ANALYZING FEMALE LABOR FORCE PARTICIPATION IN …

Ishani Desai

Li Li

Advisor: Professor Michael Callen Section Leader: Professor Michael Walton

MPA/ID Second Year Policy Analysis March 2016

Written in fulfillment of the requirements for the degree of Master of Public Administration in International Development, John F. Kennedy School of Government, Harvard University.

ANALYZING FEMALE LABOR FORCE PARTICIPATION IN AFGHANISTAN Identifying the key barriers that prevent women from entering the labor force

“ Here’s to strong women. May we know them.

May we be them. May we raise them. ”

- Unknown

Acknowledgements We would like to thank Professor Lant Pritchett for providing the seed for this project, and to our advisor, Professor Michael Callen, and our section leader, Professor Michael Walton, for helping us develop and nurture the idea. We are extremely grateful to Professor Callen for encouraging us to work on this topic, providing us with the relevant datasets and contacts, and supporting our research. We are also grateful for support from Harvard Kennedy School faculty, especially Professor Jeni Klugman, Professor Dan Levy, Professor Alberto Abadie, and Professor Matt Andrews for their teaching, guidance, and dedication to their work. We would like to thank Harvard Kennedy School’s Women and Public Policy Program and its Cultural Bridge Fellowship Program for their research support. In addition, we would like to thank Scott Guggenheim, Tarek Ghani, Lael Mohib, and Zachary Warren. Their insights, guidance, and inputs were instrumental in producing this paper. We would also like to thank our MPA/ID classmates who provided us with thoughtful feedback, technical assistance, and challenged us to think deeper – in particular our classmates Maria Schwarz and Manuel Schoenfeld for their invaluable support. And finally, we would like to acknowledge our families for their lifelong support and commitment, and for giving us the freedom to pursue our passion and taking an interest along the way.

Table of Contents

EXECUTIVE SUMMARY..................................................................................................................................1I. INTRODUCTION..............................................................................................................................................2

A.MOTIVATION..........................................................................................................................................................................2B.POLITICALCONTEXT.............................................................................................................................................................2C.INTERNATIONALCOMMUNITYINVOLVEMENT.................................................................................................................4D.POLICYQUESTIONS...............................................................................................................................................................4

II. CONTEXTUALIZING THE PROBLEM..................................................................................................5A.ISGOVERNMENTPOLICYAPARTOFTHESOLUTION?......................................................................................................5B.THEAFGHANCONTEXT........................................................................................................................................................9C.POLICYINTERVENTIONSINTHEREGION..........................................................................................................................9

III. METHODOLOGY......................................................................................................................................11IV. FINDINGS & ANALYSIS.........................................................................................................................12

A.WHOARETHEWOMENINTHELABORFORCE?.............................................................................................................13B.PREDICTIVEMODEL:DETERMININGTHEBARRIERS...................................................................................................14C.DIAGNOSTICS:DETERMININGTHEUNDERLYINGBARRIERS.......................................................................................16i.Security:IncidenceofViolence..................................................................................................................................16ii.Education/Skills.............................................................................................................................................................18iii.Caregiving.......................................................................................................................................................................20iv.Customs/Norms.............................................................................................................................................................21v.Religion...............................................................................................................................................................................23vi.LegalFramework.........................................................................................................................................................24

D.STATECAPABILITIES.........................................................................................................................................................26V. POLICY OPTIONS......................................................................................................................................27

A.OPTION1–LOCAL-LEVELFEMALEQUOTAS.................................................................................................................29B.OPTION2–INCREASEACCESSTOINFORMATIONTHROUGHTELEVISION...............................................................31C.OPTION3–BUSINESSTRAININGPROGRAM..................................................................................................................33

VI. RECOMMENDATIONS............................................................................................................................34A.SECURITYCONSTRAINTS...................................................................................................................................................34B.POLICYINTERVENTION:INCREASEACCESSTOINFORMATIONTHROUGHTELEVISION.........................................35C.DATAQUALITY....................................................................................................................................................................36D.IMPLEMENTATIONPLAN..................................................................................................................................................36

VII. CONCLUDING REMARKS...................................................................................................................39REFERENCES....................................................................................................................................................40APPENDIX...........................................................................................................................................................43SECTIONA.FIGURESANDTABLES......................................................................................................................................43SECTIONB.DATASOURCES..................................................................................................................................................54SECTIONC.DIAGNOSTICSFRAMEWORK............................................................................................................................55SECTIOND.PREDICTIVEMODEL.........................................................................................................................................56SECTIONE.QUESTIONNAIRE................................................................................................................................................62

Desai and Li Analyzing FLFP in Afghanistan

1

Executive Summary

Female labor force participation (FLFP) in Afghanistan is one of the lowest in the world. This implies that there are underlying factors that prevent women from working such as limited mobility, gender inequality through legal frameworks, lack of economic opportunities, and low household bargaining power. Improving female engagement in economic activity will not only benefit economic development, but will also promote sustainable long-term growth.

