FEMALE LABOR FORCE PARTICIPATION IN PAKISTAN: WHAT DO WE KNOW? Pakistan Gender and Social Inclusion Platform & Social Protection and Jobs teams FEBRUARY 2018 Pakistan’s development road map “Vision 2025” sets an ambitious target of an increase in female labor force participation (FLFP) from its current level of 25 percent to 45 percent by 2025. Women’s labor force participation is rising across the country; however, significant challenges remain. This Note ex- plores the dynamics of FLFP via analysis of the Enterprise Survey 2013, two rounds of the Labor Skills Survey (2013 and 2015), and multiple rounds of the Labor Force Survey. Results summarized here provide a picture of trends in FLFP in Pakistan since 1992, identify reasons for low FLFP and highlight key knowledge gaps. This Note, a collaborative product of the Pakistan Gender and Social Inclusion Platform and Social Protection and Jobs teams, is structured to complement the forthcoming Pakistan Jobs Diagnostic. It is also a precursor to an upcoming study, Women in the Workforce, that will collect primary qualitative and quantitative data on urban women’s labor force participation in urban Punjab, Karachi, Peshawar and Quetta. Patterns and trends in female labor 1992 to 25 percent in 2014 but was fueled mostly by unpaid work in rural areas (Figure 1). Yet, according force participation since 1992 FLFP in Pakistan doubled between 1992 and 2014. The gender gap is also diminishing. Yet, FLFP FIGURE 1: TRENDS IN LABOR FORCE PARTICIPATION remains low, particularly in urban areas, while it SINCE 1992 is rural unpaid work that has grown.1 Overall, only 100 half of Pakistan’s population is in the labor force (53.4 82.86 82.81 82.27 81.86 82.13 percent), as reported by the 2015 Labor Force Survey 80 82.17 82.53 81.37 81.33 (LFS).2 This represents a small increase from 1992, 70 when labor force participation (LFP) was 48.6 percent. Between 1992 and 2014, men’s LFP remained 60 Percent In the Labor Force unchanged, albeit at a high level of over 82 percent. Women’s LFP rate doubled from 13.3 percent in 50 40 30 1 Patterns described in this paragraph are analyzed in more 25.04 22.39 23.15 detail in the upcoming Jobs Diagnostic. 19.72 20 23.60 23.99 2 LFP rates should be interpreted with caution because of a 13.30 21.05 potential downward bias, particularly for women, as is discussed 10 14.24 in detail in the upcoming Jobs Diagnostic. One important driver of this downward bias is measurement error arising from a lack of 0 clarity about what constitutes “work” for women. In Pakistan this is important as women are often employed in non-standard forms of 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 2012 2014 work such as unpaid family work or home-based work. Downward bias can also occur because of misreporting: the LFS questions the Year male head of household, who may not be aware about all the eco- nomic activities carried out by women, or may under-report such Sex Men Women activities because of perceived biases against women working for pay. The ongoing Women in the Workforce study will address some of these shortcomings. Source: Authors' analysis of multiple Labor Force Surveys 82.81 82.27 81.86 82.13 82.86 to data from ILOSTAT, Pakistan’s female labor force 3 80 there is little doubt Still, 82.17 that FLFP follows 82.53 a rising 81.37 81.33 participation remains one of the lowest, not just in trajectory. Compared to cohorts born in the past, South Asia, but also globally. cohorts 70 today have higher levels of labor force participation.4 An analysis of women’s labor force Urban female labor force participation in Pakistan participation 60 at different ages across a period of four is particularly low and has risen little over the last decades shows that, for any age, women born more Percent In the Labor Force two and a half decades. While rural FLFP doubled recently 50 have higher levels of labor force participation from 16 percent to 32.9 percent over this period, ur- than did women of their age in prior cohorts. For in- ban FLFP rose only from about 7 percent to 11 per- stance, 40 about 15 percent of 30-year old women who cent. Even after taking into account a range of factors were born in 1970 were in the labor force (Figure 2). that influence FLFP in multivariate analysis, urban res- In contrast, 30 25 percent of 30-year old women who 25.04 idence is still significantly associated with lower FLFP. were born in 1980 were in the labor 22.39 23.15 19.72 force. Such23.60 co-23.99 Moreover, it has been a significant factor since at least 20 hort changes suggest that conditions of employment, 21.05 1999, with its effect increasing over time: in 1999, ur- norms 13.30and attitudes, and 14.24 women’s opportunities re- 10 ban women were 11 percent less likely to be in the la- lated to labor force participation are more favorable bor force than rural women but by 2014 urban women for women 0 today than was the case for similarly-aged were 21 percent less likely than rural women to be in women born 1996 ago. longer 1990 1992 1994 1998 2000 2002 2004 2006 2008 2010 2012 2014 the labor force (Appendix 1). While urban men also Year are less likely to be in the labor force than are rural FLFP has also been steadily rising among all age men, this effect is not as large as it is for women and in both groups over time, Sex Men urban and rural areas. Women has not increased over time. Young people, particularly between the ages of 20- 24 years, have seen some of the largest LFP increase between 1999 and 2015. However, at all ages, urban FIGURE 2: FLFP BY BIRTH COHORT 40 Born 1980 Born 1970 35 30 Percent of Women in the Labor Force 25 20 Born 1950 15 Born 1960 Born 1990 10 5 0 0 10 20 30 40 50 60 Age Source: Authors' analysis of multiple Labor Force Surveys 4 This study uses a partial cohort analysis as the LFS data avail- able for this study, as well as the length of time between survey years, are unevenly clustered over the period of analysis (1992- 3 www.ilo.org/ilostat 2014). 