WPS6616 Policy Research Working Paper 6616 Women’s Legal Rights over 50 Years Progress, Stagnation or Regression? Mary Hallward-Driemeier Tazeen Hasan Anca Bogdana Rusu The World Bank Development Research Group Finance and Private Sector Development Team September 2013 Policy Research Working Paper 6616 Abstract Using a newly compiled database of women’s property associated with higher probabilities for reform, but rights and legal capacity covering 100 countries over only in lower-income countries. With the remaining 50 years, this paper analyzes the triggers and barriers constraints as prevalent in middle- as low-income to reform. The database documents gender gaps in the countries, increased growth is not necessarily going to ability to access and own assets, to sign legal documents spark additional reforms. Clearer patterns emerge from in one’s own name, and to have equality or non- the momentum created by international conventions, discrimination as a guiding principle of the country’s such as the Committee to Eliminate All Forms of constitution. Progress in reducing these constraints has Discrimination against Women (CEDAW), women’s been dramatic—half of the constraints documented in political representation at the national level, mobilization the 1960s had been removed by 2010. However, some of women’s networks, and increasing labor force sticky areas persist where laws have not changed or have participation in sectors that provide a voice for women, even regressed. The paper analyzes potential drivers of which are positive forces for change. Conversely, conflict reforms. A significant finding is that the relationship and weak rule of law can entrench a discriminatory with a country’s level of development and the extent status quo. And much is at stake; strengthening women’s of its reforms is not straightforward. For the first half legal rights is associated with important development of the sample, there was no systematic connection; outcomes that can benefit society as a whole. only in the last 25 years have increases in income been This paper is a product of the Finance and Private Sector Development Team, Development Research Group. It is part of a larger effort by the World Bank to provide open access to its research and make a contribution to development policy discussions around the world. Policy Research Working Papers are also posted on the Web at http://econ.worldbank.org. The authors may be contacted at mhallward@worldbank.org. The Policy Research Working Paper Series disseminates the findings of work in progress to encourage the exchange of ideas about development issues. An objective of the series is to get the findings out quickly, even if the presentations are less than fully polished. The papers carry the names of the authors and should be cited accordingly. The findings, interpretations, and conclusions expressed in this paper are entirely those of the authors. They do not necessarily represent the views of the International Bank for Reconstruction and Development/World Bank and its affiliated organizations, or those of the Executive Directors of the World Bank or the governments they represent. Produced by the Research Support Team Women’s Legal Rights over 50 Years: Progress, Stagnation or Regression? Mary Hallward-Driemeier, Tazeen Hasan and Anca Bogdana Rusu 1 JEL Codes: K37; N40; O1 Key terms: Law; gender; property rights; legal capacity; CEDAW; development. 1 World Bank, Development Research Department, Finance and Private Sector Development. The authors gratefully recognize financial support from the UK’s Department For International Development (DFID). Special thanks go to Sarah Iqbal who, along with Tazeen Hasan, supervised the data collection effort, and to the team members from the Women, Business and the Law who contributed to the database: Garam Dexter, Khrystyna Kushnir, Lourdes Lizarraga Ortega, Nayda Almodovar Reteguis, Paula Tavares, Thibault Meilland and Yasmin Klaudia bin Humam. In addition, Abhinav Goel, Ashley S. Zohar, Erin F. Rogers, Lindsey Scannell, Sarah Knapp, Shazia Akhter, Tatyana Chursova, Yulia Dovgaya and Zachary Walter conducted research for the database as part of their internships with the World Bank Group. The opinions expressed here do not necessarily reflect the position of the World Bank, its Board of Executive Directors or the member countries they represent. Comments can be sent to mhallward@worldbank.org. 1 Women’s Legal Rights over 50 Years: Progress, Stagnation or Regression? “Life must be lived forward, but it can only be understood backward.” Kierkegaard “Study the past if you would define the future.” Confucius Introduction "Gender equality means having the same rights and opportunities." "It means the same rights at work and in the home." Two urban adult women, Lambayeque, Peru. 2 Discriminatory laws in many countries across the world continue to act as a significant brake on women’s entrepreneurship. Gender differences are not found in business laws, which are generally gender-neutral, but in laws that are ostensibly supposed to regulate the private space, such as family and inheritance laws. These laws can determine the amount of assets controlled and owned by women, restrict their mobility, and limit their autonomy to make everyday decisions, such as getting a job outside the home or going to court. The restrictions can trigger marriage resulting in a paradox; a married woman can have a similar legal capacity as a child when it comes to opening a bank account or signing a contract in that she cannot do it on her own. This study tracks the evolution of key constraints to women’s and girls’ equal rights to property and restrictions on their legal capacity over the past 50 years across 100 countries. The database of legal indicators represents all geographic regions, legal traditions, and income levels (see annexes 1-3 for more details on methodology and country coverage). The paper describes the nature and implications of these legal restrictions and analyzes how their prevalence has changed across countries and over time. It demonstrates that change is possible; over half of the constraints that were on the books in the 1960s were removed by 2010. Whereas 75 countries in the initial period had at least one constraint to women’s economic rights, only 47 did in 2010. Of the 75 countries with some initial inequalities, all but 18 countries had closed at least one legal gender gap. And 28 countries had eliminated all the inequalities analyzed here. The extent of progress, however, is uneven. Eastern Europe and Central Asia, a region that began with very few inequalities, no longer has any that are followed here. Many high-income OECD countries started with several inequalities, but all have removed them and countries in Latin America have also made tremendous progress. Countries in Sub-Saharan Africa, which had a higher rate of constraints than any other region, have removed more than half of the constraints that had been on the books. East Asian countries have also halved the number of constraints. However, in the Middle East and North Africa and in South Asia, much less progress has been realized. 2 World Bank Dataset 2011a. 2 This paper explores some of the factors that can trigger reforms. It examines which country characteristics and processes are associated with reforms, focusing on income growth, education, patterns of employment, conflict, women’s empowerment (e.g. women parliamentarians, ratification of CEDAW, women’s age of first marriage) and the strength of the rule of law. The relationship between reform and higher income is not straightforward. It is not the case that development itself will help close these gender gaps. While income is not associated with reforms during the first half of the period studied here, in the last 20-25 years, there is more of a positive association between income and reforms, but only in low-income countries. As the constraints that still remain on the books are as prevalent in middle-income countries as low- income countries, this is unlikely to be a strong predictor of future reforms. On the other hand, two positive trends are particularly striking. The first trend can be referred to as the “CEDAW effect.” The ratification by states of the CEDAW convention is linked to reform within a short-term period, benefiting unmarried women in particular. The average pace of reform within the 5 years after ratification is almost double the pace in the 15 years prior to ratification. The second significant trend is based on increases in women’s political representation. Increased female legislators at the national level (25% and over) has catalyzed the reform of discriminatory laws. Thus, CEDAW and higher rates of women in parliaments are associated with more reforms closing legal gaps in women’s economic rights. If active engagement is needed to move the reform agenda forward, these may be the more promising channels. The next section will give a brief background of the literature of the impact of legal reforms, motivating why it is important to remove the remaining inequalities women and girls face. The paper then provides more background on the historical and legal sources for why many of the constraints are on the books in various countries. The following section then discusses the methodology for collecting the legal indicators. Then paper examines the overall trends – where there has been improvement and where the sticky points are. Finally we analyze in greater detail some of the underlying factors behind these trends. Literature Review of the Impact of Legal reforms Reforming discriminatory laws is important for women’s economic participation. A 2013 World Bank report highlights how gender differences in legal status and property rights are a significant constraint on women’s entrepreneurship. Stronger economic rights are linked to higher ratios of female employers within a country. Weaker property rights undermine women’s ability to access assets to use as collateral for loans and grow their businesses. A study demonstrated that improving women’s property rights in Burkina Faso could increase agricultural production by 6% with no additional resources – solely by the reallocation of resources (fertilizer and labor) from men to women. 3 Improving the security of land titling can increase women’s labor 3 Udry 1996. 3 participation by liberating them from standing watch over their property, as reported in a 2007 study of slum dwellers in Peru. 4 The reform of family laws in Ethiopia in 2000, which raised the minimum age of marriage, removed the husband’s ability to deny permission to work outside the home, and required both spouses’ consent in the administration of marital property, was linked to increased female labor force participation and in more productive sectors. 5 A 2007 study highlights that laws demarcating women’s working hours or types of employment that are found today in approximately 75 countries 6 can constrain women’s employment opportunities and significantly increase gender wage gaps. 7 A 2012 World Bank study on norms and agency found that in countries such as Yemen, where women’s physical mobility is restricted, women were less attractive as potential employees. Even where they were as qualified or even more qualified than male applicants, the need for a husband’s permission to work made them less reliable in the eyes of employers. 8 Improving women’s access to assets can benefit society as a whole and can result in transformative changes. Studies using evidence from across a range of countries demonstrate that increasing the share of household income that women control either through their own income or cash transfer can result in greater investment in children. 9 Inheritance law reforms in India equalizing inheritance shares for unmarried girls resulted in delayed marriage and more education for girls. 10 A quantitative study in Benin found that women who were able to earn an income were more likely to buy a bed net to prevent malaria transmission, which could be used by children. Men were more likely to purchase bed nets for their own personal use on the basis that as the breadwinner of the family they were entitled to protection. 11 A 2009 IFPRI research paper showed that if men and women were to have equal rights, the number of malnourished children could be reduced by 1.7 million in Sub-Saharan Africa and by 13.4 million in South Asia. 12 Context for Understanding Approaches to Legal Rights With a few significant exceptions, most of the present day legal constraints are found in decades – or even centuries – old statutes that have not been reformed. Contemporary national legal frameworks did not emerge from a vacuum – like culture, language and social norms, legal 4 Field 2007. 5 Hallward-Driemeier and Hasan 2012. 6 World Bank Group. Data retrieved April 14, 2013, from Women, & the Law database (2011a). 7 Weichselbaumer and Winter-Ebmer 2007. 8 Turk, Munoz-Boudet, and Petesch 2012. 9 Doss 2006 and Quisumbing and Maluccio 2000. 10 Deininger, Goyal and Nagarajan 2010. 11 Rashed 1999. 12 Ragasa and others 2012. 4 frameworks are a melting pot of historical influence, partly indigenous and partly imported through colonial histories. Different legal systems have comparative strengths and weaknesses when it comes to gender rights (see Annex 3 for a complete list of countries by legal tradition). Historically family law was traditionally seen as non-state law that was determined by traditional authorities such as tribal chiefs and religious officials. Over time and at different times, states started assuming control over family law as part of a state-building process either through sidelining the traditional authorities or striking bargains to consolidate their power. 