WOMEN IN THE WORKFORCE IN PESHAWAR Pakistan Gender and Social Inclusion Platform and the Pakistan Poverty and Equity Program JUNE 2021 Pakistan’s female labor force participation (FLFP) remains low by regional and global standards, and major dis- parities are observed between rural and urban FLFP, with the latter being significantly lower. The World Bank’s multimethod Women in the Workforce study investigates the patterns and challenges that govern urban FLFP. This Note analyzes women’s labor market outcomes as reported by women in Peshawar city in the Khyber Pakh- tunkhwa province, using data collected as part of the World Bank’s Peshawar Urban Household Survey (PUHS). This Note was developed as a collaborative effort between the World Bank’s Pakistan Gender and Social Inclusion Platform and the Pakistan Poverty and Equity Program. Further efforts under the Women in the Workforce study will present FLFP findings from other urban cities in Pakistan. 1 Introduction barriers to women’s work (Section 4)? What are the char- acteristics and quality of the jobs held by women, and to Female labor force participation (FLFP) in urban Pa- what extent do they differ from men’s jobs (Section 5)? kistan stands out in more ways than one. Not only is it And, finally, what can be done to promote greater FLFP among the lowest in the world (International Labour Or- (Section 6)? ganization [ILO] 2020), but it is also remarkably stagnant, having hovered around 10 percent for at least 20 years 2  Measuring Women’s Work (Cho and Majoka 2020; World Bank 2018). The World Bank’s Women in the Workforce study has deployed a Among the many questions raised by the available multimethod approach to gain a nuanced understanding estimates of FLFP in urban Pakistan is whether they of the constraints to women’s work in this context. should be taken at face value. The literature warns of potential downward biases affecting the measurement of The qualitative component of the Women in the Work- FLFP, especially in low-income countries. One of the fac- force study has analyzed the labor market experiences of tors that may contribute to an underestimation of FLFP women in Quetta, Peshawar, Lahore, and Karachi (World is the widespread use of proxy respondents in household Bank 2019). Those findings have helped design the Pe- survey labor modules.1 The male household head, who shawar Urban Household Survey (PUHS), a multipurpose usually reports for other household members, may not household survey that collected information regarding a be adequately informed about women’s economic activ- range of themes, such as living conditions, labor market ities, or he may fail to report them due to implicit bias participation and economic empowerment, time use, ex- regarding women’s work. A typical example is a woman periences with harassment in public, individual aspira- who is unpaid but is supporting a family business man- tions and values, and many others (see Appendix 1 for aged by a male household member. Empirical evidence more details). The aim of this report is to present the find- on this type of measurement error is mixed and still lim- ings of the PUHS, addressing the following questions: Is ited in its coverage of different cultural contexts (Bardasi FLFP in urban Pakistan truly as low as it appears (Section et al. 2011; Benes and Walsh 2018; Desiere and Costa 2)? Why does FLFP remain low (Section 3)? What are the 2019). In general, questionnaires that directly elicit, in 1 Most household and labor force surveys do not expressly require each household member to answer directly for him/herself. Given the time constraints and difficulties of having all members present at the moment of the interview, the questionnaire is generally administered to one (maximum two) respondent(s), which in most cases corresponds to the household head (spouse). 1 separate questions, all possible forms of labor market for family use, increases FLFP by a small margin. Ex- engagement that constitute employment are found to tending the concept of employment to include people yield a more precise measurement of FLFP. But the dis- engaged in the production of agricultural goods for their cussion of measurement issues within the literature on own consumption (subsistence agriculture) generates women’s work goes far beyond data collection and into a more comprehensive estimate of labor force partic- the definition of work itself. Even when recorded with- ipation (LFP). The resulting differences with the stan- out error, the standard concept of work, which focuses dard concept are clearly gendered: Men’s LFP does not on the production of goods and services primarily for the change, while women’s LFP increases to 15.5 percent market (International Conference of Labour Statisticians (Figure 1). The small size of the difference is justified by 2013), leaves out many productive activities that are typ- the urban context—nonetheless, the difference is statis- ically carried out by women in the production of goods for tically significant. family consumption, or playing a supportive role, often times unpaid, in family businesses. The PUHS attempts 3  Social Norms and FLFP to improve the measurement of FLFP by directly asking each woman in the 15–64 age group about her labor mar- Research indicates that a myriad of interconnected ket engagement, by increasing the number of questions factors, including social and cultural restrictions on that directly spell out all possible forms of employment women’s mobility, rigid gender role ideologies, and and by allowing the accounting of production of goods for the notion of honor associated with women, greatly family consumption. limit FLFP in Pakistan. As determined by multiple stud- ies, an honor culture is strongly linked with “social image” Compared to the Labor Force Survey, the PUHS al- or reputation, that is, representation of self in the eyes lows for more accurate estimates of FLFP in urban of others. Anjum, Kessler, and Aziz (2019) have termed Pakistan. While FLFP in urban Khyber Pakhtunkhwa Pakistan as having an “honor culture.” In patriarchal so- (KP) measured by the first source in 2017 is 9.4 percent, cieties like Pakistan, in order to control the demeanor of self-reported FLFP measured in 2019 Peshawar is 13.4 women and hence protect their honor, men often limit percent (Figure 1). The difference is statistically signifi- women from leaving the home and mandate that women cant at any conventional level.2 Granted, the latter esti- in their families (or clans) limit themselves in their con- mate remains low by international standards, but it sug- nection to the outside world. When women do go out, gests that the way women’s work is measured through they must be chaperoned and be appropriately garbed. survey data matters and should be improved. Within an honor culture, women are typically expected to display shyness in their demeanor, avoid eye contact Adopting a more comprehensive definition of em- with men, refrain from loud speech or laughter (espe- ployment, which includes the production of goods cially in the presence of men), and limit their interac- FIGURE 1: URBAN LABOR FORCE PARTICIPATION RATES, DIFFERENT SOURCES AND DEFINITIONS (%) Women Men 80 75.9 74.2 74.3 60 40 20 13.4 15.5 9.4 0 Labor Force Survey Peshawar Survey Peshawar Survey, Labor Force Survey Peshawar Survey Peshawar Survey, including work for including work for own consumption* own consumption* Notes: * indicates that the estimate of labor force has been extended to include subsistence agriculture. Labor Force Survey estimates refer to urban KP Province in 2017. Peshawar Survey estimates refer to 2019. Bars indicate 96 percent confidence interval. All estimates refer to individuals aged 15–64. 2 The p-value of a two-sample t-test for the difference of the two estimates at zero is 0.0003. Note that the comparison may be muddled by confounding factors, and only experimental evidence can definitively pin down the size of respondent bias in this context. 