Financial Transfers to Husbands’ and Wives’ Elderly Mothers in Mexico: Do Couples Exhibit Preferential Treatment by Lineage?

Abstract

The aim of this study was to contrast the likelihood that a husband's elderly mother receives financial assistance from a couple with that of a wife's mother. Prior U.S.-based research has documented a strong bias toward transfers to wives’ parents. The authors aimed to extend this literature to Mexico, where financial help from adult children is a critical source of support for a rapidly aging population lacking institutional assistance. The authors’ approach to modeling competition between mothers accounted for the nature of their need. The results demonstrate that among mothers of similar financial need, a husband's mother is twice as likely to receive financial assistance as a wife's mother. In contrast, when faced with personal care needs, a wife's mother is disproportionately favored. These results reflect gender differences in Mexican adult children's responsibility for family members’ financial and physical well-being. The findings uncover new complexity in the patterns by which couples transfer money to parents of different lineage.

As Mexico continues to experience rapid increases in longevity at old ages, midlife Mexicans are likely to find themselves subject to strong claims by ascending generations (Grundy and Henretta 2006; Rosenthal, Martin-Matthews, and Matthews 1996; Uhlenberg 1993) and to competing obligations for assistance between their own aging parents and aging parents acquired through marriage. In this study, we examined financial transfers from middle-aged Mexican couples to their elderly mothers. We contrast the likelihood that a husband's mother receives monetary assistance from a couple with that of a wife's mother.

Primarily on the basis of U.S. data, prior studies have shown important differences by parental lineage in married couples’ upward transfers of resources. Consistent evidence indicates that when faced with the competing needs of husbands’ and wives’ parents, couples strongly favor wives’ kin (Soldo, Wolf, and Henretta 1999).

Using data from the Mexican Health and Aging Study (MHAS), our aim was to extend the existing literature on married adult children's support for elderly parents to the Mexican context, where the population is very rapidly aging (Palloni, Pelaez, and De Vos 2000; Wong and DeGraff 2007), and state-sponsored old-age assistance is lacking (Cutler et al. 2000). Projections estimate that the share of Mexicans aged 60 years and older will rise from 7% in 2000 to 24% in 2050, placing Mexico among the 10 countries with the largest elderly population in the world (United Nations 2001). Although the survival of midlife adults’ ascending generations of kin is not yet the norm in Mexico, it is rapidly becoming common. Compared with developed countries, the proportions of elderly Mexicans covered by social security and public health programs are extremely limited. About half the population aged 65 years and older is without formal health insurance plans, and three quarters do not receive any type of oldage pension (Instituto Nacional de Estadística, Geografía e Informática 2000). Given this lack of public assistance, Mexican parents frequently rely on adult children as their main source of support in old age. Wong and Espinoza (2002) showed that assistance from kin, primarily adult children, amounts to approximately 35% of the total income of elderly Mexicans. Beyond structural constraints leading to elderly Mexicans’ reliance on middle-aged generations, Mexico is also characterized by strong norms of filial responsibility that promote the expectation that adult children will provide emotional and material support to elderly members of their families (Bridges 1980; Hanratty 1997). Finally, relatively low levels of socioeconomic status could foster intense competition for resources between husbands’ and wives’ parents.

Couples’ monetary assistance to their elderly mothers can be motivated by a variety of needs. Therefore, central to our examination of couples’ financial transfers to husbands’ and wives’ mothers was the distinction between assistance provided in response to mothers’ financial needs and support for mothers’ health-related needs. In addition, while controlling for couples’ economic resources and competing claims on those resources, we accounted for transfers from husbands’ and wives’ adult siblings (Soldo 1996). Transfers from members of a couple's extended kin may define notions of expectations, elicit reciprocation between relatives sharing the burden of elderly care, or lead couples to diminish support to an elderly mother as she relies on assistance by others (Stoller and Earl 1983; Wolf 1994). We focused on elderly mothers because due to women's greater life expectancy and age differences between husbands and wives, Mexican mothers are more likely than Mexican fathers to survive until their children have reached middle age. In addition, Mexican mothers are typically more frail, less economically well off, and thus more dependent on assistance from adult children (Gomes 2007).

The results indicate that spouses’ preferences for monetarily assisting mothers of one or the other lineage is strongly contingent on the nature of the need being addressed. We show that among mothers of similar financial need, a husband's mother is twice as likely to receive financial assistance as a wife's mother. In contrast, when faced with personal care needs, a wife's mother is disproportionately favored. We argue that resolution of competition between a husband's mother and a wife's mother for a couple's financial support is determined largely by gender differences in Mexican adult children's propensities to care for family members’ financial and personal needs. Contrasting with prior research, our findings uncover new complexity in the patterns by which couples transfer money to parents of a different lineage.

Previous Literature

Starting with the work of Glick (1957) and Sweetser (1963, 1966) in Europe and the United States, researchers have documented asymmetry in couples’ relations with husbands’ and wives’ kin. Consistent evidence indicates that couples strongly favor exchanges with wives’ relatives over interactions with husbands’ kin (for an interesting exception, see Barbagli 1996). The asymmetry in couples’ relations with each partner's kin is unexpected given the prevailing bilateral systems of kinship in most Western nations (Kuznesof 2005), which suggest that married adult children display similar levels of interaction with both marital partners’ kin and that few distinctions are made between a wife's and a husband's relatives. A couple's priority for interaction with relatives in the wife's line of kinship in the United States has been demonstrated using a variety of measures, including exchange of various kinds of help, frequency of contact (Sweetser 1984), and willingness to assist in case of need (Rossi and Rossi 1990).

