the selection of their husband and increases the probability that women will meet their husband for the first time on their wedding day (South et al., 2016). A recent study using U.S. county-level data finds a higher proportion of men who are married in counties with comparatively few women (Schacht & Kramer, 2016), a finding consistent with the sociocultural perspective.
However, studies that examine marriage specifically and test hypotheses derived from both the sociocultural approach and demographic-opportunity theory have generally found more support for demographic opportunity than for sociocultural arguments (Trent et al., 2015; Trent & South, 2011) In China, sex ratios in the local community are positively associated with marriage for women (Trent & South, 2011). And in India, a surplus of potential mates increases the likelihood that both men and women will marry (Trent et al., 2015).
In the Mexico context, there is another way that the effect of the sex ratio on marital transitions might be gendered that is distinct from the arguments proffered by sociocultural theory. Many Mexican men who have migrated to the United States may nonetheless remain available marital partners for the women who remain in Mexico. Indeed, many young Mexican men may migrate in the hopes of earning sufficient money to return to Mexico and start a family there. Transcommunity and especially transnational social networks (Garip & Asad, 2016; Massey et al., 1998), along with migrants’ at least occasional movement to and from their community of origin, means that many former male community residents likely remain potential marriage partners to the women who remain behind (Carling et al., 2012; Dreby, 2010).
In contrast, Mexican women migrants to the United States are perhaps less likely to remain potential partners to the men who remain in Mexico. Women are less likely than men to emigrate for economic reasons and more likely to move either for reasons of family reunification or personal safety (Cerrutti & Massey, 2001). Compared to male migrants, women migrants are less likely to move to the U.S. in the hopes of earning money to return to Mexico and start a family, which may be reflected in the larger gender disparities in migration among economically disadvantaged social classes (Hamilton, 2015). Thus, for the men who remain in Mexico, the municipality sex ratio likely captures the actual supply of potential female marriage partners, but the local sex ratio may not adequately capture the supply of potential male mates available to women because many of these men reside in the U.S. This line of reasoning suggests that the effect of the municipality sex ratio on the likelihood of marrying will be stronger among Mexican men than among Mexican women.
Prior Research on Sex Ratios and Marriage in Mexico
Despite the existence of general theories implicating imbalanced sex ratios as a potential cause of a wide range of demographic behaviors and the widely acknowledged migration-induced skewed sex ratios in Mexico, little research has explored the ramifications of imbalanced sex ratios in Mexican communities. There is, of course, a large literature on the determinants of migration from Mexico to the U.S. (e.g., Donato, 2010; Durand & Massey, 2004; Massey et al., 2002, 2006). We know that this form of international migration is driven mainly by economic opportunities in the U.S. and the absence of such opportunities in Mexico, and that households in Mexico employ migration as a strategy to combat financial uncertainties with risk diversification. The household division of labor also plays a role in the decision to migrate (Nobles & McKelvey, 2015). Migration also tends to depress marriage rates in the short-term, but the economic resources gained through migration tend to raise men’s value in the marriage market upon their return to Mexico (Parrado, 2004).
However, despite this rich and growing literature on the determinants and consequences of Mexican migration to the U.S., only a few studies have examined the impact of imbalanced sex ratios on marriage in Mexico. These prior studies have often relied on small samples and report conflicting results. Parrado and Zenteno (2002) use data from the 1998 Mexican Retrospective Demographic Survey (Encuesta Demográfica Retrospectiva) to examine the impact of the sex ratio, measured at the municipality level, on union formation among women and men. Adopting a measurement strategy for the sex ratio similar to that used by Lloyd and South (1996) and focusing on three birth cohorts (1935–39; 1950–54; and 1965–69), Parrado and Zenteno (2002) estimate discrete-time event history models of the timing of transition to first union separately for women and men. No net effect of the sex ratio on the timing of marital entry is observed for either women or men.
Parrado (2004) uses data from 5687 male respondents in the Mexican Migration Project who resided in one of 43 communities in Mexico’s Western states (from which many international migrants originate) to examine the impact of individual migration experience and community out-migration on men’s marital timing. Parrado (2004) finds that the proportion of young single men from the community residing in the U.S. is inversely associated with the timing of men’s marital entry. However, this study does not directly assess the impact of the community sex ratio on men’s marriage timing.
