Evidence-based gender equality policy and pay in Latin America and the Caribbean: progress and challenges

Abstract

This paper aims to identify the main contributing factors to the observed gender pay inequality in the last 10 years in Latin America and the Caribbean. It also aims to identify the main restrictions to design evidence-based policy. The average woman in the region works 25 h more per month than the average man. Yet only half of women in the region are paid for or otherwise profit from their work. Work without formal pay leads to weak protection of human rights and limits civic participation. Moreover, gender inequality deters economic development and fosters income inequality. Research shows gender gaps result from the interactions of social norms and biological differences. The challenges to generating effective policies include the difficulty of identifying relevant mechanisms to explain gaps, the cost of collecting and analyzing relevant data, and the complexity associated with determining pertinent features of gender gaps. Collection and analysis of those mechanisms, data, and features are critical to learning how to correct gender gaps going forward.

1 Introduction

Men and women may sometimes have different roles, which impact the exercise of human rights. For example, in Latin America and the Caribbean region women work more yet get less formal compensation or benefits in exchange for it. The average woman in the region works 25 h more per month than the average man (United Nations 2015b), and half the women in the region work for no pay or profit at all (World Bank 2016). A pay gender gap implies limits to individual choice based on sex, thus violating individual human rights. Note that the word “gender” refers to a social construct. Gender denotes a group of behaviors that society considers appropriate for men and women (WHO 2015). Thus, gender gaps refer to differentials between the sexes derived from different treatment by society. The word “sex,” on the other hand, denotes the biological and physiological characteristics that separate men and women (WHO 2015).¹

Governments in the region need policies to promote equal enjoyment by men and women of human rights, including the right to reap economic and social benefits. Despite the current gender gap in Latin America and the Caribbean, there has been significant progress in the region. Women have increased their labor participation since the 1990s more than in any other region in the world. The percentage of employed women 15 years of age and older increased from 38 to 50% from 1991 to 2014 in Latin America and the Caribbean, in contrast to a decrease from 49 to 46% in the rest of the world (World Bank 2016).² However, it is still an open question how much of this labor increase constitutes a substitution for unpaid work and how much of it is actual additional labor. Economists cannot fully explain why women continue to assume unpaid work in society and bear the cost associated with it. Neither is it clear why society continues to bear the social and economic costs associated with gender segregation into occupations.

Economists have studied labor and pay gender gaps since the 1950s. The general argument built to date includes several components. Biology and social norms interact to determine skills. Skills and technology produce goods and services that individuals trade. Institutions regulate trade to reduce transaction costs. Thus, biology, social norms, technology, and institutions influence the pay gender gap. Researchers have found that these factors interact and influence the gender gap at several stages. Therefore, researchers are challenged to find simple solutions. Policymakers and researchers are working together to address the challenge. This paper aims to aid these efforts. It summarizes the progress in research, explains the rationale for the main theories to explain gender gaps, summarizes pay gap studies in the region in the last 10 years, and lists the main challenges to generating evidence-based gender equality policy in Latin America and the Caribbean.

Studying gender gaps is important for two reasons. The first is that gender gaps limit individual choice based on sex. The second article in the Declaration of Human Rights states: “Everyone is entitled to all the rights and freedoms set forth in this Declaration, without distinction of any kind, such as […] sex” (United Nations 1948). However, the observed inequality in pay between men and women is related to limits to freedoms. For instance, Duflo (2012) noted that gender would not influence spending if informal mechanisms enabled individuals to voice their preferences. However, women allocate more resources than men to the education and health of children (Thomas 1990; Duflo 2003; Duflo and Udry 2004). In Mexico, married women’s earnings reduce within-household inequality because they share their income with other household members (Campos-Vázquez et al. 2012). In the government sphere, the gender of the decision-maker also influences public spending (Yañez-Pagans 2014). Thus, pay gaps reflect gaps in the ability of individuals to voice their preferences.

The second reason why it is important to study gender gaps is that these gaps deter economic growth. Indeed, women tend to invest more than men in human capital for future generations (Duflo 2003; Duflo and Udry 2004). Gender gaps also undermine collective action. Women are more likely to be out of the labor market or to work in the informal sector (ILO 2013b, 2015b). Thus, tax collection and social coverage is particularly limited for women.³ In addition, women are less likely to participate in the democratic process as leaders (Zetterberg 2009), and gender segregation into certain occupations also contributes to deterring economic development (World Bank 2012).

