thesis on gender discrimination in pakistan

• Research article
• Open access
• Published: 22 December 2022

Perpetuation of gender discrimination in Pakistani society: results from a scoping review and qualitative study conducted in three provinces of Pakistan

• Tazeen Saeed Ali   ORCID: orcid.org/0000-0002-8896-8766 1 , 2 ,
• Shahnaz Shahid Ali 1 ,
• Sanober Nadeem 3 ,
• Zahid Memon 4 ,
• Sajid Soofi 4 ,
• Falak Madhani 3 ,
• Yasmin Karim 5 ,
• Shah Mohammad 4 &
• Zulfiqar Ahmed Bhutta 6 , 7  

BMC Women's Health volume  22 , Article number:  540 ( 2022 ) Cite this article

41k Accesses

4 Citations

8 Altmetric

Metrics details

Gender discrimination is any unequal treatment of a person based on their sex. Women and girls are most likely to experience the negative impact of gender discrimination. The aim of this study is to assess the factors that influence gender discrimination in Pakistan, and its impact on women’s life.

A mixed method approach was used in the study in which a systematic review was done in phase one to explore the themes on gender discrimination, and qualitative interviews were conducted in phase two to explore the perception of people regarding gender discrimination. The qualitative interviews (in-depth interviews and focus group discussions) were conducted from married men and women, adolescent boys and girls, Healthcare Professionals (HCPs), Lady Health Visitors (LHVs) and Community Midwives (CMWs). The qualitative interviews were analyzed both manually and electronically through QSR NVivo 10. The triangulation of data from the systematic review and qualitative interviews were done to explore the gender discrimination related issues in Pakistan.

The six major themes have emerged from the systematic review and qualitative interviews. It includes (1) Status of a woman in the society (2) Gender inequality in health (3) Gender inequality in education (4) Gender inequality in employment (5) Gender biased social norms and cultural practices and (6) Micro and macro level recommendations. In addition, a woman is often viewed as a sexual object and dependent being who lacks self identity unless being married. Furthermore, women are restricted to household and child rearing responsibilities and are often neglected and forced to suppress self-expression. Likewise, men are viewed as dominant figures in lives of women who usually makes all family decisions. They are considered as financial providers and source of protection. Moreover, women face gender discrimination in many aspects of life including education and access to health care.

Gender discrimination is deeply rooted in the Pakistani society. To prevent gender discrimination, the entire society, especially women should be educated and gendered sensitized to improve the status of women in Pakistan.

Peer Review reports

Gender discrimination refers to any situation where a person is treated differently because they are male or female, rather than based on their competency or proficiency [ 1 , 2 ]. Gender discrimination harms all of society and negatively impacts the economy, education, health and life expectancy [ 1 , 2 ]. Women and girls are most likely to experience the negative impacts of gender discrimination. It include inadequate educational opportunities, low status in society and lack of freedom to take decisions for self and family [ 1 , 3 ].

Likewise, gender discrimination is one of the human rights issues in Pakistan and is affecting huge proportion of women in the country [ 1 , 2 ]. In Pakistan, nearly 50% of the women lacks basic education [ 4 ]. In addition, women in Pakistan have lower health and nutritional status. Furthermore, most of the women are restricted in their homes with minimal or no rights to make choices, judgments, and decisions, that directly affect their living conditions and other familial aspects [ 2 ]. In contrast, men are considered dominant in the Pakistani society [ 5 ]. This subordination of women has negative influences on different stages of women’s life.

Study design

The mixed method study design was used. Systematic review was done in phase one and qualitative interviews; in-depth interviews (IDIs) and focus group discussions (FGDs) were conducted in phase two.

The objective of the systematic review

To map a broad topic, gender discrimination/inequality research in Pakistan including women undergoing any form of intimate partner violence.

Systematic review

The three authors (TSA, SSA and SN) independently performed an extensive literature search using two databases: PubMed and Google Scholar and reports from organizations such as WHO and the Aurat Foundation. Quantitative and Boolean operators were used to narrow down the search results. The following keywords and phrases were used: Intimate partner violence (IPV), domestic violence, violence against women, domestic abuse, spousal violence, and Pakistan. Articles from 2008 to 2021 were assessed. The selection criteria of the articles included: women undergoing any form of IPV (physical, psychological, and sexual); quantitative study design; English as the publication language; and articles in which Pakistan was the study setting. The shortlisted articles were cross-checked by two of the authors (TSA, and SN) for final selection. The quality of the selected articles was reviewed using a STROBE (Strengthening the Reporting of Observational Studies in Epidemiology) checklist, which ensured all articles followed a structured approach, including an introduction, methodology, results, and a discussion section. It was also determined that all selected articles are published in peer-reviewed journals and have been used nationally or internationally. The Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) chart was used for study selection (Fig.  1 ).

PRISMA Diagram to select the final articles

The selected articles were approved by one of the authors (TSA), who is an expert in the field of IPV. Articles were excluded: (i) If the study was not conducted in Pakistan; (ii) Studied spousal violence against men and (iii) Domestic violence involving in-laws or other family members. Furthermore, from the selected articles, the data were extracted by 3 authors (TSA, SSA, SN) by carefully studying the methodology and results. The methodology was entered into an extraction template in which location was summarized including the study design and sample size in the articles. The results covered: (i) The title, (ii) Authors, (iii) Publication year, (iv) Objectives of the research, (v) Population and Setting, (vi) Research design, (vii) Data collection methods, (ix) Results, (x) Perpetuating factors (xi) Recommendations and (xii) prevalence of Intimate Partners Violence (IPV) faced by women, which was further categorized into: (a) Psychological/emotional violence, (b) Physical violence, (c) Sexual violence, (d) Both combined and (e) Violence of any other type.

Qualitative data collection

Participants selection.

Purposeful sampling was done to recruit the participants for qualitative data collection. Participants included groups of married men and women aged between 18 to 49 years, groups of unmarried adolescent boys and girls aged between 14 to 21 years, and groups of healthcare professionals (HCPs), comprising of doctors, nurses, Lady Health Visitors (LHVs), Lady Health Workers (LHWs) and Community Midwives (CMWs). Ethics approval was obtained from the Aga Khan University, Ethics Review Committee.

Study sites

The selected study sites included two districts from Chitral (Upper and Lower Chitral), six districts from Gilgit (Gilgit, Ghizer, Hunza, Nagar, Astore, and Skardu), and two districts from Sindh (Matiari and Qambar Shadadkot). The following are the details of the data collection (Refer Table  1 ).

Data collection

Data were collected by conducting (IDIs) and (FGDs). The IDI and FGD interview guides were developed specifically for the study and reviewed based on the literature. IDIs were conducted with the healthcare industry administrators, Heads of the Departments (HODs), and HCPs of private and government health settings, including gynaecologists, LHWs, LHVs, and CMWs. The IDI interview guides comprised of the questions related to knowledge, sources of information, and attitudes regarding gender-based discrimination (how each gender is perceived in society and how physical and social differences in the roles of males and females affect an individual or society). The IDIs were conducted in Urdu and local language. The interviews were audio-recorded. Each IDIs lasted for 45–60 minutes.

Likewise, the FGDs were conducted using different interview guides, which were designed to assess the perception of adolescent girls and boys, married men and women and health care workers regarding gender discrimination in the society (perceptions of masculinity and femininity, and gender role expectations of a society). The FGDs were conducted in Urdu and local language. The interviews were audio-recorded. Each FGDs lasted for 60–120 minutes.

Data analysis

All interviews were audio recorded and transcribed in English. Training was provided to the data collectors, and they were supervised by the authors throughout the process to ensure transcriptions are written accurately and correctly, representing the actual data collected during interviews. Thematic analysis was carried out in four different steps. Firstly, manual analysis was done by the research team where transcriptions were thoroughly read, and codes were identified. These codes were combined according to their contextual similarity which followed the derivation of categories, based on which, themes were developed. Secondly, similar manual analysis was conducted by an expert data analyst. Thirdly, analysis was conducted using QSR NVivo 10. In the final step, all three analyses were combined and verified by the research team followed by the compilation of results.

Data integrity

To maintain the credibility or truthfulness of the data, the following strategies were used: (1) Prolonged engagement: Various distinct questions were asked related to the topic and participants were encouraged to share their statements with examples, (2) Triangulation: Data was analyzed by the author, expert data analyst and through QSR NVivo10, (3) Persistent observation: The authors read and reread the data, analyzed them recoded and relabeled codes and categories and revised the concepts accordingly, and (4) Transferability: The ability to generalize or transfer the findings to other context or settings, was ensured by explaining in detail the research context and its conclusions [ 6 ].

Ethical considerations

Ethical approval was obtained from the ethics review committee (ERC), Aga Khan University. The ERC number is 2020-3606-11,489. To ensure voluntary participation of the study participants both verbal and written consent were obtained. For those who were younger than 18 years of age were given written assent, and their parent, or guardian’ verbally consented due to literacy issues. In addition to anonymity of the study participants were maintained by assigning codes to the study participants. To avoid loss of data, interview recordings were saved on a hard drive and in the email account of the author. The data on hard copies such as note pads used during IDIs and informed consents were kept in lock and key. All the data present in hard copy was scanned and saved in the hard drive with password protection. To ensure confidentiality, only the authors had access to hard and soft data of the study.

The studies selected were scrutinized to form a data extraction template with all the relevant data such as author, publication year, study title, purpose, design, setting, sampling, main results, perpetuating factors, and recommendations (Refer Table  2 , provided in the attachment). Most of the 20 studies included in the review were conducted in Pakistan however the most frequent study design was cross-sectional ( n  = 9) followed by narrative research based on desk reviews ( n  = 8), one was a case study, and two were cross-country comparison by using secondary data. Four studies were conducted in Province Punjab, three studies were conducted in KPK, and one in both KPK and Punjab. Only one study was conducted in Sindh province. The remaining used whole Pakistan in systematic review. The maximum sample size in a cross-sectional study was ( n  = 506). Six major themes have emerged from the review which included (1) Status of Women in Society (2) Gender Inequality in Health (3) Gender Inequality in Education (4) Gender Inequality in Employment (5) Gender Biased Social Norms and Cultural Practices (6) Micro and Macro Level Recommendations.

Status of a woman in the society

The Pakistani women often face gender inequality [ 13 ]. Women are seen as a sexual object who are not allowed to take decision for self or their family. However, the male is seen as a symbol of power. Due to male ownership and the patriarchal structure of the Pakistani society women are submissive to men, their rights are ignored, and their identity is lost. Out of twenty, nine studies reported that a female can not take an independent decision, someone else decides on her behalf, mainly father before marriage then-husband and son [ 1 , 3 , 4 , 6 , 7 , 8 , 13 ]. The three studies report that women are not allowed to participate in elections or have very limited participation in politics. Furthermore, women often face inequalities and discrimination in access to health, education, and employment that have negative impact in their lives [ 1 , 2 ]. In addition, media often portrays women in the stereotyped role whose only responsibility is to look after the family and household chores [ 2 ]. Likewise, women have less access and control over financial and physical assets [ 13 ]. Similarly, in most of the low economic and tribal families’ women face verbal and physical abuse [ 8 ].

Gender inequality in health

Gender disparity in health is obvious in Pakistan. Women suffer from neglect of health and nutrition. They don’t have reproductive health rights, appropriate prenatal and postnatal care, and decision-making power for birth spacing those results in maternal mortality and morbidity [ 13 ]. Women can not take decision for her and her children’s health; she doesn’t have access to quality education and health services [ 13 , 15 ]. Furthermore, many papers report son preference [ 1 , 3 ]. Gender-based violence is also very common in Pakistan that leads to harmful consequences on the health and wellbeing of women [ 9 ].

Gender inequality in education

Low investment in girls’ education has been reported in almost all the papers reviewed. The major reason for low investment is low returns from girls, as boys are perceived to be potential head of the house and future bread winner [ 6 , 10 , 11 , 12 , 13 , 15 ]. One of the case study reports, people believe, Muslim women should be brought up in a way that they can fulfill the role of a good daughter, wife, and a mother; and education can have a “bad influence” to develop these characteristics in women [ 12 ]. If girls are educated, they become less obedient and evil and don’t take interest in household chores that is the primary responsibility of her [ 12 ]. Moreover, religious leaders have strong authority in rural areas. They often misuse Islamic teaching and educate parents that through education, women become independent and cannot become a good mother, daughter, and a wife. These teachings mostly hinder girl’s education. Other barriers in girls’ education are access to the facility and women’s safety. Five studies reported that most of the schools are on long distances and have co-education system that is perceived as un-Islamic. Parents are reluctant to send their daughters for education as they feel unsafe and threatened [ 1 , 4 , 12 , 13 , 15 ]. Poverty is another root cause of gender disparity in education, as parents cannot afford the education of their children and when there is a choice, preference is given to boys due to their perceived productive role in future. As a result, more dropouts and lower attainment of education by girls particularly living in rural areas [ 6 , 7 , 8 , 9 , 11 , 13 ].

Gender inequality in employment

Economic disparity due to gender inequality is an alarming issue in Pakistan. The low status of women in society, home care responsibilities, gender stereotyping, and social-cultural humiliated practices against women are the main hurdles in women’s growth and employment opportunities. Low education of females, restriction on mobility, lack of required skillsets, sex-segregated occupational choices are also big obstacles in the attainment of economic opportunities. Most of the women are out of employment, however those who are in economic stream are facing several challenges [ 7 ]. They face discrimination in all layers of the economy. Men are mostly on the leadership positions, fewer females are involved in decision making, wages are low for females if compared with males, workplace harassment and unfavourable work environment is common that hinders long stay in job [ 1 , 7 , 8 ]. Moreover, a study reported that in a patriarchal society very limited number of females are in business field and entrepreneurship. The main hurdles are capital unavailability, lack of role models, gender discrimination in business, cultural and local customs, and lack of training and education [ 8 ].

Gender biased social norms and cultural practices

The gender discrimination is deeply rooted in the Pakistani society. The gender disparity in Pakistan is evident at household level. It includes Distribution of food, education, health care, early and forced marriages, denial of inheritance right, mobility restriction, abuse, and violence [ 1 , 2 , 4 , 6 , 7 , 11 ]. Furthermore, birth of a boy child is celebrated, and the girl is seen as a burden. Likewise, household chores are duty of a female, and she cannot demand or expect any reward for it. On the other hand, male work has socio-economic value [ 2 , 7 , 15 ]. Furthermore, the female has limited decision making power and most of the decisions are done by male figures in a family or a leader of the tribe or community who is always a male. This patriarchal system is sustained and practiced under the name of Islamic teaching [ 2 , 12 , 13 ]. The prevalence of gender-based violence is also high, in form of verbal abuse, physical abuse, sexual assault, rape and forced sex, etc., In addition, it is usually considered a private matter and legal actions are not taken against it [ 8 ] . Moreover, Karo Kari or honor killing of a female is observed in Pakistan. It is justified as killing in the name of honor . Similarly, women face other forms of gender-based violence that include: (i) bride price (The family of the groom pay their future in-laws at the start of their marriage), (ii) Watta Satta (simultaneous marriage of a brother-sister pair from two households.), (iii) Vani (girls, often minors, are given in marriage or servitude to an aggrieved family as compensation to end disputes, often murder) and (iv) marriage with Quran (the male members of the families marry off their girl child to Holy Quran in order to take control of the property that legally belongs to the girl and would get transferred to her after marriage) [ 1 , 4 , 9 , 14 , 15 ]. Furthermore, the women are restricted to choose political career [ 13 ].

Micro and macro level recommendations

The women should have equal status and participation in all aspects of life that include, health, nutrition, education, employment, and politics [ 1 , 4 , 7 , 9 , 11 ]. Women empowerment should be reinforced at policy level [ 1 , 7 ]. For this, constitution of Pakistan should give equal rights to all citizens. Women should be educated about their rights [ 1 , 2 , 4 , 6 , 13 , 14 , 15 ]. To improve status of women, utmost intervention is an investment in girls education. If women is not educated she cannot fight for her rights. Gender parity can only be achieved if women is educated and allowed to participate in decision-making process of law and policies [ 4 , 5 , 6 , 9 , 11 , 14 ]. Similarly, access to health care services is women’s right. Quality education, adequate nutrition, antenatal and post-natal care services, skilled birth attendants, and access and awareness about contraceptives is important to improve women’s health and reduce maternal mortality.

Similarly, women should be given equal opportunities to take part in national development and economic activities of the country to reduce poverty. This is possible through fair employment opportunities, support in women’s own business, equitable policies at workplace and uniform wages and salaries. Besides these, female employees must be informed about their rights and privileges at workplace and employment [ 1 , 7 , 8 , 11 ]. Policy actions should be taken to increase the level of women’s participation in economic growth and entrepreneurship opportunities. There should be active actions to identify bottlenecks of gender parity and unlock growth potential of social institutions [ 5 ]. Another barrier for women empowerment is threatened and unsafe environment to thrive. There should be policies and legislation to protect women from harm, violence, and honor killing that ensure their health, safety, and wellbeing [ 4 , 12 ]. Educational institutions and mass media are two powerful sources that can bring change in society. Government must initiate mass media awareness campaign on gender discrimination at household level, educational institutes, and employment sectors to break discriminatory norms of patriarchal society and to reduce the monopoly of males in marketplace. Parent’s education on gender-equitable practices is also important to bring change at the microlevel. It includes gender-equitable child-rearing practices at home including boys mentoring because they think discrimination against females is a very normal practice and part of a culture [ 3 ]. There is insufficient data on women’s participation and gender parity in health, education, and employment. Thus, there is a strong need to identify effective interventions and relevant stakeholders to reduce the gender discrimination in Pakistan [ 5 ] .

Findings from primary data collection

The following are the major themes emerged from the primary data collection (Refer Table  3 ).

Theme 1: perception of women regarding gender discrimination in society

Woman as a sexual object.

Female participants highlighted that they are seen as “sexual objects” and “a mean of physical attraction” which prevents them from comfortably leaving their homes. One female participant explained this further as,

“We are asked to stay inside the house because men and boys would look at our body and may have bad intentions about us” (Adolescent girl, FGD).

Male participants echoed this narrative as they agreed that women are judged by their physical appearance, such as the shape of their bodies. A male participant stated,

“ Woman is a symbol of beauty and she's seen by the society as the symbol of sex for a man" (Male HCP, IDI).

A male participant reported,

“Women should cover themselves and stay inside the house” (married man, FGD).

One female participant verbalized,

“ We have breasts, and therefore, we are asked to dress properly". (adolescent girls, FGD).

Another stated,

“ Girls are supposed to dress properly and avoid eye contact with boys while walking on the road” (adolescent girls, FGD).

Women as dependent beings

One of the major study findings suggests the idea that women must be “helped” at all times, as they are naturally dependent upon other persons to protect them. One participant stated,

“If a woman is alone, she is afraid of the man's actions ” (adolescent girl, FGD).

Some female participants, however, agreed with this statement to some extent because they felt that men help women to fit into society. Oftentimes, judgment is passed for women without an accompanying male. Participants verbalized that wife cannot survive without husband and similarly daughter cannot live without her father. One participant mentioned,

“We are only allowed to go out when we have our father or brothers to accompany us” (Adolescent girl, FGD).

Other participants agreed with the sentiment differently. Since it is implied that men easily get attracted to women, having a male figure with female will protect her from naturally prying eyes. However, if she cannot be accompanied by a male, she must protect herself by covering fully and maintain distance with males.

Women’s autonomy

Female participants, especially young adolescent girls, shared how restrictions have affected their livelihoods. Participants expressed how easy it is for males to gain permission and leave the house, while females often have series of obstacles in front of them. A young girl stated,

“ There are lot of constraints when we see women in our culture. They must take care of everything at home, yet they must get everybody's permission to go five minutes away. Whereas a boy can go out of town and that too, without anyone’s permission. Looking at this, I wish I were a boy. I'd go wherever I want, and I could do whatever I want” (adolescent girl, FGD).

Males as an identity for females

Women are often identified through a prominent male figure in their life and are not considered to have individual personalities and identities. A female participant mentioned that,

“Woman is someone having a low status in society. People know her through their husband or father name” (married women, FGD).

Child’s upbringing responsibility

Culturally, it is expected from the female members of the family, often mothers, to rear children and take care of their upbringing. Male members, mainly fathers, are expected to look after finances. Thus, mothers usually take a greater portion of responsibility for child’s upbringing and blame in case of misconduct. A married woman explained that,

"If a girl does something, the mother is blamed for that. Even in our house, my mother-in-law talks to my mother if I argue or refuse for anything. This is the culture in my maiden home as well" (Married Woman, FGD).

Unrecognized contribution of women

Many female participants verbalized their concern for disregard they receive from their families despite contributing significantly. Women who perform major roles in maintaining the family and household chores are not recognized for their efforts. By doing cleaning, cooking and other duties, they keep family healthy and help keep costs low. One participant mentioned,

“If women don’t clean the house, it is extremely dirty. If women do not rear children, no one else would do it. We do so much for the family” (married woman, FGD).

Gender differences in daily activities

Both men and women struggle with self-expression as certain expectations from both genders hold people back from expressing their views and opinions. Men, for example, as indicated by participants, are expected to remain firm in challenging situations and not show emotions. Even for hobbies, participants shared that, parks and recreational activities are geared towards young boys and men, while girls and women are given more quiet and indoor activities. A female participant verbalized that,

“ Boys have a separate area where they play cricket and football daily but for girls like us, only indoor activities are arranged” (adolescent girl, FGD).

In places where males and females freely mix or live closely in one area, people often find themselves taking extra precautions in their actions, as to not be seen disgraceful by the community. One female participant reported,

“ Two communities are residing in our area. Events for females, such as sports day, are very rarely arranged. Even then we cannot fully enjoy because if we'll shout to cheer up other players, we would be scolded as our community is very cautious for portraying a soft image of females of our community ” (adolescent girl, FGD).

Another participant stated that,

“ After prayers, we cannot spend time with friends as people would point that girl and say that she always stays late after prayers to gossip when she is supposed to go home ” (adolescent girl, FGD).

Deprivation of women’s rights

A woman’s liberty has always struggled to be accepted and males are always favoured. Thus, women are given lower status. Participants highlighted that, in general, men are seen as superior to women. One participant stated,

“ Men are the masters of women…” (FGD married women).

On the other side, male suppress female liberty and women are unaware of their rights leaving them vulnerable to deprivation. A female participant explained,

“Women do not dominate society that's why people take away their rights from them” (married woman, FGD).

Female participants also shared that they see men as strong and dominant personalities, making them better decision makers regarding health care acquisition, family income, availing opportunities and producing offspring. One female participant verbalized,

“If there's one egg on the table and two children to be fed, it is considered that males should get it as it is believed that males need more nutrition than us” (HCP, IDI).

