Maternity and Labor Markets: Impact of Legislation in Colombia

Ramírez Bustamante, Natalia; Tribin Uribe, Ana Maria; Vargas, Carmiña O. Working Paper Maternity and Labor Markets: Impact of Legislation in Colombia IDB Working Paper Series, No. IDB-WP-583 Provided in Cooperation with: Inter-American Development Bank (IDB), Washington, DC Suggested Citation: Ramírez Bustamante, Natalia; Tribin Uribe, Ana Maria; Vargas, Carmiña O. (2015) : Maternity and Labor Markets: Impact of Legislation in Colombia, IDB Working Paper Series, No. IDB-WP-583, Inter-American Development Bank (IDB), Washington, DC, https://hdl.handle.net/11319/6841 This Version is available at: https://hdl.handle.net/10419/115516 Standard-Nutzungsbedingungen: Die Dokumente auf EconStor dürfen zu eigenen wissenschaftlichen Zwecken und zum Privatgebrauch gespeichert und kopiert werden. Sie dürfen die Dokumente nicht für öffentliche oder kommerzielle Zwecke vervielfältigen, öffentlich ausstellen, öffentlich zugänglich machen, vertreiben oder anderweitig nutzen. Sofern die Verfasser die Dokumente unter Open-Content-Lizenzen (insbesondere CC-Lizenzen) zur Verfügung gestellt haben sollten, gelten abweichend von diesen Nutzungsbedingungen die in der dort genannten Lizenz gewährten Nutzungsrechte. Terms of use: Documents in EconStor may be saved and copied for your personal and scholarly purposes. You are not to copy documents for public or commercial purposes, to exhibit the documents publicly, to make them publicly available on the internet, or to distribute or otherwise use the documents in public. If the documents have been made available under an Open Content Licence (especially Creative Commons Licences), you may exercise further usage rights as specified in the indicated licence. http://creativecommons.org/licenses/by-nc-nd/3.0/igo/legalcode Maternity and Labor Markets: Impact of Legislation in Colombia Natalia Ramírez Bustamante Ana Maria Tribin Uribe Carmiña O. Vargas Country Department Andean Group IDB-WP-583 IDB WORKING PAPER SERIES No. Inter-American Development Bank March 2015 Maternity and Labor Markets: Impact of Legislation in Colombia Natalia Ramírez Bustamante, Ana Maria Tribin Uribe and Carmiña O. Vargas 2015 Inter-American Development Bank Cataloging-in-Publication data provided by the Inter-American Development Bank Felipe Herrera Library Ramirez Bustamante, Natalia. Maternity and labor markets: impact of legislation in Colombia / Natalia Ramirez Bustamante, Ana María Tribin Uribe, Carmiña O. Vargas. p. cm. — (IDB Working Paper Series ; 583) Includes bibliographic references. 1. Women—Employment—Colombia. 2. Maternity leave—Law and legislation—Colombia. I. Tribin Uribe, Ana María. II. Vargas Riaño, Carmiña Ofelia. III. Inter-American Development Bank. Country Department Andean Group. IV. Title. V. Series. IDB-WP-583 http://www.iadb.org Any dispute related to the use of the works of the IDB that cannot be settled amicably shall be submitted to arbitration pursuant to the UNCITRAL rules. The use of the IDB’s name for any purpose other than for attribution, and the use of IDB’s logo shall be subject to a separate written license agreement between the IDB and the user and is not authorized as part of this CC-IGO license. Following a peer review process, and with previous written consent by the Inter-American Development 2015 Copyright © Inter-American Development Bank. This work is licensed under a Creative Commons IGO 3.0 Attribution-NonCommercial-NoDerivatives (CC-IGO BY-NC-ND 3.0 IGO) license ( http://creative - commons.org/licenses/by-nc-nd/3.0/igo/legalcode ) and may be reproduced with attribution to the IDB and for any non-commercial purpose. No derivative work is allowed. Bank (IDB), a revised version of this work may also be reproduced in any academic journal, including those indexed by the American Economic Association’s EconLit, provided that the IDB is credited and that the author(s) receive no income from the publication. Therefore, the restriction to receive income from such publication shall only extend to the publication’s author(s). With regard to such restriction, in case of any inconsistency between the Creative Commons IGO 3.0 Attribution-NonCommercial-NoDerivatives license and these statements, the latter shall prevail. Note that link provided above includes additional terms and conditions of the license. The opinions expressed in this publication are those of the authors and do not necessarily reflect the views of the Inter-American Development Bank, its Board of Directors, or the countries they represent. Maternity and Labor Markets: Impact of Legislation in Colombia1 Natalia Ramírez Bustamante2 Ana Maria Tribin Uribe3 Carmiña O. Vargas4 Abstract This study seeks to determine the impact on female labor outcomes of the amendment to the Colombian labor law that extended maternity leave from 12 to 14 weeks (Law 1468 of July 2011). To identify this impact, labor market outcomes of two groups of women with different fertility rates are compared. The study finds evidence that as a result of the extension of the maternity leave period, women in the high-fertility age group experience an increase in inactivity rates, informality, and self-employment. The study points to the need for a redesign of maternity protection policy that would enable the economic and social costs of bearing children to be shared by both parents and that may generate social change regarding the importance of paternal care. JEL classification: J08, J2, J3, J7, K31 Key words: maternity leave, female labor market, labor regulation 1 We would like to give special thanks to our research assistants, Andrea Paola Poveda and Daniel Rodríguez Guio. We appreciate the help of Jaime Tenjo, Oriana Alvarez, and Maria Camila Jiménez in obtaining the data for this research. Valuable comments made by Marcela Eslava, Dolores de la Mata, and Raquel Bernal were indispensable to this research. We also thank Luis Eduardo Arango, Francesca Castellani, and Eduardo Lora for their support and comments. Opinions expressed herein are those of the authors and do not necessarily reflect the views of, or imply any responsibility of, the institutions with which they are affiliated. 2 Lawyer and Philosopher, Universidad de los Andes. LL.M. and SJD Candidate in Law at Harvard University. Professor, Department of Law, Universidad de los Andes. email: [email protected]. 3 Economist, Pontificia Universidad Javeriana. M.A. in Economics, Pontificia Universidad Javeriana and Brown University. Ph.D. in Economics from Brown University. Researcher at Banco de la República (Colombian Central Bank). Email: [email protected]. 4 Economist, Universidad Nacional, M.A. in Economics, Universidad Nacional and Brown University. Ph.D. in Economics from Brown University, Researcher at Banco de la República (Colombian Central Bank). Email: [email protected].co. 1 1. Introduction The increasing participation of women in economic activity is one of the greatest social revolutions of the twentieth century. Colombia is the Latin American country with the largest increase in the rate of female participation in the labor force in the last three decades (Amador, Bernal, and Peña, 2013). Several social factors drove this process. Access by women to college education since 1934, birth control through mass access to contraceptives, growing urbanization, and the increasing need for dual-earner households, among other factors, prompted the increase in women’s participation in the economically productive sectors. Due in part to this increase, and in an attempt to spark more interest in paid work among women, starting in the 1930s labor laws began protecting the event of pregnancy for female workers, and subsequently added protections for women from gender discrimination. This same effort was reflected in the Colombian Constitution in 1991 in the equality and non-discrimination clauses, and is a central element in international treaties signed by Colombia, most prominently the Convention on the Elimination of all Forms of Discrimination against Women (CEDAW).5 However, the equality to which women are entitled by law is hampered by a series of biological events, including pregnancy, childbirth, and breastfeeding, that directly and primarily affect women.6 In recognition of this biological disparity and in order to promote female employment during the period of industrialization, several countries began implementing maternity leave programs in the early twentieth century. This trend can be seen in the adoption of the International Labour Organisation’s (ILO) Convention No. 3, which has been mandatory in Colombia since 1931,7 as well as subsequent laws that have modified the maternity protection regime. According to the laws in effect in 2014, pregnant workers in Colombia have two fundamental guarantees: first, the right to enhanced job security, which implies a prohibition 5 The United Nations General Assembly adopted CEDAW in 1979. 