2 FIGURE 3: TRENDS IN FLFP BY AGE GROUP AND FIGURE 4: FLFP BY EDUCATION AND RESIDENCE RESIDENCE 45 Rural Urban 45 1999 2014 1999 2014 60 Rural Urban 40 1999 2014 55.5 1999 2014 60 40 55.5 50 35 Force 50 Force 35 Force 40 Force the Labor 30 37.9 the Labor 40 inLabor 30 37.9 inLabor 31.7 25 30 27.4 31.7 in the Rural 26.5 in the 25.5 of Women 30 of Women 25 2014 27.4 Rural 26.5 25.5 of Women 20 20 19.2 of Women 2014 17.6 20 20 19.2 Percent 17.6 Percent 15 10.1 11.3 10.0 10 Percent 8.1 7.2 Percent Rural 5.6 6.0 11.3 6.9 15 10.1 10.0 1999 10 8.1 Rural 7.2 6.9 10 5.6 6.0 1999 0 10 No Education No Education No Education No Education Primary Secondary Post Secondary Primary Secondary Post Secondary Primary Secondary Post Secondary Primary Secondary Post Secondary 0 5 Urban 2014 No Education No Education No Education No Education Primary Secondary Post Secondary Primary Secondary Post Secondary Primary Secondary Post Secondary Primary Secondary Post Secondary 5 Urban Urban 2014 0 1999 Urban 15-19 20-24 25-29 30-34 35-39 40-44 45-49 50-54 55-59 60-64 0 1999 Education Level 15-19 20-24 25-29 30-34 35-39 40-44 45-49 50-54 55-59 60-64 No Education Education Level Primary Secondary Post Secondary Age No Education Primary Secondary Post Secondary Age Source: Authors' analysis of multiple Labor Force Surveys Source: Authors' analysis of multiple Labor Force Surveys women are less likely to be in the labor force than are ucation levels, larger proportions of rural than urban rural women (Figure 3). women are in the labor force (Figure 4). Between 1999 and 2014, the largest increases in The gender gap in FLFP rates has decreased in LFP occurred only after post-secondary education, Balochistan, Khyber Pakhtunkwa (KP), Punjab for both men and women. In the 15-year period and Sindh provinces, with some variation across between 1999 and 2014, women’s labor force partic- province. Punjab has had higher levels of FLFP than ipation rose for all levels of education, but most sig- Sindh, KP and Balochistan since at least 1992. In fact, nificantly after post-secondary education. However, even after controlling for factors such as urban-rural multivariate analysis suggests that the magnitude of residence, education and household characteristics, the relationship of post-secondary education to LFP women in Sindh, KP and Balochistan are significantly appears to be diminishing for women: in 1999, women less likely to be in the labor force than are women in with post-secondary schooling were 22 percent more Punjab. In contrast, provincial differences are smaller likely to be in the labor force than women with no for- for men (Appendix 1). Punjab has also had the small- mal schooling, while by 2014 they were only 13 per- est gender gap5 in FLFP in all time periods, in both cent more likely. In contrast, over time the beneficial urban and rural areas. By 2014, Sindh had the largest effect of schooling for men’s labor force participation gender gaps overall and in urban areas, and KP had has increased at all levels of schooling (Appendix 1). the largest gender gaps in rural areas. Balochistan had the largest shifts, going from having the worst Post secondary education has a stronger positive overall gender gap to the second lowest among all effect on urban women’s LFP than for rural wom- provinces, behind only Punjab.6 KP made the largest en; yet, at all education levels rural women are strides in lowering the urban LFP gender gap such more likely to be in the labor force than are urban women. Urban women with post-secondary educa- tion are three times more likely to participate in the 5 The gender gap here is measured as the ratio of men’s to labor force than urban women with primary education women’s labor force participation, or the percent of male LFP only. In contrast, highly educated rural women are less divided by the percent of female LFP 6 However, LFS data from Balochistan is considered less than twice as likely as rural women with primary ed- reliable than from other provinces, so this pattern needs to be ucation alone to be in the labor force. Still, at all ed- considered cautiously. 3 FIGURE 5: GENDER GAPS IN LFP BY PROVINCE OVER Normative Barriers: Marriage, Mobility, TIME AND RESIDENCE Safety and Attitudes towards FLFP Residence Province Rural Urban Marriage appears to be increasingly associated 20 with lower levels of FLFP. Multivariate analysis of Labor Force Surveys over time suggests that being Gender gap Balochistan 10 married is increasingly associated with women’s la- bor force participation, even after the effect of other 0 20 related factors such as education or urban-rural and provincial residence is considered. While marriage Gender gap KP 10 was not significantly associated with FLFP in 1999, in both 2005 and 2014 married women were 7 percent 0 less likely to be in the labor force than were unmar- 20 ried women. On the other hand, across all three years Gender gap Punjab of measurement, married men were 8-to-10 percent 10 more likely to be in the labor force than unmarried 0 men (Appendix 1), perhaps because of pressure to 20 provide for a growing family. For women, marriage may bring constraints such as the increased respon- Gender gap Sindh 10 sibilities for childcare and housework, as well as in- creased constraints on mobility and ability to make 0 independent decisions. 