Another reported that,

“There is a lack of equal accessibility of health care facilities and lack of employment equality for women” (HCP, IDI).

Theme 2: perception of men regarding gender discrimination in society

Male dominance.

Inferiority and superiority are common phenomenon in Pakistan’s largely patriarchal society. This allows men to be seen as dominant, decision-maker of family and the sole bread winner. Women, however, are caught in a culture of subordination to men with little power over family and individual affairs. A female participant said,

“If we look at our society, men are dominant. They can do anything while a woman cannot, as she is afraid of the man's reactions [gussa] and aggression” (adolescent girl, FGD).

While another reported,

"In our society, husband makes his wife feel his superiority over her and would make her realize that it is him, who has all the authority and power” (married woman, FGD).

Preference for male child

There is often an extreme desire for birth of sons over daughters, which adds to the culture of gender discrimination in Pakistan. Male children are important to the family as they often serve their parents financially, once they are able. This is one of the main reasons that parents are more inclined towards birth of a male child rather than female. Consequently, education is prioritized for male children. Female participants expressed that their desire for a male child is to appease their husband’s family and reduce the pressure on her to fit in the house. According to a female participant,

“When my son was born, I was satisfied as now nobody would pressurize me. I noticed a huge difference in the behavior of my in-laws after I gave birth to my son. I felt I have an existence in their family” (married woman, FGD).

Participants highlighted that, women who have brothers are often more protected. According to a young participant,

“Brothers give us the confidence to move within the society because people think before saying anything about us” (adolescent girl, FGD).

Lack of communication among husband and wife

Married couples often lack communication and rarely discuss important matters with each other. Men often choose not to share issues with their wives as they believe they are not rational enough to understand the situation. A male participant stated,

“ Women are so sensitive to share anything. They can only reproduce and cook food inside the home” (married man, FGD).

Men are protectors

Many female participants considered men as a source of protection, as they manage finances and ensure safety of family members. They feel confident in man’s ability to contribute to their livelihoods. One participant mentioned,

“We go out when we have our father or brothers to accompany us” (Adolescent girl, FGD).

Another highlighted,

“Men are our protectors. We can only survive in the society because of them” (Married woman, FGD).

Theme 3: factors influencing gender discrimination

The role of family head.

A tight-knit family situation, difference of opinions, cultural values and generation gap can highly affect one’s view on gender. Participants highlighted the role of elders in the family who often favor their sons and male family members. Married women expressed that daughter in-laws often struggle to raise their voice or express their concerns in such family situation. One participant mentioned,

“We don’t take decisions on when to have the child or what method needs to be used for family planning. Our mothers-in-law decide and we must obey” (married woman, FGD).

The family system that often includes three generations living closely, allows traditional norms to carry forward, as opposed to a typical nuclear family. This includes attire, conduct, and relationships. One participant mentioned,

“I live with my mother-in-law. I must cover my head whenever I had to leave the house”. (Married woman, FGD).

Media influence

Media plays an important role in disseminating gender awareness. For example, advertisements of cooking oils and spices usually show young girls helping their mothers in kitchen, while men and boys are observed enjoying something else or not present. These short advertisements are impactful in perpetuating gender conduct solely for societal acceptance. One participant verbalized,

“Every household has a radio, on which different advertisements are going on. People get messages through media” (married man, FGD).

The study reveals that women are seen as sexual objects and therefore confined to their homes. Women are often judged on their physical appearance that hinders their autonomy in various aspects of life. Many women face difficulties in leaving their homes alone and require protection from men [ 3 ]. Men are, therefore, labeled as protectors while women are regarded as dependent beings who need man’s identity. The role of men inside the house is identified as authoritative, while women need approval from male because they are considered incapable of making appropriate decisions. Women are caretaker of their families and have primary responsibility of husband, children, and in-laws. However, these contributions are mostly unnoticed. These gender power differentials are so strong in households, that many women do not know their rights. Women comply with societal and cultural values that force them to become lesser beings in the society. Girls in society grow up and eventually adopt the traditional role of women [ 8 ]. Increased education and awareness level among communities can improve status of women in the Pakistani society [ 3 ].

Moreover, males have dominant role in the society [ 1 ]. Likewise, there is discrepancy in power structures between male and female in the family system that often leads to lack of communication especially between married couples as husbands do not share concerns with their wives nor ask for their advice, considering women incapable to understand anything [ 5 ].

Furthermore, a common phenomenon observed in the Pakistani society, is the strong desire for a male child, while the birth of a female child is mourned [ 5 ]. Girls are seen as a liability, while the birth of a male child is celebrated as it is believed that males will be the breadwinner of the family in the future [ 5 ]. Thus, preference for a male child leads to illegal termination of pregnancies with female fetuses in many situations [ 9 ]. In addition, some of the studies suggest that the preference for a son is significantly high in low socioeconomic areas if compared with the middle and upper ones. Men are seen as economic and social security providers of the household. Therefore, men are tagged as manhood in the society as it is considered that hierarchal familial structures are produced from them, and all powers are attributed to men. This increases the disparity of roles between men and women leading to gender discrimination [ 5 ]. Our study also reveals that media has important influence towards gender discrimination. It is commonly observed in the Pakistani TV advertisements, that household chores are mostly performed by women while men have professional roles in the society [ 6 ].

Thus, lack of female autonomy and empowerment are recognized as the major reasons of discrimination of women in our society. They do not have the means to participate in society, neither they are allowed to speak against traditions. Therefore, interventions are required to increase female autonomy and decision-making capacity. The other significant contributor to gender discrimination is male dominance, which must be brought down to empower women. To reduce this, communication is key between spouses, family members and community members. Gender discrimination has greater influence at different levels of Pakistani society. Certain schools and television advertisements portrays stereotypes, such as allowing boys to be active outdoors and forcing girls to remain indoors. Therefore, media channels and other public systems such as healthcare facilities and schooling systems must promote gender equity and equality. In terms of Sexual and Reproductive health (SRH), the health care facilities should play an important role in providing knowledge and effective treatment to both males and females. The SRH related services are often compromised for people due to lack of resources, staff, and attention. Schools and communities should play an important role in creating SRH related awareness among youth and adults that include puberty, pregnancy, and motherhood. SRH should also be made part of curriculum in educational institutes.

The use of group interviews allowed rapport development with communities. With multiple people present sharing similar views, many were inclined to give purposeful answers and recommendations regarding gender roles in communities. Based on previous literature searches, this study, to the best of our knowledge, has not been published in Pakistan at the community level. No other study explores the views of Pakistanis on gender discrimination with inclusion of multiple community groups and across multiple districts. In limitations, due to the topic’s sensitive topic, may have held back participants from answering fully and truthfully. Thus, considerable time was taken to develop trust and rapport. Therefore, it is possible that some study subjects might not have answered to the best of their ability. Furthermore, challenges were faced due to the COVID-19 pandemic and extreme weather conditions in some areas, as some participants could not reach the venue. Also, the lockdowns following the pandemic made it very difficult to gather 10–12 people at one place for the FGDs. Interviews could not be done virtually as the information was very sensitive.

Gender roles in Pakistani society are extremely complex and are transferred from generation to generation with minimal changes since ages. This study reveals some of the factors due to which women in Pakistan face gender discrimination. The cultural and societal values place women in a nurturing role in the Pakistani society. Through reinforcement of these roles by different family members, as well as by the dominant men in the society, women face adverse challenges to seek empowerment that will help them defy such repressive roles assigned to them. Gender discrimination is evident in public institutions such as healthcare facilities and schooling systems. Thus, administrative reorganization and improved awareness in the healthcare facilities, and appropriate education in schools for boys and girls will help decrease gender discrimination in the Pakistani societies.

Availability of data and materials

On request, the data will be available by hiding the IDs.

Though we have already provided the transcripts, yet there is a need of further information then kindly contact the corresponding author. Dr. Tazeen Saeed Ali: [email protected] .

Abbreviations

Aga Khan Foundation

Preferred Reporting Items for Systematic Reviews and Meta-Analyses

Reporting of Observational studies in Epidemiology

Intimate Partner Violence

Healthcare Professionals

Lady Health Visitors

Lady Health Workers

Community Midwives

In-Depth Interviews

Focus Group Discussions

Heads of the Departments

Sexual and Reproductive health

United Nations Population Fund

Ethics Review Committee

Iqbal H, Afzal S, Inayat M. Gender discrimination: implications for Pakistan security. IOSR J Humanit Soc Sci. 2012;1(4):16–25.

Article   Google Scholar  

Ejaz SS, Ara AA. Gender discrimination and the role of women in Pakistan. J Soc Sci Humanit. 2011;50(1):95–105.

Google Scholar  

Delavande A, Zafar B. Gender discrimination and social identity: experimental evidence from urban Pakistan. FRB New York Staff Rep. 2013;593. Available at SSRN 2198386. 2013.

Bukhari MAHS, Gaho MGM, Soomro MKH. Gender inequality: problems & its solutions in Pakistan. Gov-Annu Res J Pol Sci. 2019;7(7):47–58.

Ferrant G, Kolev A. Does gender discrimination in social institutions matter for long-term growth?: cross-country evidence: EconPapers; 2016. Paris https://econpapers.repec.org/scripts/redir.pf?u=https%3A%2F%2Fdoi.org%2F10.1787%2F5jm2hz8dgls6-en;h=repec:oec:devaaa:330-en .

Kazimi AB, Shaikh MA, John S. Mothers role and perception in developing gender discrimination. Glob Polit Rev. 2019;4(2):20–8.

Alam A. Impact of gender discrimination on gender development and poverty alleviation. Sarhad J Agric. 2011;27(2):330–1.

Mahmood B, Sohail MM, Khalid S, Babak I. Gender-specific barriers to female entrepreneurs in Pakistan: a study in urban areas of Pakistan. J Educ Soc Behav Sci. 2012;2(4):339–52.

Tarar MG, Pulla V. Patriarchy, gender violence and poverty amongst Pakistani women: a social work inquiry. Int J Soc Work Hum Serv Pract. 2014;2(2):56–63.

Atif K, Ullah MZ, Afsheen A, Naqvi SAH, Raja ZA, Niazi SA. Son preference in Pakistan; a myth or reality. PakistanJ Med Sci. 2016;32(4):994.

Rabia M, Tanveer F, Gillani M, Naeem H, Akbar S. Gender inequality: a case study in Pakistan. Open J Soc Sci. 2019;7(03):369.

Shah S, Shah U. Girl education in rural Pakistan. Rev Int Sociol Educ. 2012;1(2):180–207.

Hamid A, Ahmed AM. An analysis of multi-dimensional gender inequality in Pakistan. Asian J Bus Manag. 2011;3(3):166–77.

Bhattacharya S. Status of women in Pakistan. J Res Soc Pakistan. 2014;51(1):179–211.

Butt BI, Asad AZ. Social policy and women status in Pakistan: a situation analysis. Orient Res J Soc Sci. 2016;1(1):47–62.

Khurshid N, Fiaz A, Khurshid J. Analyzing the impact of gender inequality on economic development in Pakistan: ARDL bound test cointegration analysis. Int J Econ Financ Issues. 2020;10(4):264–70.

Mahmood T, Aslam Shah P, Wasif NM. Sociological analysis of women’s empowerment in Pakistan. New Horiz. 2021;15(1):1–2.

Mumtaz S, Iqtadar S, Akther M, Niaz Z, Komal T, AbaidUllah S. Frequency and psychosocial determinants of gender discrimination regarding food distribution among families; 2022.

Li C, Ahmed N, Qalati S. Impact of gender-specific causes on women entrepreneurship: an opportunity structure for entrepreneurial women in rural areas. J Entrep Org Manag. 2022;8(1):2–10.

CAS   Google Scholar  

Ahmad A. Men’s perception of women regarding the internet usage in the Khyber agency Pakistan: an exploratory study. Global Mass Commun Rev. 2020;V(I):47–58.

Download references

Acknowledgments

The authors would like to thank the research specialist, coordinator, and research associates for data collection, and the study participants for their time and valuable data. We would also like to appreciate and thank Mr. Adil Ali Saeed for helping us with the literature for the systematic review of the paper, and Ms. Amirah Nazir and Daman Dhunna for the overall cleaning of document. We are thankful to UNFPA and AKF for providing advisory and monitoring support. We would like to acknowledgment UNFPA Pakistan that through them the funding was received from Global Affairs Canada.

Global Affairs Canada (GAC). Project No: P006434; Arrangement #: 7414620.

Role of the funder: This is to declare that there was no role of the funding agency for planning and implementation of this study.

Author information

Authors and affiliations.

School of Nursing and Midwifery, Aga Khan University, Stadium Road, P.O. Box 3500, Karachi, Pakistan

Tazeen Saeed Ali & Shahnaz Shahid Ali

Department of Community Health Sciences, Aga Khan University, Stadium Road, P.O. Box 3500, Karachi, Pakistan

Tazeen Saeed Ali

Aga Khan Health Services, Karachi, Pakistan

Sanober Nadeem & Falak Madhani

Center of Excellence Women & Child Health, Aga Khan University, Karachi, Pakistan

Zahid Memon, Sajid Soofi & Shah Mohammad

Aga Khan Rural Service Pakistan, Gilgit, Pakistan

Yasmin Karim

Institute for Global Health, Karachi, Pakistan

Zulfiqar Ahmed Bhutta

Department of Paediatrics & Child Health, Aga Khan University, Karachi, Pakistan

You can also search for this author in PubMed   Google Scholar

Contributions

All authors have read and approved the manuscript. Their contribution is as follows: TSA contributed to proposal development, interview guide development, ERC approval, data supervision, data validation, systematic review, data analysis, manuscript development, and overall supervision. SSA assisted in proposal development, data collection supervision, data validation, systematic review, data analysis, and reviewed manuscript. SN, contributed in -literature Review, analysis of literature review and write up of findings. ZM reviewed interview guides, assisted in ERC approval, filed preparation for data collection, assisted in data validation and enhancing the approval processing, reviewed data analysis, and the final manuscript. SSA, contributed to proposal development, assisted in ERC approval, overall supervision, filed preparation for data collection and training of data collectors, assisted in data validation and enhancing the approval process and review of final manuscript. FM contributed to the interview guide development, facilitated field data collection, and contributed to the validation and analysis processes. Reviewed the final manuscript before submission. YK contributed to the interview guide development, facilitated field data collection, and contributed to the validation and analysis processes. Reviewed the final manuscript before submission. SM, contributed to proposal development, field preparation for data collection, validation, and review of the final manuscript. ZB, contributed to proposal development, brought the funding, assisted in ERC approval, overall supervision, data validation and enhancing the approval process and reviewed the final manuscript. He provided overall mentorship.

Corresponding author

Correspondence to Tazeen Saeed Ali .

Ethics declarations

Ethics approval and consent to participate.

The ERC approval was taken from the Aga Khan University Ethics Review Committee for primary data collection. The ERC number is 2020-3606-11489. The written informed consent was taken from all the participants. For those who were younger than 18 years of age were given written assent, and their parent, or guardian verbally consented.

We declare that this is original research and all the authors have contributed to the proposal writing, funding management, data collection, analysis, and manuscript development.

Consent for publication

Competing interests.

Authors declare that they have no conflict of interest.

Additional information

Publisher’s note.

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/ . The Creative Commons Public Domain Dedication waiver ( http://creativecommons.org/publicdomain/zero/1.0/ ) applies to the data made available in this article, unless otherwise stated in a credit line to the data.

Reprints and permissions

About this article

Cite this article.

Ali, T.S., Ali, S.S., Nadeem, S. et al. Perpetuation of gender discrimination in Pakistani society: results from a scoping review and qualitative study conducted in three provinces of Pakistan. BMC Women's Health 22 , 540 (2022). https://doi.org/10.1186/s12905-022-02011-6

Download citation

Received : 02 March 2022

Accepted : 16 October 2022

Published : 22 December 2022

DOI : https://doi.org/10.1186/s12905-022-02011-6

Share this article

Anyone you share the following link with will be able to read this content:

Sorry, a shareable link is not currently available for this article.

Provided by the Springer Nature SharedIt content-sharing initiative

• Gender discrimination
• Women discrimination

BMC Women's Health

ISSN: 1472-6874

• Submission enquiries: [email protected]
• General enquiries: [email protected]

Advertisement

Gender Differences in Education: Are Girls Neglected in Pakistani Society?

• Published: 22 March 2023

Cite this article

• Humaira Kamal Pasha   ORCID: orcid.org/0000-0002-8894-5142 1 , 2  

3636 Accesses

2 Altmetric

Explore all metrics

Differences in education between girls and boys persist in Pakistan, and the distribution of household resources and socioeconomic disparities are compounding the problem. This paper determines education attainment (primary to tertiary level) and current enrollment and explores underlying gender differences with reference to per capita income and socioeconomic characteristics of the household by using survey data of Pakistan (2005–2019) that have never been used in this context before. The potential endogeneity bias between income and education is addressed through the two-stage residual inclusion (2SRI) method that is appropriate for non-linear models used in this study. Findings indicate that income is likely to increase and facilitate a significant transition from primary- to tertiary-level education attainment. The boys have a higher likelihood to increase tertiary-level education attainment by household income. However, the probability of current enrollment is equivalent for girls and boys after controlling for endogeneity. The gender effects of Oaxaca-type decomposition indicate higher unexplained variation that describes a strong gender gap between boys and girls. The standard deviation for education inequality and gender gap ratio confirm that higher levels of discrimination and lower economic returns are associated with girls’ education, and individual and community attributes favor boys’ education. Findings suggest policies and educational strategies that focus on female education and lower-income households to build socioeconomic stability and sustainable human capital in the country.

Similar content being viewed by others

The impacts of gender inequality in education on economic growth in Tunisia: an empirical analysis

Khayria Karoui & Rochdi Feki

Predictors of the gender gap in household educational spending among school and college-going children in India

Rashmi Rashmi, Bijay Kumar Malik, … S. V. Subramanian

Gender Bias in Education in West Bengal

Amita Majumder & Chayanika Mitra

Avoid common mistakes on your manuscript.

Introduction

According to the Education for All (EFA) report, knowledge stimulates the stock of human capital in an economy (Karoui et al., 2018 ; Kim et al., 2021 ) and increases the probability of resources being equally distributed of regardless of gender, caste, color, or region (Heb, 2020 ; de Bruin et al., 2020 ). Gender equality in education is indispensable for developing countries like Pakistan which holds rich human capital to improve economic growth (Asif et al., 2019 ). The existence of patriarchy, cultural norms, regional conflicts, son preference, and traditional notions of womanhood regarding procreation, domestic chores, and early marriage have deep roots in society (Ashraf, 2018 ). All the impediments that women face have interconnected bases in prevailing gender differences and insufficient investment in education (Kleven et al., 2019 ) at the household and state level; these also negatively impact the economic growth in Pakistan (ur Rahman et al., 2018 ).

Some educational initiatives are working effectively in Pakistan but have not completely achieved. These include alternative learning programs (ALPs) for formal schools, digital innovations programs by the collaborations between UNICEF and UNESCO targeting the attainment of Sustainable Development Goals (Ministry of Federal Education, 2022), an EU partnership to implement a 5-year development program (Education Ministry of Balochistan, 2021), the Ilm- Possible Footnote 1 Project for Zero OOSC (out of school children), and equity-based critical learning (STEM, 2021). However, 22.84 million children of secondary school age have never enrolled in formal education (UNESCO Pakistan Country Strategic Document, 2018–2022). In addition, the literacy rate has declined from 62 to 58 % (World Bank Statistics, 2022) that has increased global inequality (Paris21 Strategy Agenda, 2030). This situation raises the question as to whether existing educational policies and projects are adequate for curbing the gender inequality in different provinces of Pakistan (see Fig. 1 ).

Literacy rate by province and gender in Pakistan. Source: Author construction based on data from PSLM Bureau of Statistics, Pakistan. Figure 2 displays the trend of per capita income from 2005 to 2019, one of the inevitable indicators of educational achievement. The statistics calculate a sharp drop in per capita income after 2010, which improved in 2012 but eventually declined after 2016

The country has been ranked 151 out of 153 countries by the Gender Parity Index. It has also been found that 21 % of boys and 32 % of girls in primary education have experienced gender-based discrimination (Human Rights Watch, 2018). Likewise, boys are 15 % more likely to have the opportunity to go to school than girls, as boys are viewed as financial assets by their parents. Evidently, if household income is equally distributed, girls perform outclass in grades (Yi et al., 2015 ), provide higher marginal returns to education (Whalley & Zhao, 2013 ), and achieve sustainable environment (Heb, 2020 ). The economic benefits that result from female education are as high as those that result from male education (Minasyan et al., 2019 ; Sen et al., 2019 ), particularly in relation to the achievement of tertiary education (Alfalih et al., 2021 ; Wu et al., 2020 ). In addition, although Pakistan has the largest young population in Asia, approximately 80 % of the female population has never participated in the labor market, and 130 million girls (those aged between 6 and 17 years) have never attended any form of educational institution (World Bank, 2020). Nevertheless, the latent demand for schooling remains associated with the socioeconomic status and purchasing power of the household (Asif et al., 2019 ). Likewise, parental and household treatment effects can formulate considerable gender gap that requires thorough investigation at micro level.

The aim of this study is to examine the relationship between gender differences in education and household income in Pakistan. Measuring gender differences with the help of microdata and through the use of qualitative and quantitative approaches is not easy in studies of human capital development (Najeeb, 2020 ). Nor is the investigation of the circumstances that lead to more investment in a male child than a female child a straightforward matter. Findings in this area remain inconclusive, which demonstrates a lack of research conducted at the household level in Pakistan (Minasyan, 2019 ). In addition, many studies of the effect of household income on education suffer from bias-related issues which arise as a result of measurement errors and spurious relationships. Some studies use corresponding variables, such as permanent income (Kingdon, 2005 ), or ignore endogeneity while controlling for children’s cognitive skills (Chevalier et al., 2002 ). Others deal with potential endogeneity by examining sector- or community-based union membership (Chevalier, 2013 ), government tax changes (Paul, 2002 ), and rented or owned lands with the caution of the weak instrument (Okabe, 2016 ).