6 Although these biological events are thought to affect only women who are mothers or who hope to become mothers, as will be shown in Section 2, cultural norms and preconceptions about motherhood affect all women, regardless of their parental status or their individual preference regarding motherhood. 7 The ILO has adopted two other Conventions on maternity that expand the rights granted in Convention No. 3. These are Conventions 103 (1952) and 183 (2000). They have not been ratified by Colombia (ILO, 2001). A list of ratifications by Convention and ratifying country is available at: http://www.ilo.org/public/spanish/standards/relm/ilc/ilc90/pdf/rep-iii-2.pdf. Last accessed January 2014. 2 against dismissal due to pregnancy that protects the worker during the nine-month gestation period and maternity leave; and second, 14 weeks of paid maternity leave.8 For the ILO, these benefits aim, first, to protect the health of women and children during pregnancy and after birth; and second, to ensure that women’s reproductive role does not adversely affect their economic and job security (ILO, 2005; 2010). The most recent ILO initiative was to promote payment of maternity leave by social security systems rather than exclusively by employers in order to prevent discrimination against female workers. Indeed, the ILO has recognized that motherhood and the responsibilities associated with it still have an important influence on perceptions about women in the labor market and can be used by employers as a discriminatory criterion when recruiting workers. For this reason, the ILO stated that “the problem is how to ensure that employers do not reject candidates of reproductive age who are already taking on the heavier burden of family responsibilities and whose absence due to maternity leave or even longer periods generates organizational problems for employers, in some cases even assuming the financial burden of paying wages during such absences” (ILO, 1997). The focus of the ILO has been twofold. First, it has emphasized the importance of including women in labor markets with regard for their sexual and reproductive rights, including the right to be a mother without fear of discrimination; and second, it has highlighted the urgency of generating institutional mechanisms to reduce the costs to employers associated with workers who take maternity leave, in order to reduce the discriminatory effects that maternity leave may have on female workers. Indeed, gender equality in employment requires that women and men be treated as substantially similar subjects with similar capacities, potential, and aspirations for personal and professional growth, who, with similar training, stand on equal footing to reach the highest levels of productivity in a given job. Achieving this aim requires the elimination of differentiating economic and institutional costs that could create an incentive for employers to hire male over female workers. By the same token, male and female workers, regardless of age, sexual orientation, or parental or marital status, should assume the same burdens regarding the care of children, relatives, and the elderly and should be expected to need certain special benefits to take 8 Colombian law mandates that the worker must take one of the 14 weeks before the birth of the child and the next 13 weeks after the birth. In the case of premature births, the 14 weeks may be increased by the difference between the initial due date and the date that the child is born. In the case of multiple births, the length of maternity leave is increased by two weeks. 3 care of significant others.9 Specifically in the case of a newborn, both parents should be expected to contribute to the care and welfare of the child, which means that similar portions of paid leave should be legally recognized for both parents. Otherwise, employers could be motivated to try to offset the additional cost of female workers by not hiring them, hiring them less frequently, or firing them more often than male workers. To quantify the direct costs of maternity leave paid by employers in Colombia, Espino and Salvador (2014) use information from household surveys in 2012,10 shown in Table 1, and take into account current labor legislation. They find that the annual extra cost to the firm of providing paid maternity leave is 6.73 percent of the average annual female wage.11 This extra cost is directly associated with providing maternity leave and hiring a replacement worker during the leave period. There are, however, other important costs assumed by the employer that are not taken into account in this calculation. These are related to the necessary adjustments in the organization due to the absence of the worker on leave, including the costs of posting the job, searching for an appropriate replacement, and productivity losses either during the training and adjustment period or throughout the leave period if the replacement is not as productive as the worker on maternity leave. Unfortunately, measuring these costs is very difficult due to lack of appropriate data. These factors show the potential differential impact of maternity and labor legislation designed to protect it on women’s employability. For these reasons, examining this type of regulatory intervention is especially important for understanding labor market differences by gender, but the scope of its impact is not yet fully understood. Our research seeks to determine the effect of the most recent reform of pregnancy legislation, which extended maternity leave from 12 to 14 weeks (Law 1468 from 2011), on women’s employment status. We believe that research on these variables could help to identify some of the determinants of female employment and unemployment in Colombia, and, in subsequent research, could serve as a benchmark for comparison with labor markets with similar economic and institutional contexts 9 In addition to maternity leave policies, several countries recognize the right to parental leave to attend to children’s emergencies or continued illnesses without the fear of job loss or lost pay. 10 Specifically, they use Gran Encuesta Integrada de Hogares (monthly household surveys) and Encuesta Nacional de Calidad de Vida (quality-of-life household surveys), both administered by DANE, the national institution responsible for collecting statistics in Colombia. 11 The annual cost for a worker with no maternity leave is 100. For a woman with maternity leave, the employer must cover 73.06 percent of her annual salary (remaining wages are covered by the social security system). 30.31 percent corresponds to the salary of the replacement worker, and 3.37 percent corresponds to the value of one paid hour per day for nursing for three months after woman returns from maternity leave. 4 as those of Colombia, which is the case of most countries in the Latin American and Caribbean region. To understand the change in legislation and its effect on women’s labor market outcomes, we exploit the differential effect on women who are associated with different fertility rates because of their age. That is, we compare women of childbearing age (between 18 and 30 years of age) to women between the ages of 40 and 55. These two groups have very different fertility rates. Thus, Law 1468 affects women between 18 and 30 years of age (the treatment group), but not women between 40 and 55 years of age (the control group), who have passed their peak of fertility. We use the difference-in-differences approach. The results show that since 2011, women of childbearing age (treatment group) experience worse labor market outcomes than the group of women with low fertility rates (control group). Women in the treatment group are more likely to enter into inactivity, informality, and self-employment after 2011 relative to women in the control group. We explore the possibility that our results might be tainted by a period of adjustment, or some cyclicality responses, around the time of implementation of the law. For this, we run regressions excluding data from some months before and after implementation of the changes in the maternity leave period. We also check for the possibility that our results might be affected by the cohort composition of the groups. Therefore, we run the same regressions for men only, and also using men ages 18 through 30 as the control group. Finally, we estimate placebo treatment effects using data from pretreatment years. We show that our results are robust across demographic groups and time periods, suggesting a causal effect of the increase in the maternity leave period. Following Autor, Donohue, and Schwab (2006), we stress that our paper does not attempt to provide an overall assessment of maternity protection laws. We do not evaluate the benefits of such laws to workers or the broader public. The fact that there are some effects on the labor market for high-fertility