13 Socialist systems were the most progressive on paper when it comes to gender rights and these legal traditions typically do not have legal constraints. The socialist revolution in the Soviet Union in 1917 deliberately diminished the role of the church by liberalizing divorce laws and requiring that all marriages be civilly registered. A 2003 study highlights that the earliest reforms of family law did not take place in the most developed economies and the most democratic political regimes. Some countries were very early reformers – most of the reforms in the Nordic countries took place during the early 1900s. By 1930, the laws regarding divorce and matrimonial property were gender equalized. The explanation offered for this is a higher degree of securalization and the presence of two dominant ideological movements. The combination of liberal left ideologies with secularist ideas was the backbone of this reform. 14 Civil law systems historically deriving from Roman law are found for example in Latin America, Francophone and Lusophone ex-colonies in Africa, the Middle East and North Africa, East Asia Pacific and the former colonizing countries in Europe such as France, Spain, Portugal, Germany and the Netherlands. Roman law shaped the 1804 Napoleonic Code that was adopted in Latin America and continental Europe. The 1900 German Civil Code was heavily influenced by the Napoleonic Code but was differently structured and influenced the family codes in Korea and China. 15 Both codes were advantageous when it came to division of marital property 16 and inheritance rights but they also tended to bestow greater control to the husband over assets and the wife’s actions than common law systems. The need for marital authorization still existed in some countries in Europe until the latter part of the twentieth century. The head of household provision was reformed in France in 1974 and the husband’s control over marital property was removed finally in 1985. The need for spousal permission to work remained on the books up to 1978 in Spain and 1984 in Switzerland. Significant reforms to the wife’s control over her personal property took place in the early 20th century in some countries in Latin America. Reforms to the husband’s sole administrative control over joint marital property and over family affairs took longer to achieve in many countries; many of these later changes that took place are captured in the database. 13 Htun and Weldon 2011. 14 Antokolskaia 2003. 15 Htun and Weldon 2011. 16 It should be noted that divorce was not permitted in many Catholic countries in Europe and Latin America until the mid to late 20th century. 5 Common law systems found in the ex-British colonies and the UK gave more autonomy to the wife but gave less access to marital assets and were weaker on inheritance rights. Historically the doctrine of couverture in the common law system had given the husband control over the wife’s assets. The wife had no independent legal personality. England repealed these laws in 1882. In contrast countries such as India colonized by Britain generally had less restrictive pre-colonial personal laws and accorded married women an independent legal personality. 17 By the time of independence the statutory marriage laws inherited by the ex-colonies in the 1940s and 1950s onwards no longer included the couverture doctrine. The British colonial framework with its emphasis on indirect rule more readily recognized plural legal systems with customary and religious law left to regulate the personal affairs of the indigenous populations. Colonial rule sometimes codified a more patriarchal and uniform version of indigenous customs and religious law that reflected the social norms of the colonizing country rather than the local practice. 18 In some instances, religious laws codified by the British in India were exported wholesale to colonies in Africa such as Kenya. 19 Customary law was sometimes recognized in francophone colonies with civil law systems to appease indigenous populations, e.g., Algeria and Niger. The strengthening of customary and religious systems during colonial rule meant that these systems continued post-independence. 20 Countries vary, however, not only in whether customary or religious laws are recognized as formal sources of law in the constitution, prevailing over marriage, marital property and inheritance, but also on whether or not customary and/or religious law is exempt from constitutional principles of non- discrimination. As discussed below, these are areas where constraints have not only been inherited from colonial powers, but have been introduced in more recent years. Hybrid systems, which are a mix of common law and civil law systems, often come out with the highest share of constraints in these indicators. This can be partly explained by the fact that the majority of ‘hybrid’ countries covered were located in Southern Africa. Their history has been marked by the rigid discriminatory aspects of apartheid rule which dominated the region until the mid-1990s. They tend both to recognize customary law and exempt it from non-discrimination, while also recognizing many of the head of household laws that limit women’s legal capacity. The historical evolution of these different legal traditions within countries have present-day implications for women. A 2013 study linked marital property regimes to land ownership in Ecuador, Ghana and Karnataka in India. Ecuador, a civil law system, had the highest level of ownership of ‘couple wealth’ -44% -which can be directly linked to its joint property regimes and egalitarian inheritance laws. This included joint ownership of marital homes and agricultural parcels. In contrast married women in Ghana and Karnataka, both common law system with 17 Deere 2013. 18 Htun and Weldon 2011. 19 Hallward-Driemeier and Hasan 2012. 20 Htun and Weldon 2011. 6 limited access to marital property and and male-biases in inheritance laws, owned just 19% and 9% of assets with men owning the majority of dwellings, agricultural parcels and other real estate. Land was mostly titled in the sole name of the husband. 21 Methodology for Collecting Legal Indicators (See Annex 1) The database provides 0-1 indicators for measures of formal property rights, restrictions on legal capacity and constitutional protections of equality and non-discrimination. The indicators are based on provision in countries’ constitutions and statutes (usually family laws or family codes, succession acts, etc.). Countries receive a “0” if the particular right is not equal for men and for women and a “1” if women and men enjoy the same right equally. The indicators do not measure the extent of the inequality, just that there is a difference between the genders. The database covers 17 indicators. Each is reported separately; the database does not aggregate or weigh the relative importance of the indictors. More detailed notes regarding reforms, and access to the underlying data, are also available on-line (wbl.worldbank.org/timeseries). All regions and income groups are represented in the 100 countries covered in the database (see Annex 2 for the list of countries by region). There is a concentration of countries in Sub-Saharan Africa to build on substantial legal data already collated relating to women’s economic empowerment and reflecting operational interest in the region (see Hallward-Driemeier and Hasan 2012). We track the laws for the main business city (not necessarily the capital city) and the majority of the population. In Nigeria, we cover Lagos and exclude federal laws of other states as well as laws applicable to the minority communities within the province of Lagos. The starting point of 1960 was selected as this era encompassed the recent independence of many of the ex-colonies in Africa, East Asia and the Pacific, the Middle East and South Asia, with some exceptions. Certain countries achieved independence later after periods of conflict either splitting or reunifying – Bangladesh split from West Pakistan in 1971 and North Yemen reunified with South Yemen in 1990. The majority of Latin American countries had become independent much earlier in the 1820s and all of the ECA countries 22 apart from Turkey achieved independence and legal autonomy after the fall of the Soviet Union in the late 1980s. For countries that have merged, we track the laws of the ‘winning’ state from 1960 onwards. For instance we covered the laws of North Yemen for the period 1960 to 2010 which dominated the legal framework of the entire country after reunification with South Yemen in 1990. For countries that divide, they share the same indicators for the period when they were the same country and then their own separate indicators among their independence. 21 Deere 2013 22 Some countries such as Bulgaria were technically independent from the Soviet Union but fell under its sphere of influence. 7 Table 1: Regional and Income Representation of Countries (1960) Region Number of Countries Low Income 1960 Low Middle Income 1960 High Middle Income 1960 High Income 1960 EAP 12 5 3 3 1 ECA 10 0 4 6 0 LAC 15 0 1 11 3 MENA 8 0 5 3 0 OECD 17 0 0 2 15 SAR 5 5 0 0 0 SSA 33 16 14 2 1 Total 100 26 27 27 20 Selection of Indicators – Why They Are Important The indicators selected impact on women’s economic empowerment but are not intended to be exhaustive. Although the categories overlap, they can be broadly classified under the following categories: Table 2. List Of Questions Tracked From 1960 To 2010 Category Question Property Rights and Access to Do unmarried men and unmarried women have equal ownership Assets rights over immoveable property? Do married men and married women have equal ownership rights over immoveable property? Do sons and daughters have equal inheritance rights over immoveable property from their parents? Do female and male surviving spouses have equal inheritance rights over immoveable property? What is the default marital property regime? If joint titling exists for married couples, is it the default for marital property? Legal Capacity: Restrictions on Can an adult married woman legally be designated as "head of Agency and Mobility for household" or "head of family" in the same way as an adult married Married Women man? Does an adult married woman legally need permission from her husband in order to get or keep a job or pursue a trade or profession; or can a husband legally stop his wife from working? 8 Can an adult married woman legally open a bank account in the same way that an adult married man can? Are there any legal provisions that prevent an adult married woman from signing any type of contract in the same way that an adult married man would? Do adult married women need permission from their husbands in order to initiate legal proceedings in court in civil cases? Constitutional Approach to Does the Constitution guarantee all citizens equality before the law? Gender Equality and Plural Legal Systems Is there a non-discrimination clause covering gender or sex in the Constitution? Are customary law and/or religious law a valid source of law under the Constitution? Are customary and/or religious law invalidated if they violate Constitutional provisions on equality and/or nondiscrimination? These questions provide a snapshot as to where a country was and is on the spectrum of gender- friendly laws. Equal rights to property provide opportunities for women seeking to use those assets to generate income or grow their businesses through collateral. Unhampered access to assets can be a safety net in times of crisis and reinforce exit options for women particularly in the context of gender violence. We focus on intestate inheritance laws which apply where there is no will – this can be the most common scenario in many countries especially in Sub-Saharan Africa. Inheritance laws are an important transmission mechanism of assets between generations and can determine household income and earnings capacity of the next generation. Such laws can marginalize women further by allocating smaller shares of the estate or excluding them altogether and reinforce gender gaps for decades to come. The impacts of marital property regimes are generally less understood and operate in more subtle ways to limit access to assets for women. They determine whether property acquired before or during the marriage is shared by the spouses or not. As such they can have implications for marital property within marriage as well as on divorce and inheritance; Annex 4 discusses them in more detail. They often stem from particular legal traditions (separate property regimes are prevalent in common law traditions and where polygamy is recognized, while community of 9 property is more prevalent in civil law countries). As such, the default regime 23 is not often changed, so much of the analysis done here will not apply to this indicator. Turning to the restrictions on agency and mobility, permissions to carry out everyday activities are clear-cut constraints. The range of benefits accorded to a head of household vary across countries – in some countries such as Cameroon, the role of head of household gives tangible rights such as control over household finances and in Jordan, exclusive access to state benefits. In other countries such as Tunisia, the provision is more symbolic which nevertheless can reinforce patriarchal social norms. All these laws impacting mobility are particularly challenging for de facto female headed-households e.g. women whose husbands are migrated or are separated but not officially divorced. Constitutional frameworks can be interpreted as signposts regarding a country’s approach to gender equality. In particular, constitutional exemptions for customary and religious law from the provisions of non-discrimination in the realm of family and inheritance laws. Although there still may be statutes limiting the application of customary law and religious law on the grounds that they are unfair, the lack of constitutional oversight is a weakness in the legal framework. Women’s groups have used constitutional provisions on equality and non-discrimination to challenge discriminatory statutory and customary laws. In Botswana, women’s groups supported a landmark constitutional challenge, the Unity Dow case, which successfully overturned discriminatory citizenship laws and set in motion a whole range of statutory and constitutional reforms. The legacy continues today in Botswana with the 2013 Mmusi case, which successfully challenged customary laws favoring male heirs as unconstitutional. 24 This is in spite of the fact that the Botswana constitution specifically exempts customary law from principles of non- discrimination. 