The Inter-American Development Bank presents ... pushing for progress. Women, work, and gender roles in Latin America.

Nopo, Hugo

The world, and particularly Latin America, has experienced important changes regarding the role of women and men during the last three to four decades. Women's visibility at home, at school, in the labor markets, and in society in general, has evolved significantly. Concurrently, men's role has evolved as well. This article will discuss these trends, highlighting the pending issues towards a better participation of females in the labor markets and outlining some possible courses of action.

Female school achievement has increased more than that of males. For those born by 1940, males achieved on average an extra year of schooling over females (six vs. five years respectively), while for those born by 1980, females attended an extra quarter of a school year more than males (9.5 vs. 9.2 years). The gender gap in schooling for the cohort born in 1968 reversed from being male-dominated to being female-dominated. The only countries for which the academic gender gap has not reversed are Bolivia and Guatemala, two countries with large indigenous populations, which may suggest important linkages between gender and ethnic disparities among these countries.

It is also interesting to note that this improvement in female schooling occurred at the higher end of the schooling distribution. While by 1992, 16.35 percent of working females and 10.66 percent of working males had some (complete or incomplete) tertiary education in the region overall, by 2007 those percentages were 26.05 percent and 17.26 percent respectively. The global phenomenon of higher schooling achievement among females began earlier in Latin America than in the rest of the world.

Female labor force participation has increased, while that of males remains almost constant. Whereas by the beginning of the 1990s one-half of Latin American women participated in the labor markets, either working or looking for a job, nowadays it has reached almost two out of three women in most of the region. The most recent studies at the Inter-American Development Bank reveal that most of the increase in female labor force participation can be explained by the increase of participation amongst young married women, suggesting a cohort effect. During the same period, male labor force participation remained almost constant in the region. In other regions, male labor force participation even dropped (as in OECD countries, for instance) However, males still dominate labor markets to the point that three out of five workers in the Latin American region are males. Occupational segregation by gender is still largely prevalent in the region as well.

The share of female-headed households has also shown substantial increases. By the beginning of the 1990s, females headed 1.2 percent of complete households (those where there are both husband and wife present) and 79.8 percent of single-headed households. Such percentages have increased to 9.2 percent and 82.3 percent respectively by the late 2000s, with the most significant increase in complete households. Besides the challenges that represent the measurement of male and female headship, the data trends show that a picture of increasing female headship, even in those households where there is a father present, is gaining space in the region. Also, the data shows that female-headed households are at both extremes of the income distribution. Some household-head females correspond to the profile of a single young woman in professional or managerial positions with young children (age one or two at most) at the upper extreme of the income distribution. Some other household head females correspond to the profile of a low-educated single mother, with many children (three or more), and an informal job in the service and commerce sectors, around or below poverty lines.

Fertility has dropped and the region has been characterized by notorious cross-country heterogeneity, not only in terms of fertility rates, but also in the speed at which these rates have been dropping. In Argentina and Uruguay fertility rates started to decrease as early as in the 1930s. Countries like Bolivia, Guatemala, Honduras, Nicaragua, and Paraguay, in contrast, still showed high fertility rates by the mid 1990s, but recent years show a decrease as well. This is reflected in the statistics about the presence of children in Latin American households. While by the beginning of the 1990s one in nine working women (11 percent) lived in a household where there was at least one child of age six or younger, by the end of the first decade of the 2000s this percentage almost halved. A similar situation took place for males: the percentage living in households with children dropped from 16 percent to 9 percent. This phenomenon, which has been linked to delays in women's age at first child and higher prices in the labor markets, suggests an alleviation of household responsibilities, at least those related to childbearing and childrearing. For the countries that began the demographic transition early, however, the responsibilities are shifting to the care of the elderly and the sick.

[ILLUSTRATION OMITTED]

Marriage, education, and work decisions have evolved. As mentioned above, most of the increase in female labor force participation has been due to an increase in participation among married women. Furthermore, the most significant increase can be seen among those who married more educated husbands than themselves and, not surprisingly, those with no children and no elderly relatives at home. Also, recent work by Ricardo Hausmann and his coauthors reveals that skilled women are less likely to be married (or cohabiting) than unskilled women and skilled men. It is less likely for skilled women who are married to work than unmarried, skilled women. Also, skilled women in Latin America are more likely to marry a less skilled husband than in other regions, simply as a result of relative abundance and scarcity in the mating markets. When marrying a husband less skilled than themselves, women are more likely to Work than when marrying an equally skilled husband.