Matrilateral asymmetry in couples’ interactions with relatives reflects both variations in the patterns of family obligations and strong gender asymmetry in husbands’ and wives’ involvement with kin. Rossi and Rossi (1990) reported that family members obtained by marital ties evoke lower levels of obligation than consanguineal kin in comparable positions. These findings are consistent with those of a recent study of gender differences in support to parents and parents-in-law indicating that women have significantly greater contact with their parents than men but that men have more frequent contacts with in-laws than women (Lee, Spitze, and Logan 2003). Although women's greater contact with relatives is mostly limited to their original kin groups, men get drawn into frequent contact with their in-laws, through their wives’ influence (Rossi and Rossi 1990). Various studies have argued that the bond between daughters and mothers is particularly strong and long lasting (Fischer 1983; Merrill 1993). Fischer (1986) proposed that although the lives of mothers and daughters are linked from the time of birth, in-law relationships have shorter histories and are contingent on the endurance of the marriage. This results in less familiarity, intimacy, and emotional involvement in a wife's relationship to her husband's parents than to her own.

Compounding variations in marital partners’ perceived obligations to their own parents and to parents obtained by marriage is wives’ greater involvement in relations with relatives. Daughters have been described as “kin keepers” charged with maintaining ties with couples’ extended families (Fischer 1983; Willson, Shuey, and Elder 2003). They are also significantly more likely to assist parents than sons (Stoller 1983). Researchers have accumulated evidence of a strong gender division in the provision of care to parents, especially regarding daughters’ role as the main providers of regular personal assistance to frail elderly parents (Matthews and Rosner 1988; Stoller 1990).

Empirical research directly investigating competition between a husband's and a wife's elderly parents for a married couple's assistance is relatively scarce. As far as we are aware, no study has addressed the issue in the Mexican context. Some studies have examined competition between marital partners’ parents for coresidential space with the couple in various contexts (Aykan and Wolf 2000; Soldo et al. 1999; Sweetser 1984). Sharing of coresidential space between a couple and both partners’ parents is uncommon, even in more traditional societies (Aykan and Wolf 2000). Thus, a couple is typically required to opt for one or the other partner's parents. Using data from the Health and Retirement Study, Soldo et al. (1999) found that although coresidence with an elderly parent is unusual, a wife's unmarried mother is significantly more likely to share a household with a couple than a husband's unmarried mother. Even in cases in which a husband's mother is frail or does not have other alternatives for coresidence, and the wife's mother does, the odds of coresidence with the couple remain in favor of the wife's mother.

In the only study of differences by lineage in couples’ transfers of assistance in the form of money and time to elderly parents, Shuey and Hardy (2003) showed that couples favor wives’ parents in their transfers of help with basic activities of daily living (ADLs) and monetary gifts or loans. This is particularly true among Hispanic and African American couples. Unfortunately, the study's contribution to understanding the specific patterns of couples’ financial assistance to husbands’ parents and wives’ parents is limited. Because of restrictions of the data, the authors did not distinguish between transfers of money and transfers of personal assistance to elderly parents. Although the balance of transfers of personal care favors elderly parents in the United States, intergenerational transfers of financial resources tend to strongly benefit adult children, and upward financial transfers are relatively rare (Sloan and Zhang 1995).

Thus, prior research on couples’ transfers to each spouse's parents has directly or indirectly focused on couples’ provision of personal care, a markedly female currency of intergenerational exchanges. Although coresidence involves both members of a couple, it typically follows increases in a parent's level of physical or cognitive disability and is often initiated by the elderly parent's need for personal assistance (Bishop 1986; Wolf 1990). The gender division of caregiving and kin-keeping roles suggests that such assistance from couples is most likely to flow along matrilineal rather than patrilineal lines (Shuey and Hardy 2003).

Because of the distinctively gendered nature of parental personal care, we broaden our understanding of competition between a husband's and a wife's parents by focusing on a couple's financial transfers. Couples’ monetary assistance can serve both to relieve elderly parents’ financial burden and as a substitute for caregiving when adult children are unable to assist with their parents’ personal needs. For instance, Couch, Daly, and Wolf (1999) showed that high-earning couples that incur a high time price rely relatively more on providing cash transfers to assist aging parents with the purchase of personal care than on providing personal care themselves.

Most middle-aged Mexicans have at least one sibling with whom to share filial responsibilities. Situating couples’ decisions to transfer money to mothers within a broad array of intergenerational exchanges has the conceptual advantage of providing insight into the distribution of parental burdens across midlife adults linked by blood or by marriage (Soldo 1996). Recent empirical evidence suggests that monetary assistance to each partner's mother is likely to be affected by the transfer behavior of his or her siblings. Antman (2007) provided evidence that Mexican adult children's time and money contributions to elderly parents are strongly influenced by their siblings’ contributions. The effect of siblings’ contributions on respondents’ transfers of money to elderly parents varies with the nature of siblings’ assistance. Siblings’ monetary gifts to their parents elicit increased monetary transfers from respondents. However, siblings’ transfers of time to their parents result in decreased monetary assistance from respondents. Following Antman, we distinguished between the types of assistance contributed by a marital partner's siblings.

Methods

We analyzed data on adult children and their elderly mothers collected in the MHAS. The analytical sample consisted of couples interviewed in 2001 during the first wave of the study. We examined recent monetary assistance from a marital unit to a husband's and a wife's mother. Our analytical strategy consisted in estimating a modified logit model that accommodated situations in which only one spouse's mother was living as well as circumstances in which both mothers were alive. While controlling for a number of confounding variables, we examined the effect of elderly mothers’ characteristics on couples’ likelihood of monetary assistance.