Evidence suggests that there is an internal shift relating to Afghan women’s role in society. Based on the 2015 Survey of the Afghan People, around two-thirds of Afghans say that women should be able to work outside the home. In addition, the Government of Afghanistan (GoA) has made a series of commitments to address women’s issues. The Ministry of Labor and the Ministry of Women’s Affairs have been drafting the National Action Plan for Women’s Economic Empowerment Plan “to create an enabling environment for women to participate in economic activities to benefit their families, communities, and the country”.

After determining the underlying barriers that prevent women from entering the labor force, we recommend effective policy actions for the GoA. Some key FLFP insights from Afghanistan are: • Security is an underlying barrier limiting women’s participation in the economy. High

violence not only diminishes women’s willingness to leave their home, but it also raises concern in men, which further restricts women’s mobility.

• Norms is another key underlying barrier that limits FLFP. Prevailing norms have resulted in a weak legal framework, low autonomy and mobility for women, and men’s biased attitude toward the role of women in the family, community, and society.

• A supporting environment for women’s education is instrumental, especially support from men towards equal education opportunities. Women also self-report that lack of education, illiteracy and professional trainings are the largest challenge facing them.

• Information matters, and the channel in which women receive information matters. Obtaining information from the outside world provides exposure and can update women’s prior beliefs.

• Given the state capabilities, programs that seek to be effective have to take place from both top-down and bottom-up. Afghanistan is a weak state, but has a long history of strong informal ruling through community jirgas and shuras, which have power to exercise local control.

Given the increase in television viewership over time, we recommend the GoA to use television programs to provide exposure to the outside world and address a key underlying barrier, norms. We recommend this after analyzing the technical aspects, political supportability, and administrative capabilities. Using the Problem Driven Iterative Adaptation (PDIA) approach, we stress that implementation must involve pilots, testing, and validation of the proposed theory of change. We hope to help the GoA start a long-term dialogue to empower not only women, but all of Afghanistan.


2

I. Introduction

a. Motivation

The current female labor force participation (FLFP) rate in Afghanistan is around 16% – one

of the lowest in the world.1 Afghanistan ranks 171st among 188 countries in the UN’s Gender

Inequality Index and, in 2011, it was the most dangerous country in the world for women.2 In

2015, UN Women estimated that only 5% of Afghan businesses are female owned.

These dismal statistics have serious implications for the country – for socioeconomic

inclusivity, poverty reduction, and for overall growth and productivity. Improving female

engagement in economic activity will not only benefit economic development, but will also

promote sustainable long-term growth.3 Low FLFP is a problem in itself as it affects potential

growth, but it also implies that there are other underlying factors that prevent women from

working such as limited mobility, gender inequality through legal frameworks, lack of economic

opportunities, and low household bargaining power.

Evidence suggests that Afghan women are willing to work. In the 2015 Survey of Afghan

People, 74.5% of Afghan women state that women should be allowed to work outside the home.

Since 2012, Afghan women have been stating “lack of job opportunities” as one of the top three

largest problems facing women. In the broader South Asian context, traditional norms have

played a large role in dictating gender roles; nevertheless, indicators for women in neighboring

countries have improved. There is a role for well-designed policy to improve the status of

women in Afghanistan.4

b. Political Context Since the fall of the Taliban, indicators for women have dramatically improved. The

Government of Afghanistan (GoA) has made a series of critical commitments to women, such as

ensuring equality in its Constitution, signing international conventions on gender issues,

integrating gender concerns throughout its core national development strategy, and promoting

1 World Development Indicators. World Bank. 2015. 2 Ghaisy, Richard, Jiayi Zhou, and Henrik Hallgren. Afghanistan's Private Sector: Status and Ways Forward. Rep. Stockholm: SIPRI, 2015. 38-42. 3 Duflo, Esther. “Women Empowerment and Economic Development.” Journal of Economic Literature 50, no. 4 (December 1, 2012). 4 Ghaisy, Richard, Jiayi Zhou, and Henrik Hallgren. 2015.


3

services for women within many of its key service delivery programs.5

Since 2001, while primarily focusing on security, the international community has helped

drastically improve female education and health indicators. In addition, perceptions about

women’s role in the economy are starting to change, as are the trends on the ground. Based on

the 2015 Survey of the Afghan People, around two-thirds of Afghans say that women should be

able to work outside the home. The percentage of Afghans who say that female family members

contribute to household income has increased from 13.9% in 2009 to 22.6% in 2015.6 In

addition, 23.4% of women reported increased access to schools compared to two years ago. This

evidence suggests that there is an internal shift relating to women’s role in Afghan society.