However, women are still lagging behind men in one crucial factor: labor earnings. Women have achieved important progress in their participation in the labor markets and the human capital they bring into it, but the markets still fail to reward them appropriately. Or, in a dynamic perspective, the speed at which gender earnings disparities are reducing does not seem to align with the speed at which women gain schooling and human capital. Current prime-age workers are precisely the cohort for which the gender schooling gap reversed from being male-dominated to a female-dominated one. The gender earnings gap, however, is not only far from a similar reversal, but also far from parity.

Gender earnings disparities

By the beginning of the 1990s, the region's average indicated that for every 100 monetary units females earned in the labor markets, males would earn 116.32, a 16.32 percent gap. By 2007, this gap was 8.8 percent. The gap has dropped almost eight percentage points in almost two decades, but it still favors males and it does not reflect the favorable human capital characteristics that women have. However, these gender earnings gaps could well be the result of gender differences in individuals or in the job characteristics that the labor markets reward. Hence the observed gender disparities are somewhat misleading as the correct comparison of earnings disparities should be made on the basis of comparing earnings of individuals with similar characteristics.

A sequence of work done and forthcoming at the Inter-American Development Bank has been documenting the extent to which gender earnings disparities correspond to gender differences in observable individual and job characteristics. The first result to highlight from this line of research is the role of education in explaining the earnings differentials. It is somewhat paradoxical that besides having more schooling, women still earn less than men, as will be discussed later in this article. When comparing males and females with the same age and schooling, the earnings gaps are actually higher than those originally observed, and they show a smaller drop than the one shown by the unconditional gaps (25 percent by the beginning of the 1990s and 22 percent by the end of the first decade of the 2000s respectively).

[GRAPHIC OMITTED]

Another variable that plays an important role in the analysis of gender earnings disparities is part-time work, defined as working 30 hours or fewer per week. The inclusion of such a variable also increases the gender earnings gap. A refined exercise was made such that male and female earnings with the same age, schooling, presence of kids at home, presence of other income-generator at home, the urban/rural divide, type of employment (distinguishing self-employed, employees, and employers), and time worked (distinguishing part-time, full-time, and over-time workers) were compared. The results showed that the gender earnings gap was 34 percent in 1990s and 30 percent by the mid 2000s, more than double what the unconditional gap shows.

These average gender earnings gaps also show important heterogeneity within the labor markets. Females' underpayment with respect to males with the same characteristics is more pronounced among poor and low-educated workers, those in small firms or self-employed, and diose working part-time or with no formal labor contract. The good news, however, is that these same segments of the labor markets where the gender disparities are more pronounced are also the segments that have shown the biggest reductions in the earnings gaps. When looking at particular countries within the region, it is in Brazil where the gender earnings disparities are the highest, but it is also in this country where the drop in the gaps has been the largest. The gaps are declining and this has been happening where they are wider. The speed of this reduction, however, has not been reflecting the speed at which related characteristics, such as females' education, have improved. There is still much work to be done.

[ILLUSTRATION OMITTED]

Persistent disparities

If women's credentials are improving, what explains such resilient gender earnings disparities? Traditionally, two of the most prominent explanations have been discrimination and gender roles. Recent evidence suggests that information can be an effective antidote against discrimination. Gender roles, deeply embedded in the culture, may require a longer-term approach.

There are good indications that discrimination can be alleviated with information. Joseph Altonji and Charles Pierret pioneered this with the idea that discrimination, if present, should fall with tenure (as information about real productivity reveals with tenure), validating it with US data. More recently for Latin America, specifically Peru, Castillo, Petrie, and Torero have shown in an experimental setup that individuals tend to stereotype. In the absence of information, when pressed to make decisions regarding group formation, individuals use observable characteristics such as gender, skin color, and height as proxies for productivity. When real information about productivity is revealed, individuals do not use these proxies anymore and base their decisions on the productivity data instead. Information trumps discrimination. Also, the Bank's finding that the gender earnings disparities are most pronounced among the self-employed gives less credence to the idea of employer discrimination in the Latin American labor markets.

The finding that female underpayment is more pronounced in part-time work, small firms, and informal jobs is also revealing. All of these point to an employment characteristic that females prefer: flexibility. As a matter of fact, while one in four women work part-time in Latin America, it is only one in ten men who do so. Flexible labor force participation is a predominantly female terrain. Such preference for flexibility is not just an innate characteristic of women; rather, it arises as a result of the constrained decision-making women have to make in order to participate in labor markets. Given the traditional division of labor within households, and hence females' greater responsibilities at home, many women can only participate in the labor market if the position offers flexibility. This reduces their bargaining power for higher earnings, configuring the situation just outlined.

Available data on time use substantiates the explanation above. Between 70 percent and 80 percent of the time devoted to domestic work in Latin American households is the contribution of women, far more than women's contribution in other societies. For instance, in the United States and the United Kingdom, women's contribution is around 60 percent. When a woman in Latin America works part-time, her share of domestic work rises to roughly 80 percent, while her share for full-time work is between 60 percent and 70 percent. This, as well as the number of hours devoted to work weekly, has not changed much during recent years.