Data and Analytical Sample

MHAS is an ongoing panel study aimed at investigating a broad range of issues affecting the Mexican elderly population, including intergenerational transfers, health, migration, and economic well-being. MHAS is nationally representative of the roughly 13 million Mexicans born before 1951. In 2001, 15,186 Mexicans aged 50 years and older and their spouses or partners were interviewed. Proxy interviews were sought when poor health or temporary absence precluded direct interviews (7% of all interviews were conducted with proxy respondents). All interviews were conducted in person. The survey was based on a multistage stratified sample identified in conjunction with the 4th Quarter 2000 National Employment Survey/Encuesta Nacional de Empleo (ENE), a nationally representative survey conducted by Instituto Nacional de Estadística, Geografía e Informática, the counterpart of the U.S. census. ENE covers both urban and rural residents in all 32 Mexican states. The entire MHAS sample was drawn from the 64,475 ENE households, 40% of which contained at least 1 person eligible for MHAS. The response rate was 90.1%, a figure similar to that achieved in comparable U.S.-based surveys, such as the Health and Retirement Study (Wong and Espinoza 2005). In addition to information on themselves, each interviewee provided information regarding a parent's survival status, demographic background, health, disability, socioeconomic status, and living arrangements.

MHAS interviewed 5,307 couples who were married or in a consensual union. The analytical sample was restricted to couples in which at least one of the spouses’ mothers was alive. To take into account siblings’ assistance in our investigation of couples’ upward financial transfers, we further retained only couples for whom both the husband and the wife had at least one living sibling. These restrictions yielded a sample of 1,757 couples. For 450 of these couples (26%), only the husband's mother was alive; for 877 couples (50%), only the wife's mother was alive; and for 430 couples (24%), both the husband's and the wife's mothers were alive.

Modeling Strategy

Our measure of couple-to-mother financial assistance covered recent financial transfers. Respondents were asked, “In the last two years, have you (and/or your spouse) given financial assistance to your parent(s)? Include help to pay costs such as rent; exclude shared housing or shared meals.” The formulation of the question implies that respondents reported on explicit financial transfers aimed at assisting parents with relatively large financial burdens.

We estimated a modified multinomial logit model to examine the relationship between financial transfer to elderly mothers and various independent variables. Our model simultaneously accommodated all possible patterns of survivorship among a husband's and a wife's mother (for examples using the modified multinomial logit model, see Aykan and Wolf 2000; Ofstedal 1995; Soldo et al. 1999; Tomassini, Wolf, and Rosina 2003; Wolf and Soldo 1988). Rather than focusing only on the small proportion of couples in which both spouses had living mothers, this approach used useful information on transfers by couples with a single living mother. Furthermore, our modeling strategy allowed us to generalize our findings to the entire Mexican population of married adults with surviving mothers.

Each couple potentially faced up to four options for a financial transfer to their mothers. The outcome variable represented unordered categories distinguishing between these possible options: a transfer to the wife's mother only, a transfer to the husband's mother only, transfers to both mothers, or a transfer to neither mother. A couple's option set was a function of whether either or both a husband's mother and a wife's mother were alive. If only one of the mothers was alive, a couple faced two options: either providing financial assistance to her or providing no assistance. Conversely, if both mothers were alive, a couple had four options: a transfer to the husband's mother only, a transfer to the wife's mother only, transfers to both mothers, or a transfer to neither mother.

The probability that couple i was in category j of n categories of the outcome variable (n = 2 or 4) was given by

$$\mathrm{Pr}(Y_i=j)=\frac{M_j{e}^{{\theta }_j}}{1+M_{\mathrm{h}}{e}^{{\theta }_1}+M_{\mathrm{w}}{e}^{{\theta }_2}+M_{\mathrm{h}}{M}_{\mathrm{w}}{e}^{{\theta }_3}},$$ (1)

where M_h and M_w are dummy variables taking a value of 1 if the husband's mother and the wife's mother were alive, respectively, and a value of 0 otherwise. The mothers’ survivorship dummies ensured that the model accommodated in a single expression all possible survivorship circumstances of the mothers. If both partners’ mothers were alive (M_h = 1 and M_w = 1), the above model was a four-category multinomial model (n = 4). For a couple with, for example, a surviving wife's mother and a deceased husband's mother, M_h would take a value of 0, resulting in the disappearance of two of the terms in the denominator and thus simplifying the model into a binary logit (n = 2). Our model assumed that the effects of one mother's characteristics on the outcome variable were independent of the survival status of the other mother.

In equation 1, the arguments of the exponential functions relate to the vector X of explanatory variables as follows:

$$\begin{array}{cc}\hfill {\theta }_1& ={\beta }_1\mathbf{X},\hfill \\ \hfill {\theta }_2& ={\beta }_2\mathbf{X},\hfill \\ \hfill {\theta }_3& ={\beta }_1\mathbf{X}+{\beta }_2\mathbf{X}+{\beta }_3\mathbf{X}.\hfill \end{array}$$ (2)

Hence, β₁ is the effect of X on the log odds of a couple's making a financial transfer to the husband's mother only versus assisting neither. Similarly, β₂ is the effect of the explanatory variables on the log odds of a couple's making a financial transfer to the wife's mother only relative to making no financial transfer. The interpretation of β₃ is slightly more complicated. It is the additional effect, beyond that of β₁ and β₂, of a particular variable on the log odds of a couple's transferring money to both mothers relative to neither. Statistically significant β₃ coefficients indicate that financial transfer to one of the mothers is dependent on financial transfer to the other mother.