Improving the status of women and ensuring they have a greater role in the economy is on

the agenda of President Ashraf Ghani and the GoA. Democratically elected and inaugurated in

September 2014, Ghani publically thanked his wife for her support towards the women of

Afghanistan.7 He has also tweeted: “Experience has shown that if we invest on a girl, it will

change the next 5 generations. Hence, we must understand the importance of women”.8 His

focus on women’s issues predicts well for the future of Afghanistan; however, Ghani has

political opposition (see Figure 1A) and it is unclear if the majority of Afghans will embrace his

strategy. Given this, it is critical to ensure that policymakers focus on designing a policy that is

not contentious and is internally supportable to drive long-term reform.

In addition, since 2015, the Ministry of Labor, Social Affairs, Martyrs and Disabled and the

Ministry of Women’s Affairs (MoWA) have been drafting the National Action Plan for

Women’s Economic Empowerment Plan (NEEP) “to create an enabling environment for women

to participate in economic activities to benefit their families, communities, and the country”.9

Driven by the GoA, this initiative aims to improve participation of women in markets and the

private sector.

5 World Bank. Afghanistan: Women's Role in Afghanistan's Future - Taking Stock of Achievements and Continued Challenges. Rep. no. AC34474. Washington DC: World Bank, 2013. 6 This could be due to increase household production or simply more visibility around women’s work. 7 Loh, Katherine. “Does the Ghani Administration Mean a Greater Role for Women in the Afghan Economy?” In Asia. The Asia Foundation, 19 Nov. 2014. <http://asiafoundation.org/in-asia/2014/11/19/does-the-ghani-administration-mean-a-greater-role-for-women-in-the-afghan-economy/>. 8 Ghani, Ashraf. “Twitter Feed.” Twitter. 8 Nov. 2014. 9 Afghanistan. Ministry of Labor, Social Affairs, Martyrs and Disabled. Draft of National Action Plan for Women's Economic Empowerment in Afghanistan 1395-1399. Ver. 2. Kabul: 2015.


4

c. International Community Involvement

International stakeholders have made women’s issues central to the objectives of aid

commitments. In some areas, these commitments have been translated into demonstrable

progress, while in others, advancement is less visible and requires reexamination. Since 2001,

the US has played an active role in women’s issues in Afghanistan. From the USAID website:

Thirteen years ago, virtually no girls attended school. Women died every day from preventable pregnancy complications and were restricted from contributing to the economy. Today, millions of girls are enrolled in schools…. Maternal mortality rates have declined more than threefold…. Women have more opportunities to receive job training and apply for loans….10

Table 1A summarizes a subset of programs introduced by USAID, the World Bank, and other

international organizations to empower women. Of relevance is USAID’s five-year $216 million

Promote Program – the largest women’s empowerment program of its kind – aimed at advancing

opportunities for Afghan women through training and increasing the number of women in

decision-making positions within the GoA.

Regarding progress on women and employment, the 2013 World Bank Report mentions that

the current picture today is less clear – at least when compared to education and health

indicators.11 The majority of programs have focused on education and skills training to younger

women, along with mentorship to women who have access to higher education. There has been

less focus on older women and women in rural areas, perhaps due to higher costs associated with

reach. These programs are also costly for the GoA to implement and monitor independently.

Despite the past efforts of the international community, indicators around women’s participation

in the labor force have not changed significantly over the years; however, this could also be

attributed to lack of data.

d. Policy Questions

Despite large data collection efforts, issues surrounding women’s role in the economy have

not been carefully examined. Recent waves of both the National Risk and Vulnerability

Assessment (NRVA) and the Survey of Afghan People have included more comprehensive

women’s modules; however, for several reasons, these modules do not appear to have influenced

10 “Gender.” Afghanistan. USAID, 26 Jan. 2016. <https://www.usaid.gov/afghanistan/gender-participant-training>. 11 World Bank. Afghanistan: Women's Role in Afghanistan's Future - Taking Stock of Achievements and Continued Challenges. Rep. no. AC34474. Washington DC: World Bank, 2013.


5

program design.12 Before designing and implementing programs, it is important for the GoA to

understand if their intended reforms address the underlying barriers women face to enter the

labor force.

The World Bank reports that policymakers should “support agricultural value chains where

women's contribution is strongest and most valued, strengthen quotas and incentives for women

to participate in the public sector, and target key barriers to women's employment”; however, it

is unclear what these barriers may be.13 In order to bring sustainable long-term improvement to

FLFP, it is critical to determine the key barriers women face and use policy to effectively target

these barriers. To date, there has not been a comprehensive analysis to study the issues Afghan

women face to join the labor force.

We aim to determine the underlying barrier(s) that prevent women in Afghanistan from entering the labor force and to find effective actions the GoA can take to affect reform and

monitor progress in the Afghan context. In order to delve into these issues, we will assess

relevant literature (Section II), design a predictive model to understand the differentiating

characteristics of women in the labor force (Section IV), present a diagnostics to decompose the

various factors affecting FLFP (Section IV), and assess Afghanistan’s state capabilities (Section

IV). Our goal is to design the most effective policy to remove the underlying barrier(s) and thus,

encourage women to join the labor force.