The relationship between women’s empowerment and economic development is not trivial. Evidence shows that simply endowing women with resources does not always result in economic gains. For example, Udry (1996) found that men’s farming plots were more productive than women’s plots in Africa. Banerjee et al. (2015) found no evidence the gender of the recipient of microcredit loans in Mexico mattered for returns. Overcoming gender gaps can be difficult, and perpetuating gender gaps is likely to deter some paths to economic growth.⁴

This study focuses on the gender gap in pay and how it is related to other important gender issues in the region. There are three features of the scope and limitations of this study. First, the study focuses on economic labor models of individual choice. One concern related to gender issues is that social norms drive gender gaps. Therefore, other studies could consider studying gender gaps from a collective decision-making perspective. Second, this study discusses aggregate evidence that allows for establishing relationships among contributing factors. As a result, it leaves out a rich body of literature on case studies or descriptions of behavior based on small samples. Third, this study will discuss policies related to gender issues in the last 10 years. A historical view demands a broader discussion and is beyond the scope of this work.

2 Gender gaps in pay

This section describes the formal pay gap related to gender, highlighting three ways into which gender pay gaps can be split: the pay/no-pay choice, the occupational choice conditional on receiving pay, and the wage gap conditional on an occupation.

First, women disproportionately spend more time in non-remunerated activities. Columns (A) and (B) of Table 1 show the minutes worked per day by sex in developing countries. Women spent 1 h and 41 min on unpaid work for every hour of paid work in 2010 (United Nations 2015a). The equivalent for men was 13 min (United Nations 2015a). Overall, women work 50 more minutes per day than men both in paid and unpaid work. This difference adds up to 25 h or about one day of work per month for women. However, women are less likely to get paid than men for all of that work because they spend more time producing goods and services for household or family consumption.⁵ For example, women spent 14 h preparing food and 10 h cleaning the house per week in 2014 in Mexico (INEGI 2015). In contrast, men spent four hours preparing food and four hours cleaning the house per week (INEGI 2015).

Table 1 The gender gap in paid work, employment, and earnings in developing countries and Latin America and the Caribbean.

The employment rate for persons ages 15 or older in Latin America and the Caribbean is informative as to the share of men and women working in paid or for-profit activities. The employment rate was 50% for women and 75% for men in 2014 (Table 1, column (D)) (World Bank 2016).⁶ Thus, women spend more time on household and family activities than men and are less likely to hold jobs for pay or profit.

Employment gaps between men and women are not new, and they used to be larger. Gender gaps in employment closed in Latin America and the Caribbean between 1991 and 2014. Figure 1 shows employment of men and women in the region and in the rest of the world. Employment among women decreased from 49 to 46% worldwide but increased from 38 to 50% from 1991 to 2014 in Latin America and the Caribbean (World Bank 2016). Employment among men declined slightly from 78 to 75% in Latin America and the Caribbean, and from 76 to 72% in the rest of the world (World Bank 2016). At this average rate of change in employment among men and women, the gap between them would close by 2056, reaching a global employment rate of 71%.

Fig. 1

Percentage of men and women employed, Latin America and the Caribbean versus the rest of the world, 1991–2014. Source: Author’s calculations based on data from World Bank (2016). Percentages are computed as a share of the total population. For the countries, economies, and territories included in Latin America and the Caribbean and the rest of the world. LAC Latin America and the Caribbean

The second way to look at gender pay gaps is by considering that people who work for pay segregate into occupations according to gender. Worldwide, women tend to work in jobs related to education, health, welfare, and the arts and humanities, while men segregate into agriculture, engineering, manufacturing, construction, science, and services (World Bank 2012). Segregation into occupations is related to pay. For example, using data from Bolivia, Chile, Ecuador, Honduras, Nicaragua, El Salvador, Uruguay, and Paraguay from 2007, Ñopo (2012) found that women held only 33% of top-paying jobs related to business, law, health, computer science, government, and science. The International Labour Organization (ILO) estimates that 60% of chief executive officers in Latin America and the Caribbean are men, while 74% of them are men are worldwide (ILO 2015a). Moreover, only 40% of firms are owned by women in Latin America and the Caribbean, while 37% of firms are owned by women worldwide (ILO 2015a). Thus, women may receive less money for their work because they do not work in the best paid occupations.