MHAS oversampled six Mexican states accounting for 40% of all outmigration to the United States (Zacatecas, Guanajuato, Michoacán, Jalisco, Nayarit, and Durango) at a rate of slightly less than two to one. In preliminary models, we controlled for a couple's residence in either of these states. Because the effect of couple's location was not statistically significant, we excluded these variables from our final model. Furthermore, following DuMouchel and Duncan (1983), we present results of unweighted models. Unweighted models ensure greater statistical efficiency because there are no differences in couples’ likelihood of financial transfers to each spouse's mother across states with high and low outmigration.

Explanatory Variables

The multinomial model included three sets of variables measuring characteristics of each elderly mother, the couple, and each spouse's set of siblings. We were interested primarily in the effect of a mother's financial and disability characteristics on her own odds of receiving financial assistance (direct effect). In addition, we examined the effect of characteristics of one spouse's mother on the couple's likelihood of transferring money to the other spouse's mother (cross-effect).

Mother variables

Transfers from adult children to elderly parents have been found to increase with parental needs in a wide variety of settings (Cox, Eser, and Jimenez 1998; Lillard and Willis 1997). For instance, in Mexico, elderly mothers in poor health are significantly more likely to receive assistance from adult children than elderly mothers in good health (Petrova 2004). Conditional on her survival, a mother's socioeconomic resources were represented by whether she received any formal education and by a categorical measure of her financial situation. We determined a mother's disability status on the basis of her reported need for assistance with basic ADLs, such as dressing, eating, and bathing. Mother's age was included in our model because it is an important factor in determining overall health status and subsequent need for assistance by adult children. We also controlled for the marital status of the mother. Among elderly Mexicans, husbands are much more likely than wives to continue working for pay or to receive pensions in old age, thus making the presence of a spouse a strong correlate of an elderly women's financial well-being in later years (Gomes 2001, 2007). The availability of a husband to assist with physical tasks and to provide emotional support may also reduce adult children's perceptions of filial responsibility toward their mothers.

In our analysis, we considered potential substitution between monetary assistance and other types of assistance provided by a couple to elderly mothers by including two measures of the couple's nonfinancial transfers to both partners’ mothers. Transfers of time involve help with an elderly mother's personal care provided by either member of a couple. In addition, we controlled for each mother's current coresidence with a couple.

Couple variables

Several variables reflected a couple's economic and financial resources as well as the burden placed upon them. A couple's economic resources were measured as the husband's education, the couple's net worth, and past migration to the United States. We measured education as the number of years spent receiving formal schooling. We included only husbands’ education because of strong correlations between spouses’ educational attainment. A couple's net worth is a broad measure of its economic well-being, including income and pension; total assets such as bank accounts, houses, and land; and debt. Our model included a measure of whether either marital partner ever worked or lived in the United States. An extensive literature on migration from Mexico to the United States documents the importance of financial transfers from U.S. migrants to nonmigrating family members, particularly to elderly parents (Gomes 2001; Wong, Soldo, and Capoferro 2000). Petrova (2004) found that the effect of the number of children working in the United States on financial transfers to elderly parents is much greater than the effect of the number of children working in Mexico. The inclusion of variables measuring a couple's socioeconomic status relied on the assumption that financial transfers from adult children are most likely when resources are abundant and claims on those resources are limited (Lillard and Willis 1997). In general, we expected to find a positive relationship between couples’ resources and financial transfers to elderly mothers. Direct claims on a couple's resources were represented by a simple count of the couple's total number of children (Soldo et al. 1999). We included in this figure the couple's surviving biological, adopted, and stepchildren.

We controlled for measures of the relative distribution of financial decision power within a couple. These included whether the wife had equal or the most say in important family decisions (the reference was that the husband had the most say) and whether she ever participated in paid employment. We expected both measures to be negatively associated with the odds of making a financial transfer to the husband's mother. We included husband's age in the model as a proxy for the couple's life cycle. Young couples may carry a fair amount of responsibility for their parents in the early years of their marriages. However, they may shift that burden to other family members as their own familial responsibilities grow and later return to caring for parents (Gomes 2007; Hancéoglu 1985).

Sibling variables

Although individual data on each spouse's siblings are not available, MHAS contains information on the size of each partner's sib-ship. Increases in a marital partner's number of siblings can be expected to reduce a couple's burden of financial assistance to that spouse's mother. In addition, each marital partner reported on collective transfers from his or her siblings to their parents. We included three measures of transfers from siblings to their respective mothers: current coresidence with a mother, financial assistance with a mother's financial burdens, and help with basic personal activities related to a mother's difficulties performing ADLs. Finally, we tested for the effect of an agreement to care for elderly parents between a spouse and his or her siblings on a couple's likelihood of assisting a husband's or a wife's mother. Respondents were asked, “Have you and one or more of your siblings agreed to share the financial responsibilities for your parents?” Such an agreement may allow a couple to free resources for the other spouse's mother, or it may bind them into preferentially supporting the beneficiary of the agreement. Because we expected that assistance from extended kin would influence couples’ financial transfers to each partner's mother, we were interested in both direct effects and cross-effects of the sibling variables. Direct effects of sibling variables represent the effect of one partner's siblings’ assistance on a couple's odds of financially assisting that partner's mother. Cross-effects allowed us to examine the importance of transfers from one partner's siblings to his or her own mother on the couple's odds of making a financial transfer to the other partner's mother.