II. Contextualizing the Problem

Despite dramatic changes across the world in women’s access to employment, education, and

political participation, progress has been uneven.14 In order to examine women’s role in the

economy, we need to understand why this gap exists. In this section, we review the empirical

evidence behind FLFP and analyze the context of developing countries to understand the factors

that affect women’s participation in the economy.

a. Is government policy a part of the solution?

FLFP in developing countries is quite complex to study since there are many interrelated

factors to consider. Researchers have highlighted a few dimensions that affect women’s ability

12 Section B in the Appendix provides additional information about these data sources. 13 Ibid. 2013. 14 World Bank. World Development Report 2012: Gender Equality and Development. Washington, DC: World Bank, 2012.


6

to engage in the labor market: level of economic development, educational attainment, social

dimensions (e.g., norms influencing marriage, fertility, and women’s role outside the household),

institutional setting (e.g., laws, protection, and benefits), access to credit and other inputs, and

household/spousal characteristics.15 We will begin with one that is studied most frequently:

economic development.

As countries develop, the U-shaped female labor force hypothesis states that labor force

participation of married women initially declines and then rises.16 When incomes are extremely

low and agriculture dominates, women are highly active in the labor force – sometimes as paid

laborers, but more often as unpaid workers on family farms. As incomes rise, because of market

expansion and the introduction of new technology, women’s labor force participation falls – this

is the income effect. This is due to the reduction in the relative price of home produced goods

and a decrease in the demand for women’s labor in agriculture. Even when women’s relative

wages rise, married women may be barred from formal employment by social customs and/or

employer preferences.17 As education improves, women move back into the paid labor force,

which suggests that the value of women’s time in the labor market has increased. At this point,

the substitution effect (substituting home production with labor market production) dominates

the income effect. This theory argues that economic growth alone can affect FLFP.

Figure 1 reveals where Afghanistan stands with respect to the U-shaped hypothesis.

Given the level of development, Afghanistan has an extremely low FLFP rate as compared to the

predicted rate. In most of the countries analyzed for the U-shaped hypothesis, political freedoms

occurred decades before economic change was apparent.18 In conflict-ridden Afghanistan,

political freedoms are a recent phenomenon so it is difficult to apply the U-shaped hypothesis.

The countries examined for the hypothesis were never as poor on a per capita basis as the poorest

countries are today. They never had extreme religious views so the paths of women may differ;

however, the role of education can still have similar effects in spurring change for women.19

15 Gaddis, Isis, and Stephan Klasen. “Economic Development, Structural Change, and Women’s Labor Force Participation: A reexamination of the feminization U hypothesis.” Journal of Population Economics J Popul Econ 27.3 (2013). 16 Goldin, Claudia. “The U-Shaped Female Labor Force Function in Economic Development and Economic History.” NBER Working Paper Series (1994). 17 Ibid. 1994. 18 Ibid. 1994. 19 Ibid. 1994.


7

Figure 1: Afghanistan’s position in the U-shaped hypothesis

Source: WDI (2015)

One of the strongest determinants of labor market outcomes in both developed and

developing countries is educational attainment, suggesting that education levels will be a critical

predictor to FLFP.20 There is a visible U-shaped relationship between educational attainment and

FLFP; that is, with very low levels and very high levels of education, there appears to be high

FLFP. Using data from the National Sample Survey (NSS) in India, there is a threshold

suggesting a payoff in the labor market only for women with more than a secondary education.21

On the other hand, it is argued that women’s empowerment and economic development are

closely intertwined. While development alone can play a role in reducing gender inequality,

empowering women can actually spur economic development.22 Duflo discusses childcare,

marriage, cultural norms, nutrition, female aspirations, education, opportunities, legal rights,

property rights, and other micro topics to highlight the complexities in empowering women. In

the past, policymakers have focused on growth and poverty reduction without adopting strategies

at improving the condition of women. Economic development does reduce poverty and

improves the condition of everyone, including women; however, it is not enough to bring

equality - policy action is needed.23

In terms of policy instruments, top-down options like quotas, institutional reform, and

conditional transfer programs have been common. Duflo suggests that quotas alone may not be

20 Cazes, S., and S. Verick. The Labour Markets of Emerging Economies: Has Growth Translated into More and Better Jobs? Geneva and Basingstoke, UK: ILO and Palgrave Macmillan, 2013. � 21 Verick, Sher. “Female Labor Force Participation in Developing Countries.” IZA World of Labor Izawol (2014). 22 Duflo, Esther. 2011. 23 Ibid. 2011.