Choi (2011) uses survey data from the Encuesta Nacional de Dinámica Demográfica, linked to municipality-level census data, to examine the impact of the community sex ratio on women’s marital timing. In partial contrast to the findings of Parrado and Zenteno (2002), Choi finds that Mexican women in this sample marry earlier when they are exposed to a relatively large number of potential husbands, a finding consistent with demographic-opportunity theory. Choi and Mare (2012) explore the impact of international migration on patterns of assortative mating among Mexican married couples using data from the 2000 Mexican census. Although this study does not directly measure the sex ratio of a municipality, it finds that individuals residing in sending communities with high levels of international migration are more likely to marry someone from outside their education category relative to comparable individuals in (sending) communities with low levels of migration. White and Potter (2013) use 2000 census data from 314 municipalities in seven Mexican states to examine the impact of male out-migration—and through this, the community sex ratio—on women’s marriage (and fertility) rates. This strictly aggregate analysis finds that low male-to-female sex ratios are positively associated with lower proportions of women married (and lower total fertility rates), findings also consistent with demographic-opportunity theory.
Raphael (2013) uses state-level data from the 1960 through 2000 Mexican censuses to examine how decadal changes in the sex ratio are associated with decadal changes in the proportion of young women who are married. Conventional OLS, instrumental variable (IV), and fixed effect regression analyses show that in states that experienced an increase in the ratio of men to women (measured using procedures similar to Goldman et al.’s (1984) availability ratio), the proportion of women in three different age groups (16–19, 20–25, and 26–30) who were married increases. However, this study does not examine the effect of the sex ratio on men’s marital status.
In sum, demographic-opportunity and sociocultural perspectives—and accompanying empirical research—suggest that the community-level sex ratio has an important impact on women’s and men’s likelihood of marrying, but these theoretical approaches diverge in important ways. Both demographic-opportunity theory and the sociocultural thesis predict a positive relationship between the local male-to-female sex ratio and women’s probability of marriage. However, these theories diverge regarding the influence of the local sex ratio on men’s marriage propensities. While the essentially gender-neutral demographic-opportunity theory suggests that high local sex ratios (indicating a numerical deficit of women) will lead men to postpone or even forego marriage, the sociocultural argument suggests that high sex ratios will encourage men to marry because they will shun marriage when women are plentiful. And because the municipality sex ratio likely captures the actual supply of potential mates for men but not women, we expect the effect of the sex ratio—whether positive or negative—on the likelihood of marrying will be stronger among Mexican men than among Mexican women.
Data and Methods
Data
This analysis examines the association between the community sex ratio and the transition to marriage by attaching municipality-level data drawn from the 2000 Mexican population census to the individual records of respondents to the Mexican Family Life Survey (MxFLS). The MxFLS is a nationally representative survey of Mexican households and their constituent members (Rubelcava & Teruel, 2006). The first wave of interviews, conducted in 2002, collected data from 8441 households containing about 35,000 individuals (including children). A follow-up survey was conducted in 2005–2006. The MxFLS collects data on a wide range of topics, including but not limited to household consumption patterns, educational attainment, labor force behavior and earnings, social and economic transfers, migration, and criminal victimization. At both waves, information is collected on adult respondents’ marital status, and in the latter wave information is collected on changes in marital status since the prior wave.
The 2000 Mexican census is used to create a measure of population sex composition and other community characteristics for the 136 municipalities inhabited by respondents to the MxFLS. We draw on public use microdata from the 2000 Mexican census available from Integrated Public Use Microdata Samples—International (IPUMS) (Minnesota Population Center, 2019). The 2000 Mexico IPUMS is a 10.5% sample of the Mexican population, containing information on roughly 10 million individuals. We extract from the Mexican IPUMS municipality-level counts of the population by sex and age and then attach these figures to the individual MxFLS records using the municipality codes.
In Mexico, municipalities are the second-level administrative division after states. (In Mexico City, boroughs serve as the second-level division). Mexican municipalities are roughly similar to U.S. counties. Municipality-level measures, including the sex ratio, are commonly thought to capture important contextual influences on a variety of social and demographic behaviors in Mexico, including marital transitions (e.g., Parrado & Zenteno, 2002; White & Potter, 2013). The 136 municipalities covered by the MxFLS have an average population size of about 200,000 persons.