2 BOX 1  PESHAWAR AT A GLANCE Peshawar is the sixth most populous city in Pakistan with an estimated population of 1.97 million according to the 2017 census (PBS, 2017 & Finance Division, Government of Pakistan, 2018). It is situated in a valley surrounded by hills on three sides and opens into the plains of Punjab on the fourth. Ethnicity of the population is mainly Pashtun with a significant representation of Hindko speakers and Afghan refugees. The average household size is 9.4. Under the KP-FATA merger, Peshawar district was merged with Khyber agency and other smaller regions in 2018, which have only recently emerged from a decade-long conflict. In the past, also, the city has seen several incidents of conflict and terrorism. At the household level, male and female spaces are normally segregated into the zanana for women, designed to keep them secure from interaction with people outside of the family, and hujra for men to socialize. Tribal ties of kinship and blood relations form a strong backbone of households in KP, with extended families commonly liv- ing together. In Pashtun families men have disproportionately more control than women regarding public and private life matters. Violence against women is often perpetuated on the pretext of honor or the culture and not disclosed outside the household owing to its perception as a family matter. Hence, the notion of women carrying family honor is a sensitive one, and male members go to great lengths to protect it by restricting women so that family honor cannot be besmirched (Sanauddin 2015). Women in Khyber Pakhtunkhwa (KP) experience some of the highest rates of gender-based violence in Pakistan. The prov- ince of KP has one of the highest rates of women experiencing spousal violence (52% vs. 32% in Punjab and 18% in Sindh). According to the Pakistan Demographic and Health Survey (PDHS) 2017–18, 63% of women in KP agree that wife beating is justified for at least one of the following reasons: burning food, arguing with her husband, leaving the house without her husband’s permission, neglecting children or in-laws, and refusing sexual intercourse with husband. This higher accep- tance and prevalence of violence indicates a deep-rooted structure of restrictive gender roles. tions and conversations with out-of-family males to when they expressed interest in unconventional job roles, necessary topics only. This results in restrained speech and men opined that workplaces where free mixing of and movement for women (Sanauddin 2015), an effect the sexes was common were in defiance of local norms that is significantly pronounced in Peshawar. See Box 1 (World Bank 2019). For jobs outside the home, women for more details. may also have to restrict their job search to proximal employers or locations where it is convenient for male Women typically abide by the honor code, and are heavily household members to accompany them. These trends influenced by it in terms of their decision making, mo- are observed and confirmed in the data from the PUHS. bility, and interaction with spaces outside the home. Any violation of the code leads to severe repercussions. By Women’s employment remains mainly limited to the restricting women’s mobility and access to the public household setting due to mobility restrictions and sphere, the honor code has a profound impact on the ex- the burden of having the sole responsibility for home- tent and quality of women’s LFP. Asadullah and Wahhaj making and childcare. In Pakistan, as in other parts of (2016) found that community practices such as purdah South Asia, social norms around the division of labor at have a negative impact on women’s likelihood to partici- home are relatively inflexible. Women tend to perform pate in paid employment. Since women often can’t leave most household and care work. Exploring the Pakistan home, they seek employment opportunities that can be Time Use Survey 2007,3 Field and Vyborny (2015; cited managed at home. Earlier qualitative research pertain- by Tanaka and Muzones 2016) found that women who ing to FLFP in urban areas indicates that the traditional are out of the labor force still spend a lot of hours per day honor culture also influences the sectors in which women working on household chores, and that employed women seek employment and creates a barrier when it comes to on average spend more time per day on household and them exploring jobs outside of fields that are considered care work compared to employed men. The latter could socially acceptable. Hence, women are frequently en- be, in part, because men typically work longer hours for gaged in home-based work or in the education sector, and a wage than women, but it is also possible that the rea- very few choose to work in service industries. Women son women spend fewer hours earning a wage is because reported facing restrictions from male family members they have to balance market-work time with household 3 https://www.pbs.gov.pk/sites/default/files/other/tus2007/tus2007.pdf. 3 responsibilities. Indeed, when women are asked in labor in Figure 2, women with postsecondary education are force surveys why they are not available for work, the over-represented among working women. In fact, while majority say they have home responsibilities that do not only 11 percent of working-age women have completed allow them to do so (Field and Vyborny 2015). postsecondary education, the share increases to a whop- ping 25 percent among employed women.5 These mechanisms are hard to pin down quantita- tively, but they certainly play an important role in de- Women’s lack of agency further contributes to re- termining women’s agency. The instrument developed ducing female labor force participation. Low levels of for the PUHS includes a battery of questions aimed at FLFP contrast with women’s beliefs concerning work for eliciting social norms. As expected, one’s positive attitude pay. Overall, 85 percent of working-age women believe toward women’s work and one’s perceived autonomy in that women should work for pay, but only 7.6 percent of the decision to work outside the home are positively as- women are able to decide autonomously whether they sociated with the probability of participating. However, can work for pay outside the house. Table 1 shows that the addition of controls for social norms to the baseline most women indicate that their husband or father is the specification does not affect the stability of the coefficient primary decision maker about whether or not they can of factors analyzed so far.4 Going beyond the regression work for pay, be it from home or not. Given the promi- setting proves more useful to understanding the role of nent role of the husband in making decisions regarding culture in influencing women’s behavior. a spouse’s labor market engagement, it is worth noting that one in every four men believes that women should 4  What Constrains Women’s Work? never work for pay. Women’s lack of agency encompasses all aspects of life. Even strictly individual matters—for in- Education attainment is an important determinant stance, those regarding one’s own political participation of FLFP, and the limited education of women in Pe- or medical issues—often see women excluded from de- shawar correlates with their low participation in the cision making. labor market. Working-age women in Peshawar have a strikingly low level of education, with 54 percent hav- Traditional gender roles assign women a caregiving ing less than primary education (vs. 27 percent of men) role in the home. As shown in Table 2, most women and and only 29 percent of them having education beyond men subscribe to the belief that women’s rightful place this level (vs. 45 percent of men). The low level of wom- is in the home, indicating that patriarchal gender norms en’s human capital endowment is reflected in the over- are entrenched uniformly across society. Not surprisingly, all lower rate of LFP compared to men and in the highly the production of services for family use occupies much of skewed education profile of working women. As shown women’s time—and almost none of men’s time. Time use FIGURE 2: EMPLOYMENT BY GENDER AND EDUCATIONAL ATTAINMENT (COMPLETED GRADES) Women Men 60 53.6 48.3 40 31.3 28.5 24.5 27.0 28.4 20 17.7 16.8 16.0 14.8 15.4 11.2 10.8 12.5 11.5 10.0 6.6 7.7 6.9 0 All Employed All Employed below primary completed primary completed lower secondary completed upper secondary completed postsecondary Notes: All estimates refer to individuals aged 15–64. Below primary = did not complete grade 5; completed primary = completed at least grade 5, but not grade 12; completed lower secondary = completed at least grade 10, but not grade 12 (may include a vocational diploma obtained after middle or metric school); completed upper secondary = completed at least grade 12, but not the second year of university (may include a vocational diploma obtained after grade 12); completed postsecondary = completed at least the second year of university. 