8

sufficient to ensure that women policy makers act differently than men, especially if women are

politically weak and less influential in society. In India, however, Beaman et al. find differences

in outcomes: The bias against women and women’s leadership is diminished and, in men’s

response to the speeches, the biases against women were erased.24 Beaman et al. also found that

that parents and girls changed their aspirations. In areas with a female leader, even without

additional investment in education infrastructure, there was an increase in educational attainment

for girls, reducing the gap with boys.25 This suggests that even if direct outcomes are not

influenced, there are effects on future generations of girls simply through female political quotas.

Other policy avenues include the legal environment – divorce laws, property laws, workplace

laws, etc. In many developing countries, women are very poorly protected in the case of divorce,

and stand to lose assets/property and the custody of their children. Even when divorce laws

exist, it is frowned upon in society, and remarrying is difficult – cultural norms can oppress

women. There has been little research on the effect of these institutions in developing countries.

Furthermore, by expanding women’s work opportunities through policy, women have more

hope – if women work outside home, there may be a perception that they do need to be as strong

and healthy as men and that they do need a formal education.26 Evidence suggests that

conditional transfer programs targeted to women seem to make a difference, even when they are

both temporary and small.27 However, it seems difficult to assess if these programs affect

household bargaining power or other dimensions of women empowerment.28

Even if these top-down policy measures do not bring about radical changes in the way

women are perceived in society, evidence presented suggests that properly designed policies

targeted towards women can have immediate consequences. Duflo takes an extreme position

when she states that “equity between men and women is only likely to be achieved by continuing

policy actions that favor women at the expense of men, possibly for a long time”.29

24 Beaman, Lori, Esther Duflo, Rohini Pande, and Petia Topalova. “Female Leadership Raises Aspirations and Educational Attainment for Girls: A Policy Experiment in India.” Science 335, no. 6068 (February 3, 2012). 25 Verick, Sher. 2014. 26 Duflo, Esther. 2011. 27 Lundberg, Shelly J., Robert A. Pollak, and Terence J. Wales. “Do Husbands and Wives Pool Their Resources? Evidence from the United Kingdom Child Benefit.” The Journal of Human Resources 32.3 (1997). 28 Duflo, Esther. 2011. 29 Ibid. 2011.


9

b. The Afghan Context

While there is a clear need for policy to enforce equality, promoting equality has been

challenging in countries with long-standing traditions that sustain discrimination in all facets of

life.30 Even with political will, these policies are often insufficient to change deeply embedded

gender stereotypes – such is the reality for Afghanistan.

For over a decade, the international community has spent millions of dollars with top-down

programs yet indicators remain among the worst in the world for Afghan women. Some argue

that these programs have provoked Afghan men into taking more conservative positions.31 Bahri

suggests that efforts to promote gender equality could be bolstered by including men in programs

and addressing the issue through an Islamic perspective as opposed to an ethnocentric approach.

In addition, he suggests that programs to encourage women’s participation in society should

focus on the benefits to the family instead of economic reasons. 32

There is some hope: Evidence suggests that certain policies can work to some degree even in

environments where women are subjected to very high levels of discrimination. A randomized

experiment finds that gender quotas in Afghan village councils improves outcomes specific to

female participation in some areas, such as increased mobility and income generation; however,

quotas produce no change (in the short term) to deeply entrenched attitudes towards the role of

women in society.33 Along with the deeply entrenched cultural practices, the heightened

insecurity imposes further constraints on female mobility in the Afghan context.34

A large limitation in women’s labor research all over the world – and especially in

Afghanistan – is data. FLFP is poorly measured and underestimated in developing countries.35

Though data collection has improved with time, this remains a major obstacle to the analysis of

household surveys, especially with surveys where men are the primary survey respondents.

c. Policy Interventions in the Region

Given the inextricable similarities and cultural linkages, examining policies to address

women’s participation in the economy in the region will provide a guideline for what could work 30 Beath, Andrew, Fotini Christia, and Ruben Enikolopov. “Empowering Women through Development Aid: Evidence from a Field Experiment in Afghanistan.” American Political Science Review 107, no. 03 (August 2013). 31 Bahri, J. “Western Gender Policies in Afghanistan: Failing Women and Provoking Men.” Gender, Technology and Development 18.2 (2014). 32 Ibid. 2014. 33 Beath, Andrew, Fotini Christia, and Ruben Enikolopov. 2013. 34 Ibid. 2013. 35 Verick, Sher. 2014.


10

in Afghanistan. Table 2A reveals a subset of policy interventions designed by academics to

address women’s issues and their outcomes. We rely primarily on programs from India,

Pakistan, Bangladesh, and Saudi Arabia.

A business training program to poor self-employed women in Ahmedabad finds that among

Hindu women, training increased borrowing and business income for those facing more

restrictions on mobility and social interactions; however, Muslim women failed to benefit from

the training program altogether.36 There could be several reasons for this – one theory is that the

training helped women whose businesses had been held down by social restrictions, but women

subject to extreme restrictions had too little agency to easily change their aspirations or

activities.37 While these results may not be directly applicable to the Afghan context, they are

important to consider when designing business training programs because they might not be right

vehicle for all women.