Sample Selection
The analysis selects MxFLS respondents who, as of the first wave in 2002, were not married or in a cohabiting union and who were between the ages of 15 and 39. The legal age to marry in Mexico is 14 for females and 16 for men, provided parental permission is given. However, the MxFLS only asks about marital status, and inter-wave marital status changes, among respondents 15 years and older. We use age 39 as the upper limit to capture remarriages (and re-partnering) upon union dissolution and because migration to the U.S. and subsequent return to Mexico is likely to delay marriage among these international migrants. Although marriage/partnering rates drop fairly precipitously after the mid-twenties, almost six percent of the at-risk respondents ages 35 to 39 reported beginning a partnership between the first and second waves of the MxFLS. In contrast, very few respondents older than 39 entered a partnership during this period. After omitting the few cases with missing values on the covariates, our effective sample consists of 2119 women and 1706 men.
Measures
The dependent variable is a binary indicator of whether the unmarried (and non-cohabiting) respondents as of wave 1 reported being married or cohabiting at wave 2. Specifically, respondents were asked in the wave 2 interview whether their marital status had changed since wave 1 and, if so, whether they had married or “got together.” Sixteen percent of the unmarried, non-cohabiting respondents as of the wave 1 interview reported marrying or entering a cohabiting union by wave 2. For stylistic convenience, we refer to respondents who initiated a union as “marrying.”
The focal independent variable is a measure of the sex composition of the respondent’s wave 1 municipality, computed from the 2000 Mexican IPUMS. So that the measure is symmetrical for women and men, rather than compute a sex ratio we instead use the percentage of the population that is male (Ancona et al., 2017). Because partner selection is circumscribed by age, we limit this measure to persons between the ages of 15 and 39, inclusive, consistent with the age restriction on the sample of MxFLS respondents. Thus, our key independent variable is the percentage of the municipality population ages 15 to 39 that is male.
The models include several other individual-level and municipal-level characteristics that might influence the likelihood of entering a marital or cohabiting union. All of these measures are computed from information available in the wave 1 interview. Respondent’s age is measured in years, and the square of age is included to capture the decline in the risk of marrying at older ages. A dummy variable distinguishes respondents who report belonging to an indigenous ethnic group from members of the ethnic majority. The measure of respondent’s educational attainment consists of three categories tapping their highest level of schooling that the respondent attended: less than high school, high school, and college. A separate dummy variable distinguishes respondents who, at the wave 1 interview, were attending school.
Respondents who were working for pay are distinguished from non-employed respondents by a dummy variable. Household income is the sum of all adult household members’ annual income from earnings, social programs, and monetary transfers; this variable is logged to reduce skewness. Respondents living in an urban area, defined as a locality of 2500 or more inhabitants, are distinguished from rural residents by a dummy variable.
The probability of marrying is likely to be influenced not only by the relative number of opposite-sex potential partners but also by the socioeconomic characteristics of the marriage pool (e.g., Lichter et al., 1992; Lloyd & South, 1996). Accordingly, we include as predictors the educational attainment and income of the opposite-sex members who may serve as potential partners. To characterize women’s marriage pool, we compute from the IPUMS the municipality percentage of men ages 15 to 39 who have at least 12 years of schooling along with the mean monthly income of these men. To characterize men’s marriage pool, we compute the municipality percentage of women ages 15 to 39 who have at least 12 years of schooling and their mean monthly income. Finally, we include as a predictor municipality population size, measured in thousands of persons.
Analytic Strategy
We estimate logistic regression models of the likelihood that the at-risk MxFLS respondents married between wave 1 and wave 2, focusing primarily on the effect of the percentage of males in the respondents’ municipality. Because the sex ratio is expected to have different effects for women and men, we estimate sex-specific models. Although the individual observations are nested within municipalities, the level of clustering—on average about 15 respondents per municipality—is insufficient for estimating multilevel models. However, we adjust for the non-independence of observations by computing robust standard errors based on the clustering of respondents within municipalities.
Results
Descriptive Statistics for Variables Used in Analysis of Transition to Marriage: Mexican Family Life Survey 2002 and 2005/06.
As would be expected based on sex-selective migration from Mexico to the U.S., the MxFLS respondents tended to experience a numeric deficit of men in their residential municipality. For the women respondents, on average, men comprised 47.37% of the municipality population between the ages of 15 and 39. This corresponds to a sex ratio of 90 men per 100 women. For the male respondents, on average men comprised about 47.72%—and hence women comprised 52.28%—of the municipality population between the ages of 15 and 39. This corresponds to a male-to-female sex ratio of 91.
The respondents are typically in their early twenties when first observed. Ten percent of the respondents report belonging to an indigenous ethnic group. About 40% of respondents attended at least high school, with 12% of the women and 13% of the men receiving some college education. Almost one-third of the respondents were still attending school at the time of the wave 1 interview.