1992 2014 1992 2014 Province MARRIAGE CAN INHIBIT FLFP Balochistan KP Punjab Sindh Source: Authors' analysis of multiple Labor Force Surveys that the urban gender gap in KP more than halved In both 2005 and 2014 married women were between 1992 and 2014 (Figure 5). This shift in ur- ban LFP is likely due to province-specific factors, for example, the natural disasters and security crises that 7% have prompted migration from rural to urban areas.7 Additionally, these crises brought increased attention of humanitarian agencies and government to women, leading possibly to greater access to education and economic opportunities that, in turn, facilitated higher urban labor force participation for women in KP.8 less likely to be in the labor force than were Barriers to female labor force unmarried women participation Data and analysis are limited on barriers to women’s Women are labor force participation, and – even more so – on the constraints that women face, once employed, in being 5% able to work effectively and to advance in their jobs. Here we deepen the understanding on some of the barriers to FLFP to the extent allowed by existing sur- veys, viz., the LFS, LSS and Enterprise Survey. 7 Zaidi Y., Farooq S. et al. 2016. “Women’s Economic Partic- ipation and Empowerment in Pakistan - Status Report 2016”. UN more likely to participate in the labor force if men in Women Pakistan. Islamabad, 57. the household agree that married women should be 8 Ibid, 156. allowed to work outside the home. 4 Limited mobility is also associated with women’s likely to be in the labor force, even after controlling for ability to participate in the labor force. A significant a range of other factors (Appendix 2). proportion of women respondents in the 2013 Labor Skills Survey (LSS) reported that they could not travel Safety concerns dampen women’s economic activ- alone for basic services, social reasons, or to the local ities. Closely related to mobility is the perception of market. For example, only about 30 percent of women safety when one is outside. Less than half the wom- could go to local markets or to a local health facili- en in the 2013 LSS reported that they feel safe walk- ty alone. About one-fifth of women reportedly nev- ing around in their neighborhood, whether during er go to the local market, while 13 percent say they the day or at any other time. This perception seems never go to local health facilities. This lack of mobility to matter when it comes to labor force participation: for women constrains their flexibility to travel to work women who feel safe walking alone outside in their and conduct business, thus affecting their labor force communities or neighborhoods are more likely to participation. Women with greater mobility, at least in work (17 percent) than those who do not feel safe (11 terms of being able to go to local markets alone or percent). The relationship is also confirmed by regres- accompanied, tend to be more likely to be in the labor sion analysis, which shows that women who feel safe force. Thus, 17 percent of women who could travel to walking alone in their community at least during the local markets alone are in the labor force compared to day are significantly more likely (3 percent) to be in 9 percent of women who reported they could never the labor force than women who do not feel safe, even go to the market. Multivariate analysis of the LSS con- after controlling for characteristics of women and their firms this pattern: women allowed to go to local mar- households (Appendix 2). Both safety concerns and kets alone and unaccompanied were 3 percent more lack of mobility also partially explain why women in Pakistan are disproportionately engaged in home- based work. LIMITED MOBILITY CONSTRAINS LABOR FORCE PARTICIPATION WOMEN’S PERCEPTION OF SAFETY Only about INFLUENCES THEIR LABOR FORCE PARTICIPATION 30% MOST WOMEN DO NOT FEEL SAFE OUTSIDE THE HOME In 2013, of women can go alone to local markets and/or to Less than half of surveyed local health facilities women reported that they feel safe walking around WOMEN WITH GREATER MOBILITY ARE MORE LIKELY in their neighborhood, whether during the day or TO BE IN THE LABOR FORCE at any other time 17% of women who can travel to of women who local markets alone are in 17% feel safe walking the labor force alone outside their communities or compared to 9% neighborhoods are more likely to work, compared to 11% of those who are never allowed to go to the of those who do not local market alone feel safe 5 FIGURE 6: ATTITUDES TOWARDS WOMEN WORKING Household Constraints: The Burden of OUTSIDE THE HOUSE Housework and Childcare Married women should be allowed If wife works, husband should A range of housework and childcare responsibil- to work outside the home help with housework ities inhibit women’s ability to work outside the All Rural Urban All Rural Urban home, even in urban areas. Most men work outside 82.3 80 79.3 79.0 the home and the primary reason for men not to be in 77.1 the labor force is because they are students. This pat- Percent of Respondents Who Agree With Each Statement 70.7 72.2 70 65.8 66.5 tern is little changed since 1999. In contrast, the major- 60.1 63.3 ity of women who are not in the labor force attribute 60 their absence to housework. This pattern too has hard- 51.3 50 ly changed, declining only from 89 percent in 1999 to 46.4 83 percent of women who do not work in 2014 citing 40 housework as the main reason. Even among women who work, the Jobs Diagnostic shows that over 61 per 30 cent of women in urban and 45 per cent in rural areas 20 work from their dwelling, likely because of the contin- ued pressure of household and reproductive tasks. 