Even within specific occupations, women tend to get paid less than men, which constitutes the third way of looking at gender pay gaps. Ñopo (2012) found the top 10 paying jobs in his study paid US$479 to men per month but only US$302 to women, an approximate 30% difference. Wage gaps persisted even when considering age, education, number of children, presence of another income earner at home, employment, and number of hours worked per week. Column (E) in Table 1 shows that women earned 87 cents for every $1 that men earned in Latin America and the Caribbean in 2013. The relationship is not exclusively related to wage gaps. For example, Angelucci, Karlan, and Zinman (2015) found that microcredit for women to start a business did not always increase their income in Mexico. The evidence suggests that women tend to get less money for their work relative to men even when both groups are compared within a specific occupation.

In summary, men tend to work in paid occupations, while women tend to work in unpaid ones. When women work for pay, they tend to work in lower-paid occupations relative to men. Within a specific occupation, women are paid less than men, although these gaps have been closing over time.

3 Causes of gender gaps in pay

This section summarizes the main theories to explain gender gaps in pay. Since Ricardo (1821), economists have modeled gaps in individual skills and abilities to explain gaps in labor. In 1965, Becker proposed a model that became standard to analyze occupational choice (Becker 1965). In a basic model, individuals choose between paid and unpaid work by considering the trade-offs—or benefits and costs—associated with the returns derived from each activity. Benefits and costs depend on individual characteristics, social preferences, technology, and institutions.

3.1 Individual characteristics

Economists have specifically explored gender gaps related to education, experience, and preferences since the 1970s (Mincer and Polachek 1974). Both individual biological features and education influence individual skills and productivity. Gender norms influence education and occupational choice. These conclusions have been widely accepted since debate on the topic started (Becker 1971; World Bank 2012; United Nations 2015a).

3.1.1 The role of nature

Biological predisposition in individuals influences their productivity in certain positions. For example, certain positions in construction may demand lifting heavy objects. The average man can lift heavier objects than the average woman. Another biological difference between the sexes is that of pregnancy, which affects the physiology of women around the time of childbirth. Women experience hormonal changes after giving birth that trigger processes such as milk production. Such processes may induce a transitory change in preferences. For example, women become more sensitive and responsive to infant cry after childbirth (Marlin et al. 2015; Levine et al. 2007; De Pisapia et al. 2013; Noriuchi, Kikuchi, and Senoo 2008). By undergoing such changes, women may face a higher cost than men to forgo the care of infants during a sensitive period. Biology may also play a role in shaping non-transitory preferences. For example, exposure to testosterone before birth influences competition, dominance, and risk preferences in adults (Baron-Cohen 2003; Dreber and Hoffman 2007). Moreover, the menstrual cycle may influence risk preferences, altruism, trust, and reciprocity (Buser 2012; Pearson and Burkhard 2013). Indeed, laboratory experiments find differences across genders in these dimensions (for a review, see Bertrand 2011). Biological processes may induce differences in skills and transitory and non-transitory preferences.

3.1.2 The role of nurture

Education is an important way to acquire skills.⁷ As of 2010, 110 girls were completing secondary education in Latin America and the Caribbean for every 100 boys (UNESCO 2016a, b). Currently, women in Latin America and the Caribbean attain more years of education than men (World Bank 2016).

Despite achievements in years of education, differences in skills between the sexes remain. Consider the average gaps in certain subjects in the eight Latin American and Caribbean countries that participated in the 2012 Programme for International Student Assessment (PISA).⁸ In math, the average gap was 0.19 standard deviations, favoring boys (OECD OECD 2012). In reading, the average gap favored girls by 0.27 standard deviations (OECD 2012).⁹

Consider a crude estimation of the potential impact of the math gap on earnings in the region. Focus only on men and women who work for pay. Recall that in 2013, women received US$87, while men received US$100 for the same type of work (Economic Commission for Latin America and the Caribbean (ECLAC) 2015). Hanushek (2011) estimated that a one standard deviation above average can lead to an expected increase in earnings of up to 15%. Closing the mathematics gender gap would imply that a woman would earn US$87*[1 + (0.19*0.15)] = US$89 for every US$100 a man makes, instead of US$87. Closing the skill gap in math would thus reduce the earnings gap by 15% [(89 − 87)/13]. Closing the reading skill gap would increase the earnings gap.