Forty-two percent of the women and 56% of the men were working for pay in the year preceding the wave 1 interview. Household income is roughly similar for women and men. Slightly less than 60% of respondents lived in an urban area.
Twenty-four percent of the men in women’s pool of eligible partners (i.e., the men in the numerator of the sex ratio) have completed at least twelve years of education. They have a mean monthly income of 2180 pesos. From men’s standpoint, one-quarter of the women in their marriage market have completed high school, and these women have a mean monthly income of 920 pesos. The average respondent resided in a municipality containing about 300,000 people.
Logistic Regression Analysis of Transition to Marriage for Females: Mexican Family Life Survey 2002 and 2005/06.
Note. Standard errors adjusted for the clustering of respondents with municipalities.
†<.10 *p < .05 **p < .01 ***p < .001.
Model 2 adds the individual-level covariates. The coefficients for age and age-squared show that the odds of marrying increase with age but then level off at older ages. Women who attended high school (but not college) are significantly less likely to marry than women who did not attend high school. At a borderline significance level, women who are attending school are less likely to marry than women who were not in school. Controlling for these individual-level characteristics has little impact on the coefficient for percent male, which remains small and statistically nonsignificant.
Model 3 of Table 2 adds the three municipality-level controls. Only the coefficient for men’s educational attainment is significant: when the men in women’s pool of eligibles have high levels of educational attainment, women’s odds of marrying are lower. Again, however, the coefficient for percent male remains nonsignificant, which supports neither the prediction from sociocultural theory nor the expectation from demographic-opportunity theory.
Logistic Regression Analysis of Transition to Marriage for Males: Mexican Family Life Survey 2002 and 2005/06.
Note. Standard errors adjusted for the clustering of respondents with municipalities.
†<.10 *p < .05 **p < .01 ***p < .001.
Model 2 of Table 3 adds the individual-level covariates. As with women, men’s odds of marrying first increase with age and then decline at older ages. In addition, attending school reduces men’s risk of marrying and employed men are significantly more likely than non-employed men to marry. Controlling for these predictors slightly increases the strength of the association between municipality percent male and men’s odds of marrying.
Model 3 adds the municipality-level controls. The coefficient for the percent of women ages 15 to 39 in the municipality who have completed twelve of more years of education is negative and significant while the coefficient for the mean income of these women is positive and significant at a borderline level. Thus, all else equal, men are less likely to marry if the women in their pool of eligibles are highly educated, but they are more likely to marry if these women have comparatively high earnings. Controlling for these marriage market characteristics diminishes slightly the coefficient for percent male, but it remains negative and statistically significant.
Figure 1 illustrates the strength of the association between municipality sex composition and men’s probability of marrying. This figure plots men’s probability of marrying by values of the municipality percent male at ages 15 to 39, holding all other covariates at their respective means. Predicted probability of men’s transition to marriage by the municipality-level percentage of the population ages 15 to 39 who are male: Mexican Family Life Survey, 2002 and 2005/06.
This figure suggests that the municipality sex ratio is neither a trivial nor overwhelming predictor of men’s propensity to marry. At the minimum observed value of percent male (or, equivalently, the maximum value of percent female), about 21% of men would be expected to marry over a three-year period, while at the maximum observed value, about 9% of men would be expected to marry. Of course, this 12 percentage point gap is based on a short three-year window of observation. Over the full course of men’s exposure to the risk of marriage, the difference would presumably cumulate.
Discussion and Conclusion
Migration from Mexico to the United States is dominated by young adult men. This sex differential in migration results in a sex ratio imbalance in many communities that have experienced high levels of out-migration. Theory and past research in other countries suggest that community-level imbalances in the number of men and women create “marriage squeezes” of differing degrees that alter men and women’s likelihood of marriage. We examine how the relative numbers of men and women in Mexican communities are related to propensities to marry by merging prospective individual-level data from the Mexican Family Life Survey (MxFLS) with community-level data from the 2000 Mexican Census. Results from logistic regression analysis models show that men’s probability of marrying is positively associated with the percentage of the municipality population ages 15 to 39 who are women, but the sex composition of the municipal population is not significantly associated with women’s likelihood of marrying.