25 The 2010 removal of the constitutional exemption for customary law in Kenya was a culmination of a mobilization campaign by women’s networks. 26 Constitutional reform processes arising in the aftermath of political transition (Spain 1978), conflict (Kenya 2010) or the new-found independence (Fiji 1970) have provided ideal entry points to introduce or reinforce equality and non-discrimination provisions. Conversely, times of upheaval can carry the risk of regression – a newly reunified Yemen amended its constitution in 1994 to remove the non-discrimination provision. However, a lack of equality and non-discrimination provisions -- and even a lack of a constitution -- cannot be read as an unequivocal gender gap. The United States has neither of these provisions, and the UK does not have a formal constitution, but both these countries score highly across various gender indexes. It should also be noted that many countries, including the 23 The default regime is the property regime that applies automatically unless both spouses agree by declaration or in writing to opt out and choose another property regime at the time of marriage. 24 http://www.southernafricalitigationcentre.org/cases/ongoing-cases/mmusi-and-another-v-ramantele-and-another/. 25 Hasan and Tanzer 2013. 26 Hasan and Kamangu 2013. 10 US and UK, have statutory equality and non-discrimination laws (rather than constitutional provisions) which are not covered by this study. Areas for Future Analysis There are gaps and areas which are not covered in this study which could merit a similar analysis in the future. For instance, in some countries single women need the consent of their male relatives to work outside of the home and travel abroad or even within a country. Discriminatory citizenship laws can prevent married women married to non-nationals from passing citizenship to their spouse and/or their children. Discriminatory labor laws can prevent women from working in certain industries and limit working hours and include other provisions such as maternity leave and differentiated retirement ages impacting on women’s participation in the labor force. Legal recognition of non-monetary contributions (i.e. where child care and household duties can entitle the wife to a share of the marital property on divorce) is not covered here but is an important distinction in comparing the protections of various default marital property regimes (see annex 4). Case law plays a prominent role particularly in common law countries. It can fill in gaps or progressively interpret statutory laws where there are glaring inadequacies in the legal framework or reinforce negative discriminatory laws. On the positive side, there may be statutory laws on discrimination that we do not cover which offer protection in instances where the constitution is silent or there is no constitution. Finally we do not look at the implementation of laws- whether good and bad laws on the books are actually applied in practice. The ex-socialist countries in the ECA region had gender equal laws derived from the Soviet Union and so scored highly on our indicators. However, the 2013 ECA World Development companion Report on Gender Equality has found implementation of laws can be weak and in practice there are fewer female land owners than men despite the legal requirement of joint titling for married women. 27 In Vietnam, prevailing social norms that regarded men as head of household rather than a legal requirement undermined the implementation of gender-neutral land titling laws. 28 At an individual country level, Zimbabwe did carry out wide scale reforms over the past fifty years that left it with just one gender gap in the variables that we are looking at but a 2012 NGO CEDAW shadow report 29 shows that the reality on the ground is very different. 30 27 Sattar 2012. 28 Klugman, J. & Morton, M. (forthcoming). 29 A shadow report is prepared by non-state actors such as not for profits to provide feedback to the CEDAW committee which is independent of an official State Report and is an important mechanism for women’s networks to highlight areas requiring reform. 30 Zimbabwe Women Lawyers Association 2012. 11 What Do the Data Show Us? First the Good News… There are fewer discriminatory constraints in 2010 than there were fifty years ago. The overall trend is towards reform of discriminatory laws. By 2010, OECD countries had reformed all the constraints that existed in 1960. Turkey was the only country in 1960 in the ECA region that did not allow women to be head of household and where married women needed their husband’s permission to work. This was reformed in 2001 along with a beneficial switch to a default community of property regime accounting for the total lack of constraints found in 2010 in ECA. LAC and EAP and Sub-Saharan Africa have reduced the number of bad laws on the books by more than 50%. Both unmarried and married women’s property rights have improved and there are more countries with a legally mandated joint titling in 2010 than there were in 1960. Figure 1.Comparison of Constraints per Region in 1960 and 2010 In 1960, every region had at least one constraint, including among the OECD countries. More countries used to have multiple gender gaps; in 2010, almost half of all the countries have no constraints, and overall those that do have fewer gender gaps. 12 Figure 2: The share of countries with different numbers of legal constraints for women Of the 12 countries that began with the most constraints (7 or more) in the initial period, the overwhelming majority were in in Sub-Saharan Africa (10 countries), apart from Spain and Indonesia. All off these countries reformed at least one law. Nigeria and Swaziland still have 8 and the DRC has 6. However, South Africa and Benin that had had 9, now have eliminated all the gender gaps measured here. Spain had had 7 and now has none. Zimbabwe had had 10 and now has 1; Lesotho had also had 10 but now has 2. So enormous progress by some of the former worst offenders helps explain the dramatic reduction in Sub-Saharan Africa. Some constraints have almost disappeared. In 1960, there were 18 countries where women needed legal permission from their spouses to start legal proceedings. 31 Now there are just two countries that require this under statutory law – DRC and Liberia. 31 Angola, Argentina, Austria, Benin, Bolivia, Brazil, Dominican Republic, Indonesia, Lesotho, Liberia, Namibia, Paraguay, Rwanda, South Africa, Spain, Switzerland and Zimbabwe. 13 Figure 3. Permission To Initiate Legal Proceedings – 1960 And 2010 14 Now the Bad News… The overall trend since the 1960s is that discriminatory laws are being removed from the books but there are sticky points and in some cases countries have regressed. Some regions have reformed less than others. There has been little overall change in both the MENA and South Asia regions even though there have been changes at an individual country level. There has been considerable progress on indicators regarding unmarried and married women’s property rights, the need for spousal permission for work, opening a bank account, initiating legal proceedings and signing a contract. Inheritance rights for unmarried and married women and the head of household provisions are proving sticky areas particularly in MENA, South Asia and Sub-Saharan Africa. Although the number of countries with head of household provisions has reduced by almost 50%, approximately 1 in 4 countries (23 out of a100 countries covered in our study) still retain this provision. There has been little progress in some individual countries over the last years. Swaziland had the same number of gender gaps (8) in 2010 as it had in 1960. There has been a clear regression with respect to the constitutional exemption from equality and non-discrimination provisions for customary law and religious law– more countries in Sub-Saharan Africa, Middle East and North Africa and South Asia carry these constitutional exemptions in 2010 than in 1960. Figure 4. Discriminatory Constraints Across Regions in 1960 and 2010 1960 2010 15 Chile is the only Latin American country covered in this study that still legally allows the husband to control joint marital property but in 2010 there were 3 countries in Latin America where the husband is legally head of household. 32 One striking regression is Bolivia which introduced the husband’s permission to work in 1988 going against the general trend. A case brought before the Bolivian Constitutional Court in 2003 successfully challenged this anachronistic law (Article 99 of the Family Code) as unconstitutional. 33 Despite the declaration of unconstitutionality, this law has not been amended by Parliament and is still in force. Triggers and Brakes in Relation to Legal Change Our findings unsurprisingly are that there is not one single explanation as to why some reforms are more difficult than others and why some countries have reformed relatively quickly in comparison to other countries. It should also be noted that while some countries never had constraints on women’s legal economic rights and some countries have adopted reforms aimed at guaranteeing more equal rights, it is also the case that countries have adopted reforms with the contrary effect (see figures 5 and 6). Figure 5: No constraint, no change, progress or regression? 32 Chile, Honduras and Nicaragua. 33 2003 Constitutional Ruling No. 0058/2003 http://www.tribunalconstitucional.gob.bo/resolucion7096. 16 Figure 6. Different legal traditions influence likelihood of reform Girls First One pattern, however, that does stand out is that property rights for unmarried girls were granted earlier than for married women. This may be partly explained by the fact that men were initially supportive of their daughters’ rights over their wives’ - a path which led ultimately to improving property rights for married women. Research papers in the US support the premise that fathers with more daughters are likely to vote more liberally on women’s issues and that they show a greater preference for women’s rights. 34 34 See for example Fernandez 2009. 17 Figure 7. Unmarried Women’s Rights Were Equalized First a. Property b. Inheritance Colonial History and Legal Traditions Matter As outlined earlier in the report, national legal traditions (i.e. civil, common law, hybrid and socialist) have been heavily influenced by their colonial history. Each legal tradition has its relative advantages and disadvantages and the timeline graphs below illustrate how these different legal traditions fare when it comes to married women’s property rights, head of household and recognition of plural legal systems. Joint titling was historically and is currently more prevalent in socialist and civil law systems as this went hand in hand with default community of property regimes. This was and is a weakness of countries with default separate property regimes. While we did not track laws recognizing non-monetary contribution, we did cover legislation making joint titling compulsory and this is taken into account in Figure 8 below. For instance, Tanzania which has a default separate property regime introduced mandatory joint titling in 1971 to address this important gap in property rights. 18 Figure 8. Joint Titling Less Common In Common Law Systems Conversely, head of household restrictions were and are more commonly found in civil and hybrid legal traditions which historically derived from French, German, Dutch, Portuguese, Spanish, and Swiss laws (see Figure 9 below). Figure 9. Head of Household Restrictions More Common in Civil and Hybrid Systems Common law traditions deriving from British law and hybrid legal traditions were and are more likely to exempt customary law from constitutional principles of non-discrimination and equality (see Figure 10). This was partly due to the accommodative nature of British colonial rule which 19 was reluctant to interfere with indigenous laws that did not interfere with civil order and national security concerns and this approach continued in the ex-colonies after independence. For hybrid legal traditions which are mostly found in Southern Africa, the variation over time is partly explained by differing post-independence approaches – Botswana’s constitution exempts customary law whereas the South African and Namibian constitutions do not. Figure 10. Customary Law More Likely To Be Exempted From Non-Discrimination and Equality Provisions In Hybrid And Common Law Countries Customary Law exempt from non-discrimination clause .6 Share of countries .2 0 .4 1960 1970 1980 1990 2000 2010 year Civil Hybrid Common Socialist Testing Key Determinants of Reform Approach In addition to the more descriptive analysis of gender gaps in legal rights, we also explore how various country characteristics are associated with the probability of reform using regression analysis (see annex 6 for a description of the variables and their sources). We look at each of the legal indicators separately. The various variables used to explain the probability of reform are included one at a time in Table 3 (located in annex 7). This is done for two reasons. First, the time coverage of the different variables varies by series; GDP per capita, conflict and the ratio of girls to boys in primary education are available for almost all countries for all 50 years. However, other variables, such as women in parliament or law and order are only available for 20 the last 20 – 25 years. Looking at each variable separately thus allows for as many observations to be included as possible. Second, there could be feedback loops; reforms themselves could affect some of the country characteristics. This is most likely to be true for the measures of education, employment and health, but could also matter for the share of women in parliament. It is also true that some changes in the country characteristics might also take some time to translate into legal reforms. In order to address this, the regressions look at the impact of country characteristics on reforms after a 5 year interval. As all the relevant country characteristics cannot be controlled for, we also include country fixed effects to absorb all the country effects