4 Table A2 in Appendix 2 includes a second specification of the participation model, one that includes proxies of gender norms. 5 The opposite pattern emerges among men, where those with lower levels of education are relatively more likely to participate in the labor market and to be employed. 4 TABLE 1: DECISION MAKING: WOMEN’S ANSWERS TO QUESTIONS ON “WHO MAINLY DECIDES…” You and Mother/ Father/ Other family Who mainly decides… You Spouse spouse mother-in-law father-in-law member(s) If you can work outside your house for pay? 7.6 46.2 5.3 8.3 25.8 6.8 If you can work inside your house for pay? 13.4 42.6 5.6 9.9 22.5 6.1 Whether you can participate in political activities? 10.3 43.8 5.1 9.6 24.8 6.5 About buying goods like clothes/shoes for yourself? 37.5 27.1 5.3 12.4 13.6 4.2 To start or continue your education? 15.6 43.5 5.3 7.1 23.0 5.5 To whom and when you should be married? 2.8 0.0 0.3 7.7 83.9 5.2 To seek professional medical treatment? 19.6 27.8 17.0 17.9 14.1 3.5 Whether to use birth control? (married women only) 8.9 38.1 52.2 0.4 0.3 0.0 Whether to buy or sell goods? (married women only) 3.9 59.7 13.5 2.1 14.6 6.2 TABLE 2: AGREEMENT WITH TRADITIONAL GENDER ROLES Agreement with following statements: Women (%) Men (%) If parents are in need, daughters/daughters-in-law should take more caring 71.7 59.5 responsibilities than sons/son in-law. Mothers should take more childcare responsibilities of children than fathers. 88.8 88.6 A woman should do most of household chores even if the husband is not working. 80.8 58.0 It is better if the man earns and the woman takes care of the home and children. 91.8 95.2 If a woman earns more than her husband, it is likely to cause problems. 59.2 71.4 Men are better at starting businesses than women. 85.8 92.5 TABLE 3: TIME USE IN AVERAGE DAY OF PREVIOUS WEEK (HOURS), MEN AND WOMEN AGED 15–64 Men Women Employed men Employed women Physiological time 11.0 12.4 10.5 11.1 Prayers 1.9 3.3 1.7 2.8 Leisure 2.0 1.9 1.5 1.4 Market work 7.8 0.6 9.2 3.7 House and care work 0.6 5.3 0.6 4.1 Other 0.6 0.5 0.4 0.8 Note: Physiological time includes sleeping, eating, and personal care; leisure includes visiting relatives and friends, movies, games, sports, and reading; market work includes activities performed as part of one’s job, that is, work for the production of goods and services intended for market exchange, as well as time spent searching for market work; house and care work includes cooking, cleaning, shopping, and child and elder care; other includes traveling, filling out the questionnaire, and unspecified activities. data from the PUHS give further insight into forms of work Women, including those who are employed, rarely that are overlooked by traditional labor force statistics (Ta- leave the home and are accompanied when they do. ble 3). The division of labor into market work and house/ Table 4 shows that women mostly leave their own home care work is starkly gendered: men spend virtually no to go to another house (61 percent of respondents cite time on house and care work, whether they are employed this reason for having left the house during the previous or not, while women spend an average of 5.3 hours per week); less often, they visit a clinic or health worker (25 day on this kind of work, an amount that decreases only percent), or shop for groceries or clothes (17 percent). slightly when they are employed. In fact, employed women The overwhelming majority of women (including those do as much house/care work as they do market work. who are employed) observe purdah; 36 percent of them 5 TABLE 4: TIME SPENT OUTSIDE THE HOME, WOMEN AGED 15–64 Employed women Nonemployed women All working-age women Average n days respondent went outside during previous week 3.1 1.5 1.8 Average time spent outside each time, in hours (h) and minutes (m) 3 h, 50 m 2 h, 58 m 3 h, 6 m Reasons cited for leaving the home (% positive responses for each item) To visit family/friends/neighbors 48.9 62.9 61.2 To visit a clinic or health worker 17.8 26.5 24.8 To go to shops for groceries/clothes 20.4 16.6 16.9 To go to celebrations 12.1 9.1 9.5 For Quran classes/dars/other 6.0 10.3 9.5 To attend school/literacy classes 9.0 10.8 10.6 To go to work 48.3 0.5 8.0 Other 3.8 3.7 3.6 TABLE 5: PURDAH AND BEING ACCOMPANIED WHEN LEAVING THE HOME, WOMEN AGED 15–64 Employed women Nonemployed women All working-age women Do you observe purdah? Yes 86.9 95.5 94.4 Sometimes 9.1 2.9 3.7 No 3.9 1.6 1.9 Total 100.0 100.0 100.0 Who usually accompanies you? (if any) Child 27.0 21.0 21.7 Husband 17.2 28.5 27.2 Male relative 10.8 8.7 9.0 Female relative or nonrelative 42.2 40.2 40.4 Other 2.8 1.6 1.7 Total 100.0 100.0 100.0 are usually accompanied by a husband or male relative ing alone outside their neighborhood. The chosen mode when they go out, though more than 60 percent are ac- of transport (which for 20 percent of women is walking) companied by another woman or a child (Table 5). is overwhelmingly considered safe for one’s habitual movements. In this context, feeling safe when going out can be extremely important for female empowerment and The lack of information about labor market opportu- female labor force participation. Women tend to feel nities significantly hampers female labor force partic- safe within the bounds of their limited movements, but ipation. Similar to findings from the Labor Force Survey, there is evidence that they would feel much less comfort- female unemployment6 is very low in Peshawar. Only 1 able expanding their mobility. In Peshawar, as many as percent of women reported being willing to work and, at 30 percent of all women reported having experienced the same time, to be looking for employment opportuni- some form of sexual harassment when leaving the home ties. Interestingly, about 16 percent of women who are (Table 6). Table 7 shows the vast majority of women con- out of the labor force reported being willing to work even sider walking alone in their own neighborhood to be safe, though they were currently not looking for a job. Among though 38.3 percent of women would not feel safe walk- them, the reason reported for not searching for a job by 6 According to ILO: Unemployed refers to a person without a job during a given week; available to start a job within the next two weeks; actively having sought employment at some time during the last four weeks or having already found a job that starts within the next three months. 6 TABLE 6: WOMEN WHO HAVE EXPERIENCED SEXUAL HARASSMENT OUTSIDE THE HOME (%) Any type of harassment 30.8 Inappropriate staring/comments 28.1 Gestures/actions of sexual nature 11.0 Inappropriate use of phone/email 7.5 Other harassment 0.4 TABLE 7: FEELINGS OF SAFETY Percent of respondents who feel safe, at least during the day… Employed women (%) Nonemployed women (%) All working-age women (%) When walking alone in their own neighborhood 89.0 81.6 82.3 When walking alone outside their own neighborhood 71.1 60.6 61.7 When using their chosen mode of transport 95.2 91.2 90.4 the majority of women was lack of knowledge about labor by gender, with women engaging in different types of market functioning (49 percent), followed by cultural and employment (Figure 3). In particular, although there are family prohibition (15 percent) and care responsibilities roughly equal numbers of working men and women in (12 percent). The lack of knowledge is mostly related to the labor market, more men are self-employed or oper- the job search process: 41.5 percent of women willing to ate their own business, and more women support fam- work reported not to be looking for a job because they did ily businesses, often without pay. On the one hand, in not know how to, while 7.5 percent reported not knowing terms of employment status by educational attainment, which type of job they could do for pay, possibly proxying the overwhelming majority of highly educated women for a lack of specific skills. Interestingly, analysis of PUHS work as employees (76 percent), while no such pattern data shows that addressing these knowledge constraints emerges for men or low-educated women. On the other could increase FLFP in Peshawar from 13.4 to 20.4 percent. hand, women with low education attainment are more than three times as likely to be self-employed compared 5  Characteristics and Quality to highly educated women. This further reflects a lower of Women’s Jobs representation of women in high-skilled professions and indicates that self-employment held by women tends to be low skilled (small scale and home based). The profile of jobs differs substantially by gender, with