women underscores the fact that legal protections do not come without cost. We make public policy recommendations which we believe would correct the distortions created by the legislation, so that women can take advantage of the 14 weeks of maternity leave without being punished by the labor market. This document is divided into eight sections. Section 2 reviews the relevant literature. Section 3 briefly describes Colombian legislation on maternity protection. Section 4 presents a 5 social services and non-discrimination. Finally, efforts have been made to extend the spectrum of protection to workers who had previously been excluded (Ramírez, 2008). Currently, Colombian labor law includes, within the general social security system, a package of protection for pregnant workers comprising the following benefits: (i) prohibition against dismissal of the worker on account of pregnancy during the pregnancy and maternity leave;18 (ii) a paid leave of 14 weeks around the time of birth; (iii) at the end of maternity leave, the reinstatement of the employee in the previous post, and (iv) two breaks of 30 minutes each for feeding the child during the first six months of the infant’s life. Maternity leave is paid for by the health insurance system to which the worker is affiliated, out of the contributions made by the worker and her employer as per the employment contract. In contrast to this protection package, men rely exclusively on a paid leave of eight working days after the birth of the child.19 The general rule was that the above-described measures benefited formal sector workers tied to an employer by an employment contract for an indefinite term. However, finding that employers were offering women short-term contracts to circumvent the protection of workers during pregnancy, starting in 1997 the Constitutional Court extended similar protections for workers with fixed-term contracts.20 Under this judicial interpretation, if the agreed term comes to an end and the worker is pregnant, the employer must obtain authorization from a labor 18 A pregnant worker may be dismissed if there is fair cause for termination of the employment contract and if authorization from a Labor Inspector has previously been granted. 19 Law 755 of 2002. 20 The Court has argued, for example, that “the arrival of the date of termination of the contract is not always a just cause for termination of the employment relationship, because if at the contract’s date of expiration the causes for hiring remain and the subject still stands, and moreover, the worker has fully complied with her obligations, the employer “must ensure the renewal of the contract.” Therefore, to terminate employment of a pregnant worker who has notified the employer of her pregnancy, and the worker has been complying with all the obligations of her contract, [the Labor Inspector] must investigate whether the causes that originated the contract remain, and if they do, then the employment relationship cannot be terminated.” Colombian Constitutional Court, T-326 of 1998. Similar arguments have been put forth in the following decisions: T-426/98; T-375/00; T-764/00; T-664/01; T206/02; T-113/03; T-895/04; T-1236/04. 12 inspector to proceed with the termination.21 Similar protection has been extended to workers hired through at will contracts.22 In summary, the jurisprudence of the Colombian Constitutional Court shows a continuous effort to ensure that the constitutional rights of pregnant workers are protected. To do so effectively, the Court had to harmonize the rights of freedom of contract with the special protections of motherhood under the Constitution and labor law, which leads to recruitment and firing constraints. 4. Theoretical Model The legislation on maternity leave is known as a variety of accommodation mandates, where the beneficiaries are entitled to a set of accommodations to facilitate their participation in a given social or economic setting. The characteristic of these mandates is that they apply to a clearly identifiable group. Generally, this type of legislation combines with anti-discrimination rules, and its consequences on labor market outcomes depend on its effect on labor demand and supply and the incentives imposed by anti-discrimination rules. Therefore, when evaluating the efficiency of these policies, their effect on wages as well as on employment must be considered.23 In this study, we use a stylized model to explore the consequences of requiring the employer to provide maternity leave to female employees. We closely follow the model by Acemoglu and Angrist (2001). This is a standard competitive model with two types of workers: men and women. The objective is to discuss how maternity leave could reduce the level of employment of women by increasing the cost of hiring them. The female labor supply function is given by the function 𝑛𝑛𝑓𝑓(𝑤𝑤𝑓𝑓) and the one for men is given by 𝑛𝑛𝑚𝑚(𝑤𝑤𝑚𝑚), where 𝑤𝑤𝑖𝑖 is the wage 21 These rules are not applicable to employment contracts of very short duration. In another established line of jurisprudence, the Constitutional Court has stated that the right to stability in employment of the pregnant worker is dependent on the existence of a true and grounded prospect to continue on the job on the part of the employee. According to the Court, such certainty cannot be attained in contracts of short duration. Constitutional Court, T206/02. 22 In the United States, at will contracts are those in which the worker can quit at any time or the employer can fire the worker at any time without the need to show cause. A similar type of arrangement in the Colombian context are agreements for the provision of services which do not extend to workers the same protections granted by the labor law regime. 23 Summers (1989) is a seminal contribution in the literature relating the effects of mandates directed to workers as a whole on labor outcomes. An important contribution is found in Jolls (2000), which adapts the Summers framework to the case of accommodation mandates. 13 received by worker type 𝑖𝑖,𝑖𝑖=𝑓𝑓,𝑚𝑚. The functions 𝑛𝑛𝑖𝑖(. ) are increasing in wages. All workers are infinitely lived, risk neutral, and exhibit a discount factor β<1. There are Z firms in the labor market that never exit, and a sufficiently large number of potential firms that could enter if they pay the cost Γ.24 This assumption allows us to characterize a market with free entry of firms (when Z→0) as well as one where the number of firms is fixed (Z>0 y Γ→∞). Every firm is risk neutral and discounts the future at the rate β. Each firm has access to the production function 𝐺𝐺(𝑀𝑀𝑡𝑡,𝑒𝑒∗𝐹𝐹𝑡𝑡), where 𝑀𝑀𝑡𝑡 is the number of male workers at time t, 𝐹𝐹𝑡𝑡 is the number of female workers at time t, and 𝑒𝑒≤1 is the relative efficiency of female workers as perceived by the firm. This characteristic includes the case in which firms discriminate against women because of preferences (taste), as in Becker (1971). The function 𝐺𝐺(. ) exhibits decreasing returns to scale. In each period t, there is a probability s that the productivity of a worker in its current firm falls to zero. These are shocks for the specific combination worker-firm that we call compatibility shocks. Therefore, quantities 𝐹𝐹𝑡𝑡 and 𝑀𝑀𝑡𝑡 in 𝐺𝐺 include only those workers that do not receive the compatibility shock. A female worker who gets fired could sue the firm with probability 𝑞𝑞𝑓𝑓 for compensation that implies for the firm a cost 𝜙𝜙𝑓𝑓. For a male worker, the values are 𝑞𝑞𝑚𝑚 and 𝜙𝜙𝑚𝑚, respectively. Therefore, the expected value of firing a worker is 𝑓𝑓𝑖𝑖=𝑞𝑞𝑖𝑖∗𝜙𝜙𝑖𝑖. We are going to consider the simple case in which the cost 𝑓𝑓𝑖𝑖 is paid by the firm, but it is not received by any other economic agent. We assume that (1 −𝛽𝛽)𝑓𝑓𝑖𝑖<𝑤𝑤𝑖𝑖 so that it is optimal for the firm to fire the fraction s of its employees that receive the negative compatibility shock. Following the current legislation in this economy, firms must provide maternity leave. This leave is given only to female workers who are pregnant and give birth, which occurs with probability δ per female worker. This probability captures information about the percentage of female workers who are fertile, as well as about fertility rates per age.25 The firm has to pay a cost 𝐶𝐶 per female worker who takes maternity leave. This assumption intends to capture the costs of recruiting and training a person to replace the woman on maternity leave, as well as adjustments in organization and production and other costs incurred during the leave period. 24 Z is the minimum number of active firms in the market that would have non-negative benefits in equilibrium, such that the entry cost for a potential firm is higher than the profits if enters. 25 In our empirical exercise, this probability would be determined by the percentage of women between the ages of 18 and 30, with their respective fertility rates, relative to the population of women between 40 and 55 years old with their fertility rates. 