This study determines education achievement using ordered logit and logit models by two outcome variables: education attainment (categorical variable) and current enrollment (binary variable). It seeks to examine the causes of the prevailing gender differences in Pakistan by examining the per capita income and socioeconomic characteristics of households. This study attempts to deal with underlying potential endogeneity using a novel approach for a non-linear model and examines extant inequalities and gender effects within households. This study finds a positive and robust relationship between gender and education attainment, and the significant transformation from primary- to tertiary-level education by per capita income of the household; this contradicts the results of Munshi ( 2017 ). The findings are significantly negative with regard to the relationship between gender and current enrollment, which is opposite to the findings of the study by Maitra ( 2003 ). After dealing with potential endogeneity using the two-stage residual inclusion (2SRI afterwards) method, the results contradict those of prior studies (Chevalier et al., 2002 ; Maitra, 2003 ), and they establish a clear link between education and income along with other socioeconomic characteristics. The findings show that inequalities in education, at the micro level, exert a more powerful impact on girls than boys in relation to reducing education attainment and current enrollment. Gender decomposition reveals that individual and community attributes favor boys’ education over that of girls.

This study contributes to the literature in the following ways. Firstly, there is a risk that the factors that influence education achievement remain mis-specified due to the fact that limited information is available about children’s environments and family structures. This is why it is vital to focus on the determinants of human capital at the micro level. Most existing studies focus on the role of education and the impact of gender inequalities in relation to their impact on economic growth across countries (Assoumou-Ella, 2019 ; Evans et al., 2021 ), within country at the macro level (Rammohan et al., 2018 ), and focus on only one education level (Lloyd et al., 2005 ). This study is the first to attempt to highlight the importance of the gender gap in relation to education attainment and current enrollment and confirm whether it exists or not. It does so by examining the link between per capita income and the socioeconomic characteristics of households using a repeated cross-sectional dataset that has not achieved much academic attention from scholars in relation to the country of Pakistan. Secondly, this study develops an empirical strategy for non-linear model to address the potential endogeneity by using 2SRI approach that remain ignored mostly. It exploits exogenous variations using income shocks, windfall income, and non-labor resources to examine the potential endogeneity between income and education (Banzragch et al., 2019 ; Chevalier et al., 2002 ). Lastly, while previous studies have argued that gender inequality influences economic growth (Kopnina, 2020 ), some of these studies contain troubling contradictions (Sirine, 2015 ), some do not find that gender inequality affects economic growth to a considerable degree (Maitra, 2003 ), and some investigate its unidirectional effect (Tansel & Bodur, 2012 ). This study captures discriminations effect in education investment in boys and girls by education inequalities and gender decomposition estimated at household level. It also adopts alternative specifications of gender inequalities to examine economic returns on education.

The rest of this study is structured as follows. The “Literature Review” section explains the importance of gender equality with reference to previous studies. The “Methodology and Data” section describes the methodology and the data used in this study. The “Empirical Results and Discussion” section presents the results and analysis, and the “Conclusion and Policy Implications” section concludes the study while also discussing policy implications and the limitations of the study.

Literature Review

Education is an essential element of the Cobb–Douglas production function (Saleem et al., 2019 ) that can improve human capital, promote economic growth, and curb poverty in the long term (Arshed et al., 2019 ). Many countries have experienced improvements in enrollment rates; however, their economic growth appears difficult to achieve. This mechanism of human capital can be revisited and revised by focusing on the equal distribution of education in economic and sustainable approach (Livingstone, 2018 ). The study of Duflo et al. ( 2021 ) examines the impact of free secondary education on gains in economic welfare after the completion the target of UPE (universal primary education). They use data relating to secondary high schools from 54 districts in Ghana to examine 1500 students enrolled in a scholarship program. They find that the program increased secondary-level education attainment by 27 % and further resulted in better learning skills and lower rates of early marriage and reduced fertility rates among girls. This suggests a potential movement toward the more equal treatment of the genders within households. However, they did not find any significant influence of education attainment on future employment. Using the Barro-Lee dataset of education attainment, Evans et al. ( 2021 ) estimate the gender gap and its effects on long-term economic growth. Instead taking the gender gap ratio, it prefers to employ difference of the education attainment between men and women. Their findings indicate that low levels of education in women are the reason why the gender gap has become so pronounced in many countries. This gap is revealed to be highly correlated with the age of the women and per capita income.

The study by Kopnina ( 2020 ) discusses the sustainable educational goals that are indispensable for progressive universal education and economic growth. It reveals alternative measures that might influence the circular economy and argues that gender differences will decrease as a result of investment in female education. It endorses the use of the term “empowerment education,” and particularly to refer to females who remain unempowered with regard to their financial independence and social status. They propose the direct influence of female education on the food patterns, efficient consumption of household and natural resources, and renewable energy that can handle growing population in a sustainable approach. Likewise, the study of de Bruin et al. ( 2020 ) finds that education and income can promote sustainability and reduce gender inequality. They use age, education, and different types of work to analyze the gender-differentiated impact of these factors on economic change.

Another study, that of Rammohan et al. ( 2018 ), examines gender disparity in education using district-level data in India and ordinary least squares (OLS) regression. To do so, they use data related to the gender gap between male and female education attainment, GDP per capita, and ethnicity. Their study finds that those living in wealthier districts are more inclined toward educating their daughters than those living in poor ones. Sahoo and Klasen ( 2021 ) focused on female participation in the STEM streams by using the variables: female, siblings, age, parental education, test scores, household size, and ethnicity. They reveal that girls are 20 % less likely to enroll in STEM streams than boys. The plausible explanation for lower female participation is associated with parental preferences and income disparity in the household. Maitra ( 2003 ) uses a probit model and a censored probit model simultaneously and finds that there is no gender difference in the current enrollment rates of boys and girls (6–12 years) but that there is a higher gap in relation to grade attainment for girls (13–24). The data used is from the Matlab Health and Socio-Economic Survey (MHSS) of rural Bangladesh, which surveys 149 villages. The explanatory variables include religion, household size, number of siblings, the head of the household’s education level and occupation, a log of per adult household expenditure, and household characteristics such as the number of bedrooms, access to water and a toilet, and the availability of electricity. The endogeneity issue of the income has dealt by taking the residual term of the log of the adult expenditure in the household.

The study of Davis et al. ( 2019 ) uses the World Value Survey (1981–2014) to capture individual effects on women’s status. They argue that individual decision-making can increase women’s education attainment, their choice to bear a child, and advance economic sustainability such as urbanization and the provision of basic necessities. The above effects provide economic benefits that further support gender equality and discourage the traditional role of women in the society. Robb et al. ( 2012 ) examine the gender differences in education attainment using data about university graduates and an ordered probit model. They find that female students perform better than their male counterparts but that they are less likely to obtain a first-class degree. It is shown that factors such as the type of institution, individual abilities, and the choice of subjects are not the reason for gender inequality; however, the effects of these factors increase the gender gap in relation to educational performance. The predict probabilities of their study explain that the likelihood that female students will attain a first-class degree is 5 %, compared with 8 % for male students. Other studies also advocate that reducing gender differences in education achievement can have transitional and long-term effects on women’s empowerment (Kabeer, 2021 ), legal protection (Durrani et al., 2018 ), employment (Najeeb et al., 2020 ), and sustainable growth (Heb, 2020 ).

Prior Literature in the Context of Pakistan

In the context of Pakistan, Ashraf et al. ( 2018 ) apply Dickey–Fuller generalized least squares (DF-GLS) to examine the impact of secondary school attainment on gender inequality. They employ multiple sources of data about Pakistan, namely, economic surveys, the National Assembly of Pakistan database, and the Pakistan Social and Living Standards Measurement (PSLM) survey. They use the Gender Inequality Index (GII) as the dependent variable. The findings show that economic deprivation can decrease women’s participation in the labor force and their education attainment. Notably, the external or spillover effects of education attainment on gender inequality are also crucial to understanding the lower purchasing power of the household. Qureshi ( 2007 ) conducts a bivariate regression analysis using the Learning and Educational Achievements in Punjab Schools (LEAPS) dataset to investigate whether the education attainment of an older sister impacts on the education attainment of younger children in the household. Mainly, it describes a spillover effect in education that remains unnoticed to receive its maximum economic benefit. It takes into account age, the father’s education level, the household head’s education level, the number of children, the infrastructure of the household, the regional languages, and the number of the districts in the province. The findings reveal 0.2 % of years of schooling increases in youngers boys by the older educated sister that can be the potential human capital in the future labor market. However, their study fails to analyze the spillover effect that an educated older brother has on a younger sister.

The study of Asif et al. ( 2019 ) demonstrates that the strong and significant impact of investment in education without gender bias creates other avenues for sustainable growth in Pakistan. Likewise, some studies investigate education investment to explore other dimensions including welfare gains in relation to eradicating hunger (Ali et al., 2021 ), the awareness of climate change by energy consumption and recyclable goods (Ali et al., 2019 ), the transformation of society into one with equal rights and zero violence (Durrani et al., 2018 ), the female leadership in entrepreneurial decision-making (Shaheen and Ahmad, 2022 ), and the voluntary effort toward food security and patterns of daily life (Qazlbash et al., 2021 ). Mahmood et al. ( 2012 ) use time-series data (1971–2009) to investigate the relationship between human capital investment and economic growth. In their work, autoregressive distributed lag (ARDL) and OLS models show a positive relationship between high enrollment rates in education and economic growth rates in the short and long term.

A similar strategy is proposed by Zaman ( 2010 ), who also suggests that there is a correlation between female education and economic development. Interestingly, Lloyd et al. ( 2005 ) find that parents tend to prefer that girls and boys attend separate schools; however, availability of primary schools and type of school (public or private) also play key roles. The study of ur Rahman et al. ( 2018 ) finds that a solution to the vicious cycle of poverty comes in the form of increasing the education level of a household. By using logistic regression, they find a negative relationship between education and poverty in Pakistan. They emphasize the role of education plays in providing potential human capital for the labor market and even generating new and improved employment opportunities that result in better living standards and economic well-being. However, a key issue with regard to the aforementioned studies is that they do not propose well-specified econometric strategies with that can be implemented to tackle gender differences in education, while others fail to address the potential endogeneity in non-linear models and some remain unable to decompose gender effects within the household.

Methodology and Data

Data and variables.

This study uses repeated cross-sectional data from the PSLM survey conducted by the Pakistan Bureau of Statistics (PBS) of the Government of Pakistan for the seven fully available rounds from 2005 to 2019 (2005–2006, 2007–2008, 2010–2011, 2011–2012, 2013–2014, 2015–2016, and 2018–2019). The survey was designed to provide social and economic indicators at the provincial and district level; starting in 2004, the survey aims to accurately describe the country. The sample size of the PSLM surveys is approximately 80,000 households. The total number of observations after pooling the data is 1,011,849.

This study uses two models for alternative measurements of the education achievement of boys and girls; the first model is education attainment (the highest completed schooling; aged 9–24 years), and the second model is current enrollment (aged 5–24 years). The boys and girls are restricted in first model to the 9–15, 16–19, and 20–24 age groups for primary-, secondary-, and tertiary-level education attainment, respectively. The following criteria are considered: additional year for class repetition by the students, late admission into schools, the completion standards of the Pakistan education system, and traditional age requirements for entering in school adopted in past studies (Maitra et al., 2003 ). In addition, boys and girls are restricted to not having the status of head or working person. In the first model, education attainment is a categorical outcome variable examines by the ordered logit model that can be defined as:

In the second model, current enrollment is a dichotomous outcome variable examines by logit model that can be described as:

The explanatory variables include dummy variables of the gender and age of boys and girls depending on models. Age is represented by a linear and quadratic term to control for birth cohort effects and capture non-linearity effects on education achievement. As age is directly proportional in contributing to cognitive skills and human capital, age square indicates marginal returns from age that decrease over time.

Other explanatory variables include the marital status of the household members (Kingdon, 2005 ). This study uses a series of dummy variables for the education level of individuals including the head of the household, parents, and other members of the household (i.e., those older than 24). This is because a joint-family structure is the majority form of family structure in Pakistan, and the head of the household is usually not the father but rather any elderly family member. Likewise, head’s personal treatment and decision-making influence on the education achievement. In addition, using parental education instead of maternal education is also feasible for gender difference analysis to avoid the issue of multicollinearity. Several tests are run to check for multicollinearity, including the variance inflation factor (VIF) and correlation matrix. The VIF for each predictor variable should be less than 10. It is 7.02 for the education attainment model and 4.12 for current enrollment model. Footnote 2 The siblings’ variable is used to control the reciprocal relationship between quantity and quality of education (Hazarika, 2001 ; Maitra, 2003 ). Occupational heterogeneity is controlled by different household members’ professions (McNabb, 2002 ) ranging from high-salaried (officer) to low-salaried (laborer) professions.

The variable of interest in this study is the per capita income of the household. It represents the household’s possible investment in education, which can maximize economic returns and minimize gender inequality. The availability of electricity, gas, and broadband internet access is a proxy for household infrastructure and technology advancement. The latter is of interest as it may impact on digital education, sustainable development, and the urgency of the micro- and macro-crisis such as health. The high demand to shift education from formal to virtual platforms during the COVID-19 pandemic has opened up new dimensions with regard to the acquisition of skills and knowledge. The other control variables consist of the dependency ratio, household size, ownership of house and any establishment other than agricultural land (ur Rahman et al., 2018 ), and ownership of the cultivating land for the personal use of the household (Sawada, 2009 ). Finally, community characteristics are controlled by including dummy variables for locations and number of the provinces of the country (Hazarika, 2001 ).

Empirical Strategy

The concept of the ordered logit model for education attainment is to incorporate intermediate continuous variable says y in the latent regression accompanied by the observed ( x i ) explanatory variables and the unobserved error term ( ε i ). The range of y is divided in adjacent intervals that comprise four categories—namely, 0 = no education, 1 = primary education, 2 = secondary education, and 3 = tertiary education—related to latent variable ( Y ∗ ). The structural model for latent education is:

where β is the vector of the parameters to be estimated; ε is the disturbance term, which is assumed to be independent across observations; and y ∗ can take value with observations.

For the discrete choices, the following are observing as:

Where Y is the category of education attainment, and τ denotes the threshold parameters, explaining the transition from one category of education attainment to another category. Consequently, τ must satisfy the rule according to τ 0  <  τ 1  <  τ 2  <  τ 3 , as the ε i is logistically distributed. The resulting probabilities can be observed as:

Hence, the probability of outcome can imply as:

whereas the log likelihood function for ordered logistic regression is:

The function formulates the ordered logit model with multiple equations, whereas each equation presents the logit model (Williams, 2005 ). The econometric model is therefore:

Endogeneity Bias

The main econometric challenge is to identify the endogeneity problems. There is the possibility that variable per capita income is likely to be related to unobservable factors that affect education achievement in many ways not included in the regression. There may be errors in measuring per capita income that bias the results. In addition, a causal relationship may exist between income and education achievement. This relationship might also be influenced by parental economic circumstances, social status, and any spurious third factor such as personal preferences. Reverse causality occurs when the poor educational performance of the boy or girl might lower household income and vice versa. Therefore, the model may suffer from omitted variable bias and (reverse) causality issues. The literature also explores per capita income as an endogenous variable that has instrumented by parental and household characteristics including employment, education, and farming activities (Bratti, 2007 ; Hoogerheide, 2012 ). Other studies examine its causal relationship with income shock (Coelli, 2005 ), the difference in households’ incomes, rainfall, and climate change in relation to productivity concerns (Fichera et al., 2015 ).

In the first model of education attainment, income shock such as head unemployment and non-labor resources of grandparents in the household are used as instruments for per capita income (Behrman et al., 1997 ). If the head of the household is unemployed, this is unlikely to influence the total years of schooling undertaken by boys and girls when there is a joint family structure where the parents are responsible for meeting educational expenditures. Similarly, the permanent or non-labor income of the grandparents is an exogenous and strong instrument that does not directly affect the total years of schooling undertaken by boys and girls (Bratti, 2007 ). However, these instruments may affect the current enrollment of boys and girls, thus necessitating the exploration of other exogenous variables. Therefore, the potential endogeneity in the current enrollment model is captured by another set of exogenous variables; first, the difference in per capita income between households (included in the PSLM survey) and country, and second, windfall income. The difference in per capita income is a proxy for income shock that does not relate to agricultural goods but rather a retrospective analysis of households having or not having wages. This difference may represent the transitional effect of the financial condition of the household (Björkman-Nyqvist, 2013 ; Sawada, 2009 ) in developing countries such as Pakistan. Similarly, windfall income comprises mainly of the unearned income of the household or non-labor income that includes lottery wins, inheritances, gifts, unexpected charity payments, and irregular sources of income (Kingdon, 2005 ; Powdthavee et al., 2013 ), which are exogenous.

Another source of endogeneity might arise due to the relationship between education spending and current enrollment in the logit model regression. The literature provides instruments for education spending such as community-, labor-, or industry-union membership of the household’s head that in unavailable in PSLM dataset while some studies refer to the head’s occupations (Maitra, 2003 ). The estimation results after instrumenting education spending with the head of household’s occupation show that the null hypothesis of homogeneity is not rejected, as it has a p -value of 0.93. However, this study tries to control educational spending through the addition of dummy occupational variables, home ownership, and land cultivation (Maitra, 2003 ; Shea, 2000 ). Other individual and socioeconomic characteristics are considered as exogenous. The OLS regression (for instrument validation) and alternative approaches to capture potential endogeneity—such as the control function approach, two-stage least squares (2SLS) (ignoring the nature of the outcome variable), and the IV probit model (splitting the outcome variable into a binary variable where necessary)—are also examined.

To apply the 2SRI method, the first step is to find exogenous variables; however, this method is different from the standard IV estimation method. The strategy behind choosing variables is that variables predict a possible definition of homogeneity. The argument behind this method (Terza, 2018 ) is based on the inappropriateness of the traditional linear instrumental variable estimator for the correction of the endogeneity problem. The core advantage of this method is that the estimated coefficients associated with the residuals from the first-stage regression significantly express the presence of endogeneity in the model (Huasman, 1978 ). In this method, the first stage consists of the OLS regression and predicts the endogenous variable by using the instruments and the rest of the explanatory variables. The second stage is estimated using the ordered logit model with the inclusion of the first-stage residuals. In the final stage, the whole program is set to be bootstrapped. The latent model will be established by splitting the explanatory variables into exogenous and endogenous variables, say X ex and X en , and the equation becomes:

The first-stage equation of the 2SRI method is estimated for income using all the exogenous variables and instruments in the OLS regression. It takes the form as:

where E ( X en ,  Z ) ≠ 0 and E ( ε i ,  Z ) = 0; β and γ are coefficient parameters; and v i and ε i are error terms, respectively. The second stage of the 2SRI method estimates outcome variable using the residuals obtained from the first-stage equation taken as control variables along with other explanatory variables. The model is described as:

This method is a simple test of endogeneity: if the residuals of the first stage are statistically significant, then the results will be biased in the first model, refer to control the endogeneity issue (Terza, 2018 ; Akarçay-Gürbüz & Polat, 2017 ).

Education and Inequality Parameters

The Gini coefficient for education, average years of schooling, and standard deviation are the inequality parameters that have been considered by observing the education system and structure of the country, the efficiency of learning performance, and variations in gender-specific education investment (Digdowiseiso, 2010 ; Thomas et al., 2001 ). Footnote 3 The consideration of these inequalities might help to reveal the socioeconomic and intrahousehold factors behind the different treatment for girls’ education. Therefore, the extended model can be described as:

The gender decomposition examines while using the basic models of each specification by the mean, the coefficient (Kingdom, 2005 ), and the interactions of the boy dummy variable (Maitra, 2003 ). Furthermore, the results are decomposed for gender effects by variant type Oaxaca decomposition (Dong et al., 2009 ; Golsteyn et al., 2014 ; Pal, 2004 ). This approach is generally used to examine the gender effects related to economic returns and the wage gap (Oaxaca, 1973 ). In this study, however, the standard approach has been modified to examine the gender effects related to education achievement. The probability of education attainment determines, say AT, separately for girls and boys with other characteristics, say X g and X b , respectively. Assuming \(\Pr \left( AT,{X}_i,{\theta}_i^{\ast}\right)\) is the expected probability of AT and \({\theta}_i^{\ast }\) is the vector if the maximum likelihood estimates of the parameters of the ordered logit model for i  =  g , b for girls and boys, respectively, the expected AT for any individual would be:

Using expected education attainment for the boys’ and girls’ samples, respectively, one can decompose the boy-girl differential in alternative ways as follows:

In brief, the explained variation is attributable to the different characteristics of boy-girl, while the unexplained variation is attributable to the different treatment of boys and girls in the household. This is achieved by allowing the parameters to vary while the characteristics are held constant. A similar approach was adopted for current enrollment as well.

The alternative specifications explore the impact of gender inequalities in education achievement on the household income using OLS regression. This study uses three different measurements of gender difference. Considering education attainment, the first indicator—the gender gap Footnote 4 —is calculated as the difference in illiteracy rates between girls and boys (Cooray, 2011 ). The second indicator, gender difference, Footnote 5 measures the difference in education attainment between girls and boys (Baliamoune–Lutz & McGillivray, 2015 ), while the final indicator—the gender gap ratio Footnote 6 —is constructed based on the difference between boys’ and girls’ education attainment (Digdowiseiso, 2010 ). Similar inequalities in current enrollment for boys and girls (5–24) are also estimated. The alternative specification estimates in the linear regression model are defined as:

Furthermore, the robustness tests for education achievement are examined using several other specifications including ordered probit and probit models, another explanatory variable—per capita expenditure of the household, and provincial heterogeneity.

Descriptive Statistics

The detailed descriptive statistics of the selected variables are exhibited in Table 1 . On average, 10 % of boys and girls attain a primary level of education, and 2.1 % attain a tertiary level of education. On average, the variable gender signifies 49 % girls in first model of education attainment and 48.8 % in second model of current enrollment. On average, 38.9 % of boys and girls are currently enrolled in education, and per capita income (in the log) is 8.8 (see Fig. 2 ). Overall, the age of the household has a nonlinear effect; as with the increase of age of the household’ members, there is decrease in the education level (see Fig. 3 ). Meanwhile, this study uses the age of the boys and girls, according to the models’ criteria. The mean age in the first model is 15.95 years whereas it is 13.59 in the second model. This study observes a higher ratio of low-salaried occupations (for example, machine operators); thus, the dependency ratio is also higher at 41.6 %. A total of 44.5 % of the population lives in urban areas, 80.6 % receive electricity, and 37.7 % have access to gas supplies. Among other provinces, the highest population locates in the province of Punjab.

Household’s income in Pakistan. Source: Author construction based on data from PSLM Bureau of Statistics, Pakistan. Figure 2 displays the trend of per capita income from 2005 to 2019, one of the inevitable indicators of educational achievement. The statistics calculate a sharp drop in per capita income after 2010, which improved in 2012 but eventually declined after 2016

Education attainment by age (2005–2019). Source: Author construction based on data from PSLM Bureau of Statistics, Pakistan. Figure 3 expresses the predictive margins between the age of the persons living in the household and their education levels. The probability of primary education attainment decreases after 25 years of age, whereas it is the opposite for the tertiary level. Meanwhile, with the increase in age, it is more likely to achieve secondary education

Empirical Results and Discussion

Determining education attainment and current enrollment levels.

