Unequal Before the Law: Measuring Legal Gender Disparities Across the World

Several economies have laws that treat women differently from men. This study explores the degree of such legal gender disparities across 167 economies around the world. This is achieved by constructing a simple measure of legal gender disparities to evaluate how countries perform. The average number of overall legal gender disparities across 167 economies is 17, ranging from a minimum of 2 to a maximum of 44. The maximum possible legal gender disparities is 71. The measure is found to be correlated with other measures of gender inequality, implying the measure does capture gender inequality while also differing from preexisting measures of gender inequality. A high degree of legal gender disparities is found to be negatively associated with a wide range of outcomes, including years of education of women relative to men, labor force participation rates of women relative to men, proportion of women top managers, proportion of women in parliament, percentage of women that borrowed from a financial institution relative to men, and child mortality rates. Subcategories within the legal disparities measure help to uncover specific types of legal disparities across economies.

Several economies have laws that treat women differently from men. This study explores the degree of such legal gender disparities across 167 economies around the world. This is achieved by constructing a simple measure of legal gender disparities to evaluate how countries perform. The average number of overall legal gender disparities across 167 economies is 17, ranging from a minimum of 2 to a maximum of 44. The maximum possible legal gender disparities is 71. The measure is found to be correlated with other measures of gender inequality, implying the measure does capture gender inequality while also differing from preexisting measures of gender inequality. A high degree of legal gender disparities is found to be negatively associated with a wide range of outcomes, including years of education of women relative to men, labor force participation rates of women relative to men, proportion of women top managers, proportion of women in parliament, percentage of women that borrowed from a financial institution relative to men, and child mortality rates. Subcategories within the legal disparities measure help to uncover specific types of legal disparities across economies. Africa, 2004). Details of the indices are summarized in table A3. There are several issues with using outcome-based measures. First, typically income is included in these measures, which tends to have the highest degree of variation. This is problematic as the variable with the most variation gets the highest weight (Dijksra, 2001). Second, several of the outcome data are sparse, measured differently across economies, and mix different aspects such as empowerment and well-being (Klasen and Schueler, 2011). For instance the Gender-Related Development Index relies on income measures which are derived from a gender breakdown of labor force participation data and nonagricultural earnings data. Labor force participation data are unreliable and difficult to compare across countries. Furthermore earnings data are sparse and may come from sectors that do not represent the whole working population. Thus these measures can be problematic.

2005) and the African Gender Status Index (Economic Commission for
Measures that are more in line in with this study include the Social Institutions and Gender Index (SIGI;Branisa et al., 2014) and two indices from the CIRI Human Rights Data project -Women's Economic Rights Index (WECON), and the Women's Social Rights Index (WOSOC).
WECON concentrates on women's equal rights in the labor market while WOSOC focuses on women's rights in the social sphere including education, marriage, travel etc. The information on laws for these indices is based on the data collected by the Women, Business and the Law (WBL) project of the World Bank Group.
There are certain differences between this measures and what this study proposes. Both WECON and WOSOC capture a subset of laws and provide scores in the range of 0 to 3 based on the premise of whether internationally recognized human rights for women are included in the laws.

6
The narrow range of the scores from 0 to 3 limits the possibility of capturing the heterogeneity of laws across countries given that there will be a lot of clustering around certain scores. The SIGI index goes beyond legal institutions and attempts to capture social institutions in general by combining data on laws from WBL with other measures from the OECD Gender, Institutions and Development (GID) database. In contrast the WBL measure of legal gender disparities proposed in this study relies only on data on laws and provides a wide range of scores given the measure is based on disparities in a wide number of laws.
There are several advantages of the WBL measure of legal gender disparities. First by narrowing down the focus to legal institutions, many conceptual and empirical issues are cast aside. For instance the problems of using earnings measures that include inconsistent methodologies of data collection across countries and patchy data are entirely avoided as laws are more precisely measureable. Conceptually what this measure tries to capture is well defined, so there is no combination of measures of different elements that may lead to some loss in clarity of interpretation. Furthermore, the measure presented is not an index. It is a simple summation of legal disparities, which on one hand does away with a lot of nuances, but on the other hand presents a measure that is easily replicable, and malleable to any particular need. Complexity and difficulty in replication are common problems with established indices, which then become inaccessible for various institutions including NGOs (Charmes and Wieringa, 2003). It is important to note that the measure is maintained by WBL, who are the primary data generators of legal disparities that are used by other indices. This is a unique advantage as it ensures the measure will be at the cutting edge of uncovering new laws and updating existing laws that may have important implications for gender equality.
To summarize, the contribution of the study is as follows (i) constructs a measure of legal gender disparities including subcategories, (ii) elucidates the patterns of legal gender disparities across the world, (iii) validates the measure by exploring correlations with other gender inequality measures, and (iv) analyzes the association with gender legal disparities and several genderrelated outcomes. Section 2 provides the methodology, section 3 unveils the patterns of legal gender disparities, section 4 shows the validity of the measure with other similar measures, section 5 shows the association of the measure with various outcomes, and finally section 6 concludes.