that do not vary over time. Table 3 provides the regression results of the probability of reform, with each cell representing a separate regression. Each explanatory variable enters on its own, with GDP per capita, a country fixed effect and a time trend also controlled for. Thus, time invariant characteristics are controlled for and the effect of the explanatory variable is identified off changes in its value over time rather than in comparison across countries. This provides a stronger test of how changes in conditions within a country predict whether or not it reforms. Table 4 then allows for the possibility of non-linear effects in both GDP per capita and in conflict. Thus it tests whether the effect of raising GDP per capita is different at higher and lower levels of income. For conflict the test is whether the presence and/or the intensity of conflict play roles in the reform process. As the signing of CEDAW was significant in Table 3, it is also included to capture the role of international norms and pressures on the probability of enacting reforms. Later tables then include several additional explanatory variables simultaneously. However, due to data availability constraints, this implies many of the earlier reform episodes will be excluded. Before moving to these results, therefore, we test for the importance of composition effects, both across time periods and countries. Table 4 is repeated for the first half of the sample (1960-1985) and the second half of the sample (1986-2010). As GDP per capita itself is not available for 17 countries prior to 1990, we also rerun the later periods on the 83 countries included in the first half of the data. The country composition effects are not significant, but there are significant differences regarding the role of GDP per capita and reforms. Whereas in the first half of the sample growth is not correlated with reform, the relationship becomes significant in the more recent period, particularly for lower income countries. This finding suggests the drivers of the earlier reformers may be different from that of the late reformers – with implications for being able to predict the reform drivers for those countries that still have not closed the gender gaps in their laws. Table 5 includes a number of explanatory variables simultaneously, focusing on the legal indicators where legal reforms were more likely. Tables 6 and 7 then explore if there are broader non-linear effects of GDP per capita and ‘rule of law’. The tables show where there are such 21 differential effects, both for measures of property rights and legal capacity. GDP per capita is the most common variable to be significant – but only if its effect is allowed to vary by level of development. Stronger growth helps encourage reforms, but the effect diminishes as incomes rise. The impact of the rule of law is significant in fewer instances, in part because the measure is only available for a large subset of the countries (83 of the 100) and since 1985. However, it does have a role to play in strengthening property rights, particularly in reinforcing CEDAW in countries with otherwise weak rule of law. The sections below discuss by thematic variable the results of these tables. Income In this section we look at whether economic growth on its own is likely to translate into equal legal economic rights for women or if countries need to employ a more proactive agenda in order to achieve legal economic equality. The top row of Table 3 shows the effects of GDP per capita on each of the legal indicators. It shows that a country’s income in and of itself is not a sufficient factor to generate change in the area of legal economic rights for women analyzed in the present study. Strikingly, income is only significant in two of the regressions – and it enters negatively. Higher growth, on its own, is not a good predictor of reforms. The second line, however, gives an indication that there may be more to the role of income. Allowing for a non-linear effect, there are four regressions where the results are significant, property rights for unmarried women and three constitutional measures. These results would indicate that if changes in incomes are associated with reforms, the effect is strongest at lower levels of income, and could even be negative at high enough levels of income. This is explored more in Table 4, and GDP is allowed to have non-linear effects in all the other tables. One question is whether the role of GDP per capita and growth varies over time. Table 4 is repeated using only the first half of the sample (Table 4B) and the second half of the sample (Table 4C). The results do show significant differences. In the earlier years, higher incomes are associated with a greater probability of recognizing equal inheritance between sons and daughters and in having equality as a principle in the constitution – effects that are stronger among lower income countries. And higher income is associated with a higher probability of recognizing women as heads of household—in higher income countries. But for most of the legal indicators, GDP per capita is not significant. However, in Table 4C, GDP per capita is far more significant for the measures of property rights and inheritance as well as most of the legal capacity variables. The effects are again non-linear, most significant in lower income countries and diminishing as incomes rise. These findings indicate that development has had a larger role in encouraging reform—in the more recent period and for lower income countries. This has 22 implications for its likely role in encouraging countries that have not yet reformed, many of which are middle income already. Table 5 confirms that the relationships with GDP per capita are robust to the inclusion of additional variables. As discussed above, the inclusion of these variables often results in eliminating the earlier time periods as these variables were not generally measured then. Table 6 then pushes the effects of income one step further – looking to see if it interacts with other variables of interest. The non-linear effects of GDP per capita remain. There are also some examples where the additional interactions are significant, but the effects are not systematic. Thus, for married property rights, there is a positive effect for greater enrollment of girls in primary school on the probability of reform, particularly in lower income countries. For inheritance rights and the ability to work without requiring one’s husband’s permission, the results show that CEDAW can be effective – but only in lower income countries. One important conclusion coming out of these estimation models is that a simple view that economic growth translates into more equal legal economic rights for women does not hold. Income was not much of a predictor of early reformers. Over the whole period, there is some greater role when non-linear effects are included, but not systematically. Only looking at the last 25 years or so, there does seem to be more of a role for income. The effect is strongest among lower income countries and not significant for most middle income countries. As many of the constraints now remaining are as common in middle income countries as in low income countries, it is unlikely that continued growth in this group of countries will be sufficient to achieve the legal reforms that would give women equal economic rights. Going forward, most countries will need to be proactive in reforming such laws. Conflict Next we look at conflict, both its presence and its intensity, and the extent to which this indicator is likely to influence the probability of reform in the areas of legal economic rights analyzed here. Thus results show conflict is associated with a lower probability of recognizing equal inheritance laws, for sons and daughters as well as for husbands and wives, as well as a lower probability of recognizing customary law. The greater intensity of conflict is also associated with lower probability of removing statutes that require women to get permission to open bank accounts, entry contracts or participate in legal proceedings (Table 3). This pattern on inheritance rights is robust to including both measures together, along with the effects of GDP per capita and CEDAW (Table 4). The results on the restrictions to legal capacity are not. Table 4 also indicates countries that have undergone periods of conflict are less likely to exempt customary law from non-discrimination provisions in their constitutions, and to both recognize religious law and exempt it from non-discrimination provisions. Preservation of plural legal systems can be seen as a bargaining chip for central leaders to consolidate the state building 23 process and post conflict reconstruction. It can also be a way of protecting cultural group rights. 35 When more complex models are used to measure the impact of conflicts (Tables 5, 6 and 7), the significance of conflicts is reduced. Again, this may reflect compositional effects –the effects discussed above mattered more in earlier conflict episodes that are not included in these later tables. It is also the case that the patterns coming out of conflicts may not be uniform. There are certainly examples such as Rwanda where many positive reforms were enacted coming out of the genocide. Rule of Law Next we test whether rule of law, as measured with the ICRG’s ‘law and order’ indicator, is related to improved legal capacity measures. As the measure is only available for 83 countries, including this explanatory variable means that a sub-set of countries are excluded from the results that include this indicator. In this sub-set of countries rule of law is shown to have a significant positive impact on women’s ability to be head of household, as well as their ability to work without needing husband’s permission (Table 3). When we look at an extended set of potentially relevant indicators simultaneously with rule of law (Table 5) we observe that the rule of law indicator has a statistically significant positive impact on equal inheritance rights for surviving spouses, regardless of gender, but it does not influence equal inheritance rights for sons and daughters, or any other access to assets variable. A similar effect is found for reforms that remove the need for wives to gain permission from their husbands to work outside the home. Table 7 then explores whether the rule of law itself affects how other variables help predict reform. It finds that it reinforces the intensity of conflict in discouraging the reform of married property rights. It also indicates that CEDAW is most effective in promoting inheritance rights where the rule of law is otherwise weak. There is also some effect in reinforcing inheritance rights where girls’ primary enrollment is higher. Sector Value Added to GDP A more diversified sectorial composition of GDP could serve to encourage greater reforms, particularly as women’ labor force participation is often higher when the service sector expands. Our data show (Table 3) that a greater share of services in the economy is associated with greater property rights for married women, as well as greater legal capacity to enter contracts and legal proceedings. However, once other variables are included and accounted for– and the sample restricted to the last 25 years rather than full 50 years – the effects of GDP’s composition are less pronounced and it loses its catalyst for reform significance (Tables 5 and 6). 35 Htun and Weldon 2011. 24 The Oil Curse has been forward as an underlying factor behind women’s low labor force participation in oil-dependent countries with little sectorial diversification and high levels of occupational segregation. 36 However, the data do not show too much evidence of this effect. The share of natural resources in merchandise exports is negatively associated with the recognition of customary law and positively associated with women being able to open bank accounts without permission. The coverage of countries in the database may limit the ability to test this hypothesis fully; the database does not cover many of the oil rich countries in the Middle East. The CEDAW Effect The Convention on the Elimination of All Forms of Discrimination Against Women (CEDAW) entered into force in 1981. There are now 186 parties to the Convention. The only countries that have not signed are Somalia, Sudan, Iran, Nauru, Tonga, and the USA. 37 The treaty sets out specific obligations targeted at eliminating direct and indirect discrimination against women and to promote formal and de facto equality for women. These include equality before the law and the elimination of discriminatory customary and traditional practices. Civil law countries are mostly monist countries, which means that there is no further need for domestic enactment of the treaty obligations after ratification. Common law countries are mostly dualist countries – the country has to enact domestic legislation to enable enforcement of the treaty’s provisions within the country. 38 The pattern of ratification has been shown to be triggered by external events at an international level such as the 1985 Nairobi Third World Conference on Women and the 1995 Beijing Fourth World Conference on Women. 39 The ratification of CEDAW encouraged states to review family laws as a way of signaling their integration in the international order and world civilization. 40 Monitoring mechanisms under the Convention include a mandatory reporting procedure by governments to the CEDAW Committee, which subsequently makes recommendations for improved implementation. Civil society groups use this opportunity to engage with the government during the review process and to submit their own shadow reports. Women’s networks have also used CEDAW in strategic litigation in many countries to challenge discriminatory laws including customary inheritance laws in Kenya 41 and Botswana. 42 A number of states have ratified CEDAW with reservations that exclude the domestic applicability of specific CEDAW articles. For our purposes, the reservations that are important are those which 36 Ross 2008. 37 The US has signed but not ratified the convention. 