It is well known that development programs and projects in the developing world should actively include women in all spheres of public life (Rogers, 2005). The impetus for this demand stems from several demographic and social constraints on the achievement of the goals of the development agenda (Nwoye, 2013). On the demographic side, fertility has declined in most developing countries with the exception of a few very poor countries (Bongaarts and Casterline, 2013).

Consequently, populations tend to age alongside a shrinking labor force necessary to carry on the tasks of economic development (Demeny, 2012). This has led to deliberations on liberalizing migration policies to gain access to skilled laborers from demographic billionaires such as India and China (Gordon, 2013). A second approach has been to encourage women's participation in labor force (Molyneux, 1985). In many developing countries such as Mexico, women have responded to opportunities for work. The proportion of women in working age groups in Mexico participating in the labor force increased from 19.6% in 1990 to 33.3% in 2010 (INEGI, 2010).

Even if the increase in the proportion of women participating in the labor force is impressive, social, psychological, and cultural resistances to women's participation in the labor market remain mainly due to the persistence of strong patriarchal norms and values in society, which prescribe and limit the role of women to domestic chores (Sen and Grown, 2013). This conflict between women's domestic and occupational roles have raised new fears about the economic and social consequences, threats to women's commitment and career development in the labor force (Korpi, Ferrarini and Englund, 2013). These issues have become the subject of empirical scrutiny in the current development of literature under the rubric of women's agency (Staudt and Jaquette, 2013). Women's agency is broadly defined as a set of skills and propensities women posses and actively utilize in their own terms in order to achieve goals and objectives that enhance their well-being (Kabeer, 1999).

The development and promotion of women's agency has become an important development goal in recent years in developing countries (Masika and Bailur, 2015). Very few social consequences are as desirable as improvement in women's agency. Not only does women's agency have several desirable outcomes for their well-being, but also for members of the family and the communities as well (Mosedale, 2005). Women's agency is associated with increases in years of education, postponement of marriage and small family size (Schuler and Rottach, 2011).

Women's agency improves not only child welfare and inculcates socially desirable behaviors with lasting impact on their well being throughout their life course (Hobson and Lindholm, 1997). The design and formulation of social policies are likely to benefit from an examination of the proximate determinants of women's agency (Hassim, 1999). The purpose of this paper is to examine the effects of a few selected determinants on women's agency in Nuevo Leon, Mexico.

In most developing countries today, women suffer from lower social status than men in both private and public domains. For example, in both low and moderate countries, the enrolment rate for girls in secondary school was 35.0% in 2011, while that for boys was almost 6% higher. Women's labor force participation rates remain significantly lower than that of men in almost all the occupational sectors of developing economies.

Women compared to men spend five times more time on childcare. Women in many developing countries lack independent rights to own land, manage property and conduct business; denial of property rights, honor killing, dowry related subjugation, lack of legal protection and exploitation of underage girls characterize the life of women in most developing countries even today (Cohen, 2006). Women's low status is accompanied by several social structural factors and also a normative climate that justifies women's subjugation (Damián, 2011; Levine, 1990).

In countries such as Mexico, programs that have been launched to improve women's social status are a key to the success of social policies (Molyneux, 2006). However, the effectiveness of such programs remains widely speculated and debated in the absence of empirical research on women's agency in Mexico (Casique, 2001; Luccisano, 2004). Both journalistic and anecdotal evidences cast doubt on the effectiveness of the empowerment program's role in improving women's agency (Luccisano, 2004 and 2006). However, empirical studies on women's agency in Mexico are few and far between.

The paper is organized under three sections. This section continues with a review of recent literature on women's agency in developing countries and proposes a hypothetical model of women's agency in Mexico. The second section presents the data and the results from the test of the hypothesized correlates of women's agency. Finally, the last section discusses the social context s that promote women's agency in Mexico.

Amartya Sen's theories of capacity building provide a theoretical framework for investigating women's agency (Hennink et al., 2012). Sen (2000) proposes a broad conceptual framework linking women's labor force participation and agency based on concepts of capabilities, functionings, and freedom. These concepts have gained attention since the integration of women's concerns into the development agenda in 1970 (Howard-Hassmann, 2011). Sen (2000) suggests two types of background variables among others, structural and normative, that influence women's agency. The structural variables are related to social positions and the extent of power associated with such positions. Interventions that promote socio-economic mobility and those which redistribute power by reallocating various types of resources across social positions are expected to bring about conditions conducive to improvements in women's agency (Dipna and Sharma, 2013).