women more likely to work without pay. The la- Occupational diversity among female workers is lim- bor market in Peshawar tends to be strongly segmented ited, reflecting the concentration of employment in FIGURE 3: STATUS IN EMPLOYMENT, BY GENDER AND EDUCATIONAL ATTAINMENT LEVEL (%) 100 1.0 3.0 1.0 2.6 0.9 3.7 80 46.6 59.3 57.8 58.3 60.7 60 75.8 40 25.7 8.5 8.3 8.8 21.3 20 32.4 31.3 25.2 29.6 13.5 17.9 0 6.9 Men Women Men Women Men Women All Employed Low Education High Education Self-employed Contributing family workers Employees Other Notes: Contributing family workers are those who work in a business owned by a family member or help a family member who works for someone else. High-educated men and women have completed at least lower secondary education; low-educated men and women have not. 7 TABLE 8: THE 10 MOST FREQUENT OCCUPATIONS, BY GENDER AND EDUCATION LEVEL (ISCO-3 DIGITS) Among women % Cumul. % Among men % Cumul. % 1 Garment and related trades workers 19.7 19.7 Shop salespersons 21.6 21.6 2 Primary school, early childhood teachers 15.8 35.5 Car, van, and motorcycle drivers 9.7 31.3 3 Domestic, hotel, and office cleaners 9.1 44.6 Garment and related trades workers 5.6 36.9 4 Food preparation assistants 6.8 51.4 Machinery mechanics and repairers 5.3 42.1 5 Shop salespersons 5.6 57.0 Mining and construction laborers 4.5 46.7 6 Handicraft workers 5.0 62.0 Street and market salespersons 3.4 50.0 7 Building and housekeeping supervisors 4.4 66.3 Building frame and related workers 2.4 52.5 8 Animal producers 4.2 70.5 Protective services workers 2.3 54.8 9 Medical doctors 3.8 74.3 Cooks 2.3 57.2 10 Secondary education teachers 3.6 77.9 Business services agents 2.3 59.4 Among low-educated women Among low-educated men 1 Garment and related trades workers 27.6 27.6 Shop salespersons 17.7 17.7 2 Domestic, hotel, and office cleaners 14.9 42.5 Car, van, and motorcycle drivers 13.5 31.2 3 Shop salespersons 8.7 51.2 Garment and related trades workers 7.4 38.6 4 Food preparation assistants 8.2 59.5 Machinery mechanics and repairers 6.4 45.0 5 Handicraft workers 7.4 66.9 Mining and construction laborers 6.2 51.1 6 Building and housekeeping supervisors 7.2 74.1 Street and market salespersons 5.2 56.3 7 Animal producers 5.1 79.2 Transport and storage laborers 3.2 59.5 8 Manufacturing laborers 2.9 82.1 Cooks 3.2 62.7 9 Street and market salespersons 2.3 84.5 Building frame and related workers 3.2 65.9 10 Other sales workers 2.1 86.5 Protective services workers 2.5 68.4 Among high-educated women Among high-educated men 1 Primary school, early child. Teachers 21.6 21.6 Shop salespersons 27.4 27.4 2 Medical doctors 9.7 31.3 Car, van, and motorcycle drivers 4.2 31.5 3 Secondary education teachers 5.6 36.9 Business services agents 3.9 35.5 4 Garment and related trades workers 5.3 42.1 Machinery mechanics and repairers 3.5 39.0 5 Food preparation assistants 4.5 46.7 Finance professionals 3.4 42.4 6 Other sales workers 3.4 50.0 Garment and related trades workers 2.9 45.2 7 Other health associate professionals 2.4 52.5 Primary school, early childhood teachers 2.7 47.9 8 Client information workers 2.3 54.8 General office clerks 2.6 50.5 9 Animal producers 2.3 57.2 Medical doctors 2.2 52.7 10 Manufacturing laborers 2.3 59.4 Sales, marketing, development managers 2.1 54.8 Note: High-educated men and women have completed at least lower secondary education; low-educated men and women have not. socially accepted occupations. Almost 80 percent of fessionals (Table 8). The occupational profile of working all employed women are concentrated in the 10 most women reflects stereotypically female roles and aligns common occupations for women workers, as opposed with preferences expressed by men regarding the condi- to 60 percent of employed men concentrated in the 10 tions under which it can be acceptable for women to work most common occupations for working men. Concen- for pay (Table 9). tration increases further among low-educated workers. Women’s most frequent occupations differ greatly by The majority of women prefer home-based work, education level—more so than the most frequent occu- and the majority of employed women work at home. pations among men. Women with low education tend to Home-based work is the most important condition that be manufacturers, cleaners, and salespeople, whereas makes female employment acceptable to men and is con- high-skilled women tend to be teachers and medical pro- sidered an ideal condition for women. In fact, when asked 8 TABLE 9: CONDITIONS THAT MEN VIEW AS ACCEPTABLE FOR WOMEN TO WORK FOR PAY (%) Acceptable for women to work for pay if… Men Work is home based 57.9 Work as a teacher or nurse 10.2 No interaction with non-mahram men 7.4 Economic necessity 6.2 Can work while observing purdah 5.6 Other 3.7 If outside the home, sex-segregated workspace 3.6 Does not interfere with responsibilities at home 2.0 If outside the home, workplace should be close to home 1.5 A good salary 1.0 Access to safe transport 0.8 Medical and other benefits (e.g., childcare) 0.1 TABLE 10: LOCATION OF EMPLOYMENT, BY GENDER (%) Low-educated High-educated All employed Women At own/other’s home 88.4 28.4 65.0 On the street/road/outside 2.1 0.8 1.6 In a shop/office/factory 5.6 62.0 27.6 Other 3.9 8.8 5.8 Men At own/other’s home 6.0 3.6 5.0 On the street/road/outside 29.6 10.9 22.1 In a shop/office/factory 60.1 80.7 68.3 Other 4.4 4.8 4.6 about the characteristics of an ideal job, 57 percent of remuneration of employed men and women. A sizable women reported a preference for it being “home-based”; proportion of the already small female workforce (21 while for 35 percent, it would be a “government job.” Such percent) is unpaid and tends to work as contributing preferences are clearly reflected in the actual profile of family workers. Unpaid work also remains relatively jobs held by women (Table 10). While only 5 percent common among skilled women (15 percent), while it of men work inside the home, 65 percent of women do. is exceedingly rare (below 4 percent) among men, re- Among women with low education, the share of women gardless of skill level. The time that the average em- in home-based work is almost 90 percent. Even among ployed woman spends performing market work is high-skilled women, close to 30 percent engage in home- consistently about half that of men. This explains that based work, while many others are employed outside the women wage earners tend to earn around half as much home as teachers and medical professionals (possibly in as men per month. the public sector).7 The results presented in this section indicate that the Women tend to work in low-quality jobs compared polarization of men’s and women’s working lives ex- to men. Table 11 details the average hours worked and tends far beyond the decision to join the labor force. 7  igh-educated women who work at home tend to be on the lower end of “high” education (having not completed upper secondary education), and tend to work in H manufacturing (garments, handicrafts, food preparation). They are more likely to be contributing family workers than own-account workers or employees. 9 TABLE 11: EARNINGS AND HOURS WORKED, BY GENDER Average hours worked Average reported earn- Education level Paid/unpaid workers Observations % of workers per day ings per month (PKR) Women Low-educated unpaid 210 24.3 4.9 0 paid 409 75.7 3.8 7,255 High-educated unpaid 16 15.8 6.5 0 paid 153 84.2 4.5 22,399 All employed unpaid 226 21.3 5.6 0 paid 562 78.7 4.0 13,464 Men Low-educated unpaid 71 2.5 9.9 0 paid 2,578 97.5 9.3 18,170 High-educated unpaid 44 3.9 9.2 0 paid 1,050 96.1 8.1 36,662 All employed unpaid 115 3.1 9.7 0 paid 3,629 96.9 8.7 25,040 Note: High-educated men and women have completed at least lower secondary education; low-educated men and women have not. They also highlight that education makes a difference recommendations for alleviating the low rate of FLFP are in working women’s experiences. Most women of presented in this section. working age do not work, making employed women a mi- nority. Those who work are largely segregated in terms of Social norms seem to be the most powerful factor in their occupations and restricted in terms of job location, determining women’s interactions with the public working hours, and pay. Women’s predominantly domes- sphere and the workforce by restricting women to tic role and responsibility for their families appear to be the home or only allowing access to certain occupa- a strong influence even when women are employed, and tions deemed