38 There are a few exceptions to the general pattern. The new 2010 Kenyan constitution changes Kenya, a common law country, from a dualist to a monist system. 39 Byrnes 2011. 40 Engle-Merry 2003. 41 Rono v Rono Civil Appeal No. 66 of 2002, 1KLR (G & F) 804 (2005). 42 Mmusi and Others v Ramantele and Another (2012) BWHC 1. 25 allow states to opt out of obligations to ensure that family law is in harmony with the non- discriminatory principles of CEDAW. 43 Our study shows that the average pace of change across the ratifying countries covered 44 in reforming discriminatory constraints is almost double within the first five years after CEDAW ratification in comparison to fifteen years prior to ratification (see Figure 11). Thus CEDAW has a positive impact on equal property rights for unmarried women, as well as on inheritance rights for both daughters and surviving spouses, and a number of dimensions of legal capacity (see Table 3). The effects are robust to the inclusion of a number of other variables (Table 5). Table 6 shows that this positive effect is strongest in the countries with lower incomes. And Table 7 suggests the effects are greater where the rule of law is otherwise weak in promoting inheritance rights. Figure 11: The “CEDAW Effect” A Comparison Of Reform Before And After Ratification 15 years before CEDAW 5 years before CEDAW 5 years after CEDAW ratification ratification ratification Women in Parliament There is evidence that higher levels of political participation are linked to high levels of female labor force participation particular in traded sectors such as agricultural and manufacturing sectors. Greater cohesiveness and networks created through work can mobilize women’s movements that in turn lobby for political participation and improving legal rights. 45 Other studies have shown that women in high political office can act as role models and boost other women’s political knowledge, interest and participation. 46 Studies in India have shown that 43 Articles 15 and 16 and general reservations are taken into account. 44 Iran, Sudan and US are the three countries in the time series who have not ratified CEDAW. 45 Moghadam 1999. 46 Burns, Scholzman and Verba 2001, Hansen 1997. 26 increased participation of women in politics at a local level influences policy-making at a local level with greater investment in public services such as water 47 and has other benefits such as an increase in the reporting of gender-based violence. 48 Our research shows that increased participation of women at a national level (25% and above, see annex 5 for the list of countries) can have a significant positive effect on the removal of discriminatory laws. Table 3 shows that it has a statistically significant positive impact equal property rights for married women. The presence of women in national parliament also has a significant positive impact on all restrictions on agency and mobility for married women, with the exception of needing permission to get a job or pursue a profession. This means that the higher the percentage of women in parliament, the less likely it is that women in that country need permission from their husbands in order to open accounts, sign contracts and initiate legal proceedings. It also points out the fact that political participation of women makes it more likely for a country to recognize adult married women as head of households, the same way as an adult married man. The extension to Table 5 shows that with the addition of other variables, women in parliament is still positively and significantly associated with almost all the indicators associated with access to and control of assets as well as legal capacity. Table 6 does not include women in parliament due to the more limited time coverage, but in its more limited sample, these positive effects are strongest in lower income countries. 47 Chattopadhyay and Duflo 2004. 48 Iyer 2012. 27 Figure 12a. There Are Less Constraints Figure 12b. There Are More Constraints Where Where There Are More Women In There Are Less Women In Parliament Parliament. Conclusion What does it mean for there to be equality between a man and a woman? • Both can take independent decisions for their life • They can choose partners of their own choice • They can gain confidence to face society if they go out • They can take decision on whether they should work or not ο Urban adult women, Town of Jaipur, Odisha, India 49 Do you see any changes because of these [new gender] laws? • You see many women have now gone to school; you can find them in offices; some of them are politicians. I think that this is result of such laws. Young women's FGD, village of Zabibu In Dodoma region, 50 Tanzania 51 49 This town has the same name as the much larger city of Jaipur in Rajastan. 50 The village name has been changed to a pseudonym. 51 World Bank Dataset 2011a. 28 Substantial changes have been achieved in closing formal gaps in women's rights to access and control assets, to have full legal capacity and to have non-discrimination principles apply to key areas of marriage, property and inheritance. Indeed, the evidence of this from the last 50 years is impressive. However, more still needs to be done. And the uneven relationship between income and reforms indicate that the agenda often requires proactive engagement. Positive change has come about in various ways including the influence of international conventions, mobilization by women’s networks, increased female political participation and sectoral diversification. These factors are interlinked – for example, the ratification of CEDAW helped to mobilize women who lobbied for increased political participation. Continuing CEDAW’s momentum can continue to put a spotlight on countries to reaffirm their international commitments. Building up the capacity of women’s networks to lobby for the implementation of CEDAW through shadow reports and strategic impact litigation could also be effective. Promoting women’s legislative representation, particularly at national levels, is also associated with more statutes granting equality of rights to women and men. This paper has laid out how legal rights have changed - across legal issues, across countries and over time. It has identified some of the factors associated with reforms. The companion paper provides a complementary analysis of the links between the legal framework and women’s economic empowerment, including employment, health and educational outcomes. It is hoped that the data that have been generated will yield even more insights of the contributions legal reforms can make, and will help policy makers and advocates for gender equality better understand the pathways and entry points for reform in countries which still have discriminatory laws on the books. 29 References Antokolskaia, Maria. 2003. “Development of Family Law in Western and Europe: Common Origins, Common Driving Forces, Common Tendencies.” Journal of Family History. Vol. 28 no.1. January 2003 52-69 Burns, Nancy, Kay Lehman Schlozman, and Sidney Verba Scholzman and Verba. 2001.’ The Private Roots of Public Action.” Harvard University Press Byrnes, Andrew and Marsha Freeman.2011. The Impact of the CEDAW Convention: Paths to Equality. Background Paper to the World Development Report 2012. Charrad, Mounira. 2001. States and Women’s Rights” The Making of Postcolonial Tunisia, Algeria and Morocco. Berkeley. University of California Press Chattopadhyay, Raghabendra and Esther Duflo. 2004. “Women as Policy Makers: Evidence from a Randomized Policy Experiment in India.” Econometrica 72(5): 1409-43 Deere, Carmen, Abena Oduro, Hema Swaminathan and Cheryl Doss. 2013. “Property Rights and the Gender Distribution of Wealth in Ecuador, Ghana and India.” Journal of Economic Inequalities (2013) 11:249-265 Deininger, Klaus, Aparajita Goyal, and Hari Nagarajan. 2010. “Inheritance Law Reform and Women’s Access to Capital: Evidence from India’s Hindu Succession Act.” Policy Research Working Paper No. 5338, World Bank, Washington, DC. Doss, Cheryl. 2006. “The Effects of Intrahousehold Property Ownership on Expenditure Patterns in Ghana.” Journal of African Economies 15 (1): 149-80 Engle-Merry, Sally. 2003. “Rights Talk and the Experience of Law: Implementing Women’s Human Rights to Protection from Violence.” Human Rights Quarterly 25:2: 343-381. Fernandez, Raquel. 2009. “Women's Rights And Development” National Bureau of Economic Research Field, Erica, 2007. Entitled to Work: Urban Property Rights and the Labor Supply in Peru.” Quarterly Journal of Economics 122: 1561-602 Hallward-Driemeier, Mary, and Ousman Gajigo. 2013. Strengthening Economic Rights and Women’s Occupational Choice: The Impact of Reforming Ethiopia’s Family Law. World Bank, Development Economics, Washington, DC. Hallward-Driemeier, Mary and Tazeen Hasan. 2013. Empowering Women: Legal Rights and Economic Opportunities in Africa. Washington, DC: World Bank. Hallward-Driemeier, Mary, Tazeen Hasan and Sarah Iqbal. 2013. Historical Database of Women's Legal Capacity and Property Rights. World Bank Group: Washington, DC. Hallward-Driemeier, Mary, Tazeen Hasan, Jane Kamangu, Emilia Lobti, and Mark Blackden. 2013. Women’s Legal and Economic Empowerment Database (Women LEED Africa), Development Economics, Washington, DC. 30 Hallward-Driemeier, Mary, Tazeen Hasan and Anca Bogdana Rusu. 2013. “Women’s Legal Rights Over 50 Years: What is the Impact of Reform?” mimeo. Hansen, Susan B. 1997. “Talking About Politics: Gender and Contextual Effects on Political Proselytizing.” Journal of Politics 59(10: 73-103 Hasan, Tazeen And Jane Kamangu.2013. “Gender Equality Under The New 2010 Kenya Constitution And Emerging Legislation.” mimeo Hasan, Tazeen and Ziona Tanzer. 2013. “The Botswana Constitution: A Strategic Tool for Targeting Discriminatory Laws” Mimeo Htun, Mala and Laurel Weldon.2011. Sex Equality in Family Law: Historical Legacies, Feminist Activism and Religious Power in 70 countries. Iyer, Lakshmi, Anandi Mani, Prachi Misra, and Petia Topalova. 2012, The Power of Political Voice: Women’s Political Participation and Crime in India. American Economic Journal: Applied Economics 2012, 4(4): 165-193 Klugman, J. & Morton, M. (in press). Enabling equal opportunities for women in the world of work: The intersections of formal and informal constraints. World Bank Legal Review, v. 5. Landsberg-Lewis, Ilana (ed). 1998. Bringing Equality Home: Implementing the Convention on the Elimination of All Forms of Discrimination Against Women Moghadam, Valentine, 1999. Gender and Globalisation: Female Labor and Women’s Movements. Journal of World-Systems Research 5 (Summer): 367-88 McPhedran and others (eds). 2000. The first CEDAW Impact Study: Final Report Quisumbing, Agnes R. and John A. Maluccio.2000.”Intrahousehold Allocation and Gender Relations: New Empirical Evidence from Dour Developing Countries.” Discussion paper 84, Food Consumption and Nutrition Division, International Food Policy Research Institute, Washington DC Ragasa, Catherine, Annie Kinwa-Muzinga and John Ulimwengu.2012. “Gender Assessment of the Agricultural Sector in the Democratic Republic of the Congo.” IFPRI Discussion Paper 01201. Washington DC Rashed, S and others. 1999. “Determinants of the Permethrin Impregnated Bed nets (PIB) in the Republic of Benin: the role of women in the acquisition and utilization of PIBs.” Soc Sci Med. 1999 Oct;49(8):993-1005. Ross, Michael L. 2008. Oil, Islam and Women. American Political Science Review. Vol. 102. No. 1 February 2008 107 -123 Sattar, Sarosh. 2012. Opportunities for Men and Women in Emerging Europe and Central Asia. Washington, DC: World Bank. 31 Simmons, Beth. 2009. ‘Mobilising for Human Rights: International Law in Domestic Politics.’ Cambridge: Cambridge University Press 2009 209 Southern Africa Litigation Center. 2013. “Botwana: Women’s Inheritance Rights Under Customary Law.” http://www.southernafricalitigationcentre.org/cases/ongoing-cases/mmusi-and-another-v- ramantele-and-another/ Turk, Carolyn, Patti Petesch and Ana Maria Munoz Boudet. 2012.”On Norms and Agency: Conversations About Gender Equality with Women and Men in 20 Countries.” Udry, Christopher, John Hoddinott, Harold Alderman, and Lawrence Haddad. “Gender Differentials in Farm Productivity: Implications for Household Efficiency and Agricultural Policy.” Food Policy 20(50 407-23 World Bank. 2011a. "Defining Gender in the 21st Century: Talking with Women and Men around the World: A Multi-Country Qualitative Study of Gender and Economic Choice." Dataset. World Bank, Washington, DC. World Bank 2011. “Women Business and the Law: Removing Barriers to Economic Inclusion.” Washington, DC. Weichselbaumer, D and R. Winter-Ebmer. 2007. The effects of competition and equal treatment laws on gender wage differentials. 22 Economic Policy Zimbabwe Women Lawyers Association. 