10 While affording them flexibility in terms of hours, 0 working from home limits the type of jobs women can 2014 take and thus negatively impacts their upward mobil- Men Women Men Women Men Women Men Women Men Women Men Women ity and income. HOUSEWORK AND CHILDCARE Sex Men Women Source: Authors' analysis of the Labor Skills Survey 2013 RESPONSIBILTIES OFTEN PREVENT Attitudes towards women working outside of the WOMEN FROM WORKING OUTSIDE home seem to be somewhat favorable, though more so among women than men. More than half THE HOME of all urban and rural men and about 70 percent of 83% women who participated in the 2013 LSS agreed that married women who want to work outside the home should be allowed to do so. Urban women were the most supportive of this statement of all surveyed groups. Also, 66.5 percent of men agreed that if a wife works outside the home the husband should help of women who do not work with the housework. At the same time, men are less outside the home cite housework likely than women to agree with both statements, es- as a reason pecially regarding women working outside the home (Figure 6). Even among those who work for pay 61% Yet, it appears that while men’s support for wom- en’s employment is an enabling factor for female labor force participation, women’s own opinions on the matter do not make a significant difference. Regression analysis using the LSS shows that women in urban and are 5 percent more likely to participate in the labor 45% force if men in their household agree with the state- ment “a married woman should be allowed to work outside the home if she wants to”. This relationship is also robust across different model specifications. On the other hand, we do not observe a significant rela- in rural areas tionship between women’s opinion on this statement and their own labor force participation (Appendix 2). work from home, limiting upward mobility 6 FIGURE 7: REASONS FOR NOT WORKING ACROSS THE FIGURE 8: REASONS FOR NOT WORKING ACROSS THE LIFE CYCLE (WOMEN) LIFE CYCLE (MEN) Age Group Age Group 100 100 90 90 80 80 Percent Reporting a Reason Percent Reporting a Reason 70 70 60 60 50 50 40 Age Group Age Group 40 100 100 30 90 30 90 20 80 20 80 10 Percent Reporting a Reason Percent Reporting a Reason 70 10 70 0 60 0 60 15-19 20-24 25-29 30-34 35-39 40-44 45-49 50-54 55-59 60-64 15-19 20-24 25-29 30-34 35-39 40-44 45-49 50-54 55-59 60-64 50 50 40 Reason Housework Student Reason Housework Student 40 30 Source: Authors' analysis of multiple Labor Force Surveys 30 Authors' analysis of multiple Labor Force Surveys Source: 20 20 10 While men’s reasons for not working change across straints. 10 In rural areas, the proportion Year reporting educa- the0life cycle, women not in the labor force simply tion Urban/Rural 0 as a reason for not 1999 working tripled between 2014 1999 transition from school to housework. For men not in and 2014. Still, even in 2014 housework constrains a 15-19 20-24 25-29 30-34 35-39 40-44 45-49 50-54 55-59 60-64 15-19 20-24 25-29 30-34 35-39 40-44 45-49 50-54 55-59 60-64 the labor force, the reasons change by life stage. For large three-quarters8.80 of young rural women from join- example, in the 2015 Reason LFS, for men Housework in their teens and Student ing the labor forceReason (Figure 9). 9 Housework 24.13 Student 20s the primary reason for not being in the labor force Rural is schooling. About one-quarter of men between the FIGURE 9: REASONS 89.19 FOR NOT WORKING ACROSS THE ages of 30 and 40 years report that they are not in the LIFE CYCLE (WOMEN AGED 15-24)9 73.66 labor force because of housework responsibilities. Year Beyond age 40, by and large men outside of the labor Urban/Rural 1999 2014 force and not students are either disabled or retired. In contrast, until about age 20 most women outside 38.888.80 44.70 Urban the labor force are students. Between 20-24 years of 24.13 age, still about 11 percent of such women are stu- 59.74 53.34 dents while 86 percent are housewives. After age 25, Rural 89.19 almost all women not in the labor force are reportedly 73.66 housewives (Figures 7 and 8). Reasons For Not Being in The Labor Force Housewife Student There does appear to be some generational change, with housework reportedly less of a con- 38.88 straint to joining the labor force among young Urban 44.70 women in 2014 compared to 1999. Urban-rural 59.74 53.34 patterns differ, however. Among young women be- tween the ages of 15 and 24 outside of the labor force, the importance of housework as the main constraint Reasons For Not Being in The Labor Force Housewife Student shows some – albeit slow – decline between 1999 and 2014, while the role of education as the main reason Source: Authors' analysis of multiple Labor Force Surveys for not being in the labor force is rising. In 1999 a little over one-third of urban young women outside the la- 9 The pie charts do not add up to a 100% as they only repre- bor force gave education as their constraint and about sent "housework" and "student" as reasons. The minor remaining 60% reported housework. By 2014, these two reasons percentage highlights other reasons such as "disabled" and were closer to being more equally reported as con- "retired." 7 Human Capital and Economic Barriers: Regardless of education level women have few- Education, Occupation, Firm Preferences er, and less diversified, occupational choices than men. Over 50 per cent of women are engaged in and Workplace Laws unpaid work, primarily driven by the over-concentra- More than half of all Pakistani women have not at- tion in agriculture, according to the Jobs Diagnostic. tended school. Only a tiny minority has had access Additional analysis of the LFS for this Note further to higher education. This scenario means a shaky shows that it takes several years of schooling to en- foundation on which to build skills and training able women to move out of low-skilled and unpaid for high-quality employment, and low returns for occupations. In fact, women are concentrated in low working women. Over the last two and a half de- skilled sectors through secondary levels of education. cades the proportion of women with no education It is only at post-secondary education levels that we has declined from over three-quarters of all women see a shift of women into the professional sector, from (77 percent) to slightly more than half of all working about 17 percent of employed women with a sec- age women – still a high proportion of uneducated. ondary education reportedly working in professional The percent of men with no formal education has de- occupations to 82 percent of employed women with clined from about 47 percent in 1992 to just over 25 post-secondary schooling. However, options other percent of men in 2014. The gender gap in