Gender gaps in skills can result from differential treatment starting early in life (Qian 2008; Barcellos et al. 2014). The brain is malleable to experiences, especially in the first years of life (Wood et al. 2008; Teasdale and Owen 1984; Wilson 1983). Studies on twins show that even if about 50% of the variance in child development is due to genetic factors, a child’s genetic expression is influenced by environmental inputs (Teasdale and Owen 1984; Wilson 1983). Bando et al. (2016) studied language and socio-emotional skills of young children and found that girls outperformed boys as early as seven months of age in Chile and Nicaragua.¹⁰ This suggests that the gaps in language observed on the PISA tests for 15-year-olds may start as early as infancy. The gender gap in language in PISA is present in all countries despite variation in culture and educational structures (World Bank 2012). However, these differences are small relative to the difference in performance across children from different socioeconomic status.

3.1.3 Identity effects

Later in life, gender norms can affect gender gaps via psychological effects. Akerloff and Kranton (Akerlof and Kranton 2002) noted that people adapt behavior to follow the ideals of the social category with which they identify. They called this phenomenon “identity effects.” In other words, individuals adjust their behavior to comply with the norms of the group to which they perceive themselves as pertaining, regardless of skill. This behavior starts around the age of six (Wetherell 1996). For example, a woman who chooses an occupation to feel feminine or a man who chooses an occupation to feel masculine are displaying identity effects.

Gender roles in Latin America and the Caribbean reflect the Catholic influence in the region. Indeed, the share of people of Catholic descent in Latin America and the Caribbean is larger than that of any other region in the world. In 2014, 84% of adults in the region declared they had been raised as a Catholic (Pew Research Center 2014). In the Catholic religion, the ideal of femininity is modeled by the Virgin Mary, the mother of Jesus. Marianismo considers women to be nurturing and morally superior to men, and holds that women should be willing to engage in self-sacrifice to benefit society, especially the family. Thus, social norms declare women to be better suited to meet family needs, with special emphasis on the education and rearing of children (Bachrach 1991; Cianelli et al. 2008). On the other hand, social norms in Latin America and the Caribbean define men to be better suited to defend their self-image as strong and dominant leaders. Machismo places expectations on men to take on high-risk behaviors, be aggressive, and show immunity to emotional pain (Lavrin 1987; Cianelli et al. 2008). These gender norms are likely to impose differential social benefits and costs to promote a fit between sex and gender when an individual chooses an occupation (World Bank 2012; ILO 2016).

3.2 Social preferences

Gender norms can affect occupational choice via demand reflecting social preferences. The psychology literature has proposed gender stereotypes affect the way people behave and treat others (Ellemers 2018). Indeed, suppose many people were willing to pay a higher price for a good or service when provided by a person of a specific sex. This phenomenon is clear in performance of some forms of arts and sports. If supply is not different between the two groups, then preferences will lead to different wages via demand. There is no evidence on how stereotypes influence wages. However, there is evidence that show stereotypes may influence demand. Ganguli, Hausmann, and Viarengo (2010) analyzed marriage and skill patterns in Latin America and concluded that men assign a high value to having a stay-at-home wife. They found that skilled women are less likely to marry than unskilled ones. Another piece of evidence is provided by Ñopo (2012) found that newspapers and commercial magazines in Colombia display individuals in occupations to match gender stereotypes and are tailored to social preferences to maximize sales. In this way, gender norms influence social preferences, which in turn influence demand. Demand for specific attributes in terms of the sex of a person who provides a good or service influences the employment gender gap.

Becker (1971) proposed “taste discrimination” as a cause for gender gaps. Taste discrimination refers to the willingness of society to pay a lower price when a good is made or a service is provided by a person of a minority group. According to Becker’s theory, discrimination derives from an individual distaste for either women or men. There is no evidence on how individual taste for labor provided by persons of a specific sex affects employment.

3.3 The role of technology

Technology influences how social norms and biology influence labor outcomes. Economists have long postulated this idea. Hicks (1932) and Roy (1951) explored the role of technology in the relationship between individual skills and occupational choice. There is evidence of a positive association between technology and gender equality. Cubas (2016) found that the price of household appliances and access to infrastructure explains cross-country female employment in Latin America. Another technological advance that influences gender gaps is family planning. Bailey, Hershbein, and Miller (2012) estimated that access to the birth control pill in the 1960s and 1970s allowed young women to earn an 8% higher hourly wage premium by age 50 in the USA.¹¹

3.4 The role of institutions

Economics has long acknowledged the role of institutions. For example, Coase (1960) highlighted the importance of legal rules to regulate labor. Institutions are especially important to protect individual, property, and labor rights.