Results

Descriptive Results

Consistent with well-documented age differences between marital partners, a greater proportion of wives (74%) than husbands (50%) in our sample reported having surviving mothers. Although overall, large proportions of mothers received monetary assistance from their married children, mothers related through the husbands received such assistance in notably greater proportions (57%) than mothers related through the wives (46%).

Table 1 presents descriptive information for the explanatory variables used in the multivariate model, separately for couples in which the husband's mother was alive and couples in which the wife's mother was alive. The wives’ and husbands’ mothers in our sample were, on average, 75 and 79 years old, respectively. Reflecting these age differences, husbands’ mothers were slightly more likely to require assistance with ADLs than wives’ mothers. We found that levels of coresidence between married adult children and either of the partners’ mothers were relatively low, comparable with those found in the United States (Soldo et al. 1999). Coresidence of married adult children with elderly parents is not the norm in Mexico, because Mexican children typically leave their parents’ homes upon marriage and establish independent households, often in proximity to their elderly parents (Gomes 2007). There were some important differences between husbands’ and wives’ mothers’ receipt of personal care from their married children. Although husbands’ mothers required slightly more personal care than wives’ mothers, the proportion of husbands’ mothers receiving caregiving from the couples (7%) was approximately half the proportion of wives’ mothers receiving personal help from the couples (13%).

Table 1.

Summary Statistics for Explanatory Variables Used in the Multinomial Logit Model, Mexican Health and Aging Study, 2001

Variable	Husband's Mother Alive	Wife's Mother Alive
Couples’ characteristics
Husband's years of schooling	6.9	6.4
Net worth quartile	2.7	2.7
Either partner ever migrated to United States	15.4	15.9
Number of children	4.5	5.0
Husband's age	54.9	56.7
Wife has most or equal say in family decision	78.5	78.6
Wife ever held paid job	34.0	34.9
Mothers’ characteristics
Any formal education	54.2	55.8
Poor financial situation	15.2	18.0
Fair financial situation	57.0	55.5
Good, very good, or excellent financial situation	27.8	26.5
Needs assistance with ADLs	23.7	22.0
Age	79.3	74.7
Married	33.7	39.2
Personal care from couple	7.2	13.4
Coresident of the couple	3.4	2.4
Siblings’ characteristics
Size of sibship	6.2	5.5
Monetary transfer to mother	74.4	71.4
Personal care to mother	26.4	25.2
Coresident of mother	62.2	53.9
Agreement to care for mother	39.5	38.9
Number of couples	880	1,307

Note: Means are shown for continuous variables, and percentages are shown for dummy variables. Variables describing siblings refer to husband's siblings in the first column and to wife's siblings in the second column. Couples with both surviving mothers are accounted for twice, once in each column. ADL = activity of daily living.

Each marital partner in our sample had on average five to six siblings with whom to share a mother's financial burden. Significant proportions of spouses’ siblings were involved not only in providing financial help for their elderly mothers but also as personal caretakers. Consistent with findings by Gomes (2007), a nontrivial proportion of all respondents had agreements with their siblings to share their mothers’ financial burdens (39%).

Figure 1 shows changes in couples’ patterns of upward financial transfers according to variations in either mother's financial and disability related needs. We classified mothers into four categories: mothers in poor financial situations, mothers in fair or better financial situations, disabled mothers, and mothers without disabilities. The proportions receiving monetary assistance were appreciable, even when mothers were in little financial need or when they were not disabled. At its highest, the proportion of mothers receiving monetary help was 63% among poor mothers related through husbands. Not taking into account competition between spouses’ mothers, we found that husbands’ mothers received monetary assistance in significantly greater proportions than wives’ mothers, regardless of need. Compared with mothers with no disability-related needs, the proportions receiving monetary assistance increased only among disabled wives’ mothers but remained constant among disabled husbands’ mothers.

Figure 1.

Proportions of mothers receiving monetary assistance from their married children according to financial and physical needs, by lineage, Mexican Health and Aging Study, 2001.

Multivariate Analysis

The results of the multivariate analysis are reported in Table 2. Coefficients with positive signs indicate increased odds of couples’ financial transfers, whereas negative signs indicate decreased odds of couples’ financial support. The likelihood of a couple's monetary assistance to each partner's mother decreased with improvements in mothers’ financial situations. However, our results indicate that married children were more sensitive to husbands’ mothers’ financial needs than to wives’ mothers’ financial needs. The coefficients for the effect of financial situation on the transfer of money to wives’ mothers were appreciably larger in magnitude than those for transfers to husbands’ mothers. This suggests that married children reduced their financial transfers to well-off mothers related through wives to a greater extent than they did to well-off mothers related through husbands. In addition, we found that couples decreased their monetary transfers to wives’ mothers at lower levels of financial well-being relative to husbands’ mothers. In contrast, we found that couples responded financially to wives’ mothers’ need for assistance with ADLs but not to husbands’ mothers’ need for personal care.

Table 2.