suitable. Given the extent of stigma and the need for reconciliation of house care and work for pay discouragement women face from families and society is reflected by the characteristics of their jobs. Although when pursuing jobs outside the home, it is clear that a this is true across the board, educated women are far definitive change in widely held views regarding tradi- more likely to work outside the home and have careers tional gender roles at the household and community that are more similar to their male counterparts; women level is required to accelerate women’s LFP. Research with low education are mostly engaged in informal home- indicates that possible interventions to influence norms based work, often in manufacturing, and they are more include strategic use of positive messaging about strong likely than men to be contributing to someone else’s busi- female role models. Furthermore, global evidence sug- ness rather than their own. gests that engaging men is crucial in changing norms surrounding women’s economic activities. For example, 6  Conclusion and Next Steps men can act as “gatekeepers” for women by providing access to capital, information, and networks. Engaging While the vast majority of women face constraints in men in these interventions also reduces the likelihood accessing work, solutions at the policy level must not of domestic violence and tensions within the home (ILO be developed using a one-size-fits-all approach. Anal- 2014). Similarly, employing women in public leadership ysis of the PUHS and earlier qualitative studies indicates positions can support an acceptance of ambitions and the need to be responsive toward the specific challenges career aspirations among women (Asian Development that women from different socioeconomic backgrounds Bank [ADB] 2016). An experiment from India demon- face, as well as the need to enact measures that facili- strated that reserved seats for women in village coun- tate working women universally. Given the strong link cils led to a 20 percent decrease among parents and a between FLFP and economic development, there is a 32 percent decrease among adolescents in gender gap pressing need to facilitate women’s economic empower- regarding educational and career aspirations. While ev- ment. Based on analysis of the survey data from Peshawar, idence did not point to improved labor market opportu- 10 nities, a shift in established gender norms was observed sourcing industry in India has demonstrated that FLFP (Beaman et al. 2012). increased by 2.4 percentage points due to recruitment strategies aimed at women (Jensen 2012). Further, quo- The high prevalence and acceptance of gender-based tas in India mandating representation of women at the violence (GBV) in KP is a factor in limiting women’s local electoral level have improved outcomes on FLFP, voice and agency and acts as a barrier for women’s entrepreneurship, enrolment of girls in school, and deci- interaction and participation in the labor force (Paki- sion making (Fletcher, Pande, and Moore 2017). Despite stan Demographic and Health Survey [PDHS] 2017–18). a general preference among women to opt for govern- Hence, it is important to invest in preventing GBV, par- ment jobs, data indicate that as of 2018–19, 17.7 per- ticularly through population-based norm and behavior cent of women were employed as grade 17 and above change efforts with women, men, girls, and boys (Arango government employees, but only 1.4 percent of women et al. 2014; Heise 2011). Such interventions may be par- as grade 16 and below (more concentration of jobs in ticularly important and effective in the context of pro- grades 1–16; Pakistan Public Administration Research grams that improve women’s status in the household Centre 2018–19). KP’s Women’s Empowerment Policy and society, such as by encouraging women to work, 2017 proposes 20 percent of jobs for women in gov- especially in male-dominated sectors and fields that are ernment services, and the Azm-e-Nau, KP Economic not considered socially acceptable for women (Fulu and Recovery plan 2020–2023, in light of the COVID crisis, Kerr-Wilson 2015). A focus on behavioral change pro- also proposes provision of jobs, training, and grants to grams at the community level can be effective in tackling women at 15–50 percent in different programs. While gender roles and attitudes toward GBV, with special at- implementation of these proposals remains a challenge, tention to the inclusion of boys and men as change agents. active efforts from the private and public sectors can The SASA! model developed in Kampala, Uganda, is a boost women’s inclusion. comprehensive example of an intervention with demon- strated results for combating mind-sets that propagate Women in Peshawar strongly lack awareness on how GBV by changing behaviors relating to power imbalances to find job opportunities or build on their skillset, a between men and women. In tandem with behavioral gap that can be remedied by greater access to infor- change interventions, implementation of legal reforms to mation. The lack of knowledge of available job opportu- promote women’s safety and rights is critical to create a nities and how to find them, as well as not knowing what conducive employment ecosystem where women can ex- skills are employable, are the major reasons why both plore opportunities for economic gain. Progressive laws high-educated and low-educated women are not cur- criminalizing domestic violence and sexual harassment rently working for pay even though they wish to. Invest- exist in KP, and their effective implementation can sup- ment in the creation of platforms and institutions that port this outcome. serve to match women job seekers with firms that are hiring can be a beneficial step. These centers or portals Currently working women are mainly distributed should focus on job requirements for women from varied among a limited number of sectors while repre- education and experience backgrounds, and they should sentation in other nontraditional sectors is very provide career counseling and opportunities for women low. Most low-educated employed women contribute to connect with mentors and networks. For example, the to the garment industry and are employed as cleaners Job Talash program has been a promising effort in this in homes and establishments; among high-educated direction and should be scaled up to other areas if eval- women, the education and medical professions are the uations of the intervention indicate positive impact. See most popular. These career fields are considered more Box 2 for more information on this program. suited to women because they are perceived to align bet- ter with traditional gender roles. Employment in other Investing in girls’ education and training is an im- fields, especially nontraditional sectors, can be espe- portant precondition to increasing FLFP. As demon- cially challenging for women. However, there is growing strated by a large body of evidence and the PUHS, women evidence to suggest that informational nudges—partic- are more likely to be involved in the labor force if they ularly those that emphasize the differential earnings be- are more educated. Addressing both demand and sup- tween female- and male-dominated occupations—can ply constraints that limit girls’ education remains a key encourage women to enroll in training programs to en- priority. Similarly, a lack of marketable skills can discour- ter male-dominated trades (Hicks et al. 2011). Similarly, age women from seeking jobs. Qualitative research con- early exposure to male role models has also been shown ducted among low-educated women reveals an untapped to improve the likelihood of women crossing over into demand for affordable government-run skills centers male-dominated sectors and occupations (Campos et where women can take courses on subjects such as beau- al. 2015). Evidence from the business-processing out- tician training, cooking, and dress designing, to foster mi- 11 BOX 2  JOB TALASH: HELPING WOMEN ACCESS EMPLOYMENT OPPORTUNITIES The gap in women’s knowledge regarding employment opportunities led to the creation of the Job Asaan Employment Fa- cilitation Center for Women under a partnership of the Punjab Commission on the Status of Women (PCSW) and the Center for Economic Research in Pakistan (CERP). The center was inaugurated in Lahore in May 2018 with the goals of providing career counseling, connecting potential employees with companies looking to hire, providing web-based assistance to those who cannot