Policy options

With this panorama of progress and the pending challenge regarding parity in the labor markets, some policy options arise. Alternative actions may seem numerous, but this article will focus on three: legislation for equality of opportunities; information diffusion to reduce stereotyping, foster attitudinal changes, and nurture egalitarian values; and effective alleviation of women's household responsibilities.

Legislation regarding gender parity is not new in the region. As a matter of fact, most countries have some sort of anti-discriminatory law and protection of women within their systems. Such legislation is full of good intentions but, unfortunately, most of them generate unintended behaviors that are not conducive to gender parity. A law that, for example, requires equal pay for all males and females performing the same job encourages employers either to prevent employees' promotions or to hire informally. This, in turn, introduces distortions to the labor markets provoking a sub-optimal use of resources at both the micro and macro level.

Legislation promoting parity should be conceptualized in a more integral way, analyzing both intended and potentially unintended consequences of the law. A common mistake of such laws is to pretend to regulate outcomes (employment, wages, or special benefits) instead of inputs (time availability, for instance). A renewed focus on inputs needs to be in place. A good example could be the promotion of paternity leaves. Currently, leaves for recent mothers notoriously surpass those for recent fathers. This, as has been argued, serves as a source of inequity as it reduces females' employability. An equalization of maternity and paternity leaves may help level the playing field regarding hiring decisions for males and females. Leave equalization may also generate positive consequences out of the labor markets as well. Encouraging men and women to devote the same time to their newborns and, hence, home, may also induce more harmonic households, with fairer intra-household bargaining and decision-making. This in turn would help to nurture a more equitable split of responsibilities, time, and opportunities within households. Of course, a single regulation like this one, even with good enforcement and conscious compliance, would not do much for a more egalitarian society if it is not accompanied by a set of concurrent measures in different spheres.

[GRAPHIC OMITTED]

Secondly, as discussed in this note, information diffu sion can be an effective tool to reduce stereotyping in day-to-day activities. Sometimes, however, these information flows cannot be produced at the speed the markets require. Not all social interactions can benefit from information flows, like the experimental example in Peru. Then, the alternative becomes generating information flows geared towards more permanent, not situation-specific, cultural and attitudinal changes. One fruitful avenue in that regard has been the introduction of placement tied with mentoring and networking programs. Getting women into the workplace has helped to change (male and female) employers' and coworkers' perceptions, trumping stereotypes. But more massive tools are needed.

Soap operas, for instance, have worked as effective instruments for such massive cultural and attitudinal changes. Work by Chong and La Ferrara has illustrated how the subtle introduction of role models in Brazilian soap operas, over the course of decades, induced changes in fertility and divorce rates in the Brazilian middle class. Massive campaigns to make people more aware about misperceptions regarding gender roles at the workplace and day-to-day life may play an important role. But bear in mind that egalitarian values are not spontaneously generated; they require some time to be nurtured.

[ILLUSTRATION OMITTED]

Such nurturing has to start at home during the early years and then continue at school. In this respect, there still are pending issues. Work done in Peru at Group for the Analysis of Development (GRADE) suggests that the schooling system nurtures gender stereotypes. An analysis of school textbooks there revealed that two-thirds of human images that appeared on 4th and 6th grades textbooks were boys while just one-third were girls. Females' images were predominantly related to leisure and domestic work, while those related to male activities were more linked to labor and schooling. Such subtle, and most likely unconscious, communication of stereotypes need to be eradicated. As already outlined in this note, more women are attending school, yes, but a pending question is "What are they learning there?" There is room for action along these lines.

The third recommendation, to alleviate women's restrictions on time use and household responsibilities, may be more costly than the previous two. Nonetheless, its effects are more likely to be seen in the short and the long run than the previous two. The reality is that women are in charge of taking care of kids at home, reducing their possibilities of fully participating in the labor markets. Hence, an effective way of alleviating those responsibilities is promoting alternative daycare options. Expanding the networks of early child development (ECD) facilities and extending the school schedules for primary students can work in the desired direction. Preliminary evidence on this reveals that the length of the school shift is important. Currently short school schedules fail to free up enough time for female participation in the labor markets. Or, they allow only part-time participation, with substantial earnings disparities, as discussed above. An effective freeing up of females' time to devote to the labor markets along these lines would not only produce short-run economic gains. This would also be a profitable long-run investment in better education.

HUGO NOPO is a Lead Research Economist for Education at the Inter-American Development Bank. His research considers the use of applied econometrics to the understanding of labor markets, educational systems, inequality and poverty, with a focus on Latin America.