2012. “Zimbabwe Civil Society’s Shadow Report to the CEDAW Committee.” Harare. http://www2.ohchr.org/english/bodies/cedaw/docs/ngos/ZCS_Zimbabwe51.pdf Cases White v White (2001) A C 596 (HL) 2003 Constitutional Ruling Bolivia No. 0058/2003 http://www.tribunalconstitucional.gob.bo/resolucion7096 Rono v Rono Civil Appeal No. 66 of 2002, 1KLR (G & F) 804 (2005) Mmusi and Others v Ramantele and Another (2012) BWHC 1 and http://www.southernafricalitigationcentre.org/1/wp-content/uploads/2013/07/Mmusi-Court-of-Appeal- Judgment.pdf 32 List of Annexes, Tables and Figures List of Annexes: Annex 1. Methodology for “50 Years of Women’s Legal Rights Database” Annex 2. Regional Representation of 100 Countries Annex 3. Countries Classified by Legal Tradition Annex 4. Marital Property Regimes Annex 5: Women’s Representation in Parliaments Annex 6. Summary of Explanatory Variables and Summary Statistics Annex 7. Regression results List of Tables: Table 1. Regional and Income Representation of Countries (1960) Table 2. List of Questions Tracked From 1960 to 2010 Table 3. Probability of Reforms Table 4. Probability of Reform Allowing for Non-linear Effects of Income and Conflict 1960-2010 Table 4B. Probabilty of Reform Allowing for Non-Linear Effects of Income and Conflict 1960 to 1985 Table 4C. Probabilty of Reform Allowing for Non-Linear Effects of Income and Conflict 1986 to 2010 Table 5. Panel Regressions with Expanded List of Explanatory Variables Table 6. Panel Regression: Non-linear Effects by GDP per Capita Table 7. Panel Regression: Non-linear Effects by “Rule of Law” List of Figures: Figure 1. Comparison of Constraints per Region in 1960 and 2010 Figure 2. The share of countries with different numbers of legal constraints for women Figure 3. Permission to Initiate Legal Proceedings – 1960 and 2010 Figure 4. Discriminatory Constraints Across Regions in 1960 and 2010 Figure 5. No Constraint, No Change, Progress or Regression? Figure 6. Different Legal Traditions Influence Likelihood Of Reform Figure 7. Unmarried Women’s Rights Were Equalized First Figure 8. Joint Titling Less Common In Common Law Systems Figure 9. Head of Household Restrictions More Common in Civil and Hybrid Systems Figure 10. Customary Law More Likely To Be Exempted From Non-Discrimination Provisions In Hybrid And Common Law Countries Figure 11. The “CEDAW Effect” A Comparison of Reform Before And After Ratification. Figure 12a. There Are Less Constraints Where There Are More Women in Parliament. Figure 12ab. There Are More Constraints Where There Are Less Women in Parliament 33 Annex 1: Methodology for “50 Years of Women’s Legal Rights Database” The objective of the database is to track legal reforms affecting women’s property rights and legal status, starting with the law applicable in 2010 and tracking changes backwards in time to 196052. For economies that came into existence after 1960, the coding general begins with the economy’s independence in its current form. However, for countries that divide over time (e.g. countries of the former Soviet Union), the individual countries are included over the whole period, each with the laws that governed the initial single country for the relevant years, and then with the country-specific laws and constitutions after independence or separation. For economies that became re-united or unified (i.e. Yemen and Vietnam) within the period between 1960 and 2010, applicable legislation is analyzed and coding is done separately for the individual economies that eventually became one, covering that entire period. The process of data collection consists of desk research using written legislation, such as Constitutions, Personal Status laws, Civil Codes, Family Codes, and Inheritance laws, applicable in the main business city of the economies covered. Online sources of legislation such as the World Bank’s Gender Law Library 53, Lexadin 54, the United Nations’ Human Rights Treaties database 55, the law library of the US Congress 56, ILO’s NATLEX database 57, the Notarius International Foreign law database 58, as well as legislation and secondary sources accessed through universities’ databases and virtual libraries such as Harvard Law School’s Law Library 59, NYU Law’s Globalex 60, and HeinOnline 61, among others, are used. Also, hard copies of historical legislation not available online were obtained through the Library of Congress Georgetown Law Library and the Jacob Burns Law Library at George Washington University. The 50 Years of Women’s Legal Rights Database analyzes the evolution of gender parity within the legislation of the covered economies through a set of 17 questions in the following areas: property ownership, marital regimes, inheritance, property titling, status and capacity, access to the judicial system, and constitutional rights. 52 This data collection was performed with the guidance of Mary Hallward-Driemeier, Lead Economist Development Research Group, and carried out under the supervision of Sarah Iqbal, Program Officer of Women, Business and the Law, and Tazeen Hasan, Development Economics and PRMGE. Team members from the Women, Business and the Law who contributed are: Garam Dexter, Khrystyna Kushnir, Lourdes Lizarraga Ortega, Nayda Almodovar Reteguis, Paula Tavares, Thibault Meilland and Yasmin Klaudia bin Humam. In addition, Ashley S. Zohar, Erin F. Rogers, Lindsey Scannell, Sarah Knapp, Shazia Akhter, Tatyana Chursova, Yulia Dovgaya and Zachary Walter conducted research for the database as part of their internships with the World Bank Group. 53 http://wbl.worldbank.org/WBLLibrary/elibrary.aspx?libid=17 54 http://www.lexadin.nl/wlg/legis/nofr/legis.php 55 http://www.bayefsky.com/ 56 http://www.loc.gov/law/ 57 http://www.ilo.org/dyn/natlex/natlex_browse.country?p_lang=en&p_country=COL 58 http://212.63.69.85/en/Links/links_international_en.htm 59 http://www.law.harvard.edu/library/research/guides/int_foreign/web-resources/foreign-law_c.html 60 http://www.law.nyu.edu/global/researchtools/globalex/index.htm 61 http://home.heinonline.org/ 34 For all questions concerning the ability of married women to carry out activities independently of their husbands, the key concern is reciprocity. Therefore, where a married man is unable to carry out the same activity without the permission of his wife, the question is not triggered because the provision is reciprocal. It is only triggered where a married man is able to carry out the activity and his wife cannot do so. To make the data comparable across economies, the following assumptions are being used 62: Assumptions about the woman • Resides in the main business city of the economy being examined; • Has reached the legal age of majority and is capable of making decisions as an adult. If there is no legal age of majority, please assume the woman is 30 years old; • Is sane, competent, in good health, and has no criminal record; • Is a lawful citizen of the economy being examined; • Where the question assumes the woman (or man) is married, the marriage is monogamous and registered with the authorities; • Where the question assumes the woman (or man) is married, the marital property regime she (or he) is married under is assumed to be the default marital property regime. • Where the question assumes the woman (or man) is unmarried, she (or he) has never been married. • Customary law is not being used to answer the questions. The only questions examining customary law are the ones that specifically address it. E.g. examining the place of customary law under the Constitution. • The year recorded in the “Year” column is the year the legislation or amendment was enacted, not the year in which it entered into force. • Where the answer differs according to the legal system applicable to the woman in question – as may be the case in economies were legal plurality exists – the answer used is the one applicable to the majority of the population. 62 For more details about the assumptions used to answer the particular questions under each sub-section, please refer to the Data Notes section of the 2012 Women Business and the Law report, pages 34 – 50, that can be accessed by visiting our website: wbl.worldbank.org. 35 Annex 2: Regional Representation of 100 Countries SAR EAP ECA LAC MENA OECD SAR SSA Banglades China Bulgaria Argentina Algeria Australia h Angola Fiji Belarus Bolivia Egypt, Arab Rep. Austria India Benin Indonesia Georgia Brazil Iran, Islamic Rep. Belgium Sri Lanka Burkina Faso Cambodia Kazakhstan Chile Jordan Switzerland Nepal Botswana Central African Lao PDR Kyrgyz Republic Colombia Morocco Germany Pakistan Republic Russian Syrian Arab Mongolia Federation Costa Rica Republic Spain Côte d'Ivoire Dominican Malaysia Tajikistan Republic Tunisia France Cameroon United Philippines Turkey Guatemala Yemen, Rep. Kingdom Ethiopia Papua New Guinea Ukraine Honduras Hungary Gabon Singapore Uzbekistan Jamaica Italy Ghana Thailand Mexico Japan Kenya Vietnam Nicaragua Korea, Rep. Liberia Peru Netherlands Lesotho Paraguay Norway Madagascar Venezuela, RB Portugal Mali Sweden Mozambique United States Mauritania Malawi Namibia Niger Nigeria Rwanda Sudan Senegal Sierra Leone Swaziland Togo Tanzania Uganda South Africa Congo, Dem. Rep. Zambia Zimbabwe 36 Annex 3: Countries Classified by Legal Tradition Civil law Hybrid Common law Socialist Argentina Sri Lanka Bangladesh Belarus Angola Botswana Australia Bulgaria Brazil South Africa Kenya Cambodia Austria Lesotho Fiji China Belgium Namibia United Kingdom Georgia Benin Philippines Ghana Hungary Burkina Faso Swaziland India Kazakhstan Bolivia Yemen, Rep. Jamaica Kyrgyz Republic Central African Republic Zimbabwe Liberia Lao PDR Switzerland Malawi Mongolia Chile Malaysia Russian Federation Côte d'Ivoire Nigeria Tajikistan Cameroon Nepal Ukraine Colombia Pakistan Uzbekistan Papua New Costa Rica Guinea Algeria Sierra Leone Dominican Republic Singapore Germany Sudan Egypt, Arab Rep. Tanzania Spain Uganda Ethiopia United States France Zambia Congo, Dem. Rep. Gabon Guatemala Honduras Indonesia Iran, Islamic Rep. Italy Japan Jordan Korea, Rep. Madagascar Mali Mauritania Mexico Morocco 37 Mozambique Netherlands Nicaragua Niger Norway Paraguay Peru Portugal Rwanda Senegal Sweden Syrian Arab Republic Thailand Togo Tunisia Turkey Venezuela, RB Vietnam 38 Annex 4: Marital Property Regimes Marital regimes are not inherently discriminatory, but have repercussions for women’s control and ownership of assets during a marriage and at the end of marriage i.e. divorce or death of the husband. The study tracks the default marital regime – the regime that applies automatically unless a couple agrees on another regime either by a declaration or in writing at the time of marriage. 63 However, as there are very few changes in the default marital property regimes covered in the database, there is little scope to analyze the determinants of reform in this area. However, as the choice of the regime does have important implications for women. The differences are thus explained in somewhat more detail here. The marital regimes are classified into five categories: a. Separation of property: under a separate property regime all property acquired by the spouses before they marry, as well as all property acquired during the marriage, remain the separate property of the acquiring spouse unless both spouses specify otherwise. This regime is usually found in common law countries or Muslim-majority countries and is applicable to all customary/religious polygamous marriages. It can also be referred to as out of community of property. b. Partial community of property: whereby assets acquired prior to marriage are regarded as the separate property of the acquiring spouse, and assets and income acquired after marriage are regarded as joint property of the couple. The only exception is property inherited or given as a gift to a spouse (unless the giver specifies that the property goes to both spouses). This regime should automatically lead to joint titling of the marital home. Joint property is automatically divided 50/50 on divorce or death of a spouse. This regime is commonly found in civil law countries, socialist and ex-socialist countries and exceptionally is the default regime in a few Muslim-majority countries e.g. Turkey and Indonesia. c. Full community of property: It is similar to a partial community regime except that all assets and income brought into the marriage and acquired during the marriage become the joint property of the couple (apart from inheritance or personal gifts). It can also be referred to as in universal community of property. It is found in South Africa deriving from the laws of the Netherlands where it still prevails. d. Deferred full or partial community of property: the rule of partial or full community of property applies at the time of dissolution of the marriage (divorce or death of one of the spouses); prior to that, separation of property applies. This is found in several LAC and Nordic countries and is derived from the German/Swiss Legal systems. e. Other, which covers regimes which can be hybrid regimes such as those found in EAP countries e.g. China, Japan where property is regarded as joint property but the court can decide the division of the property rather than an automatic 50/50 division. 63 Some countries do not allow any alternatives to a default regime. 39 Generally, community of property regimes offer the most protection to women who do not have independent access to assets or income and who have contributed to the household in other ways e.g. caring for children or looking after the household. Annex Figure 1. Benefits of A Default Partial Community Regime It should be noted that the legal recognition of non-monetary contributions such as child care or household duties in separate property regimes can also compensate women for non-paid household activities. 64 However not all countries have legally adopted this principle and even where they have, the exact shares provided to women can be arbitrary. In the UK the division of marital property on divorce has been decided as 50% by the highest court in the country after decades of ambiguity. 65 In some countries there may be specific legislation demarcating an equal property division but in other countries there is no set share which can lead to uncertainty and inconsistent result in practice. 64 This time series study does not cover legislation related to non-monetary contribution but the 2013 Women Business and the Law database and the 2012 Women-LEED- Africa database study provide data on current legislation. 65 White v White (2001) A C 596 (HL). 