14 However, providing maternity leave also generates benefits for the firm. The literature that studies the effects of providing maternity leave on the labor decisions of women find that those who have taken maternity leave are more likely to return to work after the maternity leave period. Retaining an employee who already has specific knowledge about the firm is beneficial to the firm. Furthermore, there is a hypothesis that firms that provide maternity leave are able to attract women who are more qualified and more committed to remaining in the labor market.26 In this model, we capture these benefits by assuming that each female worker (regardless of her pregnancy status) increases the firm’s revenue in the amount 𝐵𝐵. Legislation mandates that employers must provide maternity leave. If it were the case that 𝐶𝐶<𝐵𝐵, firms would provide it voluntarily even in the absence of such legislation. The fact that government regulation is required suggests that in general 𝐶𝐶>𝐵𝐵. The maximization problem for a firm at time t=0 can be written as 𝑚𝑚𝑚𝑚𝑚𝑚[𝐹𝐹𝑡𝑡,𝑀𝑀𝑡𝑡] 𝜋𝜋≡�� 𝛽𝛽𝑡𝑡[𝐺𝐺(𝑀𝑀𝑡𝑡,𝑒𝑒𝐹𝐹𝑡𝑡)−𝑤𝑤𝑚𝑚,𝑡𝑡𝐹𝐹𝑡𝑡−𝑤𝑤ℎ,𝑡𝑡𝑀𝑀𝑡𝑡 −𝛿𝛿𝐶𝐶𝐹𝐹𝑡𝑡+𝐵𝐵𝐹𝐹𝑡𝑡−𝑠𝑠𝑓𝑓𝑚𝑚𝐹𝐹𝑡𝑡−1−𝑠𝑠𝑓𝑓ℎ𝑀𝑀𝑡𝑡−1]� ∞ 𝑡𝑡=0 , where 𝐹𝐹−1 =𝑀𝑀−1= 0. The first line of the maximization problem is revenues minus wage costs. The second line introduces the costs of maternity and of terminating contracts. When 𝐹𝐹𝑡𝑡=𝐹𝐹𝑡𝑡−1 and 𝑀𝑀𝑡𝑡=𝑀𝑀𝑡𝑡−1, the number of workers is stable over time, and the firm hires 𝑠𝑠𝐹𝐹𝑡𝑡−1 women and 𝑠𝑠𝑀𝑀𝑡𝑡−1 men to replace those that got fired in the previous period. Given that costs are linear, and that there is no aggregate uncertainty, firms adjust immediately to steady state levels. For each period, 𝑀𝑀𝑡𝑡=𝑀𝑀,𝐹𝐹𝑡𝑡=𝐹𝐹,𝑤𝑤𝑚𝑚,𝑡𝑡=𝑤𝑤𝑚𝑚, and 𝑤𝑤𝑓𝑓,𝑡𝑡=𝑤𝑤𝑓𝑓. Equilibrium levels of employment and wages must satisfy: 𝜕𝜕𝐺𝐺(𝑀𝑀,𝑒𝑒𝐹𝐹) 𝜕𝜕𝐹𝐹 =𝑤𝑤𝑓𝑓+𝛿𝛿𝐶𝐶−𝐵𝐵+𝛽𝛽𝑠𝑠𝑓𝑓𝑓𝑓 𝜕𝜕𝐺𝐺(𝑀𝑀,𝑒𝑒𝐹𝐹) 𝜕𝜕𝑀𝑀 =𝑤𝑤𝑚𝑚+𝛽𝛽𝑠𝑠𝑓𝑓𝑚𝑚 26 See, among others, Berger and Waldfogel (2004); Desai and Waite (1991); and Leibowitz, Klerman, and Waite (1992). 15 To determine the equilibrium, we impose the condition that the market for men empties: 𝑛𝑛𝑚𝑚 −1(𝑧𝑧𝑀𝑀) = 𝑤𝑤𝑚𝑚 where 𝑧𝑧 is the number of firms in equilibrium. This number is determined by the conditions 𝜋𝜋≤𝛤𝛤 and 𝑧𝑧≥𝑍𝑍, which are satisfied either because profits are equal to entry costs or because there is no entry and the number of firms, 𝑧𝑧, is equal to the minimum, 𝑍𝑍. Wages perceived by women are given by 𝑤𝑤𝑓𝑓=𝑚𝑚𝑚𝑚𝑚𝑚 {𝑛𝑛𝑚𝑚 −1(𝑧𝑧𝑀𝑀), 𝜂𝜂𝑤𝑤𝑚𝑚}, where 𝜂𝜂 is a parameter equal to one if the mandates about equality of wages between men and women are effectively enforced. When there are no restrictions about women’s wages, 𝜂𝜂= 0, so that they are on their supply curve. Most likely, in reality 𝜂𝜂∈(0,1). From the equilibrium conditions, we obtain the following conclusions: 1. Legislation on maternity leave seems to have increased 𝑓𝑓𝑓𝑓 considerably more than 𝑓𝑓𝑚𝑚, first, because the probability of the firm’s being sued when terminating the contract of a pregnant worker increases, and it has to incur costs to prove that the worker was not fired because of her pregnancy; and second, because the legislation increases the amount of compensation if the court rules in her favor. Furthermore, the costs of hiring women increase by 𝛿𝛿𝐶𝐶−𝐵𝐵. Therefore, in reality, it is more likely that legislation on maternity leave decreases women’s employment and wages. 2. The mandate of wage equality between men and women27 (i.e., 𝜂𝜂> 0, and probably very close to 1) could have resulted in women’s wages higher to the one that would equilibrate their market, generating involuntary unemployment of women (they are outside of their supply curve). The mandate of wage equality also interacts with costs of dismissal and of maternity leave by preventing wages from decreasing in order to offset those costs, which has the effect of further decreasing levels of female employment. 3. If, starting from a situation in which 𝑧𝑧>𝑍𝑍 and 𝜋𝜋=𝛤𝛤, legislation results in a decline in profits for the firm, it could cause some firms to exit, thus decreasing employment and wages of both men and women. More generally, the contrast between the cases of free entry and fixed number of firms suggests that legislation further reduces female employment in firms or industries in which profits are already very close to entry costs. These are most likely the smallest firms. The theoretical discussion concludes that the net effect of maternity leave legislation depends on which mandates are more important: maternity leave or equal pay. The costs of 27 In Colombia, Código Sustantivo del Trabajo, Article 14, is a mandate of this sort (equal pay for equal work). 16 maternity leave and the costs of dismissal most likely reduce employment. If the mandate on wage equality is not effectively enforced, the equilibrium would be on the supply curve of both men and women, and the decrease in employment would be accompanied by a decrease in wages for women. In practice, however, the mandates on maternity leave generate involuntary unemployment of women. In this model we assumed that the labor supply curves are given by 𝑛𝑛𝑖𝑖(𝑤𝑤𝑖𝑖) for each type of worker, 𝑖𝑖=𝑚𝑚,𝑓𝑓. In that sense, the initial effect of providing maternity leave is an increase in women’s involuntary unemployment. In a general equilibrium analysis, however, it is very likely that the increase in unemployment decreases women’s incentives to participate in the labor market, shifting the supply curve downward. Therefore, the final effect of the legislation is to increase women’s inactivity rate. 5. Data In this study we use monthly data from the Integrated Household Survey (Gran Encuesta Integrada de Hogares, or GEIH) for the period between January 2009 and September 2013. The survey is conducted by the National Department of Statistics (Departamento Administrativo Nacional de Estadística, or DANE), and it is the main source of information about the labor market in Colombia. This survey provides data on the size and structure of the labor force as well as household and individual characteristics such as gender, education, age, marital status, and others. The baseline period for our analysis is January 2009 through June 2011 as the pre period, and July 2011 through September 2013 as the post period. The population we studied consists of respondents in the 13 metropolitan areas covered by the GEIH.28 In Table 2, we show that our database has information on 1,775,007 individuals for the entire period of analysis. Total observations are reduced to 947,844 when the sample is restricted to women only, and when considering only ages comprising the treatment and control groups, we end up with a sample comprising a total of 411,724 individuals. By selecting only a specific group of individuals, we seek to disentangle the effect of the increase in the maternity leave period on labor outcomes of women in fertile ages. 28 The 13 metropolitan areas are: Barranquilla, Bogotá, Bucaramanga, Medellín, Cali, Cartagena, Cúcuta, Ibagué, Manizales, Montería, Pasto, Pereira, and Villavicencio. 17 In Table 3, we show the distribution of our treatment and control groups for each labor force status. Most of the unemployed (72 percent) belong to the treatment group. Among the employed, most of those who are informal workers or self-employed belong to the control group (56 and 65 percent, respectively). Also, the percentage of inactive is higher for the treatment group (55 percent). In Table 4, we show some descriptive statistics for those women who are in the labor force and those who are inactive, further disaggregated by treatment and control group. Women in the treatment group who are part of the labor force report having education levels higher than primary school in a higher proportion than inactive women in the same group. It is therefore important to control for this variable, since it is expected that education encourages and facilitates active participation in the labor market. Women in the treatment group who are inactive report living with a partner at higher rates than women in the labor force. Finally, women between 18 and 30 years of age who are inactive are more likely to live in households with children under 12 than women of the same age group who are in the labor force. In summary, it can be inferred that inactivity among women in the treatment group was partially explained by individual characteristics—which are used as controls in this study—that are less valued in the labor market. Table 5 shows the general descriptive statistics for the entire sample. The average age is 35, the economic stratum is between 2 and 3, there is one child under 12 on average per household, and on average women have some high school education. 