visit the center, and supporting professional development of women who do not work currently but do wish to. Services provided include developing CVs, referrals, co-working spaces, and workshops on interview skills. The platform is now known as Job Talash and is being supported by CERP and the Foreign and Commonwealth Office (FCO). It is active in creating linkages between female job seekers and firms through a helpline and online presence. Source: Job Talash website, https://jobtalash.com.pk/. croenterprises (World Bank 2019). Emphasis on focus- women’s entrepreneurship is linked to greater well-­being, ing on these gaps at a policy level is critical to produce a improved decision making, and alleviation of poverty. workforce where women are adequately equipped and However, women are disproportionately disadvantaged represented. Higher education and higher skills will also when accessing institutional finance, a problem that is support an increase in the quality of women’s employ- exacerbated because of women’s lack of control over as- ment, potentially supporting a positive feedback mecha- sets and their concomitant inability to provide collateral. nism leading to progressively higher human capital and Financial institutions generally continue to view female socioeconomic inclusion. borrowers as risky clients and have fewer product offer- ings designed to suit women’s entrepreneurial needs. As a Professional networking opportunities, women’s result, they resort to mobilizing capital through informal groups, and strong female role models are power- borrowing channels, personal savings, or family support. ful ways to develop ambitions among women. Due to A large body of evidence, however, has deemed women to safety concerns, limited mobility, and lack of exposure to be trustworthy borrowers of microfinance loans and vari- public activities, women are more likely to miss out on ous state-run social protection and loan programs that are peer and mentor learning opportunities (Gulati, Afridi, targeted toward women. To further boost women’s access and Bandiera 2019). Educated women face difficulty to finance, regulatory frameworks need to be applied to in accessing information about opportunities through the financial sector that will facilitate women’s economic formal networks and instead have to rely on informal empowerment. Specially designed banking products with family networks, which may not operate in their favor incentives for banks to cater to female clientele can also be given the kind of opposition women face from families beneficial for banks since women are a largely unbanked when pursuing employment; whereas men are likely to potential customer base (Niethammer et al. 2007). benefit from such networks (World Bank 2019). At the enterprise level, women’s networks can help women Access to internet is a key facilitator in providing navigate career development and leadership given that women with the opportunity to expand their skills, women face challenges in reaching senior management engage in entrepreneurial outreach, and bridge roles. Among low-educated women, networks can boost knowledge gaps regarding employment options, es- confidence and create a more conducive support system pecially for women’s ability to work from home. A for work and entrepreneurship. Networks can facilitate large majority of women reported working from home the availability of material, moral, and financial support; as the most ideal work, a preference that has become they can also alleviate families’ concerns for the safety of more pronounced and unavoidable due to the COVID-19 female family members when women work together as pandemic. At such a time, access to internet can support groups (Zeb and Kakakhel 2018). greater awareness for women and provide an opportu- nity to interact with platforms for growing e-commerce A stark disparity between entrepreneurship out- and home-based jobs. For entrepreneurs and women comes for men and women indicates the need for business owners, the internet offers a host of opportuni- providing women with more convenient access to fi- ties for expansion and growth of businesses by connect- nance, internet, and business development trainings. ing sellers to markets, facilitating access to investment, The small- and medium-enterprise sector can contribute and providing access to learning materials. greatly to Pakistan’s economy and be a critical catalyst for boosting FLFP. Increased household income as a result of 12 Infrastructural reform to facilitate transportation and to their employees. Such an approach can be adopted safety of public spaces remains a key area for action to across other provinces of Pakistan as well and should be support FLFP. The security concerns and gender norms supported by regular data collection to tabulate availabil- that inhibit mobility of women directly result in stunted ity of gender-friendly facilities at the institutional level to LFP of women. While 30.8 percent of women reported be- propose solutions to fill in the gaps. At the community ing harassed in public spaces in the PUHS, other estimates level, communal childcare facilities can also ease the bur- indicate harassment of up to 85 percent of women who den. Flexible work hours and work-from-home options travel in public transport (ADB 2014). These challenges can alleviate some of the stress of dual responsibilities prevent urban women from traveling within the city and and also help women gain approval from family (World inhibit the movement of women from peri-urban and ru- Bank 2019). Provisions for separate toilets and rest/ ral areas, too. Similarly, younger women face issues when prayer areas for women are also important to establish traveling for schooling or skills trainings. Further, public comfortable workplaces for women. Additionally, mecha- transport options are not affordable or as widely available nisms to deal with sexual harassment in line with the Pro- as needed. The consequences of these mobility restric- tection Against Harassment of Women at the Workplace tions can be quite significant. Comparing wages at each Act of 2010 should be mandated and implemented to ad- education level, our survey shows that at all education lev- equately address incidents of harassment, which remain els, average earnings for women are higher if they are em- a significant barrier to women entering and continuing in ployed outside of their home compared to working from work outside the home. home. To bridge this gap, affordable, safe public transport systems responding to the specific needs of women and Data collection, transparency, and legal reform to supporting their participation in the workforce are criti- protect the rights of women workers can support the cal (ADB 2016). Additionally, pedestrian walkways should development of responsive policies to boost FLFP be adequately lit, and women should have easy access to and recognize the contribution of informal workers. reporting incidents of harassment and swift resolution of As demonstrated by the PUHS, survey techniques to mea- issues with support from law enforcement. sure FLFP should be modified to capture more robust data on productive activities by including estimates of Workplace environments where access to facilities economic contributions that are not incorporated in tra- such as childcare, flexible work hours, dedicated ditional labor force surveys. Data collection and report- transport, and separate rest areas for women are more ing at regular intervals is also critical to the monitoring conducive to women’s work. Research has consistently of gaps and challenges that limit women’s employment indicated that the burden of household responsibilities, outcomes. Further, since this report does not provide a childcare, and elder care takes up a large amount of time detailed exploration of the dynamics of female home- from women’s daily lives, resulting in challenges in en- based workers,8 the need remains to capture the extent of tering the labor market, retaining jobs, and re-­entering contributions made by this group of workers and to bet- the workforce. Global evidence suggests that early access ter understand the constraints they face. The economic to subsidized childcare increases women’s likelihood to contributions of the largely invisible home-based women work for a wage and run a business, particularly if this is workers also need attention at the policy level: 70.7 per- outside the home (Clark et al. 2017; Martinez, Naudeau, cent (Pakistan Bureau of Statistics 2019) of all women and Pereira 2012). Availability of high-quality childcare in the workforce are represented in the informal sector in Peshawar remains a challenge for women who already and face a myriad of consequences due to their “informal” bear a large burden of the strong social norms empha- status, including little or no access to organized markets, sizing their role as mothers and primary caregivers. To credit institutions, formal education and training institu- remedy that, state-subsidized childcare programs and tions, or most other public services and amenities. Fur- public-­private partnerships for daycare facilities can ther, legal reform to protect the rights of women workers support working women. The government of Punjab in the private sector can mandate conducive working serves as a working example of this initiative and has conditions with provisions for childcare, separate toilets, established a fund for providing grants to public- and flexible work schedules, and dedicated transportation private-­sector organizations to provide daycare facilities options for women. 