40 Annex Figure 2: Potential Disadvantages of A Separate Property Regime SEPARATE PROPERTY REGIME LIMITED ASSETS FOR COLLATERAL DURING MARRIAGE NO YES ??? RECOGNITION OF NO RECOGNITION RECOGNITION OF NON- OF NON-MONETARY NON-MONETARY MONETARY CONTRIBUTIONS CONTRIBUTIONS CONTRIBUTION ??? ??? LAW MANDATING LAW MANDATING 50/50 SPLIT ON 50/50 SPLIT ON INHERITANCE DIVORCE NO YES NO YES PROVE FINANCIAL 50% SHARE 50% SHARE UNCERTAIN CONTRIBUTION TO GET UNCERTAIN OF ASSETS OF ASSETS SHARE OF SHARE OF ASSETS ON SHARE OF INCLUDING INCLUDING ASSETS ON DIVORCE AND INHERITANCE HOUSE IN HOUSE ON DIVORCE INHERITANCE INHERITANCE DIVORCE The implications of the different default marital property regimes can be seen in the high correlation as to where there is joint titling of assets during marriage. Only one of the 36 countries with ‘separation of property” regimes offers joint titling as the default for married couples, and only 2 of the 43 with partial community of property does not. The pattern can also be seen in the inheritance rights. Equal inheritance rights for wives and husbands Initial period 2010 Default marital property regimes No Yes No Yes Separation of property 26 10 20 16 Deferred full or partial community of property 0 9 0 9 Partial community of property 8 35 4 39 Full community of property 2 3 0 5 Other 2 4 0 6 Total 38 61 24 75 41 Annex 5: Women’s Representation in Parliaments over 25% women in parliament in 2010 Constraints Angola 0 Argentina 0 Austria 0 Belgium 0 Belarus 0 Switzerland 0 Costa Rica 0 Germany 0 Spain 0 Ethiopia 0 Lao PDR 0 Mexico 0 Mozambique 0 Netherlands 0 Norway 0 Peru 0 Portugal 0 Sweden 0 Vietnam 0 South Africa 0 Bolivia 1 Fiji 1 Tanzania 1 Nepal 2 Rwanda 2 Uganda 2 Tunisia 3 Sudan 4 Niger 5 42 Annex 6: Summary of Explanatory Variables and Summary Statistics • World Development Indicators (WDI) is the primary World Bank collection of development indicators, compiled from officially recognized international sources. It presents the most current and accurate global development data available, and includes national, regional and global estimates. WDI was the source for indicators regarding GDP, GDP/capita, maternal and infant health, life expectancy, sectors value added to GDP and exports of natural resources. (databank.worldbank.org) • International Country Risk Guide (ICRG) - is generated by the PRS group. Updated monthly, ICRG monitors 140 countries. ICRG's statistical tables assign values to the 22 indicators underlying ICRG's business-oriented model for quantifying risk, examining such country-specific elements as currency risk, political leadership, the military and religion in politics, and corruption. The present study employs the Law and Order indicator. Law and Order are assessed separately, with each sub-component comprising zero to three points. The Law sub-component is an assessment of the strength and impartiality of the legal system, while the Order sub-component is an assessment of popular observance of the law. Thus, a country can enjoy a high rating – 3 – in terms of its judicial system, but a low rating – 1 – if it suffers from a very high crime rate of if the law is routinely ignored without effective sanction (for example, widespread illegal strikes). The maximum score is 6. (http://www.prsgroup.com/ICRG.aspx) • International Income Distribution Database (I2D2) – database of disaggregated labor market indicators derived from a large set of surveys from developing countries. The database presently includes more than 500(??) individual datasets from the period between 1984 and 2011(??). • Aiddata.org – AidData is a partnership between Brigham Young University, the College of William and Mary, and a non-profit development organization, Development Gateway. AiDA was established in 2001 by Development Gateway to serve as a current, timely registry of aid activities to improve aid coordination and effectiveness. The AidData live website and latest research release contain information on development finance activities for the period 1947-2011. The primary variables are compiled from a range of sources, including the OECD Creditor Reporting System (CRS) database, donor annual reports, project documents from both bilateral and multilateral aid agencies, data gathered directly from donor agency sources, and agency websites and databases. • Gender Statistics includes data on key gender topics. Themes included are demographics, education, health, labor force, and political participation. The data was launched together with the World Bank’s World Development Report 2012 on Gender Equality and Development. Gender Statistics was the sources for indicators such as the ratio of female to male educational enrollment and young adult literacy. (databank.worldbank.org) • Women in Parliament - The data on percentage of women in parliament is compiled by the Inter-Parliamentary Union (IPU) on the basis of information provided by National Parliaments by 1st May 2013. The percentages do not take into account the case of parliaments for which no data was available at that date. (www.ipu.org) 43 • Armed Conflict Database - This dataset is a collaborative project between the Department of Peace and Conflict Research at Uppsala University and the Centre for the Study of Civil War at PRIO. The database includes armed conflicts, both internal and external, in the period 1946 to the present. (http://www.prio.no/Data/Armed-Conflict/UCDP-PRIO/) Variables sources and definitions Source Definition Structure of the Economy GDP per capita WDI GDP per capita is gross domestic product divided by midyear population. Industry, value added to WDI Industry corresponds to ISIC divisions 10-45 and includes manufacturing (ISIC GDP (%) divisions 15-37). It comprises value added in mining, manufacturing (also reported as a separate subgroup), construction, electricity, water, and gas. Value added is the net output of a sector after adding up all outputs and subtracting intermediate inputs. It is calculated without making deductions for depreciation of fabricated assets or depletion and degradation of natural resources. The origin of value added is determined by the International Standard Industrial Classification (ISIC), revision 3 Services, value added to WDI Services correspond to ISIC divisions 50-99 and they include value added in GDP (%) wholesale and retail trade (including hotels and restaurants), transport, and government, financial, professional, and personal services such as education, health care, and real estate services. The industrial origin of value added is determined by the International Standard Industrial Classification (ISIC), rev. 3 Conflict and Rule of Law Conflict UCDP/PRIO UCDP defines conflict as: “a contested incompatibility that concerns government Armed Conflict and/or territory where the use of armed force between two parties, of which at Dataset, version least one is the government of a state, results in at least 25 battle-related deaths. 4-2009 Conflict intensity UCDP/PRIO Takes into account the temporal dimension of the conflict. It is a dummy variable Armed Conflict that codes whether the conflict since the onset has exceeded 1,000 battle-related Dataset, version deaths. A conflict is coded as 0 as long as it has not over time resulted in more than 1,000 battle-related deaths. Once a conflict reaches this threshold, it is coded 4-2009 as 1. Rule of law ICRG ICRG data: Law and Order are assessed separately, with each sub-component comprising zero to three points. The Law sub-component is an assessment of the strength and impartiality of the legal system, while the Order sub-component is an assessment of popular observance of the law. Thus, a country can enjoy a high rating – 3 – in terms of its judicial system, but a low rating – 1 – if it suffers from a very high crime rate of if the law is routinely ignored without effective sanction (for example, widespread illegal strikes). The maximum score is 6. Employment Women's labor force WDI Labor participation rate, female (% of female population ages 15+). Labor force participation participation rate is the proportion of the population ages 15 and older that is 44 economically active: all people who supply labor for the production of goods and services during a specified period. Share of women in non- WDI Share of women employed in the nonagricultural sector (% of total agricultural employment nonagricultural employment). Share of women employed in the nonagricultural sector is the share of female workers in the nonagricultural sector (industry and services), expressed as a percentage of total employment in the nonagricultural sector. Share of women who are WDI Wage and salaried workers, female (% of females employed) . Wage and salaried wage workers workers (employees) are those workers who hold the type of jobs defined as "paid employment jobs," where the incumbents hold explicit (written or oral) or implicit employment contracts that give them a basic remuneration that is not directly dependent upon the revenue of the unit for which they work. Share of women who are WDI Employers, female (% of employment). Employers refers are those workers who, employers working on their own account or with one or a few partners, hold the type of jobs defined as a "self-employment jobs" i.e. jobs where the remuneration is directly dependent upon the profits derived from the goods and services produced), and, in this capacity, have engaged, on a continuous basis, one or more persons to work for them as employee(s). Education Ratio of girls to boys in Gender Statistics Ratio of female to male primary enrollment is the percentage of girls to boys primary education enrolled at primary level in public and private schools Ratio of girls to boys in Gender Statistics Ratio of female to male primary enrollment is the percentage of girls to boys secondary education enrolled at secondary level in public and private schools Ratio of girls to boys in Gender Statistics Ratio of female to male primary enrollment is the percentage of girls to boys tertiary education enrolled at tertiary level in public and private schools Health Life expectancy (female) WDI Life expectancy at birth indicates the number of years a newborn female infant would live if prevailing patterns of mortality at the time of its birth were to stay the same throughout its life. Infant mortality WDI Infant mortality rate is the number of infants dying before reaching one year of age, per 1,000 live births in a given year. Empowerment CEDAW www.un.org/ An indicator whether a country has ratified the Convention on the Elimination of womenwatch Discrimination Against Women. Women in parliament IPU Women in parliaments are the percentage of parliamentary seats in a single or lower chamber held by women. Age of first marriage Gender Statistics Mean age at first marriage shows the average length of single life expressed in years among those who marry before age 50 45 Summary statistics Number First Last of Standard year year countries Average deviation Minimum Maximum available available Structure of the economy GDP per capita 100 7.72 1.53 4.01 10.64 1960 2010 Industry share of GDP 100 29.59 10.87 1.88 72.72 1960 2010 Services share of GDP 99 48.04 12.25 4.14 79.34 1960 2009 Share of natural resources in merchandise trade 98 25.09 37.04 0.00 788.40 1962 2009 Aid as a share of GDP 95 3.91 5.70 0.00 109.50 1980 2010 Conflict and rule of law Conflict 100 0.19 0.39 0.00 1.00 1960 2010 Conflict intensity 100 0.13 0.34 0.00 1.00 1960 2010 Rule of law 83 3.52 1.47 0.00 6.00 1985 2010 Employment Women's labor force participation 100 40.55 9.33 11.00 56.00 1990 2010 Share of women in non-agricultural 100 38.06 11.49 6.00 56.00 1990 2010 Share of women who are employers 83 2.07 1.34 0.00 8.00 1980 2010 Share of women who are wage workers 90 67.38 24.66 0.00 96.00 1980 2010 Education Ratio of girls to boys in primary education 99 89.46 15.66 17.24 149.19 1970 2010 Ratio of girls to boys in secondary education 99 85.69 26.05 17.26 309.97 1970 2010 Ratio of girls to boys in tertiary education 99 75.54 41.36 4.07 228.77 1970 2010 Ratio of young literate femalesmales (% 86 63.53 44.76 0.30 115.58 1975 2010 age 15-24) Health Life expectancy (female) 100 62.48 12.63 28.53 86.44 1960 2010 Infant mortality 100 66.99 49.21 2.00 273.00 1960 2010 Empowerment Age of first marriage 99 23.15 3.33 14.40 33.20 1980 2010 Women in parliament 100 14.76 10.21 0.00 56.30 1990 2010 46 Annex 7: Regression Results Table 3: Probability of reform ( 5 year lags) Access to and control Constitutional principles and Data Legal capacity of assets legal pluralism availability Contract without permission Religious law invalidated by Sons-daughters inheritance Customary law invalidated Legal proceedings without Husband-wife inheritance Work without permission Constitutional equality Costomary law is valid Bank account without by non-discrimination Religious law is valid Observations (5 year Unmarried property Countries with data Constitutional non- non-discrimination Head of household Married property discrimination Joint titling permission permission intervals) available Structure of the economy GDP per capita -* -* 100 859 GDP per capita ^ 2 -** -* -** -** 100 859 Manufacturing VA in GDP 98 651 Services VA in GDP * * ** 98 737 Share of natural resources in merchandise exports ** -* 97 627 Aid as a share of GDP ** 89 450 Conflict and rule of law Rule of law (ICRG) * * 83 395 Conflict -* -** -* 99 859 Intensity of conflict -** -** -** 99 859 Women's labor force participation -* 99 394 Emploment Non-agricultural employment (women) 97 243 Share of women in wage employment * 84 217 Share of women who are employers 78 185 Empower- CEDAW ** ** ** ** ** ** ** * * 99 859 ment Women in Parliament ** * * * 99 261 Age of firm marriage - female -* -* 95 241 Ratio of girls to boys in primary enrollment Education 98 561 Ratio of girls to boys in secondary enrollment -*** -*** * 98 456 Ratio of girls to boys in tertiary enrollment -** -** 95 388 Health Life expectancy of women -** -*** -* -** -*** ** -** 99 859 Infant mortality - female 59 118 47 Each cell represents a separate regression. Panel regressions were run with fixed effects and standard errors clustered at the country level. A trend was included as well as GDP per capita. Observations were included in 5 year intervals, with a 5 year lag in the explanatory variables. Shaded cells indicate insufficient reforms in the period covered to estimate effects. *** p<0.01, ** p<0.05, * p<0.1 Table 4: Probability of reform: allowing for non-linear effects of income and conflict, 1960-2010 (1) (2) (3) (4) (5) (6) (7) (8) Unmarried property without permission Head of household Married property Sons-daughters Work without Bank account Husband-wife inheritance inheritance Joint titling permission GDP/capita 0.382* 0.0285 0.225 0.244 0.105 -0.0720 0.202 0.157 (0.224) (0.202) (0.259) (0.267) (0.279) (0.190) (0.177) (0.175) GDP/capita squared -0.0272* -0.00481 -0.0172 -0.0178 -0.00605 0.00765 -0.0153 -0.0114 (0.0143) (0.0148) (0.0170) (0.0178) (0.0165) (0.0143) (0.0133) (0.0130) Conflict -0.0126 -0.00871 -0.0716** -0.0726** -0.0778** -0.0203 -0.0147 -0.0158 (0.0447) (0.0478) (0.0336) (0.0343) (0.0355) (0.0405) (0.0524) (0.0398) Intensity of conflict -0.0298 -0.0499 0.0425 0.0280 0.0773** 0.0210 0.0163 -0.0540 (0.0227) (0.0521) (0.0464) (0.0479) (0.0376) (0.0498) (0.0548) (0.0373) CEDAW 0.00932 0.0798** 0.0674** 0.0745** 0.0214 0.0624* 0.0859** 0.0511* (0.0279) (0.0343) (0.0330) (0.0347) (0.0262) (0.0346) (0.0378) (0.0307) Observations 859 853 854 854 835 862 858 863 R-squared 0.113 0.204 0.107 0.118 0.028 0.156 0.133 0.139 Number of countries 99 99 100 100 99 100 100 100 (9) (10) (11) (12) (13) (14) (15) (16) Customary law invalidated Legal proceedings without Religious law invalidated Constitutional equality Costomary law is valid by non-discrimination by non-discrimination Religious law is valid Constitutional non- Contract without discrimination permission permission GDP/capita -0.0707 0.208 0.313** 0.130 0.277 0.0900 0.172* -0.184 (0.170) (0.188) (0.125) (0.221) (0.187) (0.141) (0.0899) (0.128) GDP/capita squared 0.00417 -0.0119 -0.0214** -0.0161 -0.0235** -0.00610 -0.0115** 0.0111 (0.0128) (0.0121) (0.00843) (0.0149) (0.0111) (0.00819) (0.00555) (0.00808) Conflict -0.0210 -0.0141 -0.0316 -0.00182 -0.0698* -0.00386 0.0408** -0.0348** (0.0389) (0.0418) (0.0361) (0.0458) (0.0361) (0.0129) (0.0176) (0.0147) Intensity of conflict -0.0423 -0.0321 0.0127 -0.0196 0.0741 -0.0446* -0.0311 0.00606 (0.0269) (0.0260) (0.0472) (0.0585) (0.0531) (0.0266) (0.0265) (0.0330) CEDAW 0.0855** 0.0576** 0.0516 0.0320 0.0645* 0.00485 -0.0299 0.0319 (0.0331) (0.0286) (0.0360) (0.0479) (0.0386) (0.0180) (0.0211) (0.0203) Observations 861 871 864 864 864 863 864 863 R-squared 0.141 0.082 0.102 0.223 0.114 0.010 0.058 0.037 Number of countries 100 100 100 100 100 100 100 100 Robust standard errors in parentheses; *** p<0.01, ** p<0.05, * p<0.1 Control variables are all lagged by 5 years. 