Methodology
The construction of the WBL measure of legal gender disparities is based on three main principles. These principles are as follows. Women are considered to face unequal treatment before the law if (i) laws exist with explicit gender-based differences, or (ii) there is an absence of laws that protect the status of women e.g. absence of a non-discrimination clause in the constitution, or (iii) general laws exist that may imply gender-based legal differences or undermine existing laws that protect the status of women e.g. personal or customary law. Laws that satisfy any of these principles are marked as legal disparities for married and unmarried women. The measure of legal gender disparities can be presented by the equation below for economy j: Where U is the total number of legal disparities for unmarried women, M is the total number of legal disparities for married women, and i represents a legal disparity. As shown above there are 7 major subcategories for which laws are considered. These subcategories are Accessing institutions (AccInst), Using property (Property), Getting a job (GetJob), Providing incentives to work (IncWrk), Going to court (Court), Building credit (credit) and Protecting women from violence (Violence). Accessing institutions encompasses the legal ability of women to interact with public authorities and the private sector in the same way as men. Examples include whether married women can pursue a trade or profession in the same way as married men, and whether married women can obtain a national ID card the same was as married men. Using property explores the ability of women to own, manage, control, and inherit property. Getting a job examines restrictions on women to work, for example are women restricted from working in certain professions, tasks or even night hours. Providing incentives to work assesses differences in tax credits and deductions available for women versus men. Going to court considers whether women's testimony in court is given the same weight as men's. Building credit covers legal disparities in access to credit on the basis of gender or marital status. Finally, Protecting women from violence explores whether laws that limit domestic violence exist and the scope of such legislation.
One important implication of equation (1) is that if a law applies to both married and unmarried women, it is counted twice. For example if an unmarried woman cannot apply for a passport the same way as an unmarried man, this is counted as one legal disparity. If a married women cannot apply for a passport the same way as a married man, this is counted as another legal disparity. If the law does not prohibit discrimination by creditors on the basis of gender in access to credit, 9 then this is counted as two legal disparities as it applies to both married and unmarried women.
Thus the value of a disparity is either 1 if it only applies to married, and 2 if it applies to both married and unmarried women. There is only one specific situation where the value of the disparity is not assigned a score of 1 or 2 -differences in paternity and maternity leave. In this case the disparity is measured as (1-(P/M)), where P is length of paternity leave and M is maternity leave. A value of 0 would indicate parity in paternity and maternity leave. A value of 1 would indicate the highest degree of disparity as length of paternity leave would essentially be zero. Thus the greater the length of maternity leave relative to paternity leave, the greater the gender disparity (paternity leave does not tend to be greater than maternity leave). Since the disparity applies to both married men and women, it is multiplied by 2. The rationale behind having a continuous measure for the disparity in paternity and maternity leave is to capture the heterogeneity in the amount of leave allowed across economies. A full listing of legal gender disparities and the respective score assignment are presented in table A1in the appendix by subcategory. Data on legal disparities are available for 167 economies based on data collected for the 2016 Women, Business and Law report.

Economy Rankings and Regional Patterns
The ordering of economies from the lowest to the highest number of legal gender disparities is presented in the appendix in