8 According to the Sindh Home Based Workers Act 2018, “Home Based Worker” means any person indulging in production and manufacturing of goods and rendering of services in relation ancillary thereto in the home premises or nearby premises, garage, or any other place near the home while working in connection with the work of any Industry, establishment, undertaking or commercial establishment or any place of his choice for hire or reward or remuneration either directly or through a contractor or subcontractor or intermediary whether the terms of employment be express or implied. Sindh is currently the only province to have passed this law. 13 This Note was authored as a collaborative effort by members of the World Bank’s Pakistan Gender and Social Inclusion Platform and the Pakistan Poverty and Equity Program. The leading author for the note was Giulia Mancini, Department of Economics, University of Rome. The effort was supported by Rohini Prabha Pande, Noor Rahman, Ahmad Shah Durrani, Sundas Liaqat, and Jyostna Subramaniam from the Pakistan Gender and Social Inclusion Platform, Social Sustainability and Inclusion Global Practice (SSI GP), while Silvia Redaelli and Maria Qazi supported from the Poverty Global Practice. Maria Beatriz Orlando and Uzma Quresh provided valuable comments. The data are from the Peshawar Urban Poverty Household Survey, a collaborative study between the Gender Platform, SSI GP, and the Poverty GP. Data were collected by the survey firm RCONS. For additional information please contact: Maria Beatriz Orlando, morlando@worldbank.org, Uzma Quresh, uquresh@worldbank.org (co-Task Team Leaders of the Pakistan Gender and Social Inclusion Platform) and Silvia Redaelli sredaelli@worldbank.org (Task Team leader of the Pakistan Poverty and Equity Program). 14 Appendix 1: Technical Details on the Peshawar Urban Household Survey Data collection Fieldwork period July – Dec 2019 Mode of data capture Paper-assisted personal interviewing (PAPI), separate questionnaires by gender Sampling Sampling frame 2017 Census Primary Sampling Units (PSU) 239  Planned sample size (households) 2,400  Actual sample size (households) 2,357  Sample composition Individuals 21,506  Males 11,063  Females 10,443  Working-age men (15–64) 5,870  Working-age women (15–64) 5,629 Afghan population 11,761 Appendix 2: Regression Results TABLE A1: SAMPLE MEANS OF REGRESSORS Full regression sample (N = 5,367) FLFP = 1 (N = 828) Mean Coeff of variation Mean Coeff of variation Age 33.3 2.51 33.8 2.97 Afghan dummy 5.5 0.24 6.9 0.27 Married dummy 65.4 1.37 56.7 1.14 Own education (completed grades) less than primary 48.9 1.17 44.1 1.04 primary 29.8 0.31 21.1 0.25 secondary 7.2 0.53 6.5 0.48 tertiary or more 14.2 0.37 28.3 0.56 Relationship with head spouse of HH head 47.6 0.95 47.4 0.95 daughter of HH head 22.6 0.54 28.3 0.63 daughter-in-law of HH head 12.2 0.37 6.5 0.26 other relationship 17.5 0.46 17.9 0.47 Household size 7.2 1.79 6.3 2.04 Household composition n 0–5 0.9 0.73 0.7 0.65 n 6–14 1.5 0.88 1.4 0.89 n 15–24 1.7 1.08 1.5 1.05 n females 25–44 0.9 1.04 0.9 1.24 n males 25–44 0.8 0.88 0.6 0.76 n 45–64 1.1 1.31 1.0 1.21 n 65+ 0.3 0.48 0.2 0.45 15 (Table A1 continued)  Full regression sample (N = 5,367) FLFP = 1 (N = 828)   Mean Coeff of variation Mean Coeff of variation Nuclear family dummy 46.3 0.93 59.8 1.22 Household head education (completed grades) less than primary 38.1 0.96 47.5 1.12 primary 36.7 0.33 26.3 0.24 secondary 6.4 0.57 7.3 0.51 tertiary or more 18.8 0.46 18.9 0.47 Food adequacy dummy 82.8 2.2 73.4 1.66 Asset score 0.6 0.31 0.5 0.26 Feels safe outside own neigh. dummy 40.3 0.82 53.4 1.07 Feels transport is safe dummy 52.7 1.06 73.5 1.66 Purdah dummy 92.5 3.5 81.2 2.07 Involvement in decision making work inside home 21.7 0.53 41.5 0.84 work outside home 16.7 0.45 35.8 0.75 community activity 19.3 0.49 34.7 0.73 political activity 19.3 0.49 34.8 0.73 shopping 46.3 0.93 61.1 1.25 education 25.3 0.58 40.5 0.82 marriage 4.1 0.21 11 0.35 health 40.8 0.83 52.2 1.04 Own belief: pro female work 86.5 2.53 95.2 4.43 Note: HH = household TABLE A2: AVERAGE MARGINAL EFFECTS FROM PROBIT PARTICIPATION EQUATIONS (1) FLFP (2) FLFP Age 0.0200*** (0.0038) 0.0187*** (0.0037) Age squared −0.0003*** (0.0001) −0.0003*** (0.0001) Afghan==1 0.0554*** (0.0163) 0.0521** (0.0162) Married==1 −0.0944*** (0.0161) −0.0877*** (0.0170) n 0–5 −0.0036 (0.0058) −0.0051 (0.0057) n 6–14 0.0009 (0.0046) 0.0008 (0.0048) n 15–24 −0.0021 (0.0045) 0.0003 (0.0045) n females 25–44 −0.0124 (0.0102) −0.0091 (0.0097) n males 25–44 −0.0295** (0.0091) -0.0268** (0.0089) n females 45–64 -0.0181 (0.0196) −0.0117 (0.0187) n males 45–64 −0.0021 (0.0150) −0.0010 (0.0146) n 65+ −0.0085 (0.0150) −0.0054 (0.0146) Own education (completed grades) Reference = less than primary −0.0047 (0.0253) −0.0133 (0.0242) primary 0.0437 (0.0230) 0.0345 (0.0222) secondary 0.1981*** (0.0257) 0.1761*** (0.0256) tertiary or more −0.0282 (0.0236) −0.0326 (0.0229) 16 (Table A2 continued) (1) FLFP (2) FLFP Education of HH head (completed grades) Reference = less than primary −0.0625* (0.0257) −0.0654* (0.0265) primary −0.0835** (0.0297) −0.0971** (0.0305) secondary −0.0725** (0.0227) −0.0625** (0.0224) tertiary or more −0.0167** (0.0050) −0.0172*** (0.0049) Food adequacy==1 0.0426** (0.0161) 0.0236 (0.0163) Asset score 0.0803*** (0.0157) 0.0753*** (0.0152) Feels safe outside own neighborhood −0.1024*** (0.0263) −0.0744** (0.0277) Feels transport is safe 0.0992*** (0.0214) Purdah==1 0.0334 (0.0237) Own belief: in favor of female work 0.0776* (0.0333) Own decision: work inside 0.0010 (0.0354) Own decision: work outside −0.0205 (0.0374) Own decision: community activity 0.0272 (0.0201) Own decision: political activity −0.0029 (0.0248) Own decision: shopping 0.0210 (0.0305) Own decision: education −0.0148 (0.0191) Own decision: marriage 0.0200*** (0.0038) 0.0187*** (0.0037) Own decision: health −0.0003*** (0.0001) −0.0003*** (0.0001) Observations 5,367 5,245 Pseudo R-squared 0.168 0.2021 Notes: HH = household. All working-age (15–64) women are able to work (i.e., not in school or ill/disabled). Standard errors in parentheses. * p < 0.05, ** p < 0.01, *** p < 0.001. 17 Bibliography Abramsky, T., et al. 2014. “Findings from the SASA! of Formalization Assistance and a Bank Study: A Cluster Randomized Controlled Information Session on Business Registration Trial to Assess the Impact of a Community and Access to Finance in Malawi.” Policy Mobilization Intervention to Prevent Violence Research Working Paper (7183), World Bank, Against Women and Reduce HIV Risk in Washington, DC. Kampala, Uganda.” BMC Medicine 12: 122–39. Cho, Y., and Z. Majoka. 2020. “Pakistan Jobs Diagnostic: Anjum, Gulnaz, Thomas Kessler, and Mudassar Aziz. Promoting Access to Quality Jobs for All.” 2019. “Cross-Cultural Exploration of Honor: Report, World Bank, Washington, DC. Perception of Honor in Germany, Pakistan, and South Korea.” Psychological Studies 64: 147–60. Clark, Shelley, Caroline Kabiru, Soni Laszlo, and Stella doi:10.1007/s12646-019-00484-4. Muthuri. 2017. “Can Subsidized Early Child Care Promote Women’s Employment? Evidence Arango, D., et al. 2014. Interventions to Prevent or from a Slum Settlement in Africa.” GrOW Reduce Violence Against Women and Girls: A Research Working Paper Series. #5, Institute Systematic Review of Reviews. Washington, DC: for the Study of International Development, World Bank. McGill University, Montreal, Canada. Asadullah, M. Niaz, and Zaki Wahhaj. 