Furthermore, introducing cable television was associated with improvements in women’s

status in India.38 There were increases in reported autonomy, decreases in reported son

preference, large increases in female school enrollment, and decreases in fertility (primarily via

increased birth spacing). These results were observed without targeted appeals, such as public

service announcements, which indicate that television is an effective form of persuasion as

people emulate what they perceive to be desirable behaviors.39 Given that cable television is a

powerful medium to provide exposure to the outside world and is rapidly growing in the

developing world, these results have important policy implications for Afghanistan. In

Afghanistan, television popularity has been increasing: In 2013, 55% of Afghans obtained

information from television and in 2015, the figure increased to 62%.40 Overall, women obtain

information from television more than men.

An ongoing project in Pakistan provides women’s-only transport and vouchers in Lahore to

examine if such an intervention can give women expanded choices.41 Social norms prevent

36 Field, Erica, Seema Jayachandran, and Rohini Pande. “Do Traditional Institutions Constrain Female Entrepreneurship? A Field Experiment on Business Training in India.” American Economic Review 100.2 (2010). 37 Ibid. 2010. 38 Jensen, Robert, and Emily Oster. “The Power of TV: Cable Television and Women’s Status in India.” The Quarterly Journal of Economics 124, no. 3 (August 1, 2009). 39 Ibid. 2009. 40 Figures are from the Survey of Afghan People dataset. 41 Field, Erica, and Kate Vyborny. “Women's Mobility in Pakistan.” CERP. Center for Economic Research in Pakistan, July 2014. <http://cerp.org.pk/womens-mobility/>.


11

women from coming into contact with unrelated men and restrict women’s use of public

transport. By providing women-only transport, it is thought that women can choose to

participate in the labor force, continue education, and engage in independent activities.

If this program is effective at improving outcomes for women, there may be a strong case for

providing such a subsidy in Afghanistan. The caveat is that 70% of women in Afghanistan live

in rural areas and infrastructure may be lacking.

The results of the interventions in Table 2A will be useful in guiding policy options in

Section V as these countries have similar social barriers for women, constraints, and

implementation capabilities as Afghanistan.

III. Methodology

Participation of women in the labor force is an outcome of various micro and macro factors.

Figure 2A maps the underlying reasons women may choose not to enter the labor force. To

design effective policy, we need to understand how these factors interact with FLFP. In the

Afghan context, given the administrative and political limitations, it is not useful to provide the

Government with a long list of reforms – many of which may not be targeted at the root

underlying barriers.42

To understand the barriers women face, we develop a robust predictive model to arrive at a

group of key correlates with FLFP. This model allows us to understand the characteristics of the

women in the labor force and identify the major determinants of FLFP in Afghanistan. Using the

2015 Survey of Afghan People and data on incidence of violence in Afghanistan, we use a binary

dependent variable technique to determine the factors influencing women’s decision to enter the

labor force.43 First, we use Pearson correlation, which provides a measure of the strength of the

linear relationship between two variables, to find how well each individual indicator relates to

our variable of interest – whether the woman is in the labor force or not. Then, the most

correlated indicators are put into a cross-validated lasso-logistic regression predictive model that

allows us to understand the relative importance of each indicator. Based on the results of the

model, we select relevant indicators that are politically and administratively feasible to influence

42 Hausmann, Ricardo, Dani Rodrik, and Andrés Velasco. “Growth Diagnostics” The Washington Consensus Reconsidered (2008): 324-55. 43 See Section D in the Appendix for details on the Predictive Model.


12

in a real world setting and use a logistic model to find the grouping of variables with the best R-

squared.

This model is predictive in nature. Even though the model does not directly provide causal

linkages, our process of model validation provides strong evidence of robust relationships

between our variables of interest and FLFP.44 Due to the context and the confounding factors,

focusing solely on causation in Afghanistan is challenging. Given the difficulties around

identification, a predictive model can quickly lead us to the factors that differentiate women who

are in the labor force from women who are not, allowing us to design policy to try to affect these

key correlates.

Building on the predictive model, we use the diagnostics framework as a strategic tool to

prioritize reforms. We employ this framework to determine the key underlying barrier(s) that

prevent women from entering the economy.45 By doing so, we develop a better understanding of

how the underlying barriers on FLFP can differ from setting to setting, and use these findings to

derive policy priorities in way that efficiently uses scarce political resources.

Finally, we examine Afghanistan’s state capabilities to help bridge the gap between technical

correctness and administrative feasibility. The first part of the analysis focuses solely on the

technical aspect, but the final part delves into administrative challenges facing the country and

the various stakeholders who must be engaged in order to bring long-term sustainable change to

FLFP.