Table 2 describes the average marginal effects of the ordered logit model for primary-, secondary-, tertiary-level, and no education attainment with the help of household income per capita and various socioeconomic characteristics. In the full sample models, variable gender—girl, increases primary-, secondary-, and tertiary-level education attainment by 0.4, 0.5, and 0.2 percentage points, respectively, at the 1 % significance level; this contradicts the findings of Munshi ( 2017 ). Per capita income, on average, increases the likelihood of primary-, secondary-, and tertiary-level education attainment by 0.1, 0.2, and 0.1 percentage points, respectively. The effect of age is more likely to increase secondary- and tertiary- education attainment. As findings reveal that the transitional effect of education attainment is progressive from primary level to secondary level, however, it does not appear with same proportion from secondary level to tertiary level. The impact of the age and squared-age has non-linear effect that can be justified in two manners. Firstly, with the increase in age, the proportion of transition of the education attainment levels decreases. Secondly, there is a negative relationship between the term squared age and education attainment.

Meanwhile, the presence of an educated head of household significantly improves primary-, secondary-, and tertiary-level education attainment—by 5.0, 10.4, and 4.1 percentage points, respectively. Other household members are likely to increase secondary- and tertiary-level education attainment by 20.9 and 11.5 percentage points, respectively. The results show higher marginal effects for education attainment by technicians (low-salaried) compare to managers (high-salaried), indicating that lower occupations have strong inspiration to maximize the human resources capital of the household. In addition, the availability of electricity, internet access, and access to a gas supply are highly likely to enhance education attainment. On average, living in urban area has the likelihood to impact primary-, secondary-, and tertiary education attainment by 0.2, 0.3, and 0.1 percentage points, respectively.

From models 5 to 8, for girls, it can be seen that per capita income significantly increases each level of education attainment. However, it increases secondary-level education attainment more than other levels, by 0.2 percentage points. Age has a significant and nonlinear effect. The variable married is likely to decrease the probability of education attainment by 1.6 and 5.0 percentage points at the tertiary and secondary levels of education, respectively. Interesting, parental education has a positive influence, but it is only significant at the secondary education attainment with 23.3 percentage points. In addition, the presence of an educated head of household and other members also provides a positive and significant effect. On analyzing different occupations, the results indicate a 19.6, 23.7, 8.9, and 1.5 percentage point increase in tertiary education attainment by clerks, officers, managers, and machine operators. The household size shows an inverse relationship with girls’ education attainment, particularly at the secondary level. The household infrastructure provides positive effect on girls’ education attainment. It may exhibit that sustainable consumption of household resources including electricity and gas can exert female education that can promote gender equity and economic returns.

From models 9 to 12, for education attainment, it can be seen that the impact of the per capita income of the household is comparatively equal for boys and girls. The household income is likely to increase secondary—and tertiary—education attainment in boys by 0.2 and 0.1 percentage points, respectively. Parental education is highly unlikely to increase the probability of education attainment. The presence of an educated head of household increases education attainment by 4.9, 9.5, and 3.2 percentage points at the primary, secondary, and tertiary levels, respectively. Similarly, the presence of household members with numeracy skills and secondary education is likely to increase secondary-level education attainment by 10.9 and 22.2 percentage points, respectively. This study observes a strong impact of occupational heterogeneity on education attainment; officers and clerks significantly improve the primary- and secondary-level education attainment in boys. The clerks are highly likely to increase tertiary-level education attainment, by 15.4 percentage points. Compared to Punjab, provinces such as KPK and Balochistan are less likely to increase primary, secondary, and tertiary education attainment.

Table 3 describes the average marginal effects from the current enrollment models with the help of logit model regression.

In the full model, the estimate of the variable girl is highly significant and negative—an opposite finding to that of past studies (Maitra, 2003 )—and likely to decrease the probability of current enrollment in education by 0.8 percentage points. A unit increase in income per capita is more likely to improve the current enrollment rates for girls than it is for boys; an increase of 0.4 percentage points is observed for girls. Age has a nonlinear effect with its squared term; thus, current enrollment rates decrease with age. Additionally, variable married decreases the probability of current enrollment in education in girls by 15.6 percentage points. Current enrollment increases for boys if there are educated household members; however, this is not the case for certain professions such as clerks and machine operators.

Other indicators associated with physical capital such as ownership of establishment or land are negatively related to current enrollment rates. This indicates that education is not the primary objective among landowners, as they do not worry about employment. The educational transition from primary to higher grades is less valuable than monetary assets, and most people are reluctant to leave their ancestral profession if it is associated with land cultivation. Household infrastructure is likely to benefit girls more than boys, however when we examine the influence of living in an urban location, which is highly likely to increase enrollment rates in education for boys. The dependency ratio provides higher marginal effects for current enrollment in boys, which further supports the objective of this study. The majority of the households in Pakistan support male earners who are likely to bear all the expenditures. Therefore, the parents prefer to invest in boys’ education for potential job opportunities and financial support in the long run. Results from siblings shows a positive relation to current enrollment and reveal higher quantity-to-quality trade-offs particularly among girls. The results show a higher marginal effect in KPK province; this might be due to the new framework of free and accessible education that has been in place since 2013 (KPK Government Statistics, 2021).

Dealing with Endogeneity Bias

Table 4 shows the results of the average marginal effects using the ordered logit model regression/2SRI approach after dealing with endogeneity. In the full sample, the per capita income of the household is likely to increase education attainment at each level by a higher ratio compared to the aforementioned results. There is a drastic increase in primary-level education attainment: 11.2 percentage points. Likewise, secondary- and tertiary-level education attainment increase by 15.9 and 4.9 percentage points, respectively. Even the variable gender is almost two times higher than the previous results for secondary-level education attainment. Other indicators that illustrate higher marginal effects are educated head of household, household size, and infrastructure. The results find a positive relationship between education and urbanization by introducing income shock of head unemployment and non-labor resources. It retrieves two strong arguments; first, the income shock is likely to increase potential human mobilization for confronting household economic burden. The second, non-labor resources exert positive impact on population by increasing non-market activities, as time allocation shifts from work to leisure.

From models 5 to 8, for girls, the results are significant but with higher marginal effects than the full sample. A sharp increase in secondary-level education attainment is caused by household income: an increase of 10.8 percentage points. Results find negative relationship between married persons and education attainment of the girls, especially at primary level. It might be possible that married persons are quite young in age, particularly women, without having any education awareness and sufficient resources. These results may indicate the need of awareness programs in the household to encourage women education and discourage early-age marriages. On the other hand, a significant decrease in household size supports an increase in primary-level education attainment.

There is a higher impact of per capita income on boys’ education attainment than girls, indicating household’s preferences. The per capita income of the household is likely to increase primary-, secondary- and tertiary-level education attainment by 13.0, 17.9, and 4.9 percentage points, respectively. The presence of an educated head of household has a strong and positive effect on boys’ education attainment; however, it is the opposite for girls’ primary-level education attainment. The results show that intermediate internet access is more effective for girls than boys. Meanwhile, household size also impacts quite positively on boys’ education attainment as they are potential lone bread earners for their families. Living in an urban location results the potential career for boys, thus revealing a positive correlation with education attainment.

The average marginal effects are shown in Table 5 for current enrollment after dealing with potential endogeneity. Per capita income is four times more likely to increase the likelihood of current enrollment in the full sample than the results reported in the “Determining Education Attainment and Current Enrollment Levels” section. Its impact is 4.4 percentage points for boys and girls. The variable girl reduces the probability of current enrollment by 0.3 percentage points. The results find a significant effect of parental education on boys, thus revealing a gender bias in investment in education. Similar results are reported for the impact of educated members of the household and the occupations of those living in the household. The other results describe a wider gap in current enrollment in Sindh and Balochistan, where girls are highly unlikely to enroll in any kind of educational institution.

Estimations of Education Attainment and Current Enrollment by Inequalities

Table 6 illustrates the average marginal effects by incorporating different educational inequalities such as the Gini coefficient, years of schooling (on average), and standard deviation for education attainment by ordered logit model, as shown in panels A, B, and C. For this moment, only results with educational inequalities have been provided. Full results can be provided on demand. In girls’ sample, by examining panel A, we can see that the Gini coefficient is highly significant and indicates a sharp decrease in tertiary- and secondary-level education attainment, by 0.6 and 1.6 percentage points, respectively. Furthermore, in panel B, the average years of schooling have positive relationship with secondary- and tertiary-level education attainment. In panel C, the estimates explain that the standard deviation inequality decreases secondary- and tertiary-level education attainment by 0.1 percentage points, respectively. For boys’ sample, in panel A, the results show that the Gini coefficient decreases the secondary- and tertiary-level education attainment of boys; however, the marginal effects are slightly higher compared to those for girls. In panel B of average years of education, there is an equal improvement in secondary- and tertiary-level education attainment of boys; however, no significant effect is found in panel C.

The relationship between current enrollment and educational inequalities is shown in Table 7 . In panel A, the results indicate that educational inequalities impact both boys and girls. However, examining the marginal effects by gender, the Gini coefficient is found to be higher for boys. In panel B, the average years of schooling of currently enrolled boys and girls are higher for girls by 0.6 percentage points. This indicates that girls are almost 0.4 times more likely to enroll in school. There is an insignificant impact of standard deviation on boys’ current enrollment; however, it is the opposite for girls. A unit increase in standard deviation decreases the probability of girls’ current enrollment by 0.2 percentage points.

Explaining the Gender Gap and its Decomposition

Table 8 provides mean statistics and differences in the coefficients in relation to education attainment.

In panel A, most of the household characteristics favor girls; these include personal attributes such as age and infrastructure while per capita income, educated members, head, and urbanization provide higher mean probabilities for boys ’education attainment. The difference between boys and girls is shown in the last column by interacting the boy dummy variable with each explanatory variable as an additional regressor in the basic model of the full sample using ordered logit model regression. The estimates find favorable values for girls’ education attainment in relation to the education level of her parents and the head of the household and household characteristics. Panel B provides mean statistics and coefficient differences for current enrollment. The personal attributes such as age, married, and infrastructure have the higher mean probabilities for girls’ education attainment. The last column displays the differences between boys and girls and shows that educated head, urbanization, and provinces favor boys.

Table 9 presents the gender differences in education attainment and current enrollment by predicted probabilities using variant type Oaxaca decomposition by incorporating four scenarios. Such as (i) girls using estimated parameters obtained from girls’ equation, (ii) girls using estimated parameters obtained from boys’ equation, (iii) boys using estimated parameters obtained from boys’ equation, and (iv) boys using estimated parameters obtained from girls’ equation (Pal, 2004 ). Comparatively, boys are having approximately two times lower corresponding probabilities using girls’ parameters. Conversely, the probability of girls’ education attainment increases almost two times higher using boys’ parameters. A similar proportion of increase observes in girls’ current enrollment using boys’ parameters. While two times lower probabilities observe for boys’ current enrollment using girls’ parameters. The estimates of difference are presented with the boys’ reference. In the end, explained and unexplained variations of gender difference are estimated. While explained variation in education attainment and current enrollment are −142.8 and 41.4 %, respectively (Dong et al., 2009 ). The unexplained variation, generally considers as discrimination, has higher values in both models and highlight the different treatment between boys and girls in the household. However, this study presents such variation as gender differences that may be due to unobservable factors and imperfectly observable attributes.

Alternative Specification and Robustness Tests

In Tables 10 and 11 , the estimates are presented for education attainment and current enrollment using other models such as ordered probit and probit models (McNabb et al., 2002 ), and other variables such as per capita expenditure and permanent income (non-labor assets) are included.

In both models, the results are highly significant and provide additional evidence to support the previous estimations. The variable girl is more likely to increase education attainment at the secondary level. The unit increase in income per capita is marginally higher in the probit model regression. The findings show that per capita expenditure is likely to positively impact on girls’ education, particularly in relation to secondary-level education attainment. Considering the robust test by incorporating the permanent income of the household, the variable gender is positively significant with education attainment. A unit increase in permanent income raises primary- and secondary-level education attainment more in boys. In addition, there is sharp increase in boys’ current enrollment with a unit increase in permanent income. Other robustness tests, including provincial heterogeneity, the control function approach, IV probit, 2SLS, and the determination of education attainment and current enrollment for boys and girls from a different age group (13–24), are available on request.

Table 12 presents results for alternative specification where per capita income is the dependent variable and gender inequalities (education attainment and current enrollment) as interested variables. This specification can also be interpreted as the future earning potential of girls and boys. Considering education attainment, in panel A, the gender gap due to illiteracy decreases income by approximately 11.3 % more in girls compared to boys. In panel B and C, gender difference is likely to decrease income by 3.2 and 1.2 % in girls. Moving toward current enrollment, in panels A, B, and C, each gender inequality reduces the household income comparatively higher among girls than boys by 7.1, 3.0, and 1.7 %.

Conclusion and Policy Implications

Despite having the potential for human resource capital, Pakistan struggles with extreme poverty, socioeconomic disparity, and gender inequality at the grass-root level (Ali et., 2021 ; Asif et al., 2019 ). To address these undeniable issues, it becomes crucial to comprehend the significance of the equal distribution of household resources in education regardless of gender that builds a sustainable economic structure toward global equality (Kopnina, 2020 ). This study aims to examine education achievement and underlying gender differences using two models: education attainment and current enrollment. The findings highlight the importance of the relationship between education and income along with other household characteristics. This study deals with potential endogeneity by using the 2SRI approach and examines gender and educational inequalities at the micro level.

The findings demonstrate that household income has a significant and positive impact on education attainment and the current enrollment of boys and girls. The education attainment transition from primary to tertiary-level is successful that supports the past studies (Duflo et al., 2021 ; Wu et al., 2020 ). However, the transition from primary to secondary education is higher than that from secondary to tertiary education attainment. The community and individuals’ attributes support education investment in boys indicating household and socioeconomic preferences. Girls can improve their education with the availability of personal and household attributes (Yi et al., 2015 ). Other findings from education attainment and current enrollment models predict a demographic framework that encourages a sustainable environment with a decline in household size and dependency ratio (Heb, 2020 ; Asif, 2019 ; Fichera et al., 2015 ). These findings contradict those of past studies (Munshi, 2017 ) and establish a link between temporary residents (daughters) and different occupations of the households, whereby lower-salaried households and deprived areas can significantly improve female education attainment and current enrollment.

The findings show that there is a negative relationship between the Gini coefficient and education attainment and that this gap is wider at secondary and tertiary education levels, thus supporting the results of the basic model. The standard deviation of educational inequalities is higher for girls that further confirms the existence of gender differences in education. Likewise, the findings from the alternative specifications provide decrease in potential economic returns on education by gender inequalities. The findings support those of Pfeffer et al. ( 2018 ) with regard to discouraging wealth accumulation in terms of physical capital and increasing investment in female education (Kopnina, 2020 ). It can effectively transform the developing society of Pakistan by framing public policies for women’s empowerment (United Nations Education, 2030), gender equality (Arshed et al., 2019 ), poverty alleviation (ur Rahman et al., 2018 ), and sustainable development (Sen, 2019 ). Therefore, this study identifies some valuable recommendations for policymakers wishing to promote gender equality:

Implement cooperative projects created by federal and local governments that supply free, digital, and up-to-date education in schools, colleges, and universities to improve transition levels, with a particular focus on poor infrastructure, highly deprived regions, and mobility restricted areas.

Adopt targeted policies to minimize education and gender gaps between those enrolled and not enrolled in education by supporting low-income households through the allocation of funds, scholarships, and incentives.

Reform educational strategies to provide cost effective education in collaboration with parents, teachers, and schools with the aim of creating advanced and scientific curricula aligned with sustainable development goals.

Craft awareness campaigns to eradicate gender-specific investment in education, encourage talented females to enter tertiary-level education in particular, and address socioeconomic challenges by establishing reliable and organized educational committees in each province.

Finally, some potential limitations should be noted, as these might open up new horizons for future research. Quantitative research should be conducted to examine other household characteristics and upcoming survey rounds than those discussed in this study.

Translated as Knowledge Possible project that having contribution of US $ 1 billion for sustainable programs.

These tests are available on request.

The Gini coefficient for education is defined as the ratio to the mean (average years of schooling) of half of the average overall pairs of absolute deviations between all pairs of people (Deaton, 1997). It is further redeveloped by Thomas ( 2001 ):

where E L is the Gini coefficient based on education attainment, 𝜇 is average years of schooling, Pi and P j are the proportion of the population, y i and y j are years of schooling at different educational levels, and n is the number of levels in the attainment data. Average years of schooling and standard deviation can be calculated as follows: \(\textrm{AYS}=\sum_{\textrm{i}=1}^{\textrm{n}}{\textrm{P}}_{\textrm{i}}{\left({\textrm{y}}_{\textrm{i}}-\upmu \right)}^2\) and \(\textrm{SD}=\sqrt{\sum_{\textrm{i}=1}^{\textrm{n}}{\textrm{P}}_{\textrm{i}}{\left({\textrm{y}}_{\textrm{i}}-\upmu \right)}^2}\) .

Gender gap (illiteracy rates [9–24 years of age]) = number of girls’ illiterate – number of boys’ illiterate

Gender difference (education attainment [9–24 years of age]) = total years of schooling of girls – total years of schooling of boys.

Gender gap ratio (education attainment [9–24 years of age]) = total years of schooling of girls/total years of schooling of boys.

Akarçay-Gürbüz, A., & Polat, S. (2017). Schooling opportunities and intergenerational educational mobility in Turkey: An IV estimation using Census data. The Journal of Development Studies, 53 (9), 1396–1413.

Article   Google Scholar  

Alfalih, A. A., & Alfalih, A. A. (2021). The impact of background characteristics on graduate school attendance: Evidence from a developing economy. Journal of the Knowledge Economy, 12 (1), 363–383.

Ali, A., Audi, M., Bibi, C., & Roussel, Y. (2021). The impact of gender inequality and environmental degradation on human well-being in the case of Pakistan: A time series analysis

Ali, A., Mottaleb, K. A., & Aryal, J. P. (2019). Wealth, education and cooking-fuel choices among rural households in Pakistan. Energy Strategy Reviews, 24 , 236–243.

Arshed, N., et al. (2019). Education stock and its implication for income inequality: The case of Asian economies. Review of Development Economics, 23 (2), 1050–1066.

Ashraf, I., & Ali, A. (2018). Socio-economic well-being and women status in Pakistan: An empirical analysis. Bulletin of Business and Economics (BBE), 7 (2), 46–58.

Google Scholar  

Asif, M. F., & Pervaiz, Z. (2019). Socio-demographic determinants of unmet need for family planning among married women in Pakistan. BMC Public Health, 19 (1), 1–8.

Assoumou-Ella, G. (2019). Gender inequality in education and per capita GDP: The case of CEMAC countries. Economics Bulletin, 39 (2).

Baliamoune–Lutz, M., & McGillivray, M. (2015). The impact of gender inequality in education on income in Africa and the Middle East. Economic Modelling, 47 , 1–11.

Banzragch, O., Mizunoya, S., & Bayarjargal, M. (2019). Education inequality in Mongolia: Measurement and causes. International Journal of Educational Development, 68 , 68–79.

Behrman, J. R., Khan, S., Ross, D., & Sabot, R. (1997). School quality and cognitive achievement production: A case study for rural Pakistan. Economics of Education Review, 16 (2), 127–142.

Björkman-Nyqvist, M. (2013). Income shocks and gender gaps in education: Evidence from Uganda. Journal of Development Economics, 105 , 237–253.

Bratti, M. (2007). Parents’ income and children’s school dropout at 16 in England and Wales: Evidence from the 1970 British cohort study. Review of Economics of the Household, 5 (1), 15–40.

Chevalier, A., & Lanot, G. (2002). The relative effect of family characteristics and financial situation on educational achievement. Education Economics, 10 (2), 165–181.

Chevalier, A., Harmon, C., O’Sullivan, V., & Walker, I. (2013). The impact of parental income and education on the schooling of their children. IZA Journal of Labor Economics, 2 (1), 8.

Coelli, M. I. C. K. (2005). Parental income shocks and the education attainment of youth

Cooray, A., & Potrafke, N. (2011). Gender inequality in education: Political institutions or culture and religion? European Journal of Political Economy, 27 (2), 268–280.

Davis, L. S., & Williamson, C. R. (2019). Does individualism promote gender equality? World Development, 123 , 104627.

de Bruin, A., & Liu, N. (2020). The urbanization-household gender inequality nexus: Evidence from time allocation in China. China Economic Review, 60 , 101301.

Digdowiseiso, K. (2010). Measuring Gini coefficient of education: The Indonesian cases

Du, F., & Dong, X. Y. (2009). Why do women have longer durations of unemployment than men in post-restructuring urban China? Cambridge Journal of Economics, 33 (2), 233–252.

Duflo, E., Dupas, P., & Kremer, M. (2021). The impact of free secondary education: Experimental evidence from Ghana (Vol. No. w28937). National Bureau of Economic Research.

Book   Google Scholar  

Durrani, N., & Halai, A. (2018). Dynamics of gender justice, conflict and social cohesion: Analysing educational reforms in Pakistan. International Journal of Educational Development, 61 , 27–39.

Evans, D. K., Akmal, M., & Jakiela, P. (2021). Gender gaps in education: The long view. IZA Journal of Development and Migration, 12 (1).

Fichera, E., & Savage, D. (2015). Income and health in Tanzania An instrumental variable approach. World Development, 66 , 500–515.

Golsteyn, B. H., & Schils, T. (2014). Gender gaps in primary school achievement: A decomposition into endowments and returns to IQ and non-cognitive factors. Economics of Education Review, 41 , 176–187.

Hausman, J. A. (1978). Specification tests in econometrics. Econometrica: Journal of the Econometric Society, 46 (6), 1251–1271.

Hazarika, G. (2001). The sensitivity of primary school enrollment to the costs of post-primary schooling in rural Pakistan: A gender perspective. Education Economics, 9 (3), 237–244.

Heß, P. (2020). SDG 5 and the gender Gap in standardization: Empirical evidence from Germany. Sustainability, 12 (20), 8699.

Hoogerheide, L., Block, J. H., & Thurik, R. (2012). Family background variables as instruments for education in income regressions: A Bayesian analysis. Economics of Education Review, 31 (5), 515–523.

Kabeer, N. (2021). Three faces of agency in feminist economics: Capabilities, empowerment, and citizenship. The Routledge handbook of feminist economics, 99-107

Karoui, K., & Feki, R. (2018). The effect of gender inequality on economic development: Case of African countries. Journal of the Knowledge Economy, 9 (1), 294–300.