48 Table 4B: Probability of reform: allowing for non-linear effects of income and conflict, 1960-1985 (1) (2) (3) (4) (5) (6) (7) (8) Unmarried property without permission Head of household Married property Sons-daughters Work without Bank account Husband-wife inheritance inheritance Joint titling permission GDP/capita 0.563 -0.556 0.513* 0.163 0.469 -1.154*** -0.254 -0.306 (0.377) (0.464) (0.285) (0.235) (0.395) (0.380) (0.349) (0.412) GDP/capita squared -0.0394 0.0386 -0.0327* -0.00612 -0.0283 0.0819*** 0.0208 0.0230 (0.0249) (0.0317) (0.0168) (0.0146) (0.0271) (0.0282) (0.0251) (0.0287) Conflict -0.0340 -0.0225 -0.00182 0.000184 -0.0329 0.0285 -0.0367 -0.0143 (0.0590) (0.0587) (0.00740) (0.00647) (0.0358) (0.0496) (0.0521) (0.0587) Intensity of conflict -0.0343 0.00874 -0.00954 -0.00622 0.0313 -0.0148 0.0219 -0.00981 (0.0372) (0.0527) (0.00946) (0.00673) (0.0292) (0.0373) (0.0451) (0.0508) CEDAW 0.192 0.0830 0.236 0.234 0.243 0.0182 0.122 -0.103*** (0.195) (0.150) (0.198) (0.217) (0.204) (0.106) (0.151) (0.0385) Observations 376 372 371 371 359 377 373 378 R-squared 0.117 0.120 0.081 0.069 0.059 0.168 0.082 0.102 Number of countries 81 81 80 80 79 82 82 82 (9) (10) (11) (12) (13) (14) (15) (16) Customary law invalidated Legal proceedings without Religious law invalidated Constitutional equality Costomary law is valid by non-discrimination by non-discrimination Religious law is valid Constitutional non- Contract without discrimination permission permission GDP/capita -0.463 -0.343 1.101*** -0.182 0.354 -0.0959 0.294 -0.479 (0.352) (0.247) (0.318) (0.405) (0.439) (0.379) (0.248) (0.356) GDP/capita squared 0.0348 0.0259 -0.0611*** 0.00149 -0.0173 0.00572 -0.0230 0.0327 (0.0256) (0.0181) (0.0200) (0.0273) (0.0277) (0.0231) (0.0151) (0.0226) Conflict 0.0310 0.00680 0.0237 -0.0469 -0.0160 -0.000382 0.0471 -0.00219 (0.0478) (0.0523) (0.0355) (0.0503) (0.0125) (0.00973) (0.0444) (0.0252) Intensity of conflict -0.0444 -0.0194 -0.0531 -0.103* 0.000415 -0.00821 -0.0894* 0.0735 (0.0368) (0.0394) (0.0408) (0.0609) (0.0347) (0.0277) (0.0451) (0.0478) CEDAW 0.112 -0.00524 -0.0918* 0.115 -0.0335 0.0124 -0.0385 0.0560 (0.149) (0.0481) (0.0519) (0.165) (0.0324) (0.0260) (0.0274) (0.0385) Observations 376 385 376 376 376 375 376 375 R-squared 0.098 0.074 0.083 0.124 0.014 0.004 0.075 0.064 Number of countries 82 83 83 83 83 83 83 83 Robust standard errors in parentheses; *** p<0.01, ** p<0.05, * p<0.1 Control variables are all lagged by 5 years. GDP per capita is not available for the whole sample before the 1990s. 49 Table 4C: Probability of reform: allowing for non-linear effects of income and conflict, 1986-2010 (1) (2) (3) (4) (5) (6) (7) (8) Unmarried property without permission Head of household Married property Sons-daughters Work without Bank account Husband-wife inheritance inheritance Joint titling permission GDP/capita 0.276 0.324 0.197* 0.356** -0.116 0.423** 0.115 0.306* (0.170) (0.195) (0.106) (0.144) (0.0954) (0.199) (0.128) (0.167) GDP/capita squared -0.0210* -0.0285* -0.0167**-0.0285*** 0.00788 -0.0299** -0.00911 -0.0219* (0.0121) (0.0149) (0.00790) (0.0105) (0.00812) (0.0145) (0.00962) (0.0114) Conflict 0.0711 0.0738 -0.0980 -0.0995 -0.0797* 0.0294 -0.0368 0.00107 (0.0633) (0.0787) (0.0663) (0.0666) (0.0460) (0.0485) (0.0342) (0.0164) Intensity of conflict -0.0510 -0.0861 0.105 0.0794 0.0693 -0.0529 -0.0526 -0.0538* (0.0318) (0.0723) (0.0887) (0.0889) (0.0615) (0.0668) (0.0556) (0.0302) CEDAW 0.0484 0.0835 0.0622 0.0874* -0.00240 0.129** 0.106* 0.103** (0.0324) (0.0547) (0.0407) (0.0443) (0.0373) (0.0545) (0.0547) (0.0460) Observations 483 481 483 483 476 485 485 485 R-squared 0.072 0.135 0.097 0.117 0.018 0.139 0.121 0.125 Number of countries 99 99 100 100 99 100 100 100 (9) (10) (11) (12) (13) (14) Customary law invalidated (15) (16) Legal proceedings without Religious law invalidated Constitutional equality Costomary law is valid by non-discrimination by non-discrimination Religious law is valid Constitutional non- Contract without discrimination permission permission GDP/capita 0.312* 0.304* 0.00405 0.119 0.302 0.189 0.0494 0.122 (0.168) (0.165) (0.173) (0.231) (0.224) (0.134) (0.0504) (0.124) GDP/capita squared -0.0223* -0.0217* 0.00244 -0.0120 -0.0237 -0.0143 -0.00425 -0.00831 (0.0115) (0.0113) (0.0146) (0.0185) (0.0154) (0.00918) (0.00430) (0.00839) Conflict -0.00617 0.000901 0.00750 0.0337 -0.0153 0.0110 0.0185 0.00838 (0.0194) (0.0164) (0.0207) (0.0612) (0.0274) (0.0126) (0.0181) (0.00889) Intensity of conflict -0.0739** -0.0537* -0.0264 -0.0336 -0.122*** -0.0181 -0.0218 -0.000986 (0.0342) (0.0302) (0.0267) (0.0555) (0.0418) (0.0132) (0.0208) (0.00288) CEDAW 0.131** 0.103** 0.102** 0.0643 0.0282 -0.00268 -0.00867 -0.0102 (0.0511) (0.0461) (0.0486) (0.0630) (0.0548) (0.0347) (0.00865) (0.0102) Observations 485 486 488 488 488 488 488 488 R-squared 0.160 0.125 0.098 0.159 0.070 0.030 0.017 0.023 Number of countries 100 100 100 100 100 100 100 100 n parentheses; *** p<0.01, ** p<0.05, * p<0.1 Control variables are all lagged by 5 years. 50 Note: Results are similar if restrict to same set of countries available for the 1960-85 sample using in Table 2B. Table 5: Panel regressions with expanded list of explanatory variables (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) Access to and control of assets Legal capacity Contract withou Sons-daughters Work without Husband-wife Bank account proceedings Unmarried inheritance inheritance Joint titling permission permission permission permission household property property Married without without Head of Legal GDP per capita 0.529** 0.574** 0.00159 0.00404 0.151 0.685** 0.625** 0.746** 0.553** 0.544** (0.255) (0.275) (0.00272) (0.00366) (0.330) (0.314) (0.314) (0.294) (0.230) (0.215) GDP per capita ^2 -0.0377** -0.0489*** 0.446 0.546 -0.0124 -0.0498** -0.0500** -0.0532*** -0.0405*** -0.0401*** (0.0164) (0.0179) (0.450) (0.501) (0.0204) (0.0203) (0.0216) (0.0190) (0.0153) (0.0145) Conflict -0.00753 0.0402 -0.0286 -0.0409 -0.0429 -0.0165 -0.0227 -0.0192 -0.0479 -0.0256 (0.0553) (0.0572) (0.0324) (0.0367) (0.0302) (0.0535) (0.0612) (0.0443) (0.0543) (0.0450) Conflict intensity -0.0205 -0.0826 -0.0795* -0.0754* 0.0826* -0.0244 -0.00612 -0.0433 -0.0325 -0.0233 (0.0425) (0.0622) (0.0432) (0.0437) (0.0476) (0.0677) (0.0546) (0.0407) (0.0384) (0.0360) CEDAW 0.0196 0.0411 0.0859 0.0715 0.0120 0.0645* 0.0574 0.0546* 0.0863** 0.0689** (0.0245) (0.0379) (0.0847) (0.0883) (0.0306) (0.0374) (0.0455) (0.0316) (0.0382) (0.0328) Industry VA / GDP -0.000794 -0.00370 0.0859** 0.102** -0.00137 -0.00187 -0.000921 -0.00408 -0.00200 0.000358 (0.00242) (0.00360) (0.0361) (0.0400) (0.00325) (0.00351) (0.00331) (0.00286) (0.00231) (0.00230) Service VA / GDP -7.54e-05 -0.000806 -0.00619* -0.00202 0.00128 -0.00164 -0.00115 -0.00278 0.000848 0.00317 (0.00169) (0.00292) (0.00337) (0.00369) (0.00341) (0.00269) (0.00320) (0.00222) (0.00207) (0.00242) Primary enrollment -8.36e-05 -0.00345 -0.00193 -5.01e-05 -0.00338 -0.00145 -0.00546 -0.00266 -0.000596 -0.00139 (0.00225) (0.00230) (0.00267) (0.00284) (0.00226) (0.00231) (0.00344) (0.00189) (0.00240) (0.00193) Trend 0.00274 0.0106*** 0.00121 -0.00261 0.00172 0.00683*** 0.00756** 0.00546*** 0.00276 0.00191 (0.00178) (0.00271) (0.00200) (0.00432) (0.00226) (0.00229) (0.00291) (0.00188) (0.00180) (0.00178) Observations 521 517 518 518 506 522 521 523 521 526 R-squared 0.084 0.194 0.125 0.134 0.048 0.150 0.128 0.145 0.151 0.107 Number of ccc 97 97 98 98 97 98 98 98 98 98 Robust standard errors in parentheses; *** p<0.01, ** p<0.05, * p<0.1 Regressions use 5 year intervals with country fixed effects and clustered standard errors. Additional variables included in regressions (reported separately as sample size decreases): Women in Parliament 0.00445 0.0133*** 0.00755* 0.00729* 0.00803* 0.00830** 0.00371 0.00845** 0.00831** 0.00845** (222 obs; 96 countries) (0.00405) (0.00444) (0.00427) (0.00434) (0.00472) (0.00381) (0.00462) (0.00363) (0.00355) (0.00363) 51 Law and order 0.00773 0.0326* 0.0179* 0.0176 0.00980 0.0280 0.0341** -0.00191 0.0154 -0.00182 (316 obs; 82 countries) (0.00597) (0.0188) (0.0108) (0.0108) (0.00693) (0.0185) (0.0170) (0.00738) (0.0155) (0.00729) Table 6: Panel regression: Non-linear effects by GDP per capita (1) (2) (3) (4) (5) (6) (7) (8) (9) Access to and control of assets Legal capacity without permission without permission Head of household Married property Contract without Legal proceedings Sons-daughters Work without Husband-wife Bank account Unmarried inheritance inheritance permission permission property GDP per capita 0.596** 0.872*** 0.290 0.233 0.555 0.352 0.787** 0.470* 0.616** (0.274) (0.329) (0.507) (0.512) (0.371) (0.384) (0.324) (0.270) (0.264) GDP per capita^2 -0.0372** -0.0475*** -0.00465 -0.00654 -0.0474** -0.0425** -0.0438** -0.0335** -0.0357** (0.0165) (0.0175) (0.0320) (0.0361) (0.0195) (0.0211) (0.0167) (0.0147) (0.0156) Industry VA / GDP -0.000807 -0.00377 -0.00650* -0.00261 -0.00281 -0.00143 -0.00478 -0.00242 0.000228 (0.00256) (0.00348) (0.00342) (0.00355) (0.00356) (0.00349) (0.00298) (0.00232) (0.00250) Services VA/ GDP -0.000223 -0.00172 -0.00171 0.000702 -0.00230 -0.000613 -0.00284 0.00146 0.00296 (0.00174) (0.00291) (0.00272) (0.00271) (0.00279) (0.00350) (0.00237) (0.00220) (0.00252) Conflict 0.377 0.359 -0.372 -0.356 0.0186 0.257 -0.0423 -0.0318 0.0119 (0.287) (0.351) (0.268) (0.272) (0.259) (0.261) (0.245) (0.244) (0.227) Conflict*GDPpc -0.0577 -0.0467 0.0425 0.0401 -0.00430 -0.0431 0.00243 -0.00414 -0.00574 (0.0386) (0.0486) (0.0369) (0.0368) (0.0381) (0.0361) (0.0352) (0.0367) (0.0335) Conflict intensity -0.182 -0.296 0.548 0.579 -0.643* -0.219 -0.161 -0.0288 -0.00797 (0.226) (0.368) (0.418) (0.428) (0.356) (0.264) (0.225) (0.196) (0.192) Conf. intensity*GDPp 0.0239 0.0320 -0.0662 -0.0731 0.0919* 0.0316 0.0169 -0.00164 -0.00198 (0.0298) (0.0488) (0.0526) (0.0545) (0.0544) (0.0405) (0.0309) (0.0286) (0.0273) CEDAW 0.0716 0.0139 0.445** 0.659*** 0.0708 0.280* 0.173 0.233 0.117 (0.0930) (0.153) (0.171) (0.217) (0.150) (0.165) (0.136) (0.149) (0.169) CEDAW * GDPpc -0.00779 0.00296 -0.0513** -0.0788*** -9.71e-05 -0.0309 -0.0158 -0.0195 -0.00704 (0.0117) (0.0203) (0.0204) (0.0266) (0.0201) (0.0206) (0.0173) (0.0182) (0.0213) Primary enrollment 0.00340 0.0170* 0.00819 0.000902 -0.00764 -0.0197 0.00644 -0.00204 0.00637 (0.0106) (0.00974) (0.00868) (0.0147) (0.0114) (0.0133) (0.00738) (0.00982) (0.00766) Primary enroll*GDPp-0.000606 -0.00350** -0.00129 -0.000769 0.00111 0.00240 -0.00168 0.000119 -0.00136 (0.00155) (0.00173) (0.00142) (0.00206) (0.00214) (0.00237) (0.00130) (0.00183) (0.00121) Trend 0.0027 0.011*** 0.0018 0.0036 0.0063** 0.0064** 0.0056*** 0.0024 0.0021 (0.0017) (0.0028) (0.0028) (0.0033) (0.0025) (0.0027) (0.0020) (0.0020) (0.0018) Observations 521 517 517 517 522 521 523 521 526 R-squared 0.103 0.210 0.182 0.211 0.169 0.143 0.153 0.155 0.111 Number of countries 97 97 98 98 98 98 98 98 98 Robust standard errors in parentheses; *** p<0.01, ** p<0.05, * p<0.1 Regressions use 5 year intervals with country fixed effects and clustered standard errors. 52 Table 7: Panel regression: Non-linear effects by 'rule of law' (1) (2) (3) Access to and control of assets Sons-daughters Husband-wife Married property inheritance inheritance GDP per capita -0.00778 0.335 0.0790 (0.197) (0.232) (0.305) GDP per capita^2 -0.0121 -0.0192 -0.00371 (0.0140) (0.0134) (0.0181) Rule of law 0.235 -0.0247 0.0119 (0.179) (0.0481) (0.0672) Conflict -0.0965 -0.399 -0.364 (0.0908) (0.330) (0.327) Conflict*Rule_of_Law 0.0564 0.112 0.100 (0.0354) (0.0904) (0.0898) Conflict intensity 0.125 0.456 0.459 (0.115) (0.343) (0.342) Conf. intensity*Rule_of_Law -0.0677* -0.114 -0.118 (0.0362) (0.0887) (0.0889) CEDAW 0.102 0.241* 0.328** (0.159) (0.123) (0.145) CEDAW * Rule_of_Law -0.0193 -0.0422* -0.0581** (0.0343) (0.0231) (0.0283) Primary enrollment -0.00797 -0.00690 -0.00696 (0.00604) (0.00449) (0.00454) Primary enroll*Rule_of_Law -0.000669 0.00125 0.00194* (0.00158) (0.000909) (0.00106) Industry/GDP 0.00706 -0.00391 0.000456 (0.00791) (0.00395) (0.00528) Industry/GDP * Rule_of_Law -0.00144 -0.000158 -0.00113 (0.00200) (0.000725) (0.00111) Services/GDP 0.00370 0.000839 0.00628 (0.00490) (0.00285) (0.00485) Services/GDP * Rule_of_Law -0.00161 -0.000730 -0.00181* (0.00135) (0.000669) (0.00106) Trend 0.0103** -6.12e-05 -0.000734 (0.00397) (0.000963) (0.00120) Observations 316 316 316 R-squared 0.175 0.237 0.260 Number of countries 81 82 82 Robust standard errors in parentheses; *** p<0.01, ** p<0.05, * p<0.1 53 and clustered standard errors. Regressions use 5 year intervals with country fixed effects