Our study makes several theoretical and empirical contributions to research on sex ratios and marriage behavior. We find support for demographic-opportunity theory, but not sociocultural approaches, as an explanation for Mexican men’s marriage behavior. Men are more likely to marry, not less, when women in their community are plentiful. We also find that local sex ratio imbalances, or the relative numbers of eligible men in the community, are not associated with women’s propensity to marry, lending support for a more gendered argument for Mexican women’s marriage behavior. In many Mexican communities, marriage pools for women are likely substantially larger, and much more likely to be transnational, than they are for men. Consequently, the sex ratio of the local residential community may not adequately capture the pool of potential partners available to women because many of these men reside, if only temporarily, in the U.S.
Studying the marital consequences of skewed sex ratios in Mexico may provide stronger grounds for causal inference than comparable studies in countries such as China and India. In the cases of China and India, a shortage of women is generated primarily by a preference for sons and by a patriarchal gender structure that devalues women; these forces might also create pressures for early marriage, especially for women. Consequently, inferring a causal effect of the sex ratio on marital transitions in these cultural contexts is difficult. In Mexico, however, imbalances in the adult sex ratio result mainly from sex differences in migration, which are likely to be less closely tied to marriage behavior than is son preference. Accordingly, although inferring causal effects from observational data always entails uncertainty, we are likely on stronger ground inferring a causal effect of skewed sex ratios on marriage behavior in Mexico compared to other countries with imbalances in the numbers of women and men.
The prospective, longitudinal data used in this study may also provide greater analytical leverage than the cross-sectional survey data used in prior work on sex ratios and marriage. One difficulty in attempting to assess the impact of an imbalanced sex ratio on many family behaviors with cross-sectional survey data is that the sex ratio that respondents are exposed to at the time of the survey is not necessarily the same as the sex ratio to which respondents were exposed when they were at risk of experiencing the behavior in question such as marriage. For example, survey respondents who moved from one community to another during their young adult years may have been exposed to a much different sex ratio when they were initially at risk of marrying than the sex ratio they are observed to be exposed to at the time of the survey. Moreover, even for respondents who remain in the same community, the sex ratio itself may change over time. Thus, prior studies of the impact of imbalanced sex ratios on family behavior likely suffer from nontrivial measurement error in the sex ratio that is most relevant to the outcome. In contrast, by drawing on prospective longitudinal data from the Mexican Family Life Survey, this study largely overcomes this problem.
We acknowledge that the data used in this study are not without limitations. Mexican municipalities may not adequately circumscribe the spatial boundaries within which marital partners are chosen, and the use of alternative geographic units could generate different results. Some marriage-seekers might seek partners within a more limited geographic range, while others might seek partners outside of their municipality of residence. Moreover, these spatially circumscribed pools of marriageable mates likely vary in important ways between Mexican men and women. As suggested above, for Mexican women the municipality sex ratio may not adequately capture the supply of available men who serve as potential marriage partners because some of these potential male partners reside outside of the municipality—and particularly in the United States. Although U.S. census data do not allow us to trace the municipality origins of Mexicans who have migrated to the U.S., future research would benefit from exploring women’s marriage behavior in Mexico using data that captures both their local and wider—likely transnational—marriage pool. More generally, future research might profit by examining the impact of mate availability in geographic units both smaller and larger than the municipality.
Despite these limitations, our research on the consequences of sex ratio imbalances for marriage in Mexico begins to lay the foundation for a comparative, systematic study of how the effects of sex ratio imbalances on family behavior vary across social, economic, and cultural contexts. Some theory suggests that the effects of imbalanced sex ratios on socio-demographic behavior—and particularly gendered socio-demographic behavior—vary by other characteristics of the population being studied (Guttentag & Secord, 1983). In future research, our findings for Mexico might be combined with similar research on other countries characterized by imbalanced sex ratios, such as China and India, to explore cross-national variation in how the sex ratio shapes family-related behavior. Focusing on Mexico—a case with a much different culture and gender stratification system and in which imbalanced sex ratios are generated by migration rather than fertility-related behavior—may help us to understand the broader socioeconomic and cultural conditions under which imbalanced sex ratios matter most for family behavior.
Imbalances in the relative numbers of men and women in Mexican communities generated by high male emigration to the U.S. have the potential to not only shape critical dimensions of women’s and men’s family behavior, including their propensity to marry, but also to impact their subsequent likelihood of migrating to the U.S. or returning home to Mexico. For example, Mexican men’s relative inability to marry when women are scarce may propel these men to emigrate to the U.S. Accordingly, learning more about how imbalanced sex ratios in Mexican communities are related to Mexican women’s and men’s family behaviors may help anticipate family change in a bordering country that has direct relevance to population change in the U.S.
Footnotes
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.