Kim, S. Y. (2021). Determining critical factors of gender inequality: Evidence from 34 OECD and non-OECD countries. World Development Perspectives, 21 , 100284.

Kingdon, G. G. (2005). Where has all the bias gone? Detecting gender bias in the intrahousehold allocation of educational expenditure. Economic Development and Cultural Change, 53 (2), 409–451.

Kleven, H., Landais, C., & Søgaard, J. E. (2019). Children and gender inequality: Evidence from Denmark. American Economic Journal: Applied Economics, 11 (4), 181–209.

Kopnina, H. (2020). Education for the future? Critical evaluation of education for sustainable development goals. The Journal of Environmental Education, 51 (4), 280–291.

Livingstone, D. W. (2018). The education-jobs gap: Underemployment or economic democracy? Routledge.

Lloyd, C. B., Mete, C., & Sathar, Z. A. (2005). The effect of gender differences in primary school access, type, and quality on the decision to enroll in rural Pakistan. Economic Development and Cultural Change, 53 (3), 685–710.

Mahmood, A., Mahmood, S. T., & Malik, A. B. (2012). A comparative study of student satisfaction level in distance learning and live classroom at higher education level. Turkish Online Journal of Distance Education, 13 (1).

Maitra, P. (2003). Schooling and education attainment: Evidence from Bangladesh. Education Economics, 11 (2), 129–153.

McNabb, R., Pal, S., & Sloane, P. (2002). Gender differences in education attainment: The case of university students in England and Wales. Economica, 69 (275), 481–503.

Minasyan, A., Zenker, J., Klasen, S., & Vollmer, S. (2019). Educational gender gaps and economic growth: A systematic review and meta-regression analysis. World Development, 122 , 199–217.

Munshi, S. (2017). Arranged marriage, education, and dowry: A contract-theoretic perspective. Journal of Economic Development, 42 (1), 35.

Najeeb, F., Morales, M., & Lopez-Acevedo, G. (2020). Analyzing female employment trends in South Asia. World Bank Policy Research Working Paper, 9157 .

Oaxaca, R. (1973). Male-female wage differentials in urban labor markets. International Economic Review , 693–709.

Okabe, M. (2016). Gender-preferential intergenerational patterns in primary educational attainment: An econometric approach to a case in rural Mindanao, the Philippines. International Journal of Educational Development, 46 , 125–142.

Pal, S. (2004). How much of the gender difference in child school enrolment can be explained? Evidence from rural India. Bulletin of Economic Research, 56 (2), 133–158.

Paul, J., & Wahlberg, K. (2002). Global taxes for global priorities. In Global Policy Forum.

Pfeffer, F. T. (2018). Growing wealth gaps in education. Demography, 55 (3), 1033–1068.

Powdthavee, N., Lekfuangfu, W. N., & Wooden, M. (2013). The marginal income effect of education on happiness: Estimating the direct and indirect effects of compulsory schooling on well-being in Australia

Qazlbash, S. K., Zubair, M., Manzoor, S. A., ul Haq, A., & Baloch, M. S. (2021). Socioeconomic determinants of climate change adaptations in the flood-prone rural community of Indus Basin Pakistan. Environmental Development, 37 , 100603.

Qureshi, S. A., Khan, M., Ul Husnain, M. I., & Iqbal, N. (2007). Gender, environment, and sustainable economic growth [with comments]. The Pakistan Development Review, 46 (4), 883–894.

Rammohan, A., & Vu, P. (2018). Gender inequality in education and kinship norms in India. Feminist Economics, 24 (1), 142–167.

Robb, A. M., & Watson, J. (2012). Gender differences in firm performance: Evidence from new ventures in the United States. Journal of Business Venturing, 27 (5), 544–558.

Sahoo, S., & Klasen, S. (2021). Gender segregation in education: Evidence from higher secondary stream choice in India. Demography, 58 (3), 987–1010.

Saleem, H., Shahzad, M., Khan, M. B., & Khilji, B. A. (2019). Innovation, total factor productivity and economic growth in Pakistan: A policy perspective. Journal of Economic Structures, 8 (1), 1–18.

Sawada, Y., & Lokshin, M. (2009). Obstacles to school progression in rural Pakistan: An analysis of gender and sibling rivalry using field survey data. Journal of Development Economics, 88 (2), 335–347.

Sen, G. (2019). Gender equality and women’s empowerment: Feminist mobilization for the SDG s. Global Policy, 10 , 28–38.

Shaheen, N., & Ahmad, N. (2022). Women entrepreneurship and empowerment in Pakistan: Gender, culture, education and policy in broader perspective. International Research Journal of Education & Social Sciences, 1 (1), 25–36.

Shea, J. (2000). Does parents’ money matter? Journal of Public Economics, 77 (2), 155–184.

Sirine, M. N. I. F. (2015). Impact of inequalities on economic growth: Case of the developing countries. Journal of Sustainable Development Studies, 8 (1).

Tansel, A., & Bodur, F. B. (2012). Wage inequality and returns to education in Turkey: A quantile regression analysis. Review of Development Economics, 16( 1), 107–121.

Terza, J. V. (2018). Two-stage residual inclusion estimation in health services research and health economics. Health Services Research, 53 (3), 1890–1899.

Thomas, V., Wang, Y., & Fan, X. (2001). Measuring education inequality: Gini coefficients of education (Vol. 2525). World Bank Publications.

ur Rahman, S., Chaudhry, I. S., & Farooq, F. (2018). Gender inequality in education and household poverty in Pakistan: A case of Multan District. Review of Economics and Development Studies, 4 (1), 115–126.

Whalley, J., & Zhao, X. (2013). The contribution of human capital to China’s economic growth. China Economic Policy Review, 2 (01), 1350001.

Williams, R. (2005). Gologit2: A program for generalized logistic regression/partial proportional odds models for ordinal dependent variables. Stata, Gologit2 Manual.

Wu, L., Yan, K., & Zhang, Y. (2020). Higher education expansion and inequality in educational opportunities in China. Higher Education, 80 (3), 549–570.

Yi, J., Heckman, J. J., Zhang, J., & Conti, G. (2015). Early health shocks, intra-household resource allocation and child outcomes. The Economic Journal, 125 (588).

Zaman, K., Khan, M. M., Ahmad, M., & Ikram, W. (2010). Do female enrolment rates cause economic growth in Pakistan? Asian Social Science, 6 (11), 256.

Download references

Acknowledgements

This study benefited from the discussion with the participants of the American Economic Association Annual Meeting, 2022, USA; 49th Australian Conference of Economists CEA 2021, Australia; International Population Association Conference, 1PC 2021; 23 INFER Annual conference, 2021 Porugal; and 23rd Applied Economics Meeting, ALdE, 2021, Spain. I would like to thank Prof. Theophile T. Azomahou (CNRS- CERDI), Prof. Colin Green (Norwegian University of Science and Technology), Dr. Ababacar Sedikh, and Dr. Nestor Sawadogo (CNRS-CERDI) for their thoughtful comments. I also would like to thank the editor and two anonymous reviewers for their valuable suggestions.

Data Availability Statement

Data is available on the website of Pakistan Bureau of Statistics (PBS): https://www.pbs.gov.pk/content/microdata .

Author information

Authors and affiliations.

Université Clermont Auvergne CNRS (CERDI), Clermont-Ferrand, France

Humaira Kamal Pasha

AEI International School, Université Paris-Est Créteil, Paris, France

You can also search for this author in PubMed   Google Scholar

Corresponding author

Correspondence to Humaira Kamal Pasha .

Ethics declarations

Ethics approval statement.

Not required.

Conflict of Interest

The author declares no competing interests.

Additional information

Publisher’s note.

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Submission Declaration and Verification

I declare that the article has not been published previously and is not under consideration for publication elsewhere.

Rights and permissions

Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.

Reprints and permissions

About this article

Pasha, H.K. Gender Differences in Education: Are Girls Neglected in Pakistani Society?. J Knowl Econ (2023). https://doi.org/10.1007/s13132-023-01222-y

Download citation

Received : 26 December 2021

Accepted : 24 February 2023

Published : 22 March 2023

DOI : https://doi.org/10.1007/s13132-023-01222-y

Share this article

Anyone you share the following link with will be able to read this content:

Sorry, a shareable link is not currently available for this article.

Provided by the Springer Nature SharedIt content-sharing initiative

• Human capital
• Gender differences
• Ordered logit model
• Find a journal
• Publish with us
• Track your research

An official website of the United States government

The .gov means it’s official. Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

The site is secure. The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

• Publications
• Account settings

Preview improvements coming to the PMC website in October 2024. Learn More or Try it out now .

• Advanced Search
• Journal List
• Springer Nature - PMC COVID-19 Collection
• PMC10031191

Gender Differences in Education: Are Girls Neglected in Pakistani Society?

Humaira kamal pasha.

1 Université Clermont Auvergne CNRS (CERDI), Clermont-Ferrand, France

2 AEI International School, Université Paris-Est Créteil, Paris, France

Associated Data

Data is available on the website of Pakistan Bureau of Statistics (PBS): https://www.pbs.gov.pk/content/microdata .

Differences in education between girls and boys persist in Pakistan, and the distribution of household resources and socioeconomic disparities are compounding the problem. This paper determines education attainment (primary to tertiary level) and current enrollment and explores underlying gender differences with reference to per capita income and socioeconomic characteristics of the household by using survey data of Pakistan (2005–2019) that have never been used in this context before. The potential endogeneity bias between income and education is addressed through the two-stage residual inclusion (2SRI) method that is appropriate for non-linear models used in this study. Findings indicate that income is likely to increase and facilitate a significant transition from primary- to tertiary-level education attainment. The boys have a higher likelihood to increase tertiary-level education attainment by household income. However, the probability of current enrollment is equivalent for girls and boys after controlling for endogeneity. The gender effects of Oaxaca-type decomposition indicate higher unexplained variation that describes a strong gender gap between boys and girls. The standard deviation for education inequality and gender gap ratio confirm that higher levels of discrimination and lower economic returns are associated with girls’ education, and individual and community attributes favor boys’ education. Findings suggest policies and educational strategies that focus on female education and lower-income households to build socioeconomic stability and sustainable human capital in the country.

Introduction

According to the Education for All (EFA) report, knowledge stimulates the stock of human capital in an economy (Karoui et al., 2018 ; Kim et al., 2021 ) and increases the probability of resources being equally distributed of regardless of gender, caste, color, or region (Heb, 2020 ; de Bruin et al., 2020 ). Gender equality in education is indispensable for developing countries like Pakistan which holds rich human capital to improve economic growth (Asif et al., 2019 ). The existence of patriarchy, cultural norms, regional conflicts, son preference, and traditional notions of womanhood regarding procreation, domestic chores, and early marriage have deep roots in society (Ashraf, 2018 ). All the impediments that women face have interconnected bases in prevailing gender differences and insufficient investment in education (Kleven et al., 2019 ) at the household and state level; these also negatively impact the economic growth in Pakistan (ur Rahman et al., 2018 ).

Some educational initiatives are working effectively in Pakistan but have not completely achieved. These include alternative learning programs (ALPs) for formal schools, digital innovations programs by the collaborations between UNICEF and UNESCO targeting the attainment of Sustainable Development Goals (Ministry of Federal Education, 2022), an EU partnership to implement a 5-year development program (Education Ministry of Balochistan, 2021), the Ilm- Possible 1 Project for Zero OOSC (out of school children), and equity-based critical learning (STEM, 2021). However, 22.84 million children of secondary school age have never enrolled in formal education (UNESCO Pakistan Country Strategic Document, 2018–2022). In addition, the literacy rate has declined from 62 to 58 % (World Bank Statistics, 2022) that has increased global inequality (Paris21 Strategy Agenda, 2030). This situation raises the question as to whether existing educational policies and projects are adequate for curbing the gender inequality in different provinces of Pakistan (see Fig. ​ Fig.1 1 ).

Literacy rate by province and gender in Pakistan. Source: Author construction based on data from PSLM Bureau of Statistics, Pakistan. Figure ​ Figure2 2 displays the trend of per capita income from 2005 to 2019, one of the inevitable indicators of educational achievement. The statistics calculate a sharp drop in per capita income after 2010, which improved in 2012 but eventually declined after 2016

The country has been ranked 151 out of 153 countries by the Gender Parity Index. It has also been found that 21 % of boys and 32 % of girls in primary education have experienced gender-based discrimination (Human Rights Watch, 2018). Likewise, boys are 15 % more likely to have the opportunity to go to school than girls, as boys are viewed as financial assets by their parents. Evidently, if household income is equally distributed, girls perform outclass in grades (Yi et al., 2015 ), provide higher marginal returns to education (Whalley & Zhao, 2013 ), and achieve sustainable environment (Heb, 2020 ). The economic benefits that result from female education are as high as those that result from male education (Minasyan et al., 2019 ; Sen et al., 2019 ), particularly in relation to the achievement of tertiary education (Alfalih et al., 2021 ; Wu et al., 2020 ). In addition, although Pakistan has the largest young population in Asia, approximately 80 % of the female population has never participated in the labor market, and 130 million girls (those aged between 6 and 17 years) have never attended any form of educational institution (World Bank, 2020). Nevertheless, the latent demand for schooling remains associated with the socioeconomic status and purchasing power of the household (Asif et al., 2019 ). Likewise, parental and household treatment effects can formulate considerable gender gap that requires thorough investigation at micro level.

The aim of this study is to examine the relationship between gender differences in education and household income in Pakistan. Measuring gender differences with the help of microdata and through the use of qualitative and quantitative approaches is not easy in studies of human capital development (Najeeb, 2020 ). Nor is the investigation of the circumstances that lead to more investment in a male child than a female child a straightforward matter. Findings in this area remain inconclusive, which demonstrates a lack of research conducted at the household level in Pakistan (Minasyan, 2019 ). In addition, many studies of the effect of household income on education suffer from bias-related issues which arise as a result of measurement errors and spurious relationships. Some studies use corresponding variables, such as permanent income (Kingdon, 2005 ), or ignore endogeneity while controlling for children’s cognitive skills (Chevalier et al., 2002 ). Others deal with potential endogeneity by examining sector- or community-based union membership (Chevalier, 2013 ), government tax changes (Paul, 2002 ), and rented or owned lands with the caution of the weak instrument (Okabe, 2016 ).

This study determines education achievement using ordered logit and logit models by two outcome variables: education attainment (categorical variable) and current enrollment (binary variable). It seeks to examine the causes of the prevailing gender differences in Pakistan by examining the per capita income and socioeconomic characteristics of households. This study attempts to deal with underlying potential endogeneity using a novel approach for a non-linear model and examines extant inequalities and gender effects within households. This study finds a positive and robust relationship between gender and education attainment, and the significant transformation from primary- to tertiary-level education by per capita income of the household; this contradicts the results of Munshi ( 2017 ). The findings are significantly negative with regard to the relationship between gender and current enrollment, which is opposite to the findings of the study by Maitra ( 2003 ). After dealing with potential endogeneity using the two-stage residual inclusion (2SRI afterwards) method, the results contradict those of prior studies (Chevalier et al., 2002 ; Maitra, 2003 ), and they establish a clear link between education and income along with other socioeconomic characteristics. The findings show that inequalities in education, at the micro level, exert a more powerful impact on girls than boys in relation to reducing education attainment and current enrollment. Gender decomposition reveals that individual and community attributes favor boys’ education over that of girls.

This study contributes to the literature in the following ways. Firstly, there is a risk that the factors that influence education achievement remain mis-specified due to the fact that limited information is available about children’s environments and family structures. This is why it is vital to focus on the determinants of human capital at the micro level. Most existing studies focus on the role of education and the impact of gender inequalities in relation to their impact on economic growth across countries (Assoumou-Ella, 2019 ; Evans et al., 2021 ), within country at the macro level (Rammohan et al., 2018 ), and focus on only one education level (Lloyd et al., 2005 ). This study is the first to attempt to highlight the importance of the gender gap in relation to education attainment and current enrollment and confirm whether it exists or not. It does so by examining the link between per capita income and the socioeconomic characteristics of households using a repeated cross-sectional dataset that has not achieved much academic attention from scholars in relation to the country of Pakistan. Secondly, this study develops an empirical strategy for non-linear model to address the potential endogeneity by using 2SRI approach that remain ignored mostly. It exploits exogenous variations using income shocks, windfall income, and non-labor resources to examine the potential endogeneity between income and education (Banzragch et al., 2019 ; Chevalier et al., 2002 ). Lastly, while previous studies have argued that gender inequality influences economic growth (Kopnina, 2020 ), some of these studies contain troubling contradictions (Sirine, 2015 ), some do not find that gender inequality affects economic growth to a considerable degree (Maitra, 2003 ), and some investigate its unidirectional effect (Tansel & Bodur, 2012 ). This study captures discriminations effect in education investment in boys and girls by education inequalities and gender decomposition estimated at household level. It also adopts alternative specifications of gender inequalities to examine economic returns on education.

The rest of this study is structured as follows. The “Literature Review” section explains the importance of gender equality with reference to previous studies. The “Methodology and Data” section describes the methodology and the data used in this study. The “Empirical Results and Discussion” section presents the results and analysis, and the “Conclusion and Policy Implications” section concludes the study while also discussing policy implications and the limitations of the study.

Literature Review

Education is an essential element of the Cobb–Douglas production function (Saleem et al., 2019 ) that can improve human capital, promote economic growth, and curb poverty in the long term (Arshed et al., 2019 ). Many countries have experienced improvements in enrollment rates; however, their economic growth appears difficult to achieve. This mechanism of human capital can be revisited and revised by focusing on the equal distribution of education in economic and sustainable approach (Livingstone, 2018 ). The study of Duflo et al. ( 2021 ) examines the impact of free secondary education on gains in economic welfare after the completion the target of UPE (universal primary education). They use data relating to secondary high schools from 54 districts in Ghana to examine 1500 students enrolled in a scholarship program. They find that the program increased secondary-level education attainment by 27 % and further resulted in better learning skills and lower rates of early marriage and reduced fertility rates among girls. This suggests a potential movement toward the more equal treatment of the genders within households. However, they did not find any significant influence of education attainment on future employment. Using the Barro-Lee dataset of education attainment, Evans et al. ( 2021 ) estimate the gender gap and its effects on long-term economic growth. Instead taking the gender gap ratio, it prefers to employ difference of the education attainment between men and women. Their findings indicate that low levels of education in women are the reason why the gender gap has become so pronounced in many countries. This gap is revealed to be highly correlated with the age of the women and per capita income.

The study by Kopnina ( 2020 ) discusses the sustainable educational goals that are indispensable for progressive universal education and economic growth. It reveals alternative measures that might influence the circular economy and argues that gender differences will decrease as a result of investment in female education. It endorses the use of the term “empowerment education,” and particularly to refer to females who remain unempowered with regard to their financial independence and social status. They propose the direct influence of female education on the food patterns, efficient consumption of household and natural resources, and renewable energy that can handle growing population in a sustainable approach. Likewise, the study of de Bruin et al. ( 2020 ) finds that education and income can promote sustainability and reduce gender inequality. They use age, education, and different types of work to analyze the gender-differentiated impact of these factors on economic change.

Another study, that of Rammohan et al. ( 2018 ), examines gender disparity in education using district-level data in India and ordinary least squares (OLS) regression. To do so, they use data related to the gender gap between male and female education attainment, GDP per capita, and ethnicity. Their study finds that those living in wealthier districts are more inclined toward educating their daughters than those living in poor ones. Sahoo and Klasen ( 2021 ) focused on female participation in the STEM streams by using the variables: female, siblings, age, parental education, test scores, household size, and ethnicity. They reveal that girls are 20 % less likely to enroll in STEM streams than boys. The plausible explanation for lower female participation is associated with parental preferences and income disparity in the household. Maitra ( 2003 ) uses a probit model and a censored probit model simultaneously and finds that there is no gender difference in the current enrollment rates of boys and girls (6–12 years) but that there is a higher gap in relation to grade attainment for girls (13–24). The data used is from the Matlab Health and Socio-Economic Survey (MHSS) of rural Bangladesh, which surveys 149 villages. The explanatory variables include religion, household size, number of siblings, the head of the household’s education level and occupation, a log of per adult household expenditure, and household characteristics such as the number of bedrooms, access to water and a toilet, and the availability of electricity. The endogeneity issue of the income has dealt by taking the residual term of the log of the adult expenditure in the household.

The study of Davis et al. ( 2019 ) uses the World Value Survey (1981–2014) to capture individual effects on women’s status. They argue that individual decision-making can increase women’s education attainment, their choice to bear a child, and advance economic sustainability such as urbanization and the provision of basic necessities. The above effects provide economic benefits that further support gender equality and discourage the traditional role of women in the society. Robb et al. ( 2012 ) examine the gender differences in education attainment using data about university graduates and an ordered probit model. They find that female students perform better than their male counterparts but that they are less likely to obtain a first-class degree. It is shown that factors such as the type of institution, individual abilities, and the choice of subjects are not the reason for gender inequality; however, the effects of these factors increase the gender gap in relation to educational performance. The predict probabilities of their study explain that the likelihood that female students will attain a first-class degree is 5 %, compared with 8 % for male students. Other studies also advocate that reducing gender differences in education achievement can have transitional and long-term effects on women’s empowerment (Kabeer, 2021 ), legal protection (Durrani et al., 2018 ), employment (Najeeb et al., 2020 ), and sustainable growth (Heb, 2020 ).

Prior Literature in the Context of Pakistan

In the context of Pakistan, Ashraf et al. ( 2018 ) apply Dickey–Fuller generalized least squares (DF-GLS) to examine the impact of secondary school attainment on gender inequality. They employ multiple sources of data about Pakistan, namely, economic surveys, the National Assembly of Pakistan database, and the Pakistan Social and Living Standards Measurement (PSLM) survey. They use the Gender Inequality Index (GII) as the dependent variable. The findings show that economic deprivation can decrease women’s participation in the labor force and their education attainment. Notably, the external or spillover effects of education attainment on gender inequality are also crucial to understanding the lower purchasing power of the household. Qureshi ( 2007 ) conducts a bivariate regression analysis using the Learning and Educational Achievements in Punjab Schools (LEAPS) dataset to investigate whether the education attainment of an older sister impacts on the education attainment of younger children in the household. Mainly, it describes a spillover effect in education that remains unnoticed to receive its maximum economic benefit. It takes into account age, the father’s education level, the household head’s education level, the number of children, the infrastructure of the household, the regional languages, and the number of the districts in the province. The findings reveal 0.2 % of years of schooling increases in youngers boys by the older educated sister that can be the potential human capital in the future labor market. However, their study fails to analyze the spillover effect that an educated older brother has on a younger sister.