education, than either low-level (primarily agriculture-related) however, has not improved over this period. Further, work or high-level professional work are sorely lack- while few men have post-secondary education, this ing for women. This is in sharp contrast to men, who proportion is even smaller for women: the proportion enjoy several middle-skilled positions across all levels of women with post-secondary schooling in 2014 was of education, such as machine operation, clerks, mar- roughly equivalent to men’s post-secondary school- keting and services. ing in 1999. In urban areas, even by 2014, about one- third of all women surveyed in the LFS had no formal education and only 10 percent had post-secondary schooling. Moreover, as seen in the Jobs Diagnos- EVEN WHEN IN THE LABOR FORCE, tic, skills training that could potentially complement WOMEN FACE LIMITED OCCUPATIONAL a weak formal education system typically follows tra- ditional gendered patterns and is less diversified for DIVERSITY women than men, further limiting women’s opportu- nities in the job market. LIMITED HUMAN CAPITAL CONSTRAINS Women with WOMEN'S LABOR FORCE PARTICIPATION primary education cluster in low level agriculture related In 2014 about work 1/3 of all urban women surveyed in the LFS had no formal education while those with post-secondary schooling are and only in high-level professional work 10% had post-secondary schooling but women have limited diversity of occupations in the middle levels of educational attainment 8 OCCUPATIONAL PROFILES VARY BY SEX The top five categories of occupations make up Women are doing “feminine” jobs such as domestic work or in textiles 2/3 and apparel whereas men are doing “masculine” jobs in construction and of employment for urban women whereas there is 100 services 100 more diversity for urban men 90 90 80 80 70 70 In urban areas the top five categories of occupa- they 60 find it difficult to hire women because of wom- tions make up two-thirds or more of the share of en’s family responsibilities. Almost one-third of the Percent reporting a reason Percent reporting a reason 60 employment for women, whereas there is more surveyed 50 firms believe that having women employ- 50 diversity for men. In rural areas, this difference is ees40 ‘disrupts’ the workplace, possibly because male 40 more muted. Occupation profiles across the board colleagues and customers are hesitant to interact with 30 vary distinctly by sex. Urban men are more likely to women. An equivalent proportion deem that benefits 30 be engaged in trades such as construction and ser- and20 expenses on separate workplace facilities make 20 vices (shopkeepers) whereas urban women are en- women 10 more expensive employees. Finally, slightly gaged in traditionally ‘feminine’ jobs such as domes- less than one-third of the surveyed firms reportedly 10 tic help or in apparel and textiles. At higher levels of consider 0 it difficult to hire women because of govern- 0 education, the education sector features as the largest ment regulations on working hours for women and 15-19 15-19 20-24 25-29 30-34 35-39 40-44 45-49 50-54 55-59 60-64 employer for urban women whereas urban men work maternity leave (Figure 10). across a varied number of service-oriented jobs such Reason Housework Student as shop-keeping, accounting and clerks in addition FIGURE 10: GENDER BIASED REASONS FOR NOT HIRING to education. There is limited representation of urban WOMEN women in services and retail. Yet, the top five occupa- tions even for urban men with post-secondary school- Gender biased Reasons Urban/Rural cited by Firms ing are still middle-range, non-white collar occupa- tions. In rural areas occupational choice is limited for Disruption in work 30 32 environment both men and women regardless of education level. At the same time, the limitations seem to be larger for Family commitments 36 35 Rural women than for men in rural areas also. The top five Govt. regulations like rural occupations comprise three-quarters or more of maternity leave 29 29 women’s employment regardless of education level, Few women with compared to half to two-thirds of men’s employment managerial expertise 28 (Appendix 3). Women more expensive employees 30 34 Firms themselves express a preference for not hir- Urban ing women. About two-thirds of firms surveyed in 0 10 20 30 40 50 60 70 Pakistan’s Enterprise Survey in 2013 agree with gen- Percent of Surveyed Firms der-discriminatory attitudes as reasons for not hiring Type of employee Managerial Non Managerial women in managerial or non-managerial capacities. More than one-third of the surveyed firms report that Source: Authors' analysis of Enterprise Survey 2013 9 STRUCTURAL BARRIERS THAT CONSTRAIN WOMEN'S LABOR FORCE PARTICIPATION DISCRIMINATORY ATTITUDES OF FIRMS WOMEN’S AWARENESS OF WORKPLACE LAWS Age Group 100 About Only about 90 2/3 1/3 80 70 Percent reporting a reason 60 50 40 of firms surveyed agree with gender-discriminatory of women are aware of maternity leave, and even 30 attitudes as reasons for not hiring women in smaller numbers are aware of other laws that could managerial 20 or non-managerial capacities help ease constraints on working outside the home, 10 such as a 48-hour work week 0 15-19 20-24 25-29 30-34 35-39 40-44 45-49 50-54 55-59 60-64 55-59 60-64 A minority of women are aware of workplace laws of benefits such as sick leave and annual leave com- on entitlements that might attenuate the conflict pared to other benefits. LSS 2015 reveals that 35-45 between household work responsibilities and Housework Reason Student and percent of women are aware of maternity, annual and ease constraints on working outside the home. sick leave, but very few are aware of other laws that Both men and women of working age - whether cur- could help ease constraints on working outside the rently employed or not - are more likely to be aware home, such as a 48-hour work week (Figure 11). Even in urban areas, more than 40 percent of women sur- FIGURE 11: AWARENESS OF LEGAL WORKPLACE veyed are unaware of any workplace laws. Without such awareness, working outside the home may be an BENEFITS BY SEX even more daunting and burdensome possibility for 100 women who still have to shoulder the responsibility of all household work while maintaining a job outside Percent of Respondents Who Know About Legal Workplace Benefits 90 the home of the home. At the same time, these laws apply primarily to the formal sector, in which female 80 participation is very low. Thus, even growing aware- 50.60 ness among women can only contribute to increasing 70 60 70 48.80 FLFP if these regulations are applied more broadly 60 and if firms comply with the regulations. 