3.4.1 Institutions and human rights

Most countries in Latin America and the Caribbean have laws in place to prevent people employed or looking for employment from being discriminated against based on sex (Pagés and Piras 2010). Moreover, most countries have maternity protection, laws on day care services, and laws on leave to take care of sick children (Pagés and Piras 2010). However, only 8 out of 29 countries in Latin America and the Caribbean had laws in place to mandate equal remuneration across genders in 2015 (World Bank 2016).

A challenge in the region is that institutions may have limited reach to enforce laws. Indeed, about half of those employed in Latin America and the Caribbean work in the informal sector (ILO 2015b) and are not covered by the legal framework.¹² Women are overrepresented in the informal sector in Latin America and the Caribbean, except in Bolivia and Venezuela (ILO 2013b).

3.4.2 Statistical discrimination

Arrow (1973) suggested that discrimination could be a product of information problems. The idea behind statistical discrimination is that employers build beliefs on productivity based on statistics. For example, if employers assume that individuals perform at the same level as their group average, they will be discriminating against those who perform above the average. A woman who applies for a job in a non-traditional female occupation where a few other women have performed poorly may be assumed to be less fit for it. Moreover, if productivity changes, then a lag in data generation will extend statistical discrimination in time. Testing for statistical discrimination is difficult. Standard approaches (such as standard audit studies) confound taste discrimination and statistical discrimination. Thus, there is no evidence on whether statistical discrimination does or does not exist in Latin America and the Caribbean.

The next section discusses evidence on policies implemented in the region to address employment gender gaps. These include policies associated with providing equal opportunities to all persons, relevant technology, and laws to promote gender equality.

4 Proposed solutions to gender gaps in pay

There has been extensive work to identify effective policies to close the gender gap in pay. This section summarizes the empirical work in the last 10 years in the region. The available evidence supports the idea that women disproportionately assume the cost of raising a family. On the other hand, skill gaps or gender discrimination during the hiring process do not seem to play a major role. The implementation of laws results in changes in incentives for society, which affect gender gaps.

Table 2 summarizes the theoretical mechanisms proposed to explain the gender pay gap. The first column summarizes the theoretical causes discussed in the previous section. The second column summarizes interventions in Latin America and the Caribbean to address each theoretical factor. However, interventions can often address many theoretical causes of the gender gap. Thus, linking theory and evidence is not a straightforward exercise. The table associates theoretical causes with interventions that in principle aim to primarily address them. The table lists results on intermediate outcomes if there were no studies on labor outcomes. The evidence presented in the table is discussed in the sections that follow.

Table 2 Individual, societal, technological, and institutional factors that influence women’s paid work.

4.1 Ensure that there are no differences in opportunities across gender groups

Education is likely to play an important role in moderating biological predispositions (Cunha and Heckman 2009). Unfortunately, education is probably not yet addressing inequalities between women and men in Latin America and the Caribbean, other than access. For example, Bassi et al. (2016) found that teachers pay less attention to female students in class in Chile. Moreover, Ñopo (2012) found that textbooks depict gender roles in Peru.

Biological predispositions may demand differential job accommodations. For example, consider the birth of a child. Goldin (2014) analyzed US wage patterns and hours of work demanded by occupations from 1970 to 2011 and found that occupations with better pay demanded continuous hours of work. Corporate, legal, and financial jobs particularly penalized job interruptions. Goldin concluded that gender pay gaps could be explained by career interruptions or flexible schedules. Indeed, childbearing tends to reduce the wage of mothers worldwide (Simonsen and Skipper 2012; ILO 2015c). In Latin America and the Caribbean, Piras and Ripani (2005) studied labor patterns in Bolivia, Brazil, Ecuador, and Peru and found that in 1999, mothers with children age 7 or younger were less likely to have a paid job compared to men or women without children.

For women who work, Gamboa and Zuluaga (2013) identified a “motherhood penalty” of 1.8% in Colombia. In other words, mothers were paid 1.8% less than women not having children. For their part, Atal et al. (2009) identified a “family penalty” for both men and women. However, women were more affected because more women tend to drop out of work to raise their families.¹³ Indeed, women said work–family balance is the main obstacle to professional development (Thompson Reuters Foundation 2015). A survey by the ILO in 2015 covering more than 1300 private companies in 39 developing countries ranked family responsibilities as the main barrier to women’s leadership (ILO 2015a).