Multinomial Logit Coefficients and Standard Errors for Couples’ Financial Transfers to Elderly Mothers, Mexican Health and Aging Study, 2001

	Husband's Mother Only		Wife's Mother Only		Both Mothers
Explanatory Variable	Coefficient	SE	Coefficient	SE	Coefficient	SE
Couple's characteristics
Husband's years of schooling	–0.021	0.022	–0.005	0.017	0.079	0.037*
Net worth quartile (vs. lowest quartile)
Second quartile	0.091	0.273	0.601	0.217**	–0.780	0.494
Third quartile	0.119	0.268	0.422	0.219*	–0.724	0.467
Fourth quartile	0.540	0.274*	0.730	0.231**	–0.809	0.473
Either partner ever migrated to United States	–0.501	0.242*	–0.347	0.195	0.912	0.410*
Number of children	–0.005	0.038	–0.018	0.027	–0.022	0.068
Husband's age	–0.008	0.019	0.015	0.012	–0.045	0.035
Wife has most or equal say in decisions	–0.168	0.214	–0.085	0.174	0.287	0.364
Wife ever held paid job	0.623	0.188**	0.285	0.150	–0.173	0.338
Husband's mother's characteristics
Any formal education	–0.048	0.194	0.488	0.369	–0.706	0.429
Financial situation (vs. poor)
Fair	–0.291	0.259	0.390	0.529	0.620	0.617
Good, very good, or excellent	–0.679	0.300*	0.780	0.598	–0.201	0.712
Needs assistance with ADLs	–0.263	0.255	0.740	0.433	–0.471	0.527
Age	0.009	0.014	–0.026	0.008**	0.035	0.022
Married	–0.413	0.214*	0.154	0.364	–0.101	0.443
Personal care from couple	0.954	0.454*	1.254	0.717	–0.826	0.789
Coresident of the couple	–0.355	0.537	0.356	0.728	0.290	0.960
Wife's mother's characteristics
Any formal education	0.322	0.300	0.237	0.154	–0.585	0.372
Financial situation (vs. poor)
Fair	0.225	0.422	–0.830	0.308**	–0.258	0.528
Good, very good, or excellent	0.454	0.471	–1.205	0.337**	0.089	0.591
Needs assistance with ADLs	–0.366	0.454	0.442	0.200*	–0.329	0.543
Age	–0.015	0.007*	–0.007	0.008	0.039	0.017*
Married	–0.094	0.298	–0.165	0.291	0.584	0.427
Personal care from couple	0.382	0.531	0.443	0.247	–0.373	0.584
Coresident of the couple	1.850	0.796*	0.540	0.474	–3.633	1.319**
Husband's siblings’ characteristics
Size of sibship	–0.072	0.030*	–0.010	0.025	0.124	0.057*
Monetary transfer to mother	1.737	0.221**	–0.149	0.374	0.147	0.505
Personal care to mother	–0.070	0.256	–0.248	0.465	0.304	0.542
Coresident of mother	–0.421	0.212*	0.525	0.401	–0.287	0.460
Agreement to care for mother	0.845	0.196**	0.597	0.380	–0.578	0.432
Wife's siblings’ characteristics
Size of sibship	–0.025	0.030	–0.076	0.025**	–0.019	0.056
Monetary transfer to mother	0.575	0.319	1.842	0.180**	0.363	0.502
Personal care to mother	–0.246	0.427	–0.645	0.208**	0.353	0.499
Coresident of mother	–0.382	0.309	–0.147	0.148	0.483	0.363
Agreement to care for mother	0.030	0.328	0.581	0.151**	0.309	0.374
Constant	–0.402	1.261	–1.205	0.910	–2.856	2.519

Note:ADL = activity of daily living.

^*p ≤ .05.

^**p ≤ .01.

Further evidence of imbalance between husbands’ and wives’ mothers is suggested by the strong effects of a wife's mother's coresidence with a couple on its financial transfer to the husband's mother. According to our model, couples compensated husbands’ mothers with financial transfers when wives’ mothers shared households with them. However, the reverse was not true.

In general, we found little evidence that one mother's characteristics affected a couple's monetary transfers to the other mother. However, our results provide suggestive evidence that couples took into account one mother's age in allocating money to the other mother. However, the negative cross-effects of a mother's age were relatively small in magnitude.

Transfers from siblings were among the strongest predictors of a couple's monetary transfers to their mothers. Consistent with previous research, monetary transfers from a spouse's siblings elicited financial transfers from a couple to that partner's mother. Conversely, mothers who received personal care from siblings or coresided with siblings were less likely to receive money from the couple. Marital partners’ agreement to share a mother's financial burden increased a couple's odds of transferring money to her. There was no association between couples’ monetary transfers to elderly mothers and the transfer behavior of siblings-in-law. Because all siblings’ cross-effects were insignificant, our analysis suggests that for each mother, the kin group relevant to her receiving financial support from the couple was restricted to her own children.

Most coefficients on the remaining variables included in the model were in the expected direction. Our results confirm the anticipated role of economic resources in couples’ transfers to their mothers. Migration to the United States and husband's education were associated with increases in couples’ odds of transferring money to both mothers.

Our findings do not support the argument that increases in wives’ relative decision-making power within couples lead to increases in the odds of making financial transfers to the wives’ mothers. Controlling for other variables, we did not find a statistically significant difference in the odds of a monetary transfer to either mother between couples in which the wife had most or equal say in family decisions and couples in which the husband had the most say. Contrary to what we had anticipated, couples in which wives had ever held paying jobs were more likely to elect husbands’ mothers as the sole recipients of their monetary assistance.

To facilitate the interpretation of our multinomial results, Table 3 presents computed probabilities of couples’ transfers to husbands’ mothers, wives’ mothers, or both mothers under various scenarios. The probabilities were derived by substituting the estimated values of the parameters β₁, β₂, and β₃ (shown in Table 2), along with several alternative values of the X parameters first into equation 2 and then into equation 1. Although it was possible to estimate the probability of a financial transfer associated with any characteristic of a couple, mothers, and siblings, the scenarios presented in Table 3 were chosen because they illustrate important differences between husbands’ and wives’ mothers’ likelihoods of receiving monetary assistance.