The normative variables reflect socio-cultural values, laws and women's rights, which moderate the distribution of power across social positions. These background factors influence women's agency through autonomy (Acharya et al., 2010). The extent of autonomy women enjoy is related to their perception of power necessary to achieve desired goals and objectives (Hadjigeorgiou et al., 2012).

Agency, on the other hand, is related to perceptions of possibilities for intensification as well as expansion of power to achieve desired goals and objectives (Mishra and Tripathi, 2011). While social structural modifications often call for institutional reforms necessary to improve social mobility, normative changes may be induced by targeted legal as well as value changes (Mosedale, 2014).

Both social structural and normative changes intended for improving women's autonomy are likely to advance women's agency (Boudet, Petesch and Turk, 2013). Within the domestic sphere, norms with regard to household division of labor influence the level of women's agency (Sobritchea, 1990). As men actively participate in the production of goods and services within the household, they are more likely to partner with women in improving the levels of women's agency (Acharya et al., 2010).

Several indicators are available to operationalize autonomy in empirical research. Casique (2001) used nine variables to create a scale that reflects different dimensions of autonomy among Mexican married women. These variables measure whether women ask permission to: go out alone, go out with their children, incur daily expenses, visit friends, visit relatives, work outside the home, study, use a birth control method and participate in community activities.

Bojorquez-Chapela, Salgado and Casique (2009), using a similar survey, assess Mexican women's autonomy in four dimensions: decision-making power (women's participation in everyday decisions), movement autonomy (whether women need someone's permission to engage in some activities), political autonomy (participation in collective activities) and economic autonomy (assets and economic resources). García (2003) operationalizes autonomy in Mexico in terms of six selected characteristics: women's participation in intra-household decision making, freedom of movement, access and control of economic resources, lack of domestic violence, attitudes in favor of gender equality and being free to choose a spouse and the number children.

The concept of agency is characterized as processes of change where people start becoming aware of their power, possibilities, and influences in different groups and environments. Agency is born out of a recognition of the collective nature of human interactions and unequal relationships (Abrams, 1999). It is a process through which choices are made and attitudes generate a structure that guides future actions and constitutes people's identity (Hasselberger, 2012).

As women's access to resources through labor force participation increases, women's agency is likely to increase. Labor force participation bears a close association with the concepts of autonomy and agency. Although extra-domestic work may increase access to economic resources, in the Mexican context, women's labor force participation decisions are embedded in traditional family organizations that are conflict-ridden and slow to change (Ariza, 2000, as cited in Ariza and De Oliveira, 2003).

Mexican culture is patriarchal (Rendón, 2008; Stern, 1998). Women are encouraged to stay at home and engage in domestic activities while men assume their roles as breadwinners. Lack of women's status both at home and in the public sphere has constrained efforts to enhance their well-being (Rendón, 2008). Labor force participation remains low; in 2010, only 32.3 percent of Mexican women in working age were employed, while married, cohabiting and widowed women had the lowest participation rate, 29.8, 30.5, and 22.8 percent, respectively (INEGI, 2010).

On the other hand, the burden of household chores may constrain women's agency as time and material resources are diverted away toward fulfilling domestic roles (Goldin, 1990). In sum, the proposed correlates of women's agency are likely to include autonomy, labor force participation, and amount of household chores (Soni-Sinha, 2011; Rajasingham-Senanayake, 2004; Shaheed, 1999). See Figure 1.

We hypothesize that women's agency is influenced by labor force participation and autonomy level. Figure 1 presents the hypothesized model of women's agency. The degree of participation in labor force is directly influenced by the level of autonomy women enjoy at the household level. Levels of autonomy have both a direct and an indirect effect on women's agency through women's degree of labor force participation. Furthermore, higher levels of women's autonomy are associated with higher partner's level of involvement with household chores. Table 1 presents the variables used to measure each construct.