6. Empirical Methodology The legislation on maternity leave should have a greater effect on women in the high-fertility age group than women in the low-fertility age group. This is due to a generalized social perception that a woman in the high-fertility age group is very likely to become pregnant in the near future. Employers would tend to take that perception into consideration when calculating the expected value of hiring a woman in that group. Therefore, for our empirical strategy, the treatment group consists of women between 18 and 30 years of age, and the control group comprises women between 40 and 55 years of age. Thus, to estimate the effect of the increase in the maternity leave period on labor market 18 outcomes, we compare outcome differences between treatment and control groups in the postlegislation period with those in the pre-legislation period. 29 Table 6 shows the differences in fertility rates for women in the treatment and the control groups reported by DANE. On average, during the period analyzed—2009 to 2013—fertility rates for the treatment group hover around 11.5 percent, while those for the control group are approximately 1.18 percent. This difference in fertility rates allows us to have two comparable groups, only one of which is affected by the changes in the law. 30 In order to understand the impact of the extension of the maternity leave period on the group of women in the high-fertility age group, we propose the following empirical model: 𝑦𝑦𝑖𝑖=𝛾𝛾0+𝛾𝛾1𝑡𝑡𝑡𝑡𝑒𝑒𝑚𝑚𝑡𝑡𝑒𝑒𝑡𝑡𝑖𝑖+𝛾𝛾2𝑙𝑙𝑚𝑚𝑤𝑤2011 +𝛾𝛾3𝑡𝑡𝑡𝑡𝑒𝑒𝑚𝑚𝑡𝑡𝑒𝑒𝑡𝑡∗𝑙𝑙𝑚𝑚𝑤𝑤2011 +Γ𝑋𝑋𝑖𝑖+𝜃𝜃𝑡𝑡+𝜀𝜀𝑖𝑖 (1) where 𝒚𝒚𝒊𝒊 are variables such as labor inactivity, unemployment, informality, etc. 𝑡𝑡𝑡𝑡𝑒𝑒𝑚𝑚𝑡𝑡𝑒𝑒𝑡𝑡𝑖𝑖 is a dummy variable that takes the value of 1 if the person is a woman between 18 and 30 years old, and 0 if she is between 40 and 55. 𝑙𝑙𝑚𝑚𝑤𝑤2011 is a variable that takes the value of 1 for all months starting in July 2011, when the legislation on maternity protection was introduced, and controls for common shocks affecting the labor market outcomes of both highand low-fertility women after July 2011. To control for the bias originated by differences in characteristics between the two groups that could explain the differences in participation and employment decisions, we include regressors in the model that allow us to control for observable characteristics and help to solve this problem. In the vector of regressors 𝑿𝑿𝒊𝒊 we include age, age squared, three indicator variables (whether the woman has a high-school education or less, whether she lives with a partner or not, and whether she is the household head), the number of children in the household, the total number of household members, and the household’s economic stratum according to the household’s energy bill. We also control for fixed effects by area of residence, year, and month. It is possible that seasonal shocks affect younger workers differently than older workers. In order to control for that, we include an interaction between month and the indicator of belonging to the 29 The control group includes women 40 and over to ensure that they were not part of the treatment group at any time during the period analyzed. 30 The results of our empirical model are robust to changes in the definition of the high-fertility group. For example, the results hold when the treated group consists of women ages 25 to 30, and 25 to 35. 19 treatment group (treated). All estimates are weighted by the share of area residents age 18 to 65 in the year. We are interested in the coefficient of the interaction 𝛾𝛾3, which indicates whether the legislation considered differentially affected women in the treatment group. With the above econometric model, we want to explore the effects on the labor market of including two more weeks of maternity leave, which is equivalent to an increase of 17 percent in the leave period. We estimate the equations using probit regression analysis, except for wages, for which we use OLS regression. 7. Results In this section we report the results from our estimation exercises. We report the probit (or OLS for wages) coefficients and the corresponding marginal effects for the interaction treated*law2011.31 The marginal effects reported in the main text are calculated for a woman32 in the treatment group who lives in Bogotá on June 2012, does not live with a partner, is not the head of household, and either (i) has more than a high school education, or (ii) her level of education is high school or lower. These two effects plus the remaining six of all other combinations are reported in tables in the Appendices. In our baseline scenario, the treatment group corresponds to women ages 18 to 30, and the control group to women ages 40 to 55. The pretreatment period is January 2009-June 2011, and the post treatment period is July 2011-September 2013. In column (1) of Table 7, we report the results when analyzing the probability of inactivity. The dependent variable is a dummy that takes the value of 1 if the person indicates that she is not in the labor force, and 0 otherwise.33 The results show that the probability of 31 The calculation and interpretation of marginal effects for interactions in non-linear models must take into account the cross-derivatives of the predicted probabilities. See Ai and Norton (2003) and Norton, Wang, and Ai (2004) for a discussion of this issue. 32 The marginal effects are estimated for a woman in the treatment group, using sample means, whose age is 23.87, who lives in a household composed of 4.51 members and 1.07 children, whose economic stratum is 2.27, and who lives in Bogotá on June 2012. The woman could be either living with a partner or not, be the household head or not, and have high or low educational attainment. The combination of all these possibilities gives us a total of eight marginal effects. 33 A person is classified as inactive if on the survey she or he replies in the affirmative to at least one of the following six statements: 1. Handicapped. 2. Doesn’t want to get paid work or set up a business. 3. Wants to work, but has not made steps to search for a job or set up a business because: a. Self-reported as very young / old for work. b. Family responsibilities. c. Health problems. d. Full-time student. e. Other. 4. After his last job he or she hasn’t 20 inactivity significantly increases for women in the high-fertility age group after the increase in the maternity-leave period. The marginal effects estimation, in Table 7a, 34 indicates that for a woman with less than a high school education, the probability of inactivity increases by 0.9 percentage points. If the same woman had more than a secondary school education, her probability of inactivity increases by 0.7 percentage points. In general, all other things being equal, the increase in the probability of inactivity is greater for women (i) with low educational attainment, (ii) living with a partner, and (iii) who are not heads of household. We interpret these results as support to our hypothesis, suggested by the model presented in Section 3, that employers are less willing to hire women of high-fertility age since the law went into effect. Because it is harder for women of childbearing age to find a job, they rationally decide not to participate in the labor market: the probability of being inactive increases, even though they are willing and able to work. Column (2) of Table 7 reports the results for the probability of unemployment. There is no evidence that the increase in the maternity leave period had affected the probability of unemployment for women in the high-fertility age group relative to women in the low-fertility age group. None of the marginal effects reported in Table 7a is significantly different from zero. Column (3) of Table 7 reports the results for the probability of informality. Informal workers are defined as those workers who do not satisfy one of two conditions: (i) contributing to health insurance system; or (ii) contributing to a pension plan. It shows that the increase in the maternity-leave period resulted in a significant increase in the probability of informality for women in the high-fertility age group relative to women in the low-fertility age group. The marginal effects estimation, shown in Table 7a, indicates that the probability of informality increases by 0.8 percentage points for a woman with more than a high school education and by 0.6 percentage points for a woman with low levels of education. Ceteris paribus, the probability of informality is greater for women who are more educated and who are not living with a partner. The results with respect to the effect on the probability of self-employment are reported in column (4) of Table 7 and Table 7a. We conclude that there is a significant increase in the probability of self-employment for women in the high-fertility age group after the increase in the taken any action to find a job or set up a business. 5. During the last 12 months has not done anything to find work or set up a business. 6. He or she was not available for work. 34 These and the remaining six marginal effects are reported in Appendix A. In Appendix B, we report the estimated marginal effects for the 13 metropolitan areas. All of the results are quantitatively and qualitatively similar. 