Composition of legal gender disparities
There is a substantial variation of the composition or type of legal gender disparities by region. This is uncovered by exploring the proportion of legal gender disparities by each subcategory of the composite measure. The proportion of legal gender disparities can be interpreted as follows.
Accessing Institutions subcategory non-discrimination clause in the Constitution does not explicitly mention gender, and the law does not prohibit discrimination by creditors on the basis of marital status in the Building Credit category. For Australia, legal disparities come from absence of non-discrimination clause based on gender in the Constitution in the Accessing Institutions subcategory and absence of legal obligation for employers to provide break time for nursing mothers in the Getting a Job subcategory. For Spain, legal disparities in the Getting a Job subcategory come from a difference in the amount of paternity and maternity leave and absence of legal prohibition for employers to ask about family status during interview; in the Building Credit subcategory -from absence of prohibition of discrimination on the basis of marital status; from Protection of Women against Violence subcategory -from absence of legal criminalization of marital rape in the legislation. For Mexico, legal disparities come from a difference in the amount of paternity and maternity leave, absence of legal prohibition for employers to ask about family status during interview and absence of legal provision mandating equal remuneration for work of equal value in the Getting a Job subcategory.  There is substantial heterogeneity in legal gender disparities within regions as well. Figure A3 presents the proportions of legal gender disparities in Sub-Saharan Africa by subcategory.   Figure A5 shows the source of gender legal disparities for LAC. Haiti has the highest proportions of legal gender disparities while Mexico has the lowest. Haiti attains this status due to the lack of domestic violence legislation. Figure A6 shows the source of gender legal disparities in ECA. The Russian Federation has the highest proportions of legal gender disparities followed by Uzbekistan, mostly due to disparities in Violence. Croatia has the lowest proportions of legal gender disparities in the region.

Correlations with Other Gender Indices
The WBL composite measure of legal disparities provides a clean measure of differential treatment of men and women before the law. There are two potential critiques of this measure. The correlation between the WBL measure of legal gender disparities and other measures of gender inequality does provide some support of the view that the law in the books may influence realities in the ground, addressing the first critique of the measure. Furthermore, none of the gender-related indices are perfectly correlated with the WBL measure of legal gender disparities.
This implies that the measure developed captures a different and more specific nature of legal institutions unlike the other gender-related measures. Thus the WBL measure of legal gender disparities provides additional empirical information to the existing plethora of gender-related indices, addressing the second critique of being redundant.

Relationship with Gender-Specific Outcomes
In this section simple regression analysis is conducted to explore the relationship between the WBL measure of legal gender disparities and gender-specific outcomes. Furthermore, for each outcome, subcategories of the WBL measure are explored. The theoretical underpinning of how gender inequality before the law affects an assortment of gender-specific outcomes is straightforward. The literature has empirically shown that household bargaining models are more reflective of reality than unitary household models (Thomas, 1997). Household bargaining models are characterized by household decisions dictated by the distribution of bargaining power within the household. In contrast, unitary models assume households make decisions as a single unit. The main implication is that the greater the bargaining power of women, the more influence they will have on decisions in the household. Such bargaining power is determined by the threat point which is influenced by a number of factors including "extrahousehold environmental parameters" (McElroy, 1990). 3 Social and legal institutions can be considered salient determinants of such "extrahousehold environmental parameters" (Branisa et al., 2013). Thus, unequal gender laws reduce the bargaining power of women that then reduces their influence in household decisions. This results in worse outcomes for women such as low labor force participation, low education as well as lower access to financial resources. Furthermore if mothers' are the primary caregivers of children, lowering their bargaining power could lead to lower health outcomes for children (Maitra, 2004;Thomas, 1997).
Accordingly, the following gender-specific outcomes are considered in this study: women's education as measured by years of education of women relative to men, economic empowerment as measured by women's labor force participation relative to men as well as the proportion of women top managers, political empowerment as measured by the proportion of women in parliament, percentage of women that borrowed from a financial institution relative to men, and child mortality rates. All regressions presented are cross-country in nature, using OLS. The estimations account for the level of development using real GDP per capita, and regional factors using region-specific fixed effects. There is a possibility that legal gender disparities may be capturing the quality of institutions in general, and thus all estimations control for the Rule of

Education
Results for women's years of education relative to men are presented in table 3. Overall legal gender disparities have a negative association with women's years of education relative to men, statistically significant at the 5 percent level. Thus an increase in the laws that treat men and women differently is associated with lower years of education for women relative to men. Using the measure of women relative to men years of education ensures that factors that affect both men and women's education levels in general are accounted for. For the subcategories, legal disparities in Accessing institutions, Going to court, Providing incentives to work, and Violence protection have a statistically significant negative association with women-to-men years of education. The rest of the subcategories have negative but statistically insignificant coefficients.
Thus laws that restrict women's ability to interact with the private sector, reduce work incentives, limit domestic violence protection and have disparity of weight of testimony of women vs men are correlated with lower education of women relative to men.