Policies to reduce the cost of motherhood have led to reductions in the gender pay gap. Childcare services and flexible work schedules have increased female employment in Latin America and the Caribbean (Attanasio and Vera-Hernandez 2004; Barros et al. 2011; Rosero and Oosterbeek 2011; Berlinski and Galiani 2007; Berlinski et al. 2011; Goldin 2014).

4.2 Use technology to reduce gender gaps

Technology has probably played the most important role in the shift in the gender gap in the last 25 years in Latin America and the Caribbean. Consider the use of contraception to regulate fertility. The average prevalence of using modern contraception among women ages 15 to 49 was 53% for 12 countries in Latin America and the Caribbean in 2006, ranging from 67% in Mexico to only 25% in Haiti (World Bank 2016).¹⁴ Miller (2010) found evidence for Colombia that family planning programs reduced fertility.¹⁵ Figure 2 shows that fertility decreased at a faster rate in the region in the 14 years from 1991 to 2014 than in the rest of the world. The increase in female labor participation in Latin America and the Caribbean in the last 25 years was accompanied by a decrease in fertility. Cruces and Galiani (2007) found that having one more child decreases the probability of the mother working for pay or profit in Argentina and Mexico.¹⁶

Fig. 2

Average fertility rate in Latin America and the Caribbean versus the rest of the world, 1991–2014. Source: Author’s calculations based on data from World Bank (2016). LAC Latin America and the Caribbean

Evidence in the USA and the UK suggests that other technologies, including maternal health, infant formula, and home appliance ownership, have increased the participation of women in the labor market (Albanesi and Olivetti 2015; Cohen-Pirani et al. 2010; Cavalcanti and Tavares 2008). New technologies may be enabling different schemes for work participation. For example, Internet access is allowing greater flexibility in work schedules and facilitating work from home. It remains a question how technology affects hours of work and whether it affects pay for women and men.

4.3 Strengthen institutions to ensure human, labor, and property rights

This section summarizes the importance of institutions to avoid discrimination, provide social protection, allow for the exercise of human rights, and enhance democratic participation and good government in Latin America and the Caribbean.

4.3.1 Government intervention is key to prevent discrimination

Regulation is important when differences in skills may lead to gender statistical discrimination. Consider official learning validation and certification. UNESCO and UNHCR (2016) find that these are often ignored in the provision of quality education. There are scant standardized tests that enable employers to determine which applicants for a position have the skills they need. The completion rate for secondary education is 41.6% in Latin America and the Caribbean (UNESCO 2016a, b). However, it is still an open question whether employers’ perceptions of skills result in discrimination.

4.3.2 Government intervention is key to ensure social protection

Social protection is linked to formal work. For example, pension contribution rates in Chile are 96% among men but only 75% among women, while in El Salvador pension contribution rates are 85% among men but only 55% among women (Bosch et al. 2013). In the Latin America and the Caribbean region, the gap closes at the time pensions are received, when 65% of men and 61% of women 65 years or older receive a pension (Bosch et al. 2013). Therefore, women are more likely to demand social assistance rather than to become entitled to pensions. Thus, policies to reduce informality are likely to disproportionately benefit women.

4.3.3 Government intervention is necessary to ensure human rights

Latin America and the Caribbean has the dubious distinction of being the most violent region in the world. The homicide rate was 29.3 among men and 3.7 among women per 100,000 population in 2012 (UNODC 2013). While violence disproportionately affects men, the shape of that violence is different across the sexes. Most women are killed by people expected to care for them, while men are killed by people they may not know (UNODC 2013). Half of female homicide victims are killed by an intimate partner; in contrast, 6% of male homicide victims are killed by an intimate partner (UNODC 2013). Many countries have adapted their regulatory framework to address human rights across sex groups. For example, 18 of 29 countries in Latin America and the Caribbean have explicitly criminalized marital rape (World Bank 2016). However, there is still work to be done in the rest of the countries. In addition, enforcement of the law may not be ideal. Ensuring individual rights for women may be costlier than ensuring those rights for men, given that ensuring rights for women requires regulation within the household. More research is needed in this area.

4.3.4 Government intervention is necessary to enhance good government

The democratic process demands inclusive citizen participation. However, gender gaps in occupations extend to participation in the political process. Consider the participation of women in political positions at the national level. By 2015, all Latin American countries except Guatemala and Venezuela had laws to promote female participation (Freidenberg and Lajas 2015). In 2015, women occupied 27% of national parliamentary seats in the region (ECLAC 2015), far short of their share of the population.