21 9. Conclusions Our research explores the impact of Law 1468 of 2011, which extended the maternity leave period from 12 to 14 weeks (a 17 percent increase), on female labor outcomes. Our results show that the Law increases the probability of being inactive for women ages 18 to 30 (the treatment group) relative to women ages 40 to 55 (the control group). We also show that the probability of informality and self-employment increases for high-fertility women relative to low-fertility women. Our results are robust across demographic groups and time periods, suggesting a causal effect of the increase in the maternity leave period. Following Autor et al. (2006), our paper does not attempt to provide an overall assessment of maternity protection laws. The fact that there are some effects on the labor market for women of childbearing age indicates that legal protections come with a cost. Therefore, the law must be tied to other regulations that prevent employers from excluding the beneficiary group from the labor market. Our main recommendation is to design a “parental leave” policy that can be enjoyed by both parents in equal or very similar proportions, and to socialize the extra cost of social security of the worker who is replacing the woman on maternity leave, which is now assumed by the employer. 28 References Acemoglu, D. and J. D. Angrist. 2001. “Consequences of Employment Protection? The Case of the Americans with Disabilities Act.” Journal of Political Economy 109(5): 915–57. Ai, C. R. and E. C. Norton. 2003. “Interaction Terms in Logit and Probit Models.” Economics Letters 80(1): 123–29. Amador, D., R. Bernal, and X. Peña. 2013. “El aumento en la participación laboral femenina en Colombia: ¿Fertilidad, estado civil o educación?” Ensayos sobre política económica 31(71) Bogotá, Colombia. Autor, D. H., J. J. Donohue, III, and S. J. Schwab. 2006. “The Costs of Wrongful-Discharge Laws.” Review of Economics and Statistics 88(2-May): 211–31. Badel, A., and X. Peña. 2010. “Decomposing the Gender Wage Gap with Sample Selection Adjustment: Evidence from Colombia.” Economic Analysis Review 2(25): 169–91. Baker, M. and K. Milligan. 2008. “Maternal Employment, Breastfeeding, and Health: Evidence from Maternity Leave Mandates.” Journal of Health Economics 27(4): 871–87. Blau, F. and L. Kahn. 2013. “Female Labor Supply: Why is the US Falling Behind?” The American Economic Review 103(3-May): 251–256. Baum, C. 2003. “The Effect of State Maternity Leave legislation and the 1993 Family and Medical Leave Act on Employment and Wages” Labor Economics 10: 573–96. Becker, G. 1971. The Economics of Discrimination. Second Edition. Chicago, IL: University of Chicago Press. Berger, L. M. and J. Waldfogel. 2004. “Maternity Leave and the Employment of New Mothers in the United States.” Journal of Population Economics 17(2): 331–49. Correll, S., and S. Bernard. 2007. “Getting a Job: Is There a Motherhood Penalty?” American Journal of Sociology 112(5): 1297–1339. Corse, S. J. 1990. “Pregnant Managers and their Subordinates: The Effects of Gender Expectations on Hierarchical Relationships.” The Journal of Applied Behavioral Science. 26(1): 25–47. Cuddy, A., S. Fiske, and P. Glick. 2004. “When Professionals become Mothers, Warmth Doesn’t Cut the Ice.” Journal of Social Issues 60: 701-18. 29 Desai S., and L. Waite. 1991. “Women's Employment during Pregnancy and After the FirstBirth: Occupational Characteristics and Work Commitment.” American Sociological Review 56(4): 551–66. Ekberg, J., R. Eriksson, and G. Friebel. 2013. “Parental Leave—A Policy Evaluation of the Swedish ‘Daddy-Month’ Reform.” Journal of Public Economics 97: 131-143. England, P. and N. Folbre. 1999. “The Cost of Caring.” The ANNALS of the American Academy of Political and Social Science 1999 561:39, January. Espino, A. and S. Salvador. 2014. “Un análisis de género de los costos laborales en Colombia.” Consultancy for the Ministerio del Trabajo de Colombia (Colombian Ministry of Labor). Mimeographed document. Folbre, N. and T. Weisskopf. 1998. “Did Father Know Best? Families, Markets, and the Supply of Caring Labor.” In Ben-Ner, A. and L Putterman (Eds.). Economics, Values, and Organization. New York, NY: Cambridge University Press. Forero de Saade, M., L. Cañón, L., and J. A. Pineda. 1991. “Participación de la mujer en el trabajo.” Mujer Trabajadora. Nuevo Compromiso Social. Bogotá, Colombia: Estudios Sociales Juan Pablo II. Gornick, J. and M. Meyers. 2003. Families that Work. Policies for Reconciling Parenthood and Employment. New York, NY: Russell Sage Foundation. Gruber, J. 1994. “The Incidence of Mandated Maternity Benefits.” American Economic Review, 84 (3): 622–41. Halpert, J., M. Wilson, and J. Hickman. 1993. “Pregnancy as a Source of Bias in Performance Appraisals.” Journal of Organizational Behavior 14: 649–63. Hashimoto, M., R. Percy, T. Shoellner, and B. Weinberg. 2004. “The Long and Short of It: Maternity Leave Coverage and Women’s Labor Market Outcomes.” IZA Discussion Paper Series, No. 1207. ILO (International Labour Organisation). 1997. “Maternity Protection at Work”. Available at: http://www.ilo.org/public/english/standards/relm/ilc/ilc87/rep-v-1.htm ----. 2005. “Maternity at Work: A Review of National Legislation. Findings from the ILO’s Conditions of Work and Employment Database”. ----. 2010. “Maternity at Work.” 30 Available at: http://www.ilo.org/global/publications/books/WCMS_124442/lang-- en/index.htm Jolls, C. 2000. “Accommodation Mandates.” Stanford Law Review, 53(2): 223–306. Jolls, C. and J. J. Prescott. 2004. “Disaggregating Employment Protection: The Case of Disability Discrimination.” NBER Working Paper 10740, September. Cambridge, MA: National Bureau of Economic Research. Junguito, R. et al. 1970. “Análisis de la Estructura y Evolución de la fuerza de trabajo colombiana 1938, 1951, 1964.” Universidad de los Andes, Bogotá, Colombia. Documentos CEDE. Kamerman, S. 2000. “From Maternity to Parental Leave Policies: Women’s Health, Employment and Child and Family Well-being.” Journal of the American Medical Women’s Association 55(2): 96–99. Lai, Y., and S. Masters. 2005. “The Effects of Mandatory Maternity and Pregnancy Benefits on Women’s Wages and Employment in Taiwan, 1984-1996.” Ithaca, NY: Industrial and Labor Relations School, Cornell University. Industrial and Labor Relations Review 58(2January): 274–81. Leibowitz A., J. A. Klerman, and L. Waite. 1992. “Employment of New Mothers and Child Care Choice: Differences by Children’s Age.” Journal of Human Resources 27(1): 112–133. Molinos, C. 2012. “La Ley de protección a la maternidad como incentivo de participación laboral femenina: el caso colombiano.” Coyuntura Económica XLII(1): 93–116. Bogotá, Colombia: Fedesarrollo. Norton, E. C., H. Wang, and C. Ai. 2004. “Computing Interaction Effects and Standard Errors in Logit and Probit Models.” The Stata Journal 4(2): 154–67. Olarte, L. and X. Peña. 2010. “El efecto de la maternidad sobre los salarios femeninos.” Ensayos sobre Política Económica 28(63): 190–231. Peña, X., et al. 2013. “Mujer y Movilidad Social.” Documentos CEDE 2013-05. Peña-Parga, X., and Glassman, A. 2004. “Demand for Child Care and Female Employment in Colombia.” Documentos CEDE 2004-43. Ramírez, N. 2008. “¿Poder o desventaja? El derecho de las mujeres a no ser despedidas durante el embarazo.” Tesis de grado de maestría, Universidad de los Andes, Bogotá, Colombia. On file with the author. 31 Rossin-Slater, M., C. Ruhm, and J. Waldfogel. 2013. “The Effects of California’s Paid Family Leave Program on Mother’s Leave Taking and Subsequent Labor Market Outcomes.” Journal of Policy Analysis and Management 32(2): 224–45. Ruhm, C. J. 1998. “The Economic Consequences of Parental Leave Mandates: Lessons from Europe.” Quarterly Journal of Economics 113(1): 295–317. Schönberg, U. and J. Ludsteck. 2007. “Maternity Leave Legislation, Female Labor Supply, and the Family Wage Gap.” IZA Discussion Papers No. 1699. http://nbnresolving.de/urn:nbn:de:101:1-200804110188 Summers, L. 1989. “Some Simple Economics of Mandated Benefits.” American Economic Review, Papers and Proceedings 79(2-May): 177–83. 32 Table 1. Labor Cost Associated with Maternity Leave Take Labor Cost Percentage Nominal wage 73.06 Cost of a replacement worker 30.31 Nursing license 3.37 Total Cost 106.73 Note: Annual labor costs assumed by the employer for each female worker with maternity leave and nursing leave (Values refer to a basic salary of 100). Adapted by authors from Espino and Salvador (2014). Table 2. Sample Reduction Sample Total Observations 2009-2013 Total data 1,775,007 Only women 947,844 By age group: 18-30 214,237 40-55 197,487 Note: Source: DANE and calculations by the authors. 