Economic empowerment
Of course, improvements in women's education may not necessarily translate into improvements in economic empowerment. Therefore we explore the relationship between legal gender disparities and labor force participation of women relative to men as presented in table 4. Overall legal gender disparities have a negative association with women's labor force participation rates relative to men, statistically significant at the 1 percent level. Thus an increase in the laws that treat men and women differently is associated with a lower participation of women in the labor force relative to men. Taking the women-to-men ratio in labor force participation rates ensures that factors that affect both men and women's labor force participation rates in general are accounted for. For the subcategories, legal disparities in Accessing institutions, Using property, Getting a job, and Going to court have a statistically significant negative association with women-to-men labor force participation rates. The rest of the subcategories have negative but statistically insignificant coefficients. Thus laws that restrict women's ability to interact with the private sector, ability to own, manage, and inherit property, work restrictions, and disparity of weight of testimony of women vs men are correlated with lower labor force participation of women relative to men.
Increases in labor force participation for women may not entail economic empowerment, especially if women enter into low paying jobs that may not improve their welfare. We therefore consider an alternate measure of economic empowerment -the proportion of women top managers in formal firms. Table 5 shows that an increase in the overall legal gender disparities is negatively correlated with the proportion of women top managers. The finding is statistically significant at the 5 percent level. Regarding subcategories, legal disparities in Accessing institutions, Using property, and Going to court have statistically significant negative associations with the proportion of women top managers. The implication may be that such gender disparities in laws disempower women and thus they are less likely to advance in their careers into top managerial positions. 4 One important caveat is that data on women top managers are mainly available for developing economies from the World Bank Enterprise Surveys. Thus the findings in table 5 only apply to developing economies. However given the consistency of results between women's labor force participation and proportion of women top managers, the findings are suggestive of a negative relationship between high legal gender disparities and women's economic empowerment.

Political empowerment
Economic empowerment does not necessarily imply political empowerment, and thus we explore the relationship between legal gender disparities and the proportion of women in parliament. The findings in table 6 show that an increase in the overall legal gender disparities measure is associated with fewer seats held by women in national parliaments. The finding is statistically significant at the 1 percent level. Regarding subcategories, legal disparities in Getting a job, Building credit, and Violence protection have statistically significant negative associations with the proportion of women in parliament. The implication may be that such gender disparities in laws disempower women and thus they are less likely to be represented in parliament. Of course, reverse causality may be true as well whereby more women in parliament results in better laws for women. The data at hand are insufficient to disentangle the effects.

Access to finance
Results for women who borrowed from a financial institutions relative to men are presented in table 7. Overall legal gender disparities have a negative association with women borrowing from a financial institution relative to men. The finding is statistically significant at the 5 percent level.
Regarding subcategories, legal disparities in Accessing institutions, Using property, and Going to court have a statistically significant negative association with the percentage of women who borrowed from a financial institution relative to men. This is as expected given that disparities in in the Accessing institutions subcategory, such as whether a married women can obtain a national ID the same way as a man, could encumber women's ability to access formal financial institutions. Similarly the ability of women to own, manage, and inherit property may determine whether they have savings at all in the first place.

Child mortality
With regards to outcomes for children, mortality of children under 5 is explored. The findings are reported in table 8. Overall gender legal disparities have a positive association with child mortality, statistically significant at the 5 percent level. Thus the greater the legal disparities women face relative to men, the higher the child mortality rate. Only the Going to court and Violence protection subcategories have a statistically significant association with child mortality.
Thus, similar weights in testimonies by men and women as well as domestic violence legislation may improve women's empowerment, which in turn results in greater allocation of resources towards the children, increasing their chances of survival. It is also worth noting that unlike the previous indicators, there is no statistically significant association between legal disparities and mortality rates of boys relative to girls (results not presented). The findings, as indicated above, are only for children. The implication may be that women's empowerment improves mortality outcomes for both boys and girls equally.
The results presented provide some important findings regarding legal gender disparities and various outcomes. The subcategories of the legal gender disparities have heterogeneous effects across the outcomes, implying different elements of legal gender disparities may matter more or less for different outcomes. It is beyond the scope of this study to explain why certain forms of legal empowerment may matter more than others. One can speculate on a myriad of factors that may be influential including variation in enforcement of different types of laws, as well as institutions that may complement the effectiveness of certain laws as opposed to others.