The study of Asif et al. ( 2019 ) demonstrates that the strong and significant impact of investment in education without gender bias creates other avenues for sustainable growth in Pakistan. Likewise, some studies investigate education investment to explore other dimensions including welfare gains in relation to eradicating hunger (Ali et al., 2021 ), the awareness of climate change by energy consumption and recyclable goods (Ali et al., 2019 ), the transformation of society into one with equal rights and zero violence (Durrani et al., 2018 ), the female leadership in entrepreneurial decision-making (Shaheen and Ahmad, 2022 ), and the voluntary effort toward food security and patterns of daily life (Qazlbash et al., 2021 ). Mahmood et al. ( 2012 ) use time-series data (1971–2009) to investigate the relationship between human capital investment and economic growth. In their work, autoregressive distributed lag (ARDL) and OLS models show a positive relationship between high enrollment rates in education and economic growth rates in the short and long term.

A similar strategy is proposed by Zaman ( 2010 ), who also suggests that there is a correlation between female education and economic development. Interestingly, Lloyd et al. ( 2005 ) find that parents tend to prefer that girls and boys attend separate schools; however, availability of primary schools and type of school (public or private) also play key roles. The study of ur Rahman et al. ( 2018 ) finds that a solution to the vicious cycle of poverty comes in the form of increasing the education level of a household. By using logistic regression, they find a negative relationship between education and poverty in Pakistan. They emphasize the role of education plays in providing potential human capital for the labor market and even generating new and improved employment opportunities that result in better living standards and economic well-being. However, a key issue with regard to the aforementioned studies is that they do not propose well-specified econometric strategies with that can be implemented to tackle gender differences in education, while others fail to address the potential endogeneity in non-linear models and some remain unable to decompose gender effects within the household.

Methodology and Data

Data and variables.

This study uses repeated cross-sectional data from the PSLM survey conducted by the Pakistan Bureau of Statistics (PBS) of the Government of Pakistan for the seven fully available rounds from 2005 to 2019 (2005–2006, 2007–2008, 2010–2011, 2011–2012, 2013–2014, 2015–2016, and 2018–2019). The survey was designed to provide social and economic indicators at the provincial and district level; starting in 2004, the survey aims to accurately describe the country. The sample size of the PSLM surveys is approximately 80,000 households. The total number of observations after pooling the data is 1,011,849.

This study uses two models for alternative measurements of the education achievement of boys and girls; the first model is education attainment (the highest completed schooling; aged 9–24 years), and the second model is current enrollment (aged 5–24 years). The boys and girls are restricted in first model to the 9–15, 16–19, and 20–24 age groups for primary-, secondary-, and tertiary-level education attainment, respectively. The following criteria are considered: additional year for class repetition by the students, late admission into schools, the completion standards of the Pakistan education system, and traditional age requirements for entering in school adopted in past studies (Maitra et al., 2003 ). In addition, boys and girls are restricted to not having the status of head or working person. In the first model, education attainment is a categorical outcome variable examines by the ordered logit model that can be defined as:

In the second model, current enrollment is a dichotomous outcome variable examines by logit model that can be described as:

The explanatory variables include dummy variables of the gender and age of boys and girls depending on models. Age is represented by a linear and quadratic term to control for birth cohort effects and capture non-linearity effects on education achievement. As age is directly proportional in contributing to cognitive skills and human capital, age square indicates marginal returns from age that decrease over time.

Other explanatory variables include the marital status of the household members (Kingdon, 2005 ). This study uses a series of dummy variables for the education level of individuals including the head of the household, parents, and other members of the household (i.e., those older than 24). This is because a joint-family structure is the majority form of family structure in Pakistan, and the head of the household is usually not the father but rather any elderly family member. Likewise, head’s personal treatment and decision-making influence on the education achievement. In addition, using parental education instead of maternal education is also feasible for gender difference analysis to avoid the issue of multicollinearity. Several tests are run to check for multicollinearity, including the variance inflation factor (VIF) and correlation matrix. The VIF for each predictor variable should be less than 10. It is 7.02 for the education attainment model and 4.12 for current enrollment model. 2 The siblings’ variable is used to control the reciprocal relationship between quantity and quality of education (Hazarika, 2001 ; Maitra, 2003 ). Occupational heterogeneity is controlled by different household members’ professions (McNabb, 2002 ) ranging from high-salaried (officer) to low-salaried (laborer) professions.

The variable of interest in this study is the per capita income of the household. It represents the household’s possible investment in education, which can maximize economic returns and minimize gender inequality. The availability of electricity, gas, and broadband internet access is a proxy for household infrastructure and technology advancement. The latter is of interest as it may impact on digital education, sustainable development, and the urgency of the micro- and macro-crisis such as health. The high demand to shift education from formal to virtual platforms during the COVID-19 pandemic has opened up new dimensions with regard to the acquisition of skills and knowledge. The other control variables consist of the dependency ratio, household size, ownership of house and any establishment other than agricultural land (ur Rahman et al., 2018 ), and ownership of the cultivating land for the personal use of the household (Sawada, 2009 ). Finally, community characteristics are controlled by including dummy variables for locations and number of the provinces of the country (Hazarika, 2001 ).

Empirical Strategy

The concept of the ordered logit model for education attainment is to incorporate intermediate continuous variable says y in the latent regression accompanied by the observed ( x i ) explanatory variables and the unobserved error term ( ε i ). The range of y is divided in adjacent intervals that comprise four categories—namely, 0 = no education, 1 = primary education, 2 = secondary education, and 3 = tertiary education—related to latent variable ( Y ∗ ). The structural model for latent education is:

where β is the vector of the parameters to be estimated; ε is the disturbance term, which is assumed to be independent across observations; and y ∗ can take value with observations.

For the discrete choices, the following are observing as:

Where Y is the category of education attainment, and τ denotes the threshold parameters, explaining the transition from one category of education attainment to another category. Consequently, τ must satisfy the rule according to τ 0  <  τ 1  <  τ 2  <  τ 3 , as the ε i is logistically distributed. The resulting probabilities can be observed as:

Hence, the probability of outcome can imply as:

whereas the log likelihood function for ordered logistic regression is:

The function formulates the ordered logit model with multiple equations, whereas each equation presents the logit model (Williams, 2005 ). The econometric model is therefore:

Endogeneity Bias

The main econometric challenge is to identify the endogeneity problems. There is the possibility that variable per capita income is likely to be related to unobservable factors that affect education achievement in many ways not included in the regression. There may be errors in measuring per capita income that bias the results. In addition, a causal relationship may exist between income and education achievement. This relationship might also be influenced by parental economic circumstances, social status, and any spurious third factor such as personal preferences. Reverse causality occurs when the poor educational performance of the boy or girl might lower household income and vice versa. Therefore, the model may suffer from omitted variable bias and (reverse) causality issues. The literature also explores per capita income as an endogenous variable that has instrumented by parental and household characteristics including employment, education, and farming activities (Bratti, 2007 ; Hoogerheide, 2012 ). Other studies examine its causal relationship with income shock (Coelli, 2005 ), the difference in households’ incomes, rainfall, and climate change in relation to productivity concerns (Fichera et al., 2015 ).

In the first model of education attainment, income shock such as head unemployment and non-labor resources of grandparents in the household are used as instruments for per capita income (Behrman et al., 1997 ). If the head of the household is unemployed, this is unlikely to influence the total years of schooling undertaken by boys and girls when there is a joint family structure where the parents are responsible for meeting educational expenditures. Similarly, the permanent or non-labor income of the grandparents is an exogenous and strong instrument that does not directly affect the total years of schooling undertaken by boys and girls (Bratti, 2007 ). However, these instruments may affect the current enrollment of boys and girls, thus necessitating the exploration of other exogenous variables. Therefore, the potential endogeneity in the current enrollment model is captured by another set of exogenous variables; first, the difference in per capita income between households (included in the PSLM survey) and country, and second, windfall income. The difference in per capita income is a proxy for income shock that does not relate to agricultural goods but rather a retrospective analysis of households having or not having wages. This difference may represent the transitional effect of the financial condition of the household (Björkman-Nyqvist, 2013 ; Sawada, 2009 ) in developing countries such as Pakistan. Similarly, windfall income comprises mainly of the unearned income of the household or non-labor income that includes lottery wins, inheritances, gifts, unexpected charity payments, and irregular sources of income (Kingdon, 2005 ; Powdthavee et al., 2013 ), which are exogenous.

Another source of endogeneity might arise due to the relationship between education spending and current enrollment in the logit model regression. The literature provides instruments for education spending such as community-, labor-, or industry-union membership of the household’s head that in unavailable in PSLM dataset while some studies refer to the head’s occupations (Maitra, 2003 ). The estimation results after instrumenting education spending with the head of household’s occupation show that the null hypothesis of homogeneity is not rejected, as it has a p -value of 0.93. However, this study tries to control educational spending through the addition of dummy occupational variables, home ownership, and land cultivation (Maitra, 2003 ; Shea, 2000 ). Other individual and socioeconomic characteristics are considered as exogenous. The OLS regression (for instrument validation) and alternative approaches to capture potential endogeneity—such as the control function approach, two-stage least squares (2SLS) (ignoring the nature of the outcome variable), and the IV probit model (splitting the outcome variable into a binary variable where necessary)—are also examined.

To apply the 2SRI method, the first step is to find exogenous variables; however, this method is different from the standard IV estimation method. The strategy behind choosing variables is that variables predict a possible definition of homogeneity. The argument behind this method (Terza, 2018 ) is based on the inappropriateness of the traditional linear instrumental variable estimator for the correction of the endogeneity problem. The core advantage of this method is that the estimated coefficients associated with the residuals from the first-stage regression significantly express the presence of endogeneity in the model (Huasman, 1978 ). In this method, the first stage consists of the OLS regression and predicts the endogenous variable by using the instruments and the rest of the explanatory variables. The second stage is estimated using the ordered logit model with the inclusion of the first-stage residuals. In the final stage, the whole program is set to be bootstrapped. The latent model will be established by splitting the explanatory variables into exogenous and endogenous variables, say X ex and X en , and the equation becomes:

The first-stage equation of the 2SRI method is estimated for income using all the exogenous variables and instruments in the OLS regression. It takes the form as:

where E ( X en ,  Z ) ≠ 0 and E ( ε i ,  Z ) = 0; β and γ are coefficient parameters; and v i and ε i are error terms, respectively. The second stage of the 2SRI method estimates outcome variable using the residuals obtained from the first-stage equation taken as control variables along with other explanatory variables. The model is described as:

This method is a simple test of endogeneity: if the residuals of the first stage are statistically significant, then the results will be biased in the first model, refer to control the endogeneity issue (Terza, 2018 ; Akarçay-Gürbüz & Polat, 2017 ).

Education and Inequality Parameters

The Gini coefficient for education, average years of schooling, and standard deviation are the inequality parameters that have been considered by observing the education system and structure of the country, the efficiency of learning performance, and variations in gender-specific education investment (Digdowiseiso, 2010 ; Thomas et al., 2001 ). 3 The consideration of these inequalities might help to reveal the socioeconomic and intrahousehold factors behind the different treatment for girls’ education. Therefore, the extended model can be described as:

The gender decomposition examines while using the basic models of each specification by the mean, the coefficient (Kingdom, 2005 ), and the interactions of the boy dummy variable (Maitra, 2003 ). Furthermore, the results are decomposed for gender effects by variant type Oaxaca decomposition (Dong et al., 2009 ; Golsteyn et al., 2014 ; Pal, 2004 ). This approach is generally used to examine the gender effects related to economic returns and the wage gap (Oaxaca, 1973 ). In this study, however, the standard approach has been modified to examine the gender effects related to education achievement. The probability of education attainment determines, say AT, separately for girls and boys with other characteristics, say X g and X b , respectively. Assuming Pr A T , X i , θ i * is the expected probability of AT and θ i * is the vector if the maximum likelihood estimates of the parameters of the ordered logit model for i  =  g , b for girls and boys, respectively, the expected AT for any individual would be:

Using expected education attainment for the boys’ and girls’ samples, respectively, one can decompose the boy-girl differential in alternative ways as follows:

In brief, the explained variation is attributable to the different characteristics of boy-girl, while the unexplained variation is attributable to the different treatment of boys and girls in the household. This is achieved by allowing the parameters to vary while the characteristics are held constant. A similar approach was adopted for current enrollment as well.

The alternative specifications explore the impact of gender inequalities in education achievement on the household income using OLS regression. This study uses three different measurements of gender difference. Considering education attainment, the first indicator—the gender gap 4 —is calculated as the difference in illiteracy rates between girls and boys (Cooray, 2011 ). The second indicator, gender difference, 5 measures the difference in education attainment between girls and boys (Baliamoune–Lutz & McGillivray, 2015 ), while the final indicator—the gender gap ratio 6 —is constructed based on the difference between boys’ and girls’ education attainment (Digdowiseiso, 2010 ). Similar inequalities in current enrollment for boys and girls (5–24) are also estimated. The alternative specification estimates in the linear regression model are defined as:

Furthermore, the robustness tests for education achievement are examined using several other specifications including ordered probit and probit models, another explanatory variable—per capita expenditure of the household, and provincial heterogeneity.

Descriptive Statistics

The detailed descriptive statistics of the selected variables are exhibited in Table ​ Table1. 1 . On average, 10 % of boys and girls attain a primary level of education, and 2.1 % attain a tertiary level of education. On average, the variable gender signifies 49 % girls in first model of education attainment and 48.8 % in second model of current enrollment. On average, 38.9 % of boys and girls are currently enrolled in education, and per capita income (in the log) is 8.8 (see Fig. ​ Fig.2). 2 ). Overall, the age of the household has a nonlinear effect; as with the increase of age of the household’ members, there is decrease in the education level (see Fig. ​ Fig.3). 3 ). Meanwhile, this study uses the age of the boys and girls, according to the models’ criteria. The mean age in the first model is 15.95 years whereas it is 13.59 in the second model. This study observes a higher ratio of low-salaried occupations (for example, machine operators); thus, the dependency ratio is also higher at 41.6 %. A total of 44.5 % of the population lives in urban areas, 80.6 % receive electricity, and 37.7 % have access to gas supplies. Among other provinces, the highest population locates in the province of Punjab.

Description and summary statistics of selected variables

Household’s income in Pakistan. Source: Author construction based on data from PSLM Bureau of Statistics, Pakistan. Figure 2 displays the trend of per capita income from 2005 to 2019, one of the inevitable indicators of educational achievement. The statistics calculate a sharp drop in per capita income after 2010, which improved in 2012 but eventually declined after 2016

Education attainment by age (2005–2019). Source: Author construction based on data from PSLM Bureau of Statistics, Pakistan. Figure 3 expresses the predictive margins between the age of the persons living in the household and their education levels. The probability of primary education attainment decreases after 25 years of age, whereas it is the opposite for the tertiary level. Meanwhile, with the increase in age, it is more likely to achieve secondary education

Empirical Results and Discussion

Determining education attainment and current enrollment levels.

Table ​ Table2 2 describes the average marginal effects of the ordered logit model for primary-, secondary-, tertiary-level, and no education attainment with the help of household income per capita and various socioeconomic characteristics. In the full sample models, variable gender—girl, increases primary-, secondary-, and tertiary-level education attainment by 0.4, 0.5, and 0.2 percentage points, respectively, at the 1 % significance level; this contradicts the findings of Munshi ( 2017 ). Per capita income, on average, increases the likelihood of primary-, secondary-, and tertiary-level education attainment by 0.1, 0.2, and 0.1 percentage points, respectively. The effect of age is more likely to increase secondary- and tertiary- education attainment. As findings reveal that the transitional effect of education attainment is progressive from primary level to secondary level, however, it does not appear with same proportion from secondary level to tertiary level. The impact of the age and squared-age has non-linear effect that can be justified in two manners. Firstly, with the increase in age, the proportion of transition of the education attainment levels decreases. Secondly, there is a negative relationship between the term squared age and education attainment.

Average marginal effects for education attainment: ordered logit model regression

The dependent variable is education attainment that is categorical variable. The category 1 displays for primary, 2 for secondary, and 3 for tertiary level of education and 0 demonstrates none education. The reference category for the digital is no direct connection and any extension of broadband. Reference province is Punjab. Robust standard errors are in parentheses. Significance levels are denoted as *** p <0.01, ** p <0.05, and * p <0.1

Meanwhile, the presence of an educated head of household significantly improves primary-, secondary-, and tertiary-level education attainment—by 5.0, 10.4, and 4.1 percentage points, respectively. Other household members are likely to increase secondary- and tertiary-level education attainment by 20.9 and 11.5 percentage points, respectively. The results show higher marginal effects for education attainment by technicians (low-salaried) compare to managers (high-salaried), indicating that lower occupations have strong inspiration to maximize the human resources capital of the household. In addition, the availability of electricity, internet access, and access to a gas supply are highly likely to enhance education attainment. On average, living in urban area has the likelihood to impact primary-, secondary-, and tertiary education attainment by 0.2, 0.3, and 0.1 percentage points, respectively.

From models 5 to 8, for girls, it can be seen that per capita income significantly increases each level of education attainment. However, it increases secondary-level education attainment more than other levels, by 0.2 percentage points. Age has a significant and nonlinear effect. The variable married is likely to decrease the probability of education attainment by 1.6 and 5.0 percentage points at the tertiary and secondary levels of education, respectively. Interesting, parental education has a positive influence, but it is only significant at the secondary education attainment with 23.3 percentage points. In addition, the presence of an educated head of household and other members also provides a positive and significant effect. On analyzing different occupations, the results indicate a 19.6, 23.7, 8.9, and 1.5 percentage point increase in tertiary education attainment by clerks, officers, managers, and machine operators. The household size shows an inverse relationship with girls’ education attainment, particularly at the secondary level. The household infrastructure provides positive effect on girls’ education attainment. It may exhibit that sustainable consumption of household resources including electricity and gas can exert female education that can promote gender equity and economic returns.

From models 9 to 12, for education attainment, it can be seen that the impact of the per capita income of the household is comparatively equal for boys and girls. The household income is likely to increase secondary—and tertiary—education attainment in boys by 0.2 and 0.1 percentage points, respectively. Parental education is highly unlikely to increase the probability of education attainment. The presence of an educated head of household increases education attainment by 4.9, 9.5, and 3.2 percentage points at the primary, secondary, and tertiary levels, respectively. Similarly, the presence of household members with numeracy skills and secondary education is likely to increase secondary-level education attainment by 10.9 and 22.2 percentage points, respectively. This study observes a strong impact of occupational heterogeneity on education attainment; officers and clerks significantly improve the primary- and secondary-level education attainment in boys. The clerks are highly likely to increase tertiary-level education attainment, by 15.4 percentage points. Compared to Punjab, provinces such as KPK and Balochistan are less likely to increase primary, secondary, and tertiary education attainment.

Table ​ Table3 3 describes the average marginal effects from the current enrollment models with the help of logit model regression.

Average marginal effects for current enrollment: logit model regression

The dependent variable current enrollment is a binary variable. The category 1 displays for current enrollment in primary, secondary, or tertiary education and 0 demonstrates no current enrollment. Reference province is Punjab. Significance levels are denoted as *** p <0.01, ** p <0.05, and * p <0

In the full model, the estimate of the variable girl is highly significant and negative—an opposite finding to that of past studies (Maitra, 2003 )—and likely to decrease the probability of current enrollment in education by 0.8 percentage points. A unit increase in income per capita is more likely to improve the current enrollment rates for girls than it is for boys; an increase of 0.4 percentage points is observed for girls. Age has a nonlinear effect with its squared term; thus, current enrollment rates decrease with age. Additionally, variable married decreases the probability of current enrollment in education in girls by 15.6 percentage points. Current enrollment increases for boys if there are educated household members; however, this is not the case for certain professions such as clerks and machine operators.

Other indicators associated with physical capital such as ownership of establishment or land are negatively related to current enrollment rates. This indicates that education is not the primary objective among landowners, as they do not worry about employment. The educational transition from primary to higher grades is less valuable than monetary assets, and most people are reluctant to leave their ancestral profession if it is associated with land cultivation. Household infrastructure is likely to benefit girls more than boys, however when we examine the influence of living in an urban location, which is highly likely to increase enrollment rates in education for boys. The dependency ratio provides higher marginal effects for current enrollment in boys, which further supports the objective of this study. The majority of the households in Pakistan support male earners who are likely to bear all the expenditures. Therefore, the parents prefer to invest in boys’ education for potential job opportunities and financial support in the long run. Results from siblings shows a positive relation to current enrollment and reveal higher quantity-to-quality trade-offs particularly among girls. The results show a higher marginal effect in KPK province; this might be due to the new framework of free and accessible education that has been in place since 2013 (KPK Government Statistics, 2021).

Dealing with Endogeneity Bias

Table ​ Table4 4 shows the results of the average marginal effects using the ordered logit model regression/2SRI approach after dealing with endogeneity. In the full sample, the per capita income of the household is likely to increase education attainment at each level by a higher ratio compared to the aforementioned results. There is a drastic increase in primary-level education attainment: 11.2 percentage points. Likewise, secondary- and tertiary-level education attainment increase by 15.9 and 4.9 percentage points, respectively. Even the variable gender is almost two times higher than the previous results for secondary-level education attainment. Other indicators that illustrate higher marginal effects are educated head of household, household size, and infrastructure. The results find a positive relationship between education and urbanization by introducing income shock of head unemployment and non-labor resources. It retrieves two strong arguments; first, the income shock is likely to increase potential human mobilization for confronting household economic burden. The second, non-labor resources exert positive impact on population by increasing non-market activities, as time allocation shifts from work to leisure.

Average marginal effects for education attainment by IV approach: 2SRI/ordered logit model regression

The dependent variable is education attainment that is categorical variable. The category 1 displays for primary, 2 for secondary, and 3 for tertiary level of education and 0 demonstrates no education. The reference category for the digital is no direct connection and any extension. Reference province is Punjab. The instruments are income shock described as head unemployed and grandparents’ resources. The validity of instruments estimates with 2SLS estimators. The Hausman test provides F-statistics and test of overidentification states p -value. The value for first-stage regressions gives F-statistics. Bootstrapped standard errors are presented in parentheses. Significance levels are denoted as *** p <0.01, ** p <0.05, and * p <0.1

From models 5 to 8, for girls, the results are significant but with higher marginal effects than the full sample. A sharp increase in secondary-level education attainment is caused by household income: an increase of 10.8 percentage points. Results find negative relationship between married persons and education attainment of the girls, especially at primary level. It might be possible that married persons are quite young in age, particularly women, without having any education awareness and sufficient resources. These results may indicate the need of awareness programs in the household to encourage women education and discourage early-age marriages. On the other hand, a significant decrease in household size supports an increase in primary-level education attainment.