IV. Findings & Analysis

Using our framework, we identify the critical underlying barriers women face to design

policy to address these barriers and empower women to join the labor force. Given data

limitations, we focus solely on supply-side factors, or factors that affect the supply of labor. We

acknowledge that a key challenge in Afghanistan is also labor demand; however, as the literature

suggests, the supply and demand for labor are highly intertwined – for the purposes of this

analysis, we focus on empowering women to enter the labor force. Figure 3A maps the NEEP

interventions to many of the barriers Afghan women face. The mapping reveals that the GoA

plans to affect reform through many different channels; however, we argue, due to limited

44 See Section D in the Appendix for information on the Predictive Model. 45 See Section C in the Appendix for information on the Diagnostics Framework.


13

resources, the GoA should prioritize programs that address key underlying barriers. In this

analysis, we focus on security, education, caregiving, religion, norms, and the legal framework.

Before investigating the barriers, we examine the women in the labor force to understand their

characteristics.

a. Who are the women in the labor force?

On average, women in the labor force are 32.3 years old. 80% of these women are married

and they tend to marry young, at around 17.9 years.46 In fact, 75% of the women married before

the age of 19. The average household size for these women is eight. About 70% of married

labor force women have children less than five years old, and it is very common to have more

than one child in the household. 6.2% of the married labor force women have husbands with

multiple wives.

In terms of demographics, the breakdown of the ethnic groups is the following: Tajik

(43.7%), Pashtun (24.5%), Hazara (10.8%), and Uzbek (8.1%) – these are the four largest ethnic

groups in Afghanistan. 90% of these women speak Dari, and the next most common languages

are Pashto (35.0%) and Uzbeki (13.4%). A small portion can speak Turkmeni, English, and

Urdu. More importantly, 73% of them are from rural areas, 14% from Metro (Kabul), and the

rest from other towns and cities.

With respect to education, 45.3% of women in the labor force never went to school while

16% have secondary school education and 18.3% have tertiary education. The most common

occupation is the schoolteacher (15.4%), followed by skilled workers/artisans (6.5%) and

laborer/domestic/unskilled workers (3%). Other occupations are farmers, informal sales,

government and private office workers, and self-employed workers. This is surprising because

traditionally, in addition to taking care of the family, women have been visible in three sectors –

agriculture, animal husbandry and handicraft works.47 In fact, studies reveal that women

comprise 65% of Afghanistan’s agricultural workforce and that poorer women tend to be more

involved in agricultural tasks than wealthier women.48 In non-farm work, carpet weaving,

sewing, and tailoring are the most common. Furthermore, scholars write that Afghan women’s

46 Married includes widows. 47 Tavva, Srinivas, Malika Abdelali, Aden Aw-Hassan, Barbara Rischkowsky, Markos Tibbo, and Javed Rizvi. “Gender Roles in Agriculture: The Case of Afghanistan.” Indian Journal of Gender Studies 20.1 (2013): 111-34. 48 Parto, S.R. and Mihran, R. Understanding Gender in Agricultural Production: An Annotated Bibliography for the Case of Afghanistan. Afghanistan Public Policy Research Organization, Kabul, Afghanistan (2010).


14

contributions to agriculture have been systematically excluded in agricultural planning,

production, record keeping and training.49 Due to survey limitations, the Survey of Afghan

people currently does not capture these sectors of women’s work well, and this reflects a larger

problem in what little is known about the true share of women who are working.50

Women in the labor force belong to households with an average income of roughly 7,500

AFN per month.51 75% come from households that own a mobile phone, a house, televisions,

and sewing machines, while less than 30% own cars, refrigerators and washing machines. 44%

of the women have personal mobile phones that are not jointly used by other family members,

and 70% watch television.

b. Predictive Model: Determining the barriers

As discussed in Section III, the factors in the model should be able to (1) differentiate women

who are in the labor force from those who are not; (2) have policy relevance, i.e. be able to be

affected through policy; and (3) provide suggestions for the theory of change in the NEEP.

In this model, the key correlates provide insight into how FLFP interacts with female

characteristics, security and violence, information sources, as well as men and women’s

perceptions towards gender roles. These variables have the highest linear correlation with FLFP

(see Table 3A). Detailed technical methodology of the predictive model is included in Appendix

Section D. Using province-level and ethnic group fixed effects, we attempt to control for many

of the unobservable factors that influence women’s decision.52 Province-level and district-level

fixed effects captures the unobserved factors due to geographic characteristics, while the ethnic

group fixed effects captures the unobserved factors due to ethic norms.

Table 4A shows that the factors related to security and violence are negatively and

significantly associated with FLFP when pooling females (Regression 1). After we apply ethnic

group fixed effects, security and violence remain negatively significant (Regression 4). This

implies that security and violence is a universal concern despite underlying ethnic group

differences.

49 Dixon, R.B. Women in Agriculture: Counting the Labor Force in Developing Countries. Population and Development Review 8 (1982): 539–566. 50 This appears to be due to skip patterns in the current survey instrument. 51 This is roughly 108 USD per month, which quite low but above the poverty line. 52 See Appendix Table 4A.