To measure autonomy we utilize variables such as number of children younger than 12 years of age in the household, years in the current marital status, years of education and age at first marriage. Agency is indicated by a number of variables related to the women's opinion about their roles at both intra-household and extra-household levels; a number of variables related to household chores measure the level of their partners' participation in domestic work.

In order to test each component of the hypothesized model presented in Figure 1 and to ensure that all the latent variables are reliable and valid, each construct was evaluated separately using confirmatory factor analysis (CFA). Thus, confirmatory factor analysis is used to develop the measurement model.

All the measurement models were obtained by maximizing the goodness of fit with the help of modification indexes within limits of theoretical justification to alter their independent error structures. Most of the models were obtained by correlating at least two measurement errors based on modification indexes. The structural equation model was used to assess the causality among latent factors.

Table 3 presents the estimates from the CFA models of the four latent factors. Although the squared multiple correlation value for a few indicators is less than 0.250, the goodness of fit index (GFI) for the respective constructs suggests a high level of theoretical validity for each of the four constructs with all the respective factor loadings significant at 0.01 level.

Table 4 presents the results from the estimation of the structural equation model for both models. Increases in women's autonomy lead to higher levels of agency. As women's autonomy increases by one unit, women's agency level increased by about 55.7 units. The levels of women's participation in labor market were positively related to women's agency; this suggests that when women's capabilities increase, they are more likely to seek the fulfillment of their desires and to guide the course of their lives from their own point of view and interests.

But the transformation of autonomy into agency is influenced by specific social structural characteristics and normative values that dictate women's and men's roles. Men's involvement in household chores was assumed to be correlated with women's autonomy. As expected, the correlation between household chores and autonomy is both positive and significant. Our results also suggest that women's participation in the labor market and women's level of autonomy positively influence agency. The effect of autonomy on agency measured in terms of the path coefficient is about two times the path coefficient between labor force participation and agency.

Women's agency on matters that influence their well-being is of central concern for the provision of social as well as healthcare services (Molyneux, 1985). The purpose of this study was to assess the relationships among labor force participation, level of autonomy and women's agency. The effective exercise of women's capabilities implies the existence of freedom to pursue their desires and to choose among different available opportunities (Sen, 2000).

The significance of the relationship between agency and women's well-being in developing countries such as Mexico has not been adequately investigated (Casique, 2004; Wainerman, 2003). The impact of agency does not occur in a social vacuum (Charrad, 2010). Thus, even when women possess the skills and abilities to bring about change, theses resources do not necessarily influence their well-being. In this regard, our study highlights the importance of agency for women's well-being.

Even though the existing literature on women's well-being underscores the importance of the direct effect of autonomy, this study focused also on the neglected indirect effect of autonomy on agency mediated by women's participation in the labor force. We found the indirect effect of autonomy on agency to be significant. In this regard, our study suggests that the development of agency is crucial to the success of projects and programs. The success of many public programs such as family planning or poverty depends upon women's agency.

In general, our study finds that improvements in the development of agency can be achieved through different channels: promoting men's involvement in housework, improving autonomy, and increasing women's participation in the labor market. Since the latent factor 'autonomy' was indicated by measures such as number of years of education and number of children, improvements in these are likely to increase women's agency. Improvement in education not only enhances women's role inside and outside the household, but also improves access to better job conditions. We found the independent effect of labor force participation on women's agency to be significant and positive. At the policy level, this suggests that programs on women's agency should be pursued alongside programs and policies to improve labor force participation.

Women's entry into the labor market can be seen as an emancipating phenomenon that promotes equality between men and women both inside and outside the household. Through remunerated employment women are exposed to new interactions that redefine their roles (Baca, 2005; Goldin, 1990). Their participation in the labor force can modify their traditional practices as housewives and blur social aspects that confine them to the domestic sphere (Colón, 2010). The factor 'household chores' was used as a control in our model.

A challenge for future research is to assess the translation of women's agency into intended outcomes given their levels of autonomy, domestic chores workload, and participation in the labor market. Women are increasingly likely to be burdened with housework as their domestic workload does not decrease when they work outside the household (Del Tronco, 2008). Increases in men's involvement in the domestic sphere are likely to improve extra-domestic opportunities for women reinforcing their levels of autonomy and agency. Men's involvement in domestic tasks creates a social environment conducive to improvements in women's agency.