40.20 50 Gaps in data on female labor force 40 participation 30 Despite the large number of surveys conducted 28.10 29.50 on FLFP and the knowledge generated thereby on 44.10 45.80 20 20.00 23.10 trends, patterns, and barriers, several aspects of FLFP 35.90 remain unknown. In particular, more accurate mea- 10 sures of women’s work and a deeper understanding 11.40 8.90 9.10 9.60 of the dynamics of individual, household, firm and 0 Old Age Social Minimum 48-hour Maternity Annual Sick leave structural barriers are still lacking. Benefits security wage work week leave leave Existing surveys typically underestimate women’s Sex Men Women work, and more experimentation is needed to im- Source: Authors' analysis of Labor Skills Survey 2015 prove ways to measure all of women’s economic 10 contributions inside and outside the home. The to- into the role of husbands and other gatekeepers (par- tal amount of women’s work, such as “unpaid house- ents, parents-in-law, community elders, to name a few) hold care and services,” is typically excluded from in determining women’s labor force participation, the economic accounting and thus the full nature of wom- importance of actual experiences of safety and vio- en’s economic contribution is under-estimated. Efforts lence as barriers to FLFP, and the dynamics of other are ongoing internationally to find better ways to es- social and gendered norms such as the stigma associ- timate the economic value of such work.10 Additional ated with working outside the home, the importance in-country experimentation can test different methods of purdah, and so on. to better count women’s work in the specific context of Pakistan so that we can more fully understand their A likely critical barrier – sexual harassment on the economic participation. way to and in the workplace – remains woefully under-researched. Pakistan has strong laws against There is very limited data on women’s aspirations workplace sexual harassment in place and the insti- for their lives, how labor force participation fits tutional arrangements to implement them. Yet, there into these aspirations, and how aspirations cor- is limited understanding, practically no national data, relate with opportunities. Why women are not in and little analysis of sexual harassment at work that the labor force – or why even educated women have women might face. In particular, there is no nationally limited occupational choices – may in part relate to – or for that matter internationally – standardized defi- women’s aspirations and expectations. For instance, it nition or measurement of sexual harassment related is possible that women’s aspirations for labor market to work outside the home. participation may not align with current or potential employment opportunities in urban or rural areas. There is limited data and analysis on the gender However, there is next to no data on the aspirations dynamics in the workplace that influence women’s of women in different socioeconomic strata for their performance at work. Research on FLFP to-date has lives as economic actors, such as what types of work focused primarily on barriers to entry into the labor women want to do and why, and what motivates these force. However, there is less understanding of the aspirations. gendered barriers that employed women may face, and that may negatively impact their work perfor- A range of structural and related barriers remains mance and upward mobility. An exploration of possi- to be analyzed because of lack of nationally rep- ble enablers, such as mentors and role models, or of resentative data. Examples include the status of and initiatives to facilitate crossovers to male-dominated importance of appropriate housing for women, wom- sectors, is even more limited. Moreover, the nature en’s perceptions of and experiences with transporta- of constraints and enablers will vary by the nature of tion to work, women’s sources of information about the type, levels, and sector of employment being dis- jobs, and women’s knowledge of laws on employ- cussed, adding to the nuance required to collect and ment benefits and rules. analyze such data. Our analysis suggests that household attitudes Similarly, there is a need to focus attention on the and behavior, and social norms, play an import- relationship between women’s labor force par- ant role in determining whether, what, when, and ticipation and their own wellbeing. All labor force how women can work for pay. However, existing participation does not automatically improve wom- national datasets include only a few indicators to en’s lives in all dimensions. In fact, there are possibly measure such factors. For instance, the multivariate certain types of labor force participation that, while analysis of the LSS strongly suggests that husbands’ increasing household income, may lower women’s attitudes play an important role in women’s labor force wellbeing, for instance their health. A potential ‘dou- participation, as do mobility and perceptions of safety. ble burden’ of continuing household work together Additional representative data needs to delve further with work outside the home may also adversely affect women’s wellbeing. This makes it important to