There is a higher impact of per capita income on boys’ education attainment than girls, indicating household’s preferences. The per capita income of the household is likely to increase primary-, secondary- and tertiary-level education attainment by 13.0, 17.9, and 4.9 percentage points, respectively. The presence of an educated head of household has a strong and positive effect on boys’ education attainment; however, it is the opposite for girls’ primary-level education attainment. The results show that intermediate internet access is more effective for girls than boys. Meanwhile, household size also impacts quite positively on boys’ education attainment as they are potential lone bread earners for their families. Living in an urban location results the potential career for boys, thus revealing a positive correlation with education attainment.

The average marginal effects are shown in Table ​ Table5 5 for current enrollment after dealing with potential endogeneity. Per capita income is four times more likely to increase the likelihood of current enrollment in the full sample than the results reported in the “Determining Education Attainment and Current Enrollment Levels” section. Its impact is 4.4 percentage points for boys and girls. The variable girl reduces the probability of current enrollment by 0.3 percentage points. The results find a significant effect of parental education on boys, thus revealing a gender bias in investment in education. Similar results are reported for the impact of educated members of the household and the occupations of those living in the household. The other results describe a wider gap in current enrollment in Sindh and Balochistan, where girls are highly unlikely to enroll in any kind of educational institution.

Average marginal effects for current enrollment by IV approach: 2SRI/logit model regression

The dependent variable current enrollment is a binary variable. The category 1 displays for current enrollment in primary, secondary, or tertiary education and 0 demonstrates no current enrollment. Reference province is Punjab. The set of instruments used in these models are income shocks, first, income windfall, and second, income difference. The validity of instruments estimates with 2SLS estimators. Robust standard errors are in parentheses. The Hausman test provides F-statistics and test of overidentification states p -value. The value for first-stage regressions gives F-statistics. Significance levels are denoted as *** p <0.01, ** p <0.05, and * p <0

Estimations of Education Attainment and Current Enrollment by Inequalities

Table ​ Table6 6 illustrates the average marginal effects by incorporating different educational inequalities such as the Gini coefficient, years of schooling (on average), and standard deviation for education attainment by ordered logit model, as shown in panels A, B, and C. For this moment, only results with educational inequalities have been provided. Full results can be provided on demand. In girls’ sample, by examining panel A, we can see that the Gini coefficient is highly significant and indicates a sharp decrease in tertiary- and secondary-level education attainment, by 0.6 and 1.6 percentage points, respectively. Furthermore, in panel B, the average years of schooling have positive relationship with secondary- and tertiary-level education attainment. In panel C, the estimates explain that the standard deviation inequality decreases secondary- and tertiary-level education attainment by 0.1 percentage points, respectively. For boys’ sample, in panel A, the results show that the Gini coefficient decreases the secondary- and tertiary-level education attainment of boys; however, the marginal effects are slightly higher compared to those for girls. In panel B of average years of education, there is an equal improvement in secondary- and tertiary-level education attainment of boys; however, no significant effect is found in panel C.

Average marginal effects for education attainment with education inequalities: ordered logit model regression

The dependent variable is education attainment that is categorical variable. The category 1 displays for primary, 2 for secondary, and 3 for tertiary level of education and 0 demonstrates none education. Panel A contains Gini coefficient, panel B includes average years of schooling, and panel C includes standard deviation, calculated for education attainment (9–24). Each panel is individual estimation and contains individuals, household, community characteristics, and threshold points. Robust standard errors are in parentheses. Significance levels are denoted as *** p <0.01, ** p <0.05, and * p <0.1

The relationship between current enrollment and educational inequalities is shown in Table ​ Table7. 7 . In panel A, the results indicate that educational inequalities impact both boys and girls. However, examining the marginal effects by gender, the Gini coefficient is found to be higher for boys. In panel B, the average years of schooling of currently enrolled boys and girls are higher for girls by 0.6 percentage points. This indicates that girls are almost 0.4 times more likely to enroll in school. There is an insignificant impact of standard deviation on boys’ current enrollment; however, it is the opposite for girls. A unit increase in standard deviation decreases the probability of girls’ current enrollment by 0.2 percentage points.

Average marginal effects for current enrollment with education inequalities: logit model regression

The dependent variable current enrollment is binary variable. The category 1 displays current enrollment in primary, secondary, and tertiary education and 0 demonstrates no current enrollment. Panel A contains Gini coefficient, panel B includes average years of schooling, and panel C includes standard deviation, calculated for current enrollment (5–24). Each panel is individual estimation and contains individuals, household, and community characteristics. Robust standard errors are in parentheses. Significance levels are denoted as *** p <0.01, ** p <0.05, and * p <0.1

Explaining the Gender Gap and its Decomposition

Table ​ Table8 8 provides mean statistics and differences in the coefficients in relation to education attainment.

Gender differences of selected variables: mean, coefficient, and interaction estimations

Significance levels are denoted as *** p <0.01, ** p <0.05, and * p <0.1

In panel A, most of the household characteristics favor girls; these include personal attributes such as age and infrastructure while per capita income, educated members, head, and urbanization provide higher mean probabilities for boys ’education attainment. The difference between boys and girls is shown in the last column by interacting the boy dummy variable with each explanatory variable as an additional regressor in the basic model of the full sample using ordered logit model regression. The estimates find favorable values for girls’ education attainment in relation to the education level of her parents and the head of the household and household characteristics. Panel B provides mean statistics and coefficient differences for current enrollment. The personal attributes such as age, married, and infrastructure have the higher mean probabilities for girls’ education attainment. The last column displays the differences between boys and girls and shows that educated head, urbanization, and provinces favor boys.

Table ​ Table9 9 presents the gender differences in education attainment and current enrollment by predicted probabilities using variant type Oaxaca decomposition by incorporating four scenarios. Such as (i) girls using estimated parameters obtained from girls’ equation, (ii) girls using estimated parameters obtained from boys’ equation, (iii) boys using estimated parameters obtained from boys’ equation, and (iv) boys using estimated parameters obtained from girls’ equation (Pal, 2004 ). Comparatively, boys are having approximately two times lower corresponding probabilities using girls’ parameters. Conversely, the probability of girls’ education attainment increases almost two times higher using boys’ parameters. A similar proportion of increase observes in girls’ current enrollment using boys’ parameters. While two times lower probabilities observe for boys’ current enrollment using girls’ parameters. The estimates of difference are presented with the boys’ reference. In the end, explained and unexplained variations of gender difference are estimated. While explained variation in education attainment and current enrollment are −142.8 and 41.4 %, respectively (Dong et al., 2009 ). The unexplained variation, generally considers as discrimination, has higher values in both models and highlight the different treatment between boys and girls in the household. However, this study presents such variation as gender differences that may be due to unobservable factors and imperfectly observable attributes.

Gender decomposition by predicted probabilities

Alternative Specification and Robustness Tests

In Tables ​ Tables10 10 and ​ and11, 11 , the estimates are presented for education attainment and current enrollment using other models such as ordered probit and probit models (McNabb et al., 2002 ), and other variables such as per capita expenditure and permanent income (non-labor assets) are included.

Average marginal effects for education attainment: ordered probit model, per capita expenditure, and permanent income

The dependent variable is education attainment that is categorical variable. The category 1 displays for primary, 2 for secondary, and 3 for tertiary level of education and 0 demonstrates no education. Panel A provides results of ordered probit model, panel B demonstrates results of per capita expenditure, and panel C shows results of permanent income. Each panel is individual estimation and contains individuals, household, community characteristics, and threshold points. Robust standard errors are in parentheses. Significance levels are denoted as *** p <0.01, ** p <0.05, and * p <0.1

Average marginal effects for current enrollment: ordered probit model, per capita expenditure, and permanent income

The dependent variable is current enrollment that is binary. The category 1 displays current enrollment in primary, secondary, and tertiary education and 0 demonstrates no current enrollment. Panel A provides results of ordered probit model, panel B demonstrates results of per capita expenditure, and panel C shows results of permanent income. Each panel is individual estimation and contains individuals, household, community characteristics, and threshold points. Robust standard errors are in parentheses. Significance levels are denoted as *** p <0.01, ** p <0.05, and * p <0.1

In both models, the results are highly significant and provide additional evidence to support the previous estimations. The variable girl is more likely to increase education attainment at the secondary level. The unit increase in income per capita is marginally higher in the probit model regression. The findings show that per capita expenditure is likely to positively impact on girls’ education, particularly in relation to secondary-level education attainment. Considering the robust test by incorporating the permanent income of the household, the variable gender is positively significant with education attainment. A unit increase in permanent income raises primary- and secondary-level education attainment more in boys. In addition, there is sharp increase in boys’ current enrollment with a unit increase in permanent income. Other robustness tests, including provincial heterogeneity, the control function approach, IV probit, 2SLS, and the determination of education attainment and current enrollment for boys and girls from a different age group (13–24), are available on request.

Table ​ Table12 12 presents results for alternative specification where per capita income is the dependent variable and gender inequalities (education attainment and current enrollment) as interested variables. This specification can also be interpreted as the future earning potential of girls and boys. Considering education attainment, in panel A, the gender gap due to illiteracy decreases income by approximately 11.3 % more in girls compared to boys. In panel B and C, gender difference is likely to decrease income by 3.2 and 1.2 % in girls. Moving toward current enrollment, in panels A, B, and C, each gender inequality reduces the household income comparatively higher among girls than boys by 7.1, 3.0, and 1.7 %.

Relationship between gender differences in education and income: alternative specification/ ordinary least square regression

The dependent variable is household’s per capita income. Panel A includes gender gap in illiteracy, panel B contains gender difference, and panel C is the ratio between boys and girls, each of these inequalities calculated for education attainment of boys and girls (9–24) and current enrollment of boys and girls (5–24), respectively. Each panel contains individuals, household, and community characteristics. Robust standard errors are in parentheses. Significance levels are denote as *** p <0.01, ** p <0.05, and * p <0.1

Conclusion and Policy Implications

Despite having the potential for human resource capital, Pakistan struggles with extreme poverty, socioeconomic disparity, and gender inequality at the grass-root level (Ali et., 2021 ; Asif et al., 2019 ). To address these undeniable issues, it becomes crucial to comprehend the significance of the equal distribution of household resources in education regardless of gender that builds a sustainable economic structure toward global equality (Kopnina, 2020 ). This study aims to examine education achievement and underlying gender differences using two models: education attainment and current enrollment. The findings highlight the importance of the relationship between education and income along with other household characteristics. This study deals with potential endogeneity by using the 2SRI approach and examines gender and educational inequalities at the micro level.

The findings demonstrate that household income has a significant and positive impact on education attainment and the current enrollment of boys and girls. The education attainment transition from primary to tertiary-level is successful that supports the past studies (Duflo et al., 2021 ; Wu et al., 2020 ). However, the transition from primary to secondary education is higher than that from secondary to tertiary education attainment. The community and individuals’ attributes support education investment in boys indicating household and socioeconomic preferences. Girls can improve their education with the availability of personal and household attributes (Yi et al., 2015 ). Other findings from education attainment and current enrollment models predict a demographic framework that encourages a sustainable environment with a decline in household size and dependency ratio (Heb, 2020 ; Asif, 2019 ; Fichera et al., 2015 ). These findings contradict those of past studies (Munshi, 2017 ) and establish a link between temporary residents (daughters) and different occupations of the households, whereby lower-salaried households and deprived areas can significantly improve female education attainment and current enrollment.

The findings show that there is a negative relationship between the Gini coefficient and education attainment and that this gap is wider at secondary and tertiary education levels, thus supporting the results of the basic model. The standard deviation of educational inequalities is higher for girls that further confirms the existence of gender differences in education. Likewise, the findings from the alternative specifications provide decrease in potential economic returns on education by gender inequalities. The findings support those of Pfeffer et al. ( 2018 ) with regard to discouraging wealth accumulation in terms of physical capital and increasing investment in female education (Kopnina, 2020 ). It can effectively transform the developing society of Pakistan by framing public policies for women’s empowerment (United Nations Education, 2030), gender equality (Arshed et al., 2019 ), poverty alleviation (ur Rahman et al., 2018 ), and sustainable development (Sen, 2019 ). Therefore, this study identifies some valuable recommendations for policymakers wishing to promote gender equality:

• Implement cooperative projects created by federal and local governments that supply free, digital, and up-to-date education in schools, colleges, and universities to improve transition levels, with a particular focus on poor infrastructure, highly deprived regions, and mobility restricted areas.
• Adopt targeted policies to minimize education and gender gaps between those enrolled and not enrolled in education by supporting low-income households through the allocation of funds, scholarships, and incentives.
• Reform educational strategies to provide cost effective education in collaboration with parents, teachers, and schools with the aim of creating advanced and scientific curricula aligned with sustainable development goals.
• Craft awareness campaigns to eradicate gender-specific investment in education, encourage talented females to enter tertiary-level education in particular, and address socioeconomic challenges by establishing reliable and organized educational committees in each province.

Finally, some potential limitations should be noted, as these might open up new horizons for future research. Quantitative research should be conducted to examine other household characteristics and upcoming survey rounds than those discussed in this study.

Acknowledgements

This study benefited from the discussion with the participants of the American Economic Association Annual Meeting, 2022, USA; 49th Australian Conference of Economists CEA 2021, Australia; International Population Association Conference, 1PC 2021; 23 INFER Annual conference, 2021 Porugal; and 23rd Applied Economics Meeting, ALdE, 2021, Spain. I would like to thank Prof. Theophile T. Azomahou (CNRS- CERDI), Prof. Colin Green (Norwegian University of Science and Technology), Dr. Ababacar Sedikh, and Dr. Nestor Sawadogo (CNRS-CERDI) for their thoughtful comments. I also would like to thank the editor and two anonymous reviewers for their valuable suggestions.

Data Availability Statement

Declarations.

Not required.

The author declares no competing interests.

1 Translated as Knowledge Possible project that having contribution of US $ 1 billion for sustainable programs.

2 These tests are available on request.

where E L is the Gini coefficient based on education attainment, 𝜇 is average years of schooling, Pi and P j are the proportion of the population, y i and y j are years of schooling at different educational levels, and n is the number of levels in the attainment data. Average years of schooling and standard deviation can be calculated as follows: AYS = ∑ i = 1 n P i y i - μ 2 and SD = ∑ i = 1 n P i y i - μ 2 .

4 Gender gap (illiteracy rates [9–24 years of age]) = number of girls’ illiterate – number of boys’ illiterate

5 Gender difference (education attainment [9–24 years of age]) = total years of schooling of girls – total years of schooling of boys.

6 Gender gap ratio (education attainment [9–24 years of age]) = total years of schooling of girls/total years of schooling of boys.

Submission Declaration and Verification

I declare that the article has not been published previously and is not under consideration for publication elsewhere.