Table 3.

Predicted Probabilities of Financial Transfer From Couples to Their Elderly Mothers, by Mothers’ Financial Needs and Mothers’ Needs for Personal Care, Mexican Health and Aging Study, 2001

	Probability of Transfer
Variable	To Husband's Mother Only	To Wife's Mother Only	To Both Mothers
Baselinea	.110	.044	.003
Panel A: Effects of mother's financial needb
Poor financial situation
Husband's mother only	.200	.019	.003
Wife's mother only	.065	.138	.006
Both mothers	.130	.064	.007
Panel B: Effects of mother's need for personal carec
Needs assistance with ADLs
Husband's mother only	.083	.091	.003
Wife's mother only	.077	.070	.003
Both mothers	.056	.138	.003

Note: ADL = activity of daily living.

^aBaseline scenario: Both mothers are in good, very good, or excellent financial situations; neither mother has a need for assistance with ADLs; and neither set of siblings provided assistance to their mother. The couple is in the third net worth quartile. All other variables set to means (continuous variables) or to highest frequency (dummy variables).

^bBaseline scenario, varying mother's financial situation as indicated.

^cBaseline scenario, varying mother's need for personal care as indicated.

Baseline computed probabilities are presented in the first row of Table 3, which considers husbands’ and wives’ mothers who had no need for assistance with ADLs and who were in good financial situations. In addition, in this case, we assumed that neither set of siblings provided care or money to their mother and that neither spouse had an agreement to share his or her mother's financial responsibility with his or her siblings. In this and all other illustrations presented in Table 3, couples were assumed to be in the third net worth quartile, and all other variables were set to their mean values (continuous variables) or to the highest frequency (dummy variables) as reported in Table 1. Because they refer to mothers without either financial or physical needs, the baseline probabilities offer a convenient reference point with which to compare the effects of increases in mothers’ monetary and disability-related needs. Under these baseline conditions, a husband's mother is 2.5 times more likely to receive financial assistance from a couple than a wife's mother.

In panel A of Table 3, we alternatively consider hypothetical situations in which a husband's mother only, a wife's mother only, and both mothers are in poor financial situations. If only one mother was in financial need, her odds of receiving assistance were greater than in the baseline scenario. However, couples had a greater likelihood of assisting a poor husband's mother than a poor wife's mother when there was no competition between mothers. Assuming competition between mothers, a husband's mother was roughly twice as likely as a wife's mother in similarly dire financial need to receive assistance from the couple. Under no circumstance was a wife's mother of comparable financial need more likely than a husband's mother to receive financial assistance.

Panel B of Table 3 considers the effects of mothers’ need for assistance with ADLs on a couple's financial transfers. Compared with the baseline, a husband's mother experienced a slight decrease in her likelihood of receiving money when she was the only one to have a need for personal assistance. In contrast, when only a wife's mother had a health related need for care, her likelihood of receiving a financial transfer increased strongly. Comparing the probabilities of transfers when both mothers were in need of help with ADLs illustrates a wife's mother's greater likelihood of receiving a financial transfer used as substitute for caregiving. Relative to the baseline, when both mothers needed help with ADLs, a husband's mother's likelihood of receiving monetary assistance was roughly divided by two, whereas the wife's mother's likelihood of benefiting from a monetary transfer was multiplied by three.

Under few circumstances was the likelihood of simultaneously transferring money to both mothers appreciable. We found that across the scenarios considered in Table 3, the probability of a couple's transferring money to both a wife's and a husband's mother was between 3% and 7%.

Discussion

This study extends research on variations by parental lineage in middle-aged couples’ upward transfers to the Mexican context. Contrasting with prior research, our findings suggest substantial complexity in the patterns by which couples transfer money to husbands’ and wives’ parents. We show that married adult children are responsive to their elderly mothers’ financial and health-related needs, with poor and frail mothers being more likely to receive financial help from their married children. However, our findings indicate that Mexican couples have a very low likelihood of transferring financial assistance to both marital partners’ mothers. Our research has provided new evidence on the central role of the nature of the need being addressed by couples in shaping the likelihoods that husbands’ and wives’ mothers receive monetary assistance.

We show that among mothers of similar financial need, a husband's mother is twice as likely to receive financial assistance as a wife's mother. Our findings suggest that a wife's mother in fair or better financial situation is much less likely to receive help than a husband's mother in a similar situation, further reflecting a couple's greater likelihood of taking into account the financial needs of a husband's mother. In addition, couples reduce their financial transfers to wives’ mothers to a greater extent than they do to husbands’ mothers at similar levels of financial well-being. In contrast, when elderly mothers are faced with disability-related needs, Mexican couples disproportionately favor wives’ mothers. Our results indicate that a husband's mother's needs for personal assistance with ADLs do not elicit a couple's financial assistance.

These findings are consistent with Mexican husbands’ traditional roles as providers for their family's financial well-being and Mexican wives’ disproportionate involvement in the personal care of family members. Despite recent increases in women's participation in the labor force, Mexican husbands generally remain the primary breadwinners, while most of the child care and domestic labor falls to wives. Recent studies of transfers from adult children to their elderly parents in Mexico have reported that whereas Mexican daughters bear the primary responsibility for frail elderly parents’ personal care, Mexican sons are significantly more likely to provide monetary assistance to elderly parents than daughters (De Vos, Solís, and Montes de Oca 2004; Trujillo, Mroz, and Angeles 2007). Following Gomes (2007), we argue that in husbands’ and wives’ financial transfers to their elderly mothers, men reproduce their provider role, while women reproduce their domestic role.