33 Table 3. Distribution of Labor Market Variables between the Treatment and Control Groups. Period 2009-2013 From 18 to 30 years of age From 40 to 55 years of age Labor force (%) 50.91 49.09 Employed 46.53 53.47 Informal 44.07 55.93 self-employed 34.34 65.66 Unemployed 72.34 27.66 Inactivity (%) 54.79 45.21 lack of job search activity 31.52 68.48 Note: Source: DANE and calculations by the author. These calculations are weighted by expansion factors. Table 4. Characteristics of the Treatment and Control Groups in the Labor Force and in Inactivity. Period 2009-2013 Inactivity (%) Labor Force (%) 18 - 30 40 - 55 18 – 30 40 - 55 Education None 3.95 9.40 0.51 3.95 Primary 10.70 36.94 7.21 27.89 Secondary Education 78.85 45.05 69.09 41.80 Higher education 6.51 8.60 23.22 26.36 Economic strata 1 30.12 21.59 23.52 19.95 2 36.37 38.16 40.02 36.55 3 23.05 29.65 27.10 29.48 4 6.70 6.96 6.26 8.74 5 2.57 2.37 2.10 3.38 6 1.18 1.28 1.00 1.90 Marital status Not living with a partner 54.03 29.28 62.43 46.38 Living with a partner 45.97 70.72 37.57 53.62 Number of children None 35.69 54.05 37.33 56.44 1 or 2 children 52.99 40.81 53.72 39.61 3 or more children 11.33 5.14 8.95 3.95 Note: These calculations are weighted by expansion factors. Source: DANE and calculations by the authors. 34 Table 5. Descriptive Statistics of Control Variables Period 2009-2013 Variable Average Standard deviation Interval Years of education 10.16 4.13 0 – 26 Age 34.96 12.35 18 – 55 Stratum 2.32 1.07 1 – 6 Children <12 per household 0.86 1.04 0 – 12 Total people in the household 4.32 1.98 1 – 22 Source: DANE and calculations by the authors. Table 6. Fertility Rates by Age Period Age Groups Global Rate 20-24 25-29 30-34 35-39 40-44 45-49 1985-1990 0.17 0.15 0.12 0.08 0.03 0.00 3.34 1990-1995 0.16 0.14 0.11 0.07 0.03 0.00 3.14 1995-2000 0.15 0.13 0.10 0.06 0.02 0.00 2.86 2000-2005 0.14 0.12 0.09 0.05 0.02 0.00 2.60 2005-2010 0.13 0.11 0.08 0.05 0.02 0.01 2.45 2010-2015 0.12 0.11 0.08 0.05 0.02 0.01 2.35 Average 2005-2015 0.12 0.11 0.08 0.05 0.02 0.01 2.40 Source: DANE and calculations by the authors. 35 Table 7. Effect of Law 1468 of 2011 on Labor Market Outcomes for Women Baseline Scenario and Alternative Timings Dependent variable (1) Inactive (2) Unemployment (3) Informality (4) Selfemployment (5) Log real wages treated * Law2011 0.028*** (0.008) -.014 (0.012) 0.022* (0.011) 0.018* (0.011) -.005 (0.007) Panel a. Eliminating two months before and two months after implementation of the law treated * Law2011 0.029*** (0.009) -.014 (0.012) 0.022* (0.011) 0.022* (0.011) -.005 (0.007) Panel b. Eliminating two months before and four months after implementation of the law treated * Law2011 0.033*** (0.009) -.019 (0.012) 0.020* (0.012) 0.021* (0.011) -.004 (0.007) Panel c. Eliminating two months before and six months after implementation of the law treated * Law2011 0.039*** (0.009) -.018 (0.012) 0.022* (0.012) 0.022* (0.011) -.004 (0.008) R2 0.080 0.073 0.170 0.068 0.348 Observations 409.055 290.662 240.285 241.409 127.780 Controls: Personal characteristics Yes Yes Yes Yes Yes Household characteristics Yes Yes Yes Yes Yes Time-fixed effects Yes Yes Yes Yes Yes Note: The coefficient on treated*Law2011 is the estimated parameter 𝛾𝛾3 of equation (1) which is the DD estimate of the effect of the reform in each of the outcomes. Columns (1) to (4) are probit estimates, column (5) is OLS estimate. ***Coefficients are significant at the 1% level. **Coefficients are significant at the 5% level. *Coefficients are significant at the 10% level. Standard errors are in parentheses. 36 Table 7a. Marginal Effect of Law 1468 of 2011 on Labor Market Outcomes for Women Baseline Scenario and Alternative Timings (1) Inactive (2) Unemployment (3) Informality (4) Self-employment Marginal Effect Low Education High Education Low Education High Education Low Education High Education Low Education High Education treated * Law2011 0.009*** (0.002) 0.007*** (0.002) -.004 (0.003) -.004 (0.003) 0.006* (0.003) 0.008* (0.004) 0.005* (0.003) 0.004* (0.002) Panel A. Eliminating two months before and two months after implementation of the law treated * Law2011 0.009*** (0.002) 0.007*** (0.002) -.004 (0.003) -.004 (0.003) 0.006* (0.003) 0.008* (0.004) 0.006* (0.003) 0.005* (0.002) Panel B. Eliminating two months before and four months after implementation of the law treated * Law2011 0.010*** (0.002) 0.008*** (0.002) -.005 (0.003) -.005 (0.003) 0.005* (0.003) 0.007* (0.004) 0.006* (0.003) 0.004* (0.002) Panel C. Eliminating two months before and six months after implementation of the law treated * Law2011 0.012*** (0.002) 0.010*** (0.002) -.005 (0.003) -.005 (0.003) 0.006* (0.003) 0.008* (0.004) 0.006* (0.003) 0.005* (0.002) Note: The marginal effects are estimated for a woman in the treatment group for whom, using sample means, age is 23.87, lives in a household composed of 4.51 members and 1.07 children, whose economic stratum is 2.27, does not live with a partner, is not head of household and lives in Bogotá on June 2012.The marginal effects were estimated for two scenarios described by her educational level. “Education level” is high or low according with the years of education: the individual is considered to have a high education level if she has more than 11 years of education. ***Coefficients are significant at the 1% level. **Coefficients are significant at the 5% level. *Coefficients are significant at the 10% level. Standard errors are in parentheses. 37 Appendix A. Marginal Effect of Law 1468 of 2011 on Labor Market Outcomes for Women Baseline Scenario Marginal Effect (1) Inactive (2) Unemployment (3) Informality (4) SelfEmployment Education Level Marital Status Head of Household High 0 0 0.007*** (0.002) -.004 (0.003) 0.008* (0.004) 0.004* (0.002) High 0 1 0.005*** (0.001) -.003 (0.003) 0.008* (0.004) 0.005* (0.002) High 1 0 0.009*** (0.002) -.004 (0.003) 0.008* (0.004) 0.005* (0.003) High 1 1 0.008*** (0.002) -.003 (0.003) 0.008* (0.004) 0.005* (0.003) Low 0 0 0.009*** (0.002) -.004 (0.003) 0.006* (0.003) 0.005* (0.003) Low 0 1 0.007*** (0.002) -.003 (0.003) 0.006* (0.003) 0.006* (0.003) Low 1 0 0.010*** (0.003) -.004 (0.003) 0.005* (0.002) 0.006* (0.003) Low 1 1 0.009*** (0.002) -.003 (0.003) 0.005* (0.002) 0.006* (0.004) Note: The marginal effects are estimated for a woman in the treatment group for whom, using sample means, age is 23.87, lives in a household composed of 4.51 members and 1.07 children, whose economic stratum is 2.27, and lives in Bogotá on June 2012.The marginal effects were estimated for several scenarios described by the combination of three dummy variables: “Education level” is high or low according with the years of education: the individual is considered to have a high education level if she has more than 11 years of education; “Marital status” takes the value of 1 when the individual lives with a partner and 0 otherwise; finally, the variable “Head of household” takes the value of 1 when the individual is the head of the household and 0 otherwise. ***Coefficients are significant at the 1% level. **Coefficients are significant at the 5% level. *Coefficients are significant at the 10% level. Standard errors are in parentheses. 44 Appendix B. Marginal Effect of Law 1468 of 2011 on Labor Market Outcomes for Women All Metropolitan Areas (1) Inactive (2) Unemployment (3) Informality (4) Self-Employment Marginal Effect Low Education High Education Low Education High Education Low Education High Education Low Education High Education Barranquilla 0.010*** (0.003) 0.010*** (0.003) -.004 (0.003) -.004 (0.003) 0.003* (0.002) 0.007* (0.004) 0.006* (0.003) 0.005* (0.003) Bogotá 0.009*** (0.002) 0.008*** (0.002) -.004 (0.003) -.004 (0.003) 0.005* (0.003) 0.008* (0.004) 0.006* (0.002) 0.004* (0.002) Cartagena 0.010*** (0.003) 0.010*** (0.003) -.004 (0.003) -.004 (0.003) 0.004* (0.002) 0.008* (0.004) 0.007* (0.003) 0.006* (0.003) Manizales 0.010*** (0.003) 0.009*** (0.003) -.004 (0.004) -.004 (0.004) 0.006* (0.003) 0.008* (0.004) 0.005* (0.002) 0.004* (0.002) Montería 0.010*** (0.003) 0.008*** (0.002) -.004 (0.003) -.004 (0.003) 0.003* (0.002) 0.007* (0.004) 0.006* (0.003) 0.005* (0.003) Villavicencio 0.010*** (0.003) 0.009*** (0.002) -.004 (0.003) -.004 (0.003) 0.004* (0.002) 0.008* (0.004) 0.006* (0.003) 0.005* (0.003) Pasto 0.010*** (0.003) 0.009*** (0.002) -.004 (0.003) -.004 (0.003) 0.003* (0.001) 0.007* (0.004) 0.006* (0.003) 0.005* (0.003) Cúcuta 0.010*** (0.003) 0.009*** (0.002) -.004 (0.004) -.004 (0.003) 0.003* (0.001) 0.007* (0.004) 0.007* (0.003) 0.006* (0.003) Pereira 0.010*** (0.003) 0.009*** (0.002) -.005 (0.004) -.005 (0.004) 0.005* (0.002) 0.008* (0.004) 0.006* (0.003) 0.005* (0.003) Bucaramanga 0.009*** (0.002) 0.007*** (0.002) -.004 (0.003) -.004 (0.003) 0.004* (0.002) 0.008* (0.004) 0.007* (0.003) 0.006* (0.003) Ibagué 0.009*** (0.002) 0.007*** (0.002) -.005 (0.004) -.004 (0.004) 0.004* (0.002) 0.008* (0.004) 0.006* (0.003) 0.005* (0.003) Cali 0.010*** (0.003) 0.008*** (0.002) -.004 (0.004) -.004 (0.003) 0.005* (0.002) 0.008* (0.004) 0.006* (0.003) 0.005* (0.003) Note: The marginal effects for each metropolitan area are estimated for a woman in the treatment group for whom, using sample means, age is 23.87, lives in a household composed of 4.51 members and 1.07 children, whose economic stratum is 2.27, on June 2012, and either the women has (i) a high level of education (12 years or more), or (ii) a low level of education (11 years of schooling or less). ***Coefficients are significant at the 1% level. **Coefficients are significant at the 5% level. *Coefficients are significant at the 10% level. Standard errors are in parentheses. 