Publisher’s Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

• Akarçay-Gürbüz A, Polat S. Schooling opportunities and intergenerational educational mobility in Turkey: An IV estimation using Census data. The Journal of Development Studies. 2017; 53 (9):1396–1413. doi: 10.1080/00220388.2016.1234038. [ CrossRef ] [ Google Scholar ]
• Alfalih AA, Alfalih AA. The impact of background characteristics on graduate school attendance: Evidence from a developing economy. Journal of the Knowledge Economy. 2021; 12 (1):363–383. doi: 10.1007/s13132-020-00636-2. [ CrossRef ] [ Google Scholar ]
• Ali, A., Audi, M., Bibi, C., & Roussel, Y. (2021). The impact of gender inequality and environmental degradation on human well-being in the case of Pakistan: A time series analysis
• Ali A, Mottaleb KA, Aryal JP. Wealth, education and cooking-fuel choices among rural households in Pakistan. Energy Strategy Reviews. 2019; 24 :236–243. doi: 10.1016/j.esr.2019.03.005. [ CrossRef ] [ Google Scholar ]
• Arshed N, et al. Education stock and its implication for income inequality: The case of Asian economies. Review of Development Economics. 2019; 23 (2):1050–1066. doi: 10.1111/rode.12585. [ CrossRef ] [ Google Scholar ]
• Ashraf I, Ali A. Socio-economic well-being and women status in Pakistan: An empirical analysis. Bulletin of Business and Economics (BBE) 2018; 7 (2):46–58. [ Google Scholar ]
• Asif MF, Pervaiz Z. Socio-demographic determinants of unmet need for family planning among married women in Pakistan. BMC Public Health. 2019; 19 (1):1–8. doi: 10.1186/s12889-019-7487-5. [ PMC free article ] [ PubMed ] [ CrossRef ] [ Google Scholar ]
• Assoumou-Ella, G. (2019). Gender inequality in education and per capita GDP: The case of CEMAC countries. Economics Bulletin, 39 (2).
• Baliamoune–Lutz, M., & McGillivray, M. (2015). The impact of gender inequality in education on income in Africa and the Middle East. Economic Modelling, 47 , 1–11.
• Banzragch O, Mizunoya S, Bayarjargal M. Education inequality in Mongolia: Measurement and causes. International Journal of Educational Development. 2019; 68 :68–79. doi: 10.1016/j.ijedudev.2019.04.010. [ CrossRef ] [ Google Scholar ]
• Behrman JR, Khan S, Ross D, Sabot R. School quality and cognitive achievement production: A case study for rural Pakistan. Economics of Education Review. 1997; 16 (2):127–142. doi: 10.1016/S0272-7757(96)00045-3. [ CrossRef ] [ Google Scholar ]
• Björkman-Nyqvist M. Income shocks and gender gaps in education: Evidence from Uganda. Journal of Development Economics. 2013; 105 :237–253. doi: 10.1016/j.jdeveco.2013.07.013. [ CrossRef ] [ Google Scholar ]
• Bratti M. Parents’ income and children’s school dropout at 16 in England and Wales: Evidence from the 1970 British cohort study. Review of Economics of the Household. 2007; 5 (1):15–40. doi: 10.1007/s11150-007-9001-6. [ CrossRef ] [ Google Scholar ]
• Chevalier A, Lanot G. The relative effect of family characteristics and financial situation on educational achievement. Education Economics. 2002; 10 (2):165–181. doi: 10.1080/09645290210126904. [ CrossRef ] [ Google Scholar ]
• Chevalier A, Harmon C, O’Sullivan V, Walker I. The impact of parental income and education on the schooling of their children. IZA Journal of Labor Economics. 2013; 2 (1):8. doi: 10.1186/2193-8997-2-8. [ CrossRef ] [ Google Scholar ]
• Coelli, M. I. C. K. (2005). Parental income shocks and the education attainment of youth
• Cooray A, Potrafke N. Gender inequality in education: Political institutions or culture and religion? European Journal of Political Economy. 2011; 27 (2):268–280. doi: 10.1016/j.ejpoleco.2010.08.004. [ CrossRef ] [ Google Scholar ]
• Davis LS, Williamson CR. Does individualism promote gender equality? World Development. 2019; 123 :104627. doi: 10.1016/j.worlddev.2019.104627. [ CrossRef ] [ Google Scholar ]
• de Bruin A, Liu N. The urbanization-household gender inequality nexus: Evidence from time allocation in China. China Economic Review. 2020; 60 :101301. doi: 10.1016/j.chieco.2019.05.001. [ CrossRef ] [ Google Scholar ]
• Digdowiseiso, K. (2010). Measuring Gini coefficient of education: The Indonesian cases
• Du F, Dong XY. Why do women have longer durations of unemployment than men in post-restructuring urban China? Cambridge Journal of Economics. 2009; 33 (2):233–252. doi: 10.1093/cje/ben034. [ CrossRef ] [ Google Scholar ]
• Duflo E, Dupas P, Kremer M. The impact of free secondary education: Experimental evidence from Ghana . National Bureau of Economic Research; 2021. [ Google Scholar ]
• Durrani N, Halai A. Dynamics of gender justice, conflict and social cohesion: Analysing educational reforms in Pakistan. International Journal of Educational Development. 2018; 61 :27–39. doi: 10.1016/j.ijedudev.2017.11.010. [ CrossRef ] [ Google Scholar ]
• Evans, D. K., Akmal, M., & Jakiela, P. (2021). Gender gaps in education: The long view. IZA Journal of Development and Migration, 12 (1).
• Fichera E, Savage D. Income and health in Tanzania An instrumental variable approach. World Development. 2015; 66 :500–515. doi: 10.1016/j.worlddev.2014.09.016. [ PMC free article ] [ PubMed ] [ CrossRef ] [ Google Scholar ]
• Golsteyn BH, Schils T. Gender gaps in primary school achievement: A decomposition into endowments and returns to IQ and non-cognitive factors. Economics of Education Review. 2014; 41 :176–187. doi: 10.1016/j.econedurev.2014.04.001. [ CrossRef ] [ Google Scholar ]
• Hausman JA. Specification tests in econometrics. Econometrica: Journal of the Econometric Society. 1978; 46 (6):1251–1271. doi: 10.2307/1913827. [ CrossRef ] [ Google Scholar ]
• Hazarika G. The sensitivity of primary school enrollment to the costs of post-primary schooling in rural Pakistan: A gender perspective. Education Economics. 2001; 9 (3):237–244. doi: 10.1080/09645290110086117. [ CrossRef ] [ Google Scholar ]
• Heß P. SDG 5 and the gender Gap in standardization: Empirical evidence from Germany. Sustainability. 2020; 12 (20):8699. doi: 10.3390/su12208699. [ CrossRef ] [ Google Scholar ]
• Hoogerheide L, Block JH, Thurik R. Family background variables as instruments for education in income regressions: A Bayesian analysis. Economics of Education Review. 2012; 31 (5):515–523. doi: 10.1016/j.econedurev.2012.03.001. [ CrossRef ] [ Google Scholar ]
• Kabeer, N. (2021). Three faces of agency in feminist economics: Capabilities, empowerment, and citizenship. The Routledge handbook of feminist economics, 99-107
• Karoui K, Feki R. The effect of gender inequality on economic development: Case of African countries. Journal of the Knowledge Economy. 2018; 9 (1):294–300. doi: 10.1007/s13132-015-0341-9. [ CrossRef ] [ Google Scholar ]
• Kim SY. Determining critical factors of gender inequality: Evidence from 34 OECD and non-OECD countries. World Development Perspectives. 2021; 21 :100284. doi: 10.1016/j.wdp.2020.100284. [ CrossRef ] [ Google Scholar ]
• Kingdon GG. Where has all the bias gone? Detecting gender bias in the intrahousehold allocation of educational expenditure. Economic Development and Cultural Change. 2005; 53 (2):409–451. doi: 10.1086/425379. [ CrossRef ] [ Google Scholar ]
• Kleven H, Landais C, Søgaard JE. Children and gender inequality: Evidence from Denmark. American Economic Journal: Applied Economics. 2019; 11 (4):181–209. [ Google Scholar ]
• Kopnina H. Education for the future? Critical evaluation of education for sustainable development goals. The Journal of Environmental Education. 2020; 51 (4):280–291. doi: 10.1080/00958964.2019.1710444. [ CrossRef ] [ Google Scholar ]
• Livingstone DW. The education-jobs gap: Underemployment or economic democracy? Routledge; 2018. [ Google Scholar ]
• Lloyd CB, Mete C, Sathar ZA. The effect of gender differences in primary school access, type, and quality on the decision to enroll in rural Pakistan. Economic Development and Cultural Change. 2005; 53 (3):685–710. doi: 10.1086/427042. [ CrossRef ] [ Google Scholar ]
• Mahmood, A., Mahmood, S. T., & Malik, A. B. (2012). A comparative study of student satisfaction level in distance learning and live classroom at higher education level. Turkish Online Journal of Distance Education, 13 (1).
• Maitra P. Schooling and education attainment: Evidence from Bangladesh. Education Economics. 2003; 11 (2):129–153. doi: 10.1080/09645290210131665. [ CrossRef ] [ Google Scholar ]
• McNabb R, Pal S, Sloane P. Gender differences in education attainment: The case of university students in England and Wales. Economica. 2002; 69 (275):481–503. doi: 10.1111/1468-0335.00295. [ CrossRef ] [ Google Scholar ]
• Minasyan A, Zenker J, Klasen S, Vollmer S. Educational gender gaps and economic growth: A systematic review and meta-regression analysis. World Development. 2019; 122 :199–217. doi: 10.1016/j.worlddev.2019.05.006. [ CrossRef ] [ Google Scholar ]
• Munshi S. Arranged marriage, education, and dowry: A contract-theoretic perspective. Journal of Economic Development. 2017; 42 (1):35. doi: 10.35866/caujed.2017.42.1.003. [ CrossRef ] [ Google Scholar ]
• Najeeb, F., Morales, M., & Lopez-Acevedo, G. (2020). Analyzing female employment trends in South Asia. World Bank Policy Research Working Paper, 9157 .
• Oaxaca, R. (1973). Male-female wage differentials in urban labor markets. International Economic Review , 693–709.
• Okabe M. Gender-preferential intergenerational patterns in primary educational attainment: An econometric approach to a case in rural Mindanao, the Philippines. International Journal of Educational Development. 2016; 46 :125–142. doi: 10.1016/j.ijedudev.2015.10.006. [ CrossRef ] [ Google Scholar ]
• Pal S. How much of the gender difference in child school enrolment can be explained? Evidence from rural India. Bulletin of Economic Research. 2004; 56 (2):133–158. doi: 10.1111/j.1467-8586.2004.00193.x. [ CrossRef ] [ Google Scholar ]
• Paul, J., & Wahlberg, K. (2002). Global taxes for global priorities. In Global Policy Forum.
• Pfeffer FT. Growing wealth gaps in education. Demography. 2018; 55 (3):1033–1068. doi: 10.1007/s13524-018-0666-7. [ PMC free article ] [ PubMed ] [ CrossRef ] [ Google Scholar ]
• Powdthavee, N., Lekfuangfu, W. N., & Wooden, M. (2013). The marginal income effect of education on happiness: Estimating the direct and indirect effects of compulsory schooling on well-being in Australia
• Qazlbash SK, Zubair M, Manzoor SA, ul Haq A, Baloch MS. Socioeconomic determinants of climate change adaptations in the flood-prone rural community of Indus Basin Pakistan. Environmental Development. 2021; 37 :100603. doi: 10.1016/j.envdev.2020.100603. [ CrossRef ] [ Google Scholar ]
• Qureshi SA, Khan M, Ul Husnain MI, Iqbal N. Gender, environment, and sustainable economic growth [with comments] The Pakistan Development Review. 2007; 46 (4):883–894. doi: 10.30541/v46i4IIpp.883-894. [ CrossRef ] [ Google Scholar ]
• Rammohan A, Vu P. Gender inequality in education and kinship norms in India. Feminist Economics. 2018; 24 (1):142–167. doi: 10.1080/13545701.2017.1364399. [ CrossRef ] [ Google Scholar ]
• Robb AM, Watson J. Gender differences in firm performance: Evidence from new ventures in the United States. Journal of Business Venturing. 2012; 27 (5):544–558. doi: 10.1016/j.jbusvent.2011.10.002. [ CrossRef ] [ Google Scholar ]
• Sahoo S, Klasen S. Gender segregation in education: Evidence from higher secondary stream choice in India. Demography. 2021; 58 (3):987–1010. doi: 10.1215/00703370-9101042. [ PubMed ] [ CrossRef ] [ Google Scholar ]
• Saleem H, Shahzad M, Khan MB, Khilji BA. Innovation, total factor productivity and economic growth in Pakistan: A policy perspective. Journal of Economic Structures. 2019; 8 (1):1–18. doi: 10.1186/s40008-019-0134-6. [ CrossRef ] [ Google Scholar ]
• Sawada Y, Lokshin M. Obstacles to school progression in rural Pakistan: An analysis of gender and sibling rivalry using field survey data. Journal of Development Economics. 2009; 88 (2):335–347. doi: 10.1016/j.jdeveco.2008.03.002. [ CrossRef ] [ Google Scholar ]
• Sen G. Gender equality and women’s empowerment: Feminist mobilization for the SDG s. Global Policy. 2019; 10 :28–38. doi: 10.1111/1758-5899.12593. [ CrossRef ] [ Google Scholar ]
• Shaheen N, Ahmad N. Women entrepreneurship and empowerment in Pakistan: Gender, culture, education and policy in broader perspective. International Research Journal of Education & Social Sciences. 2022; 1 (1):25–36. [ Google Scholar ]
• Shea J. Does parents’ money matter? Journal of Public Economics. 2000; 77 (2):155–184. doi: 10.1016/S0047-2727(99)00087-0. [ CrossRef ] [ Google Scholar ]
• Sirine, M. N. I. F. (2015). Impact of inequalities on economic growth: Case of the developing countries. Journal of Sustainable Development Studies, 8 (1).
• Tansel, A., & Bodur, F. B. (2012). Wage inequality and returns to education in Turkey: A quantile regression analysis. Review of Development Economics, 16( 1), 107–121.
• Terza JV. Two-stage residual inclusion estimation in health services research and health economics. Health Services Research. 2018; 53 (3):1890–1899. doi: 10.1111/1475-6773.12714. [ PMC free article ] [ PubMed ] [ CrossRef ] [ Google Scholar ]
• Thomas V, Wang Y, Fan X. Measuring education inequality: Gini coefficients of education . World Bank Publications; 2001. [ Google Scholar ]
• ur Rahman S, Chaudhry IS, Farooq F. Gender inequality in education and household poverty in Pakistan: A case of Multan District. Review of Economics and Development Studies. 2018; 4 (1):115–126. doi: 10.26710/reads.v4i1.286. [ CrossRef ] [ Google Scholar ]
• Whalley J, Zhao X. The contribution of human capital to China’s economic growth. China Economic Policy Review. 2013; 2 (01):1350001. doi: 10.1142/S1793969013500015. [ CrossRef ] [ Google Scholar ]
• Williams, R. (2005). Gologit2: A program for generalized logistic regression/partial proportional odds models for ordinal dependent variables. Stata, Gologit2 Manual.
• Wu L, Yan K, Zhang Y. Higher education expansion and inequality in educational opportunities in China. Higher Education. 2020; 80 (3):549–570. doi: 10.1007/s10734-020-00498-2. [ CrossRef ] [ Google Scholar ]
• Yi, J., Heckman, J. J., Zhang, J., & Conti, G. (2015). Early health shocks, intra-household resource allocation and child outcomes. The Economic Journal, 125 (588). [ PMC free article ] [ PubMed ]
• Zaman K, Khan MM, Ahmad M, Ikram W. Do female enrolment rates cause economic growth in Pakistan? Asian Social Science. 2010; 6 (11):256. doi: 10.5539/ass.v6n11p256. [ CrossRef ] [ Google Scholar ]

Gender Discrimination in Higher Education in Pakistan: A Survey of University Faculty

Content maybe subject to  copyright     Report

187  citations

14  citations

View 5 citation excerpts

Cites background from "Gender Discrimination in Higher Edu..."

... The findings show that the struggle for equality is long however it is through these acts that the chains of domestic enslavement can be challenged (Shaukat et al., 2014). ...

... For instance, the younger the child or the more children the greater likelihood there is to suppress a woman’s career since she may have to, whether it is by will or necessity, take a longer career break (Shaukat et al., 2014). ...

... Despite their keenness to work, female academics often find themselves at the lower rung of middle class hierarchy and given that both partners are likely to work even when there is a maid employed to look after their children and do household duties (Shaukat et al., 2014). ...

... females quit working and opt out to care for families or stay at home in Pakistani culture (Shaukat et al., 2014). ...

... can be attributed to but not limited to the socio-cultural status of women in the society and the dynamics of the workplace (Shaukat et al., 2014). ...

9  citations

8  citations

View 1 citation excerpt

... This can be proved the existence of discrimination and unequal treatment of one gender because of cultural influences, social life, education, politics, and race [11, 12, 13, 14, 15, 16]. ...

7  citations

6,555  citations

View 1 reference excerpt

"Gender Discrimination in Higher Edu..." refers methods in this paper

... Job satisfaction, based on 4 statements drawn from the studies of Weiss, Dawis, and Lofquist (1967); Kendall (1963); Nagy (1996); Porter (1969) and Hackman and Oldham (1975). ...

2,660  citations

1,935  citations

... Professional development, based on 11 statements drawn from the studies of Hargreaves and Fullan, (1992); Arends, Winitzky, and Tannenbaum (1998); Darling-Hammond and McLaughlin (1995). ...

1,619  citations

927  citations

Related Papers (5)

Ask Copilot

Related papers

Related topics

• Harvard Division of Continuing Education
• DCE Theses and Dissertations
• Communities & Collections
• By Issue Date
• FAS Department
• Quick submit
• Waiver Generator
• DASH Stories
• Accessibility
• COVID-related Research

Terms of Use

• Privacy Policy
• By Collections
• By Departments

Gender-Based Violence in Pakistan - a Critical Analysis

Citable link to this page

Collections.

• DCE Theses and Dissertations [1188]

Contact administrator regarding this item (to report mistakes or request changes)

Advertisement

Supported by

Critic’s Pick

‘In Flames’ Review: A Patriarchy Horror Story

Set in Pakistan, the story of a young woman and her family, hemmed in by men, shifts from realism to genre, with heart-pumping consequences.

• Share full article

By Alissa Wilkinson

It takes about an hour for “In Flames” to reveal itself as proper genre horror, but trepidation lurks from the start. In Karachi, Pakistan, the 20-something Mariam (Ramesha Nawal) lives with her widowed mother, Fariha (Bakhtawar Mazhar), and her younger brother, Bilal (Jibran Khan), who’s mostly glued to his video games. The family has been financially dependent on Fariha’s father-in-law, but as the film opens, he has just died — and Fariha’s brother-in-law, Uncle Nasir (Adnan Shah Tipu), is suddenly very interested in the relatives he had been neglecting.

Fariha teaches at an elementary school, and Mariam is studying for exams that will qualify her to be a doctor. They’re smart, capable women who are less concerned with dismantling established social orders than they are with keeping their home and family intact. Yet their lives are hemmed in by the men around them, with a constriction that’s suffocating. For one, there is Uncle Nasir, who has offered to pay the family’s outstanding debts if Fariha signs some documents, which Mariam pleads with her to avoid doing. But there’s also the man who throws a brick through the car window when Mariam is driving to the library, calling her a whore. Or the man who lurks outside her window, masturbating. Or even the nice young man from the library, Asad (Omar Javaid), who won’t leave Mariam alone.

As the women scramble to save their home, the walls close in on them, and that’s the point: “In Flames,” a confident feature debut written and directed by Zarrar Kahn, is one of several recent films from around the world that frame patriarchy as a nightmare. The most recent may be “ Shayda ,” set in Iran, but even movies like “Poor Things” and “Promising Young Woman” play with the same idea, albeit with a lighter touch. This one is set in Pakistan, in the midst of debates about religious fundamentalism and gender roles, but the outlines are familiar even to audiences in very different circumstances. Men commit obvious, blatant offenses, confident the system is stacked in their favor. But even the “good guys” are locked in a culture that rewards them for refusing to listen to the women who, it’s made clear, are holding the country together.

That means the horror extends to the male perpetrators, who couldn’t find their way out of the maze of unjust systems if they tried. But there’s no question the women bear the brunt of it, whether the perpetrator is abusive, or greedy, or just clueless. To seek help is fruitless, and dangerous; being in debt to yet one more man is another way to put yourself at risk.

Kahn manages to assemble the story in a way that escapes feeling like a series of object lessons. He centers the story on Mariam, giving Nawal’s expressive eyes plenty of time to convey emotions she dares not speak aloud. Mariam’s environment signals her inner life. Sometimes the character is in claustrophobic interiors, where she can’t escape others’ prying eyes; sometimes she’s blessedly alone; and sometimes she’s experiencing brief moments of respite in expansive, beautiful scenery. She feels a pull between the freedom she craves and the responsibility she feels to her family. Technically, Mariam is a plucky heroine. But she isn’t rebellious, or even defiant. She’s just trying to survive.

Wisely, Kahn creates a world in which Mariam and Fariha cannot help but be pulled apart, ruptured by the patriarchy’s force. The only way for each to endure is to depend on yet another man to help them, which has profoundly middling results, and an element of always-present danger. When the film finally gives way to full horror, the pace picks up, and we see what the movie’s been doing all along. Oppression isn’t always blatant, and it isn’t the work of individuals acting alone. It comes like night terrors, paralyzing both oppressor and oppressed — and escape can require drastic action.

In Flames Not rated. In Urdu, with subtitles. Running time: 1 hour 38 minutes. In theaters.

Alissa Wilkinson is a Times movie critic. She’s been writing about movies since 2005. More about Alissa Wilkinson

Explore More in TV and Movies

Not sure what to watch next we can help..

Even before his new film “Civil War” was released, the writer-director Alex Garland faced controversy over his vision of a divided America with Texas and California as allies .

Theda Hammel’s directorial debut, “Stress Positions,” a comedy about millennials weathering the early days of the pandemic , will ask audiences to return to a time that many people would rather forget.

“Fallout,” TV’s latest big-ticket video game adaptation, takes a satirical, self-aware approach to the End Times .

“Sasquatch Sunset” follows the creatures as they go about their lives. We had so many questions. The film’s cast and crew had answers .

If you are overwhelmed by the endless options, don’t despair — we put together the best offerings   on Netflix , Max , Disney+ , Amazon Prime  and Hulu  to make choosing your next binge a little easier.

Sign up for our Watching newsletter  to get recommendations on the best films and TV shows to stream and watch, delivered to your inbox.

Breaking down the gender wage gap in Utah: How much of it is due to discrimination, and what can be done?

Does utah have the largest wage gap in the country that depends on how you analyze the data..

(Francisco Kjolseth | The Salt Lake Tribune) Derek Miller, President and CEO of the Salt Lake Chamber and Pat Jones, CEO of the Women's Leadership Institute, release a proposal on how to close the gender wage gap in Utah during a news conference at the Salt Lake Chamber on Friday, Nov. 30, 2018.

I think it’s important to accurately describe a problem in order to solve it.

A perfect example of that is the gender wage gap. Many of the headlines you read are about how large the wage gap is, with the most recent reports saying the median woman in Utah makes 73 cents for every dollar paid to the median man. Advocates have created a holiday called “Equal Pay Day,” representing the number of days a woman works “for free” in an average year in order to match a man’s salary; this year, it took place on March 12.

The numbers are designed to shock you and they probably should — they are true.

But conservative advocates on the other side have a valid point: Once you adjust for controllable factors like occupation, experience, and education, the difference between the sexes shrinks dramatically. And they’ve used that to argue that state laws requiring equal pay shouldn’t be passed.

Here’s my thesis: The strength of both arguments detracts from solving the real situation.

Utah has the second-largest pay gap in the nation, according to U.S. census data, but only a fraction of that is due to employer discrimination. That employer discrimination needs to be addressed. The remaining amount of the pay gap comes from various societal and cultural phenomena. Those also need to be addressed.

Let’s dig in.

The current situation

Payscale is a company that tracks pay structure and makes recommendations to businesses on how they can make their wages competitive in the labor marketplace. Overall, 627,000 people have taken Payscale’s survey over the last two years about how much they make, their occupation, and other relevant factors. That’s an impressive sample size, though one the company admits is likely skewed toward those with a college degree.

Here’s what the survey found as the median wage gap among all 50 states.

(Christopher Cherrington | The Salt Lake Tribune)

As you can see, Utah women typically make just 75 cents on the dollar compared to Utah men. That’s also quite close to the 73.1 cent average reported by the Census Bureau, which is promising from a data point of view.

But what happens if you control for every compensable factor? Match job to job, experience to experience, education to education, and so on? This is the kind of data that companies like Payscale can have that the Census Bureau doesn’t.

The gap significantly shrinks — now Utah women make 97 cents to the dollar compared to Utah men.

The whole Payscale report is worth reading. It further breaks down the controlled and uncontrolled data by job, industry, race, education, age, work-from-home status, and much more. Some jobs, like truck drivers and religious directors, still have massive wage gaps even if you account for every factor.

Removing the discrimination portion of the wage gap

There are other estimates of the discrimination portion of the wage gap in Utah that are larger than 3%. For example, one University of Utah thesis put it at 14%. I think, given the Payscale tilt toward college-educated workers, that 3% is likely an underestimation. But let’s just go with 3%, for the sake of argument.

A 3% gap in pay simply due to discrimination is still hugely significant and needs to be addressed. That Utah women are being bilked out of thousands of dollars in salary per year needs to change.

Forty-three states in the nation have an equal pay law. Utah isn’t one of them. We’re the only state in the intermountain region not to have such a law, in fact. (Utah State University has an excellent report on the laws in all 50 states.)

What Utah does have is a wage anti-discrimination law. So what’s the difference? As always, it’s in the details.

Let’s choose Idaho as the counter-example — a state just as or more conservative than Utah. Utah’s law, meager as it is, exempts specific religious entities and any employer with fewer than 15 employees, or any employer asking people to work less than 20 weeks in a year. That’s a lot of employers who are allowed to discriminate! Idaho’s law has no such exemptions.

In Idaho, those discriminated against are allowed to and encouraged to take up a lawsuit in a relevant court. Utah’s law, meanwhile, forces those with a claim to go through the Utah Antidiscrimination and Labor Division of the Utah Labor Commission.

If an employee wants to sue their employer in Utah, they would have to get permission from the U.S. Equal Employment Opportunity Commission before filing a case in federal court. But the UALD is the only process allowed under Utah law.

That’s a big deal because the UALD has proven to have trouble enforcing the law. A 2017 audit revealed that the division rules in favor of the employee just .7% of the time. In nearby states, the average is about 5.3%. That’s an especially big deal for those who lose their claims: Idaho’s law also prevents retaliation by employers on employees who seek action under their equal pay law, which Utah’s law doesn’t.

Utah’s law also has a statute of limitations, asking workers to file claims within 180 days of the discriminatory act. Idaho’s law is three years.

Maybe I can appeal to Utah’s Legislators via their competitive instincts: There is no reason for Idaho to have a better law than Utah on this matter. You’re gonna allow Idaho to keep things fair while Utah languishes? Those guys up north? Those wackadoodles play on blue football fields and are proud of potatoes! We can’t lose to them!

Addressing other factors

But the truth remains that the largest portion of Utah’s wage gap isn’t due to employer discrimination alone.

I thought this thesis from Curtis Miller at the U.’s economics school was well done. Essentially, using census data and fancy analytics, it tries to extract which causes can explain Utah’s wage gap. The different industries in which men and women work? The choice of occupation within an industry? Is it different levels of experience or other qualifications, referred to as an “endowment gap?” Overtime hours worked?

According to Miller, the endowment gap and industry are the largest factors. Interestingly, Miller’s analysis of the national data doesn’t show an endowment gap between men and women — if anything, women are more likely in their fields of interest to have experience or other qualifications nationally, but in Utah, the reverse is true. (Industry of employment differences are still a large driver of the wage gap nationwide and in Utah.)

As you probably know, Utah women are more likely to be mothers than those in other states. But in wage analysis, there’s what’s called the “motherhood penalty.” Mothers are more likely to take time off of work than fathers after the birth of a child, which in itself can lead to fewer opportunities for career advancement.

But even among those who take the same amount of time off work, the perceptions are different. The U.S. Commission on Civil Rights’s report on the wage gap in Utah noted that mothers were also nearly twice as likely as fathers to say taking time off had a negative impact on their careers. “Among those who took leave from work in the past two years following the birth or adoption of their child, 25 percent of women said this had a negative impact at work, compared with 13 percent of men,” the report says.

There’s an education gap in Utah that doesn’t exist in other states. Nationally, more women get graduate degrees than men, by a 13% to 12.4% score. In Utah, 9.3% of Utah women and 14.1% of Utah men earn graduate degrees. Those with higher degrees generally make more money. This, too, explains a statistically significant part of the gap.

Finally, the choice of an occupation within a profession represents a small portion of the gap in Utah. The U.S. Commission on Civil Rights’s report notes that women make up two-thirds of those who make minimum wage or just over it, for example.

But age, citizenship status, overtime hours worked, and public vs. private sector status didn’t measurably contribute to Utah’s wage gap, according to Miller’s work.

In the end, we can make inroads in these issues. Utah should get with the times and pass a comparable equal pay law to other states. Then, it should make strides in supporting women in achieving higher levels of education and higher-paying industries.

Breaking down the wage gap into its component parts doesn’t serve to minimize it — in fact, it serves to identify places where we can make real changes, especially at the legislative level. Let’s push these improvements forward.

Editor’s note • This story is available to Salt Lake Tribune subscribers only. Thank you for supporting local journalism.

Donate to the newsroom now. The Salt Lake Tribune, Inc. is a 501(c)(3) public charity and contributions are tax deductible

The Triple Team: Kenneth Lofton Jr. lifts the Utah Jazz to victory again

The nhl is likely headed to utah. here’s how jazz owner ryan smith became first in line for arizona coyotes, rick pitino backs mark pope’s hire, says he will ‘go on to greatness’ at kentucky, panguitch lake dam evacuation order relaxed as water level drops, bagley cartoon: footloose, as byu searches for basketball coach, one prime candidate already appears to be off the board, featured local savings.