Given husbands’ greater financial independence, we expected increases in wives’ decision-making power to heighten the likelihood that wives’ mothers would receive assistance from the couples. Our results suggest that the implications of husbands’ traditional role as breadwinners for couples’ upward transfers are resilient to variations in wives’ economic or decision-making power. We found no evidence that an increase in a wife's relative decision-making power within a couple is associated with a change in the couple's transfer behavior. Contradicting our prediction, our results indicate that a wife's past paid employment increases a couple's likelihood of a financial transfer solely to the husband's mother. It is possible that having a paid job does not necessarily enhance a wife's economic autonomy, especially for women who work out of dire financial need (Sathar and Kazi 1990). In addition, past employment may not necessarily translate into wives’ increased access to and control over common financial resources.

We confirm the strong importance of siblings’ transfers in shaping adult children's own financial transfers to elderly parents. We have evidence that siblings’ financial transfers increase couples’ upward financial assistance but that siblings’ transfers of time and space decrease couples’ upward financial help. Because nonfinancial assistance to elderly parents overwhelmingly involves daughters, there is a suggestion that marital partners are unlikely to compensate for sisters’ caregiving to elderly mothers. This is consistent with findings from earlier research suggesting that siblings exhibit a tendency to cut back on their own assistance to elderly parents in proportion to the number of sisters they have (Wolf, Freedman, and Soldo 1997). Unfortunately, because of limitations of the data, we were not able to control for characteristics of each partner's sibship beyond its size. More research is required to understand the relevance of sibships’ gender composition to a couple's monetary assistance to a husband's or to a wife's mother. Finally, our results suggest that couples’ financial assistance to a given mother is influenced primarily by her own kin group and that characteristics of the other mother are generally irrelevant.

Several further limitations of this study suggest avenues for future research. First, although the dependent variable used in our model measured monetary assistance with large financial burdens faced by elderly mothers, it explicitly excluded relatively common and smaller financial support in the form of shared housing or shared meals. Therefore, our results may underestimate the prevalence of couples’ financial assistance to their mothers. Moreover, if sons are more likely to support their mothers in these various ways than daughters, our results underestimate couples’ financial transfers to husbands’ mothers and consequently their preference of husbands’ mothers over wives’ mothers. More research is required to explore variations in couples’ likelihood of transferring small sums of money to husbands’ and wives’ parents.

An important limitation of this study relates to the cross-sectional nature of the data, which leads us to ignore the role of past and future transfers from elderly parents either separately to each marital partner or to a couple. Future research should explore the role of past and expected transfers from parents to children in shaping couples’ preference for transfers to mothers related through husbands compared with mothers related through wives. Previous work has shown that in a variety of contexts, including Mexico, prior receipt of assistance by adult children is a strong determinant of subsequent transfers from adult children to aging parents (Henretta et al. 1997; Noël-Miller 2007). Differentials in elderly parents’ past transfers favoring sons over daughters (e.g., parents’ investments in education) may in turn lead to couples’ greater likelihood of reciprocating by financially assisting husbands’ rather than wives’ mothers. Furthermore, it has been suggested that wives are the main beneficiaries of ascending generations’ provision of time, money, and child care (Trujillo et al. 2007). If this is indeed the case, wives’ expectations of future retributions for their assistance to elderly parents are likely to shape couples’ financial transfers to elderly parents.

Finally, we focused on couples in which both spouses were residing in Mexico. More research is needed to understand the patterns of remittances from couples in which one or both members are migrants to relatives related through the husband and to kin related through the wife. Such a research agenda will require data on spouses in migration destinations such as the United States as well as information on elderly parents remaining in Mexico.

Faced with continuing rapid increases in life expectancy at old age and strong declines in fertility, future generations of elderly Mexicans will rely on fewer adult children, who will be required to provide assistance to their parents for longer periods of time. Therefore, competition between elderly parents related by marriage and elderly parents related by blood for married partners’ economic resources is likely to increase in the coming decades. Whether midlife Mexican couples continue to resolve the competition between husbands’ mothers and wives’ mothers on the basis of traditional gender roles and responsibilities is likely to be informed by changes currently under way in gender relations among younger cohorts. To the extent that sustained increases in women's educational levels and strong increases in rates of participation in the labor market are fueling less differentiation in husbands’ and wives’ responsibilities within couples, future generations of midlife Mexicans’ financial transfers to elderly parents may be informed less by traditional gender norms and more by elderly parents’ financial burdens.

Biography

Claire Noël-Miller is an honorary fellow at the Center for the Demography of Health and Aging at the University of Wisconsin–Madison. She received her PhD in demography from the University of Pennsylvania. Her research focuses on elderly couples, with a particular emphasis on elderly husbands’ and wives’ intergenerational transfers, the consequences of spousal loss for the use of long-term care, and concordance in elderly spouses’ health.

Rania Tfaily is an assistant professor at Carleton University. She received her PhD in demography from the University of Pennsylvania. Her research focuses on the development of the family, with special focus on intergenerational transfers, nuptiality and fertility, stratification in health, and social inequality in education. Recent publications include contribution to Redefining Retirement? How Will Boomers Fare? (Oxford University Press, 2007) and articles in Demographic Research and the International Journal of Social Research Methodology.