One frequent approach to promoting female political participation is to set gender quotas. Zetterberg (2009) finds that gender quotas have a limited impact on political trust, knowledge, or interest among women. He concludes that women do not perceive quotas as empowering in Latin America, and he argues that informal nomination procedures favor women with close ties to particular leaders. Thus, the process sends the message that procedures are unfair and that they discourage qualified women.¹⁷

Table 3 summarizes the goals and evidence of related gaps discussed in this section and lists related articles in the Declaration of Human Rights. The solution to the gender gap is a combination of policies to enhance equality of opportunities, guarantee equal access to salaried positions, promote the use of technology, and strengthen institutions. Taken together, these policies are directed to enhancing the exercise of human rights.

Table 3 Proposed solutions to gender gaps, related gaps, and human rights.

To sum up, formal institutions are important to ensure that citizens can exercise their human rights and be free of discrimination. Formal institutions play a central role in facilitating coordinated and collective action and providing social protection. Participation in the democratic process provides incentives for shaping and maintaining good government. The main challenge is to enforce regulations for individuals trading within the household or in informal markets. Another challenge is how to promote participation by women in the democratic process. Policies to address these issues are likely to promote gender equality. Addressing these issues would benefit not only to women but also men in similar circumstances. The next section discusses why identifying more specific policies may be challenging.

5 Challenges to identifying evidence-based gender equality policies

Promoting gender equality based on evidence is not easy. There are important areas where better evidence is needed to design policy. This section lists the main challenges to identifying evidence-based policies to promote gender equality.

The first policy challenge is to better understand individual choices on occupations. There is a need to understand the role of nature versus nurture in this choice, as well as the mechanisms through which society influences these choices. The second challenge is to better understand how to use technology responsibly. Indeed, technology may bring controversial results, such as facilitating abortion or increasing the double burden on individuals (i.e., the workload of people who have jobs and are also responsible for unpaid domestic labor). There is a need to better understand the effects of programs that provide family planning, as well as technological advances such as Internet access that facilitate working from home. The third challenge is to better understand how to strengthen institutions and make them more inclusive while respecting privacy. Indeed, transactions within the household limit the ability of the government to act.

A better understanding of individual choices, the use of technology, and the reach of institutions would aid international efforts to identify effective evidence-based policy. However, research related to gender gaps faces many challenges. The first is to identify specific mechanisms relevant for policymaking. For example, consider vouchers to subsidize wages. Vouchers have resulted in a higher probability of employment among women in Argentina (Galasso et al. 2004). However, it is difficult to determine a long-term solution using vouchers. The subsidy could address several contributing factors to gender inequality, including shortcomings in productivity, compensation for discrimination, or higher regulatory costs. But it is difficult to tell if such subsidies contribute to close the gender gap over the long term or are just a temporary fix.

Two mechanisms especially difficult to separate are those related to individual and social behavior. Individuals adapt their behavior according to what others do. For example, married men in the Philippines changed their expenditure choices when their wives knew they received money (Ashraf 2009). Gender interventions are prone to triggering broad social effects. In these cases, the researcher will not be able to separate the group effect from the sum of aggregated individual effects. This problem is known as the reflection problem, and it was first proposed by Manski (2000).

A second challenge is the difficulty of obtaining relevant data. For example, current public data do not allow for ascertaining how much of the increase in women’s labor is a substitution for unpaid work. Many organizations are currently working to collect these data through time-use and employment surveys. For example, the Organization for Economic Co-operation and Development (OECD), United Nations, and World Bank are starting to put together data collected across countries.¹⁸

Another data-related challenge is that men and women differ in dimensions other than sex, including developmental history.¹⁹ In addition, gender gaps are related, in part, to behavior that takes place within the household and to labor not traded in the labor market. In addition, observers may induce unconscious bias in research. Thus, measurement may be costly and prone to bias.