Conclusion
In this study legal gender disparities across the world are explored with the construction of a composite measure, generally created as a summation of gender legal disparities. Subcategories of this measure are used to elucidate the heterogeneity of the type of legal gender disparities across several regions and economies. The composite measure is then shown to be correlated with other gender-related measures, albeit not perfectly correlated thereby indicating the composite measure does capture a different dimension than other gender-related measures.
Finally the composite measure is found to be correlated with important gender outcomes. A larger number of legal gender disparities is found to be negatively associated with years of education of women relative to men, labor force participation rates of women relative to men, proportion of women top managers, proportion of women in parliament, percentage of women that borrowed from a financial institution relative to men, and child mortality rates. These findings stand after accounting for the level of development, the rule of law, and region-specific factors.
The benefit of the legal gender disparities composite measure is that it offers a clean focus on the legal dimension of gender equalities that is precisely measured and emanates from the same data source. This ensures consistency across countries and accuracy. Furthermore, the measure is easily replicable, and through its subcategories can assist governments and policy makers to focus on specific gender legal disparities that exist in their economies.
There are certain limitations of the proposed measure. First, given the nature of the focus, the measure cannot provide any indication of the degree of enforcement of the laws. However the positive association uncovered with various gender outcomes and other composite measures may indicate there is some degree of enforcement otherwise none of these results is likely to hold.
Second, certain laws typically apply to the formal economy but not the informal economy, and therefore the implications of legal gender disparities in the informal economy is beyond the scope of this measure. Third, while the underlying data are rich and collected through the rigorous efforts of the Women, Business and the Law project of the World Bank, the laws explored are not the universe of all possible laws. Some laws may be omitted due to difficulty in obtaining a consistent indicator across economies, or simply due to the fact that they may be covered in future reports. Finally the data are cross-sectional in nature, however a panel dimension can eventually be developed. Thus in time, it can be expected that the proposed measure will be more comprehensive. This will be a boon for future research, as well as policy makers who wish to measure progress of gender equality in relation to the Sustainable Development Goals (SDG).

WBL Legal Gender Disparity Composite:
A "no" response to questions below is considered a disparity and therefore assigned a score

Accessing Institutions
If there is a nondiscrimination clause in the constitution, does it mention gender? 2 If customary law is recognized as a valid source of law under the constitution, is it invalid if it violates constitutional provisions on nondiscrimination or equality? 2 If personal law is recognized as a valid source of law under the constitution, is it invalid if it violates constitutional provisions on nondiscrimination or equality? 2 Can an unmarried woman apply for a passport in the same way as an unmarried man? 1 Can a married woman apply for a passport in the same way as a married man? 1 Can an unmarried woman obtain a national ID card in the same way as an unmarried man? 1 Can a married woman obtain a national ID card in the same way as a married man? 1 Can an unmarried woman travel outside the country in the same way as an unmarried man? 1 Can a married woman travel outside the country in the same way as a married man? 1 Can an unmarried woman travel outside her home in the same way as an unmarried man? 1 Can a married woman travel outside her home in the same way as a married man? 1 Can an unmarried woman get a job or pursue a trade or profession in the same way as an unmarried man? 1 Can a married woman get a job or pursue a trade or profession in the same way as a married man? 1 Can an unmarried woman sign a contract in the same way as an unmarried man? 1 Can a married woman sign a contract in the same way as a married man? 1 Can an unmarried woman register a business in the same way as an unmarried man? 1 Can a married woman register a business in the same way as a married man? 1 Can an unmarried woman open a bank account in the same way as an unmarried man? 1 Can a married woman open a bank account in the same way as a married man? 1 Can an unmarried woman choose where to live in the same way as an unmarried man? 1 Can a married woman choose where to live in the same way as a married man? 1 Can an unmarried woman confer citizenship on her children in the same way as an unmarried man? 1 Can a married woman confer citizenship on her children in the same way as a married man? 1 Can an unmarried woman be head of household or head of family in the same way as an unmarried man? 1 Can a married woman be head of household or head of family in the same way as a married man? 1 Can a married woman confer citizenship to a non-national spouse in the same way as a man? 1 Are married women required by law to obey their husbands? 1

Using Property
Who legally administers marital property? 1 Does the law provide for the valuation of nonmonetary contributions? 1 Do unmarried men and unmarried women have equal ownership rights to property? 1 Do married men and married women have equal ownership rights to property? 1 Do sons and daughters have equal rights to inherit assets from their parents? 2 Do female and male surviving spouses have equal rights to inherit assets? 1

Going to Court
Does a woman's testimony carry the same evidentiary weight in court as a man's? 2