under- 10 See, for example, Donehower, Gretchen, Alexia Fürnk- stand the time use dynamics of not just women out of ranz-Prskawetz, Ronald D. Lee, Sang-Hyop Lee, Andrew Mason, Tim the labor force, as afforded by the Labor Force Sur- Miller, Germano Mwabu, Naohiro Ogawa, and Adedoyin Soyibo. veys, but also of economically active women. On the 2017. “Counting Women’s Work: Measuring the gendered econo- other hand, certain kinds of employment are likely to my in the market and at home.” National Transfer Accounts Bulletin No. 11. http://ntaccounts.org/doc/repository/NTA%20Bulletin%20 improve women’s wellbeing and empowerment by 11.pdf increasing their decision-making power and status in 11 the household. Better data and analysis is needed to select metropolitan areas in urban Pakistan, specifical- understand the conditions under which labor force ly, Lahore and other urban areas in Punjab province, participation can improve women’s wellbeing in the Karachi in Sindh province, Peshawar in KP province Pakistani context. and Quetta in Balochistan province. Finally, we know little about how conflict and un- The Women in the Workforce study will address certainty impact women’s aspirations, opportuni- barriers at different points of the job cycle, name- ties and experience of labor force participation ly, barriers that women seeking to get into the la- and employment. Terrorism, other forms of con- bor market face, as well as barriers that employed flict, and consequent insecurity have ramifications women have to deal with in their daily work and for women’s role as economic actors. While conflict in efforts to improve their jobs and careers. A key inflicts suffering on everyone, women are particular- aspect of this analysis will be a more nuanced un- ly affected in ways that influence not just their overall derstanding of social and gender norms influencing safety and quality of life but also their need for and women's economic participation. We will also gain a potential to engage in meaningful work. Women better understanding of women’s own aspirations for living in conflict-prone and insecure environments economic participation by asking questions such as: struggle to support their families when the traditional What kinds of jobs and working situations do wom- breadwinners – husbands and sons – are caught up en ideally want? What do women perceive as the key in fighting, or are dead. However, space to challenge barriers to working outside the home, and consider entrenched gender norms may open up and create workable solutions to address these barriers? Gender new roles and opportunities for women – including in dynamics in the workplace will be measured and an- the labor market - if conflict upsets the existing social alyzed keeping in mind that such constraints differ for order. Understanding women’s experiences of insecu- women in different levels and types of occupations rity and conflict in different parts of Pakistan and its and professions. Finally, we will seek to investigate implications for women's hopes and opportunities for overarching issues related to labor force participation engaging in meaningful work is crucial. for all women, such as the conditions under which labor force participation can enhance women’s well- What’s next? being, or the role that conflict and uncertainty play in household’s and women’s own decision-making An ongoing study on urban FLFP in Pakistan, the choices for economic engagement. Women in the Workforce study, seeks to address many of the gaps in data and analysis identified Finally, the Women in the Workforce study will ex- above. The Pakistan Gender and Social Inclusion Plat- periment with ways to better measure women’s form, in collaboration with the Poverty Global Prac- work. While women’s economic production activities tice, is preparing to field a qualitative and quantitative for market use may be measured by surveys, women study to collect data on several of the key issues for in Pakistan and other countries are also engaged in a which we do not yet have adequate information. Giv- large amount of “household production” work that is en the particularly low levels of female labor force par- rarely, if ever, measured. The upcoming study will field ticipation within urban areas, the study will focus on different ways to try and measure such contribution to the household by women in urban areas. This Note was co-authored by Saman Amir, Andy Kotikula and Rohini Prabha Pande from the Social De- velopment Team (Gender and Social Inclusion Platform for Pakistan), and Laurent Loic Yves Bossavie and Upasana Khadke from the Social Protection and Jobs team. Duina Reyes provided support in formatting, layout and graphics. For additional information please contact: Maria Beatriz Orlando,  morlando@world- bank.org and Uzma Quresh,  uquresh@worldbank.org (co-Task Team Leaders of the Pakistan Gender and Social Inclusion Platform) or Victoria Strokova, vstrokova@worldbank.org (Task Team Lead Jobs Diagnostic, Social Protection and Jobs). 12 APPENDIX 1: TRENDS IN FACTORS ASSOCIATED WITH FEMALE LABOR FORCE PARTICIPATION: 1999-2014 WOMEN MEN 1999 2005 2014 1999 2005 2014 Area of residence Urban (2013) -0.11*** -0.14*** -0.21*** -0.04*** -0.04*** 0.04*** (0.01) (0.01) (0.01) (0.00) (0.00) (0.00) Province (reference: Punjab) Sindh -0.11*** -0.16*** -0.14*** -0.02*** 0.01*** 0.03*** (0.01) (0.01) (0.01) (0.00) (0.00) (0.00) Khyber Pakhtunkhwa -0.10*** -0.18*** -0.18*** -0.06*** -0.05*** 0.04*** (0.02) (0.02) (0.01) (0.01) (0.01) (0.00) Balochistan -0.15*** -0.16*** -0.15*** -0.06*** 0.00 0.00 (0.01) (0.01) (0.02) (0.01) (0.01) (0.01) Highest education level achieved (reference: no education) Primary -0.07*** -0.06*** -0.11*** -0.02*** -0.02*** 0.01*** (0.01) (0.01) (0.01) (0.00) (0.00) (0.00) Secondary -0.03*** -0.10*** -0.17*** -0.16*** -0.14*** 0.14*** (0.01) (0.01) (0.01) (0.01) (0.01) (0.01) Post-secondary 0.22*** 0.12*** 0.13*** -0.15*** -0.15*** 