The third challenge is that factors that influence the gender gap may not be detected with a single statistic. Consider the distribution of skills across the sexes. Assume that both men and women have a standard normal distribution with a common mean. However, suppose the variance is different among groups. Heckman (1998) notes this case will result in different shares across a common threshold value other than the mean. Thus, using a common rule for both groups may result in gaps.²⁰

Despite these limitations, there are opportunities to generate evidence. Various policies already in place lack a formal evaluation. For instance, gender firm certification programs aim to ensure equal treatment across genders in several dimensions, including recruiting, professional development, pay, work–life balance, and sexual harassment. Programs currently operate in Argentina, Brazil, Chile, Costa Rica, Mexico, and Uruguay (Inter-American Development Bank 2015), yet they have not been evaluated. Programs with the goal of changing social norms also need more evaluation. For example, the Somos Diferentes, Somos Iguales Program in Nicaragua, and Program H and Program M in El Salvador and Brazil are only recently being evaluated (Inter-American Development Bank 2015; Promundo 2016; United Nations 2011; Bando et al. 2019). Programs that provide multiple services in a single geographical point also need to be studied. For instance, Ciudad Mujer in El Salvador provides labor intermediation, health, psychology, legal, and childcare services. Researchers are currently studying this program (Inter-American Development Bank 2015).

Another opportunity to generate evidence is to consider gender in broader development research. Evaluating sex-specific delivery of goods and services should aid learning and improve understanding of how to promote inclusive development. Better data and a better understanding of those data promote innovation. For example, in Latin America electricity is more expensive for residential use than commercial use. Women are more likely to own a small home business. Thus, considering other inputs for production as equal, women face a higher price for production (Komives et al. 2005). This finding indicates that a flat rate for electricity affects opportunities for one group more than the other.

Including gender in policies to promote development may also prevent unexpected adverse effects. Policies that ignore gender could result in unequal access, violate human rights, or increase violence. However, flexible works schedules may result in women working more hours to meet expectations. Another example is that of conditional cash transfers in Mexico, where transfers to women resulted in short-term increases in domestic violence (Bobonis et al. 2013).

In summary, there are important areas to explore: the acquisition of skills and occupational choice, the use of technology, and the strengthening of institutions. Future work to identify policies to close gender gaps based on evidence will face challenges. First, research must pay attention to the many mechanisms through which an intervention affects gender gaps. Second, men and women differ in dimensions other than sex. Third, men and women may differ in statistics other than the mean. Despite these challenges, research in development could inform knowledge on how to address gender gaps and help avoid potential adverse gender effects.

6 Conclusion

The percentage of women employed has increased faster in Latin America and the Caribbean than in any other region in the world in the last 25 years. However, many women do not see the benefits of their work. Women work more hours than men, but half of women in Latin America and the Caribbean do not get pay or profit. The high share of women without pay leaves them with limited entitlement to benefits. Most women work at home or in the informal sector. Thus, they are less likely to benefit from legal protection. On average, women face more obstacles to benefiting from certain human and labor rights and rights as citizens.

Gender gaps result from biological predispositions and expectations by society as to what occupation persons of a given sex should choose and how much they should be paid. In Latin America and the Caribbean, gender roles have it that women are better suited than men to pursue activities that directly benefit the household or the family. Gender roles influence occupational choice and skills. Other factors such as discrimination, regulations, and technology also influence this choice.

Policies to reduce gender gaps can be grouped into three categories: policies that aim to ensure equality of opportunity across genders; policies directed toward the use technology; and policies to ensure individual rights.

Policy design and determining whether policies should be tailored across sexes demands clear answers to some outstanding questions. There is a need to understand the dynamics and long-term effects of human capital accumulation across sexes and how it affects individual choices. Thus, studies over long periods of time would be insightful. There is a need to learn how to protect labor and human rights within the household and in informal markets; to find ways to make social services more inclusive; to find ways to promote participation by all in the democratic process; and to understand how social norms modulate the use of technology to reduce gender gaps. In summary, given that men and women are not the same, there is a need to understand how both can enjoy equal opportunities and equal rights, and to find ways to spread the cost of unpaid labor more equally.

Research in this area faces several challenges. First, policies affect behavior through several pathways, so one needs to understand the roles of those pathways and how they interact in order to design effective policies. Second, researchers may face challenges in finding men and women with similar characteristics to construct a proper sample. Indeed, men and women are raised differently. Thus, long-term research may be necessary to understand a specific intervention net of current differences. Third, gender differences are likely to matter for policy design in distributional features other than the mean. Thus, detailed analysis is necessary to inform policy.

Interventions for development can further understanding about gender even when gender is not the central issue. Indeed, many current development studies provide key information about gender gaps. Consider studies on pensions, informality, and democratic participation. These development issues relate to the gender gap, so they represent an opportunity to generate evidence. Therefore, when possible, such development studies should include a careful gender analysis. These exercises would likely bring Latin America and the Caribbean closer to finding the answer to when and how to address gender gaps.