2016. “Missing Desiere, S., and V. Costa. 2019. “Employment Data in from the Market: Purdah Norm and Women’s Household Surveys: Taking Stock, Looking Paid Work Participation in Bangladesh.” IZA Ahead.” Policy Research Working Paper, World Discussion Papers 10463, Institute of Labor Bank, Washington, DC. Economics (IZA). https://ideas.repec.org/p/ iza/izadps/dp10463.html. Dildar, Y. 2015. “Patriarchal Norms, Religion, and Female Labor Supply: Evidence from Turkey.” Asian Development Bank (ADB). 2014. “Rapid World Development 76: 40–61. Assessment of Sexual Harassment in Public Transport and Connected Spaces in Karachi.” Field, E., and K. Vyborny. 2015. “Female Labor Force Report, Asian Development Bank, Manila. Participation in Asia: Pakistan Country Study.” Unpublished Report, Research and Bardasi, E., K. Beegle, A. Dillon, and P. Serneels. 2011. Development Technical Assistance 8620 “Do Labor Statistics Depend on How and to on Economic Analysis for Gender and Whom the Questions Are Asked? Results from Development, Asian Development Bank. a Survey Experiment in Tanzania.” The World Bank Economic Review 25(3): 418–47. Finance Division, Government of Pakistan 2018. “Pakistan Economic Survey. 2017-2018: Beaman, Lori, Esther Duflo, Rohini Pande, and Petia Population, Labor Force and Employment.” Topalova. 2012. “Female Leadership Raises Chapter 12, Government of Pakistan. https:// Aspirations and Educational Attainment www.finance.gov.pk/survey/chapters_18/12- for Girls: A Policy Experiment in India.” Population.pdf Science 335(6068): 582–86. https://science. sciencemag.org/content/335/6068/582. Filmer, D., and L. H. Pritchett. 2001. “Estimating Wealth Effect without Expenditure Data—or Tears: Benes, E., and K. Walsh. 2018. “Measuring Employment An Application to Educational Enrollments in in Labour Force Surveys: Main Findings from the States of India.” Demography 38(1): 115–32. ILO LFS Pilot Studies.” Department of Statistics, International Labour Organization, Geneva. Fletcher, Erin K., Rohini Pande, and Charity Troyer Moore. 2017. “Women and Work in India: Boserup, E. 1970. Woman’s Role in Economic Descriptive Evidence and a Review of Potential Development. London: Allen and Unwin. Policies.” CID Faculty Working Paper No. 339, Harvard University, Cambridge, MA. Campos, Francisco, Markus Goldstein, and David McKenzie. 2015. “Short-Term Impacts 18 Fulu, E., and A. Kerr-Wilson. 2015. “What Works to International Labour Organization (ILO). 2014. Prevent Violence Against Women and Girls “Engaging Men in Women’s Economic Evidence Reviews. Paper 2: Interventions to Empowerment and Entrepreneurship Prevent Violence against Women and Girls.” Development Intervention.” Issue Brief. https://www.whatworks.co.za/documents/ https://www.ilo.org/wcmsp5/groups/public/- publications/35-global-evidence-reviews- --ed_emp/---emp_ent/---ifp_seed/documents/ paper-2-interventions-to-prevent-violence- briefingnote/wcms_430936.pdf. against-women-and-girls-sep-2015/file. International Labour Organization (ILO). 2020. Goldin, C. 1995. “The U-Shaped Female Labor Force ILOSTAT Database, Prime-Age Labour Force Function in Economic Development and Participation Rate by Sex, Household Type and Economic History.” In Investment in Women’s Rural / Urban Areas. Website database. https:// Human Capital and Economic Development, ilostat.ilo.org/data/. edited by T. P. Schultz, 61–90. Chicago: University of Chicago Press. Jensen, R. 2012. “Do Labor Market Opportunities Affect Young Women’s Work and Family Decisions? Gulati, Nalini, Farzana Afridi, and Oriana Bandiera. 2019. Experimental Evidence from India.” The “Making the Workplace Work for Women.” Quarterly Journal of Economics 127(2): 753–92. Growth Brief, International Growth Centre, London. https://www.theigc.org/reader/ Klasen, S., & J. Pieters. 2015. “What Explains the making-the-workplace-work-for-women/. Stagnation of Female Labor Force Participation in Urban India?” The World Bank Economic Heise, L. 2011. “What Works to Prevent Partner Review. (Originally published in 2013, IZA Violence? An Evidence Overview.” London: DP No. 7597, Institute for the Study of Labor, Department for International Development. http://ftp.iza.org/dp7597.pdf.) Hicks, Joan Hamory, Michael Kremer, Isaac Mbiti, and Loko, Boileau, and Mame Astou Diouf. 2009. “Revisiting Edward Miguel. 2011. “Vocational Education the Determinants of Productivity Growth: Voucher Delivery and Labor Market Returns: A What’s New?” Working Paper Vols. 9–225, Randomized Evaluation among Kenyan Youth.” International Monetary Fund, Middle East and Report for Spanish Impact Evaluation Phase II, Central Asia Department. World Bank. Martinez, Sebastian, Sophie Naudeau, and Vitor Pereira. Hlasny, V., and S. AlAzzawi. 2019. “Asset Inequality in the 2012. “The Promise of Preschool in Africa: MENA: The Missing Dimension?” The Quarterly A Randomized Impact Evaluation of Early Review of Economics and Finance 73: 44–55. Childhood Development in Rural Mozambique.” World Bank Group & Save the Children. Inter Agency Standing Committee. 2015. “Guidelines https://www.savethechildren.org/content/ for Integrating Gender-Based Violence dam/global/reports/education-and-child- Interventions in Humanitarian Actions: protection/ecd-report-2012-moz.pdf. Reducing Risk, Promoting Resilience and Aiding Recovery.” Global Protection Cluster: Mujahid, Noreen, and Naeem Uz Zafar. 2012. “Economic Geneva. https://gbvguidelines.org/wp/ Growth-Female Labour Force Participation wp-content/uploads/2015/09/2015-IASC- Nexus: An Empirical Evidence for Pakistan.” Gender-based-Violence-Guidelines_lo-res.pdf. Pakistan Development Review 4(51): 565–85. International Conference of Labour Statisticians. Niethammer, C., T. Saeed, S. Mohamed, and Y. Charaf. 2013. “Resolution Concerning Statistics 2007. “Women Entrepreneurs and Access to of Work, Employment, and Labor Finance in Pakistan.” Women’s Policy Journal of Underutilization.” Resolution by the Harvard, John F. Kennedy School of Government 19th International Conference of Labour 4: 1–12. Statisticians. Geneva: International Labour Organization. Olivetti, C., and B. Petrongolo. 2016. “The Evolution of Gender Gaps in Industrialized Countries.” Annual Review of Economics 8: 405–34. 19 Pakistan Bureau of Statistics (PBS) 2017. “Population Tanaka, S. , and M. Muzones. 2016. “Policy Brief on and Housing Census, KP District 2017”. Female Labor Force Participation” in Pakistan. Government of Pakistan, Ministry of Statistics. ADB Brief No. 70, Asian Development Bank. https://www.pbs.gov.pk/sites/default/files// https://www.adb.org/sites/default/files/ population_census/KP%20District%20Wise. publication/209661/female-labor-force- pdf participation-pakistan.pdf. Pakistan Bureau of Statistics 2019. Employment World Bank. 2018. “Female Labor Force Participation Trends 2018. Government of Pakistan, Ministry in Pakistan: What Do We Know?” Policy Note. of Statistics. https://www.pbs.gov.pk/sites/ Pakistan Gender and Social Inclusion Platform default/files//Pakistan%20Employment%20 & Social Protection and Jobs teams, World Bank Trend%20%20Reprt%202018%20Final.pdf Group, Washington, DC. Pakistan Demographic and Health Survey (PDHS), World Bank. 2019. “Labor Force Aspirations, 2017–18. https://dhsprogram.com/pubs/pdf/ Experiences and Challenges for Urban, FR354/FR354.pdf. Educated Pakistani Women: Discussions in Four Metro Cities.” Pakistan Gender and Social Pakistan Public Administration Research Centre. Inclusion Platform and Center of Gender 2018–19. Annual Statistical Bulletin of Federal and Policy Studies (CGaPS), World Bank, Government Employees for 2018–19. https:// Washington, DC. establishment.gov.pk/SiteImage/Downloads/ Annual%20Bulletin%202018-19.pdf. Zaidi, Y., S. Farooq, et al. 2016. “Women’s Economic Participation and Empowerment in Pakistan— Sanauddin, Noor. 2015. “Proverbs and Patriarchy: Status Report 2016.” UN Women Pakistan, Analysis of Linguistic Sexism and Gender Islamabad. Relations Among the Pashtuns of Pakistan.” PhD thesis, University of Glasgow. http:// Zeb, Arooj, and Shahid Jan Kakakhel. 2018. “Impact of theses.gla.ac.uk/6243/. Formal Networks on Women Entrepreneurship in Pakistan Islamia College University, Seneviratne, P. 2019. “Married Women’s Labor Peshawar.” Journal of Islamic Thought and Supply and Economic Development: Evidence Civilization 8: 105–28. doi:10.32350/jitc.81.08. from Sri Lankan Household Data.” Review of Development Economics 23(2): 975–99. 20 21 22