0.09*** (0.03) (0.02) (0.01) (0.01) (0.01) (0.01) Marital status Married 0.01 -0.07*** -0.07*** 0.10*** 0.08*** 0.10*** (0.02) (0.01) (0.00) (0.01) (0.00) (0.00) Household characteristics Household size -0.01*** -0.00 0.00 0.01 0.01*** 0.00*** (0.00) (0.00) (0.00) (0.00) (0.00) (0.00) Number of children under age 5 years -0.00 -0.00 -0.00 0.00 0.00 0.00*** (0.00) (0.00) (0.00) (0.00) (0.00) (0.00) Any elderly household members over 64 years -0.00 0.00 0.01** 0.00 -0.01** 0.01*** (0.01) (0.00) (0.00) (0.00) (0.00) (0.00) Age Age 0.03*** 0.02*** 0.02*** 0.05*** 0.05*** 0.05*** (0.00) (0.00) (0.00) (0.00) (0.00) (0.00) Age-squared -0.00*** -0.00*** -0.00*** -0.00*** -0.00*** 0.00*** (0.00) (0.00) (0.00) (0.00) (0.00) (0.00) Total observations 28,273 58,177 71,651 28,988 59,235 71,749 Outcome: Female Labor Force Participation; logit regression with marginal effects Standard errors in parentheses; *** p<0.01, ** p<0.05, * p<0.1 Source: Authors’ analysis of Labor Force Surveys. 13 APPENDIX 2: NORMATIVE FACTORS ASSOCIATED WITH FEMALE LABOR FORCE PARTICIPATION: 2015 Women’s attitudes on household decision-making and women’s work (2013) -0.01 Important decisions in the family should be made by men (0.017) 0.01 A married woman should be allowed to work outside the home if she wants to (0.020) Men’s attitudes on household decision-making and women’s work (2013) -0.01 Important decisions in the family should be made by men (0.021) 0.05*** A married woman should be allowed to work outside the home if she wants to (0.018) 0.00 If a woman is working outside the home, her husband should help with chores (0.020) Women’s mobility and perceptions of safety (2013) 0.03* Respondent (woman) is allowed to go to the local market alone (unaccompanied) (0.019) 0.03** Respondent (woman) feels safe walking alone outside in her community in the day (0.017) Respondent’s age, marital status and residence -0.00 Current age (0.001) 0.01 Currently married (0.022) -0.04** Urban residence in 2013 (0.020) Household wealth (reference: poorest wealth quintile) (2013) 0.00 Respondent’s household is in 2nd poorest wealth quintile (0.031) -0.07** Respondent’s household is in 3rd wealth quintile (0.029) -0.06** Respondent’s household is in 4th wealth quintile (0.031) -0.10*** Respondent’s household is in highest wealth quintile (0.032) Respondent’s highest level of education (reference: no education) and cognitive score (2013) -0.03 Primary education (0.024) -0.03 Middle school education (0.030) 0.08** High school and above (0.035) 0.00** Raven’s cognitive score (0.001) Total observations 1,608 Outcome: Female Labor Force Participation; logit regression with marginal effects Standard errors in parentheses; *** p<0.01, ** p<0.05, * p<0.1 Source: Authors’ analysis of Labor Skills Surveys 2013 and 2015 14 APPENDIX 3: OCCUPATIONAL SEGREGATION BY EDUCATION NO EDUCATION   URBAN MALE URBAN FEMALE   RURAL MALE RURAL FEMALE   Building, construction Mixed crop and animal 10.96 Domestic cleaners and helpers 19.68 26.58 Livestock and dairy producers 42.85 labourers producers Tailors, dressmakers, furriers Field crop and vegetable Mixed crop and animal Shop keepers 9.6 19.17 14.58 20.28 and hatters growers producers Sewing, embroidery and Building, construction Street food salespersons 5.18 12.58 9.43 Crop farm labourers 13.04 related workers labourers Manufacturing labourers not Field crop and vegetable 3.68 Livestock and dairy producers 12.52 Livestock and dairy producers 7.06 9.22 elsewhere classified growers Mixed crop and animal Tailors, dressmakers, furriers Car, taxi and van drivers 3.54 4.49 Crop farm labourers 5.98 3.36 producers and hatters   32.96 68.44   63.63 88.75 Primary Education   Tailors, dressmakers, furriers Mixed crop and animal Shop keepers 14.34 37.8 21.66 Livestock and dairy producers 43.3 and hatters producers Building, construction and Sewing, embroidery and Field crop and vegetable Mixed crop and animal 7.36 14.45 10.86 15.05 repair related workers growers producers Building, construction Tailors, dressmakers, furriers Shop sales assistants 5.72 Domestic cleaners and helpers 10.34 9.16 12.32 labourers and hatters Tailors, dressmakers, furriers 4.62 Livestock and dairy producers 6.59 Shop keepers 7.96 Crop farm labourers 10.51 and hatters Motor vehicle mechanics and Handicraft workers in textile, Field crop and vegetable 4.06 3.29 Livestock and dairy producers 4.53 4.38 repairers leather and related materials growers   36.1 72.47   54.17 85.56 Secondary Education   Tailors, dressmakers, furriers Mixed crop and animal Shop keepers 19.73 27.78 22.62 Livestock and dairy producers 28.88 and hatters producers Tailors, dressmakers, furriers Shop sales assistants 7.33 Primary school teachers 19.1 Shop keepers 11.96 15.2 and hatters Tailors, dressmakers, furriers Sewing, embroidery and Field crop and vegetable Mixed crop and animal 4.43 8.98 7.41 13.85 and hatters related workers growers producers Building, construction Car, taxi and van drivers 3.85 Secondary education teachers 6.05 4.78 Primary school teachers 9.12 labourers Motor vehicle mechanics and Beauticians and related 2.6 3.46 Livestock and dairy producers 3.29 Crop farm labourers 6.94 repairers workers   37.94 65.37   50.06 73.99 Post-Secondary Education   Shop keepers 10.79 Secondary education teachers 35.75 Primary school teachers 16.66 Primary school teachers 42.92 Accountants 6.09 Primary school teachers 27.58 Secondary education teachers 12.01 Secondary education teachers 28.95 Secondary education University and higher educa- Mixed crop and animal 5.91 5.62 8.45 Livestock and dairy producers 5.07 teachers tion teachers producers Generalist medical practi- Primary school teachers 4.58 4.57 Shop keepers 8.24 Teaching professional n.e.c. 4.91 tioners Teaching professionals not Field crop and vegetable University and higher educa- General office clerks 3.84 3.68 4.05 2.05 elsewhere classified growers tion teachers   31.21   77.2   49.41   83.9 15