45 Appendix C. Marginal Effect of Law 1468 of 2011 on Labor Market Outcomes for Women Alternative Timings Marginal Effect (1) Inactive (2) Unemployment (3) Informality (4) SelfEmployment Education Level Marital Status Head of Household Panel A. Eliminating two months before and two months after implementation of the law High 0 0 0.007*** (0.002) -.004 (0.003) 0.008* (0.004) 0.005* (0.002) High 0 1 0.005*** (0.001) -.003 (0.003) 0.008* (0.004) 0.005* (0.003) High 1 0 0.009*** (0.003) -.004 (0.003) 0.008* (0.004) 0.006* (0.003) High 1 1 0.008*** (0.002) -.003 (0.003) 0.008* (0.004) 0.006* (0.003) Low 0 0 0.009*** (0.002) -.004 (0.003) 0.006* (0.003) 0.006* (0.003) Low 0 1 0.007*** (0.002) -.003 (0.003) 0.006* (0.003) 0.007* (0.003) Low 1 0 0.011*** (0.003) -.004 (0.003) 0.005* (0.003) 0.007* (0.003) Low 1 1 0.010*** (0.003) -.003 (0.003) 0.005* (0.003) 0.008* (0.004) Panel B. Eliminating two months before and four months after implementation of the law High 0 0 0.008*** (0.002) -.005 (0.003) 0.007* (0.004) 0.004* (0.002) High 0 1 0.006*** (0.001) -.005 (0.003) 0.007* (0.004) 0.005* (0.003) High 1 0 0.011*** (0.003) -.005 (0.003) 0.007* (0.004) 0.005* (0.003) High 1 1 0.009*** (0.002) -.004 (0.003) 0.007* (0.004) 0.006* (0.003) Low 0 0 0.010*** (0.002) -.005 (0.003) 0.005* (0.003) 0.006* (0.003) Low 0 1 0.008*** (0.002) -.005 (0.003) 0.005* (0.003) 0.007* (0.003) Low 1 0 0.012*** (0.003) -.005 (0.003) 0.005* (0.003) 0.007* (0.004) Low 1 1 0.011*** (0.003) -.005 (0.003) 0.005* (0.003) 0.007* (0.004) Panel C. Eliminating two months before and six months after implementation of the law High 0 0 0.010*** (0.002) -.005 (0.003) 0.008* (0.004) 0.005* (0.002) High 0 1 0.007*** (0.001) -.004 (0.003) 0.008* (0.004) 0.005* (0.003) High 1 0 0.013*** (0.003) -.005 (0.003) 0.008* (0.004) 0.006* (0.003) High 1 1 0.011*** (0.002) -.004 (0.003) 0.008* (0.004) 0.006* (0.003) Low 0 0 0.012*** (0.002) -.005 (0.003) 0.006* (0.003) 0.006* (0.003) Low 0 1 0.010*** (0.002) -.005 (0.003) 0.006* (0.003) 0.007* (0.004) Low 1 0 0.014*** (0.003) -.005 (0.003) 0.005* (0.003) 0.007* (0.004) Low 1 1 0.013*** (0.003) -.004 (0.003) 0.005* (0.003) 0.008* (0.004) Note: The marginal effects are estimated for a woman in the treatment group for whom, using sample means, age is 23.87, lives in a household composed of 4.51 members and 1.07 children, whose economic stratum is 2.27, and lives in Bogotá on June 2012.The marginal effects were estimated for several scenarios described by the combination of three dummy variables: “Education level” is high or low according with the years of education: the individual is considered to have a high education level if she has more than 11 years of education; “Marital status” takes the value of 1 when the individual lives with a partner and 0 otherwise; finally, the variable “Head of household” takes the value of 1 when the individual is the head of the household and 0 otherwise.. ***Coefficients are significant at the 1% level. **Coefficients are significant at the 5% level. *Coefficients are significant at the 10% level. Standard errors are in parentheses. 46 Appendix D. Marginal Effect of Law 1468 of 2011 on Labor Market Outcomes for Men Marginal Effect (1) Inactive (2) Unemployment (3) Informality (4) SelfEmployment Education Level Marital Status Head of Household High 0 0 -.002 (0.002) 0.004 (0.003) 0.009** (0.003) -.001 (0.002) High 0 1 -.002 (0.001) 0.003 (0.002) 0.009** (0.003) -.001 (0.002) High 1 0 -.001 (0.002) 0.003 (0.002) 0.009** (0.003) -.001 (0.002) High 1 1 -.001 (0.002) 0.002 (0.001) 0.008** (0.003) -.001 (0.002) Low 0 0 -.002 (0.002) 0.004 (0.003) 0.007** (0.003) -.001 (0.003) Low 0 1 -.001 (0.002) 0.003 (0.002) 0.008** (0.003) -.001 (0.003) Low 1 0 -.001 (0.003) 0.003 (0.002) 0.008** (0.003) -.001 (0.003) Low 1 1 -.001 (0.002) 0.002 (0.001) 0.009** (0.003) -.001 (0.003) Note: The marginal effects are estimated for a man in the treatment group for whom, using sample means, age is 23.79, lives in a household composed of 4.48 members and 0.77 children, whose economic stratum is 2.24, and lives in Bogotá on June 2012.The marginal effects were estimated for several scenarios described by the combination of three dummy variables: “Education level” is high or low according with the years of education: the individual is considered to have a high education level if he has more than 11 years of education; “Marital status” takes the value of 1 when the individual lives with a partner and 0 otherwise; finally, the variable “Head of household” takes the value of 1 when the individual is the head of the household and 0 otherwise. ***Coefficients are significant at the 1% level. **Coefficients are significant at the 5% level. *Coefficients are significant at the 10% level. Standard errors are in parentheses. 47 Appendix E. Marginal Effect on Labor Market Outcomes for Women Placebo Experiment Marginal Effect (1) Inactive (2) Unemployment (3) Informality (4) SelfEmployment Education Level Marital Status Head of Household High 0 0 0.006* (0.003) -.008 (0.005) 0.002 (0.006) 0.001 (0.004) High 0 1 0.005* (0.002) -.007 (0.005) 0.002 (0.006) 0.001 (0.004) High 1 0 0.008* (0.004) -.008 (0.005) 0.002 (0.006) 0.001 (0.004) High 1 1 0.007* (0.004) -.007 (0.004) 0.002 (0.006) 0.002 (0.005) Low 0 0 0.007* (0.004) -.008 (0.005) 0.001 (0.005) 0.002 (0.005) Low 0 1 0.006* (0.003) -.007 (0.005) 0.001 (0.004) 0.002 (0.005) Low 1 0 0.009* (0.005) -.008 (0.005) 0.001 (0.004) 0.002 (0.006) Low 1 1 0.008* (0.004) -.007 (0.005) 0.001 (0.004) 0.002 (0.006) Note: The placebo pretreatment period is January-December 2009. The placebo post treatment period is JanuaryDecember 2010. The marginal effects are estimated for a woman in the treatment group for whom, using sample means, age is 23.87, lives in a household composed of 4.51 members and 1.07 children, whose economic stratum is 2.27, and lives in Bogotá on June 2010.The marginal effects were estimated for several scenarios described by the combination of three dummy variables: “Education level” is high or low according with the years of education: the individual is considered to have a high education level if she has more than 11 years of education; “Marital status” takes the value of 1 when the individual lives with a partner and 0 otherwise; finally, the variable “Head of household” takes the value of 1 when the individual is the head of the household and 0 otherwise. ***Coefficients are significant at the 1% level. **Coefficients are significant at the 5% level. *Coefficients are significant at the 10% level. Standard errors are in parentheses. 48 Appendix F. Marginal Effect of Law 1468 of 2011 on Labor Market Outcomes for Women Control Group Are Men Ages 18-30 Marginal Effect (1) Inactive (2) Unemployment (3) Informality (4) SelfEmployment Education Level Marital Status Head of Household High 0 0 -.001 (0.001) 0.004* (0.002) 0.013*** (0.003) 0.005* (0.003) High 0 1 -.0008 (0.000) 0.002* (0.001) 0.012*** (0.003) 0.005* (0.002) High 1 0 -.002 (0.002) 0.003* (0.002) 0.012*** (0.003) 0.006* (0.003) High 1 1 -.001 (0.001) 0.002* (0.001) 0.011*** (0.003) 0.006* (0.003) Low 0 0 -.001 (0.001) 0.004* (0.002) 0.014*** (0.004) 0.006* (0.003) Low 0 1 -.0007 (0.000) 0.002* (0.001) 0.014*** (0.004) 0.006* (0.003) Low 1 0 -.001 (0.002) 0.003* (0.002) 0.014*** (0.004) 0.006* (0.003) Low 1 1 -.001 (0.001) 0.002* (0.001) 0.014*** (0.004) 0.006* (0.004) Note: The marginal effects are estimated for a woman in the treatment group for whom, using sample means, age is 23.87, lives in a household composed of 4.51 members and 1.07 children, whose economic stratum is 2.27, and lives in Bogotá on June 2012. The marginal effects were estimated for several scenarios described by the combination of three dummy variables: “Education level” is high or low according with the years of education: the individual is considered to have a high education level if she has more than 11 years of education; “Marital status” takes the value of 1 when the individual lives with a partner and 0 otherwise; finally, the variable “Head of household” takes the value of 1 when the individual is the head of the household and 0 otherwise. ***Coefficients are significant at the 1% level. **Coefficients are significant at the 5% level. *Coefficients are significant at the 10% level. Standard errors are in parentheses. 49