15

As suggested in Section II, the most predictive factor is years of education – the higher the

education years, the more likely the woman participates in the labor force. Having more years of

education will increase the likelihood of women entering the labor force, when holding other

factors constant (Regression 1). The association remains strong even when other unobserved

geographic factors and ethnic norms are controlled for (Regression 2, 3, and 4). This association

makes practical sense and can also be affected by cultural norms, i.e. girls dropping out of school

early to get married or do housework.

Interestingly, men’s education has a reverse and significant effect on FLFP. As average

years of education for men at the district level increases, women associated with that district are

less likely to enter the labor force, holding for other factors constant (Regression 1). The

prediction power remains despite underlying differences in ethnic groups (Regression 4). This

suggests that the quality and content of education matters, and that just because men in an area

are educated does not imply greater mobility for women.

In addition, exposure to the outside world is an important indicator for FLFP. Watching

television and obtaining information from community shuras is positively and significantly

correlated with FLFP. The significance of obtaining information from community shuras

remains after unobserved ethnic norms are considered. The predictive power of both information

sources is robust (see Table 10A and 11A) as well as their individual linear correlations with

FLFP (see Table 3A). All these imply that television and the community shura are channels for

woman to receive information outside the home, which provides a policy space for intervention.

Women’s own empowerment is also a critical predictor. Consistent with the diagnostic, the

idea that women should be allowed to work outside home is a robust predictor even when

unobserved geographic differences and ethnic norms are considered (Regression 1, 2, 3 and 4).

There is also a strong evidence of a causal link of the perception of allowing women’s mobility

outside home to entering the labor force (see Table 8A). In addition, the predictive model also

suggests that secularism does not positively influence women’s decision to be in the labor force,

which implies that religion does not play a key role.

Finally, men’s perception toward women’s work is a powerful predictor for FLFP. The

model suggests that women are more likely to be in the labor force when men in the region are

more supportive for equal education opportunities (Regression 3) and are less strict on where

women can work, for instance, co-ed school (Regression 4). The relation of FLFP and men’s


16

acceptance of women to work in co-ed schools has two implications (Regression 4). First, men’s

perceptions matter and appear to influence women’s empowerment. Second, by having a woman

act in a more authoritative role (i.e. as a teacher), boys might be influenced to changing their

normal perceptions on gender roles. Given that there is no causality, the linkages can go in both

directions.

The predictive model reveals some key findings that have a variety of policy implications,

but before examining policy, we will look more deeply into some of these barriers (e.g., security,

education, caregiving, religion, norms, and the legal framework) to determine the key underlying

barriers that we should focus on in Afghanistan.

c. Diagnostics: Determining the underlying barriers

i. Security: Incidence of Violence Security has deeply affected all aspects of society in Afghanistan and is a major source of

concern. In 2014, the total incidence of violence, including criminal events, explosive hazards,

non-combat events and threats, is 32,033 incidences.53 Violence appears to be higher along the

border with Pakistan and in the Kabul area (see Figure 4A). Overall, FLFP in Afghanistan has

been fairly constant (and low) so it is difficult to understand on a national level how security has

affected FLFP overtime. Given this, we use province level security and labor force data.

Unsurprisingly, there is a large variation of FLFP across regions – the rate varies from as low as

5.8% in Khost province to as high as 76% in Takhar province (see Figure 5A).54

Consistent with the predictive model (see Table 4A), FLFP is negatively correlated with

incidence of violence (see Figure 5A and 6A). Table 3A reveals the strong correlations between

FLFP and key security and geographic indicators. Figure 2 (below) shows that the southwest

provinces, Nimruz, Helmand and Kandahar, have the highest violence levels and the lowest

FLFP rate, whereas the northwest and northeast areas have the opposite trend. According to the

2015 Survey of Afghan People, the share of people who report that someone in their household

‘has been physically injured as a result of violence, such as by a bullet, rocket, or unexploded

ordinance’ is highly negatively correlated with FLFP by region (Figure 2). This could imply that

security concerns prohibits women from entering the labor force.

53 Incidence of Violence in Afghanistan Data. Provided by Professor Mike Callen. 2014. 54 Estimates from 2015 Survey of Afghan People.


17

Figure 2: FLFP and violence

Source: Survey of Afghan People (2014) and violence data (2014) from Professor Mike Callen.

Data reveals that the share of women who believe they should be allowed to work outside

home is high in areas where violence is relatively low. For example, in 2014, the central region,

where 86% of women believe that they should be allowed to work outside home has relatively

few incidences of violence (see Figure 3). It is also important to consider the fact that more

dangerous areas will have poor data quality.

Figure 3 suggests that women’s willingness to leave their home is related to national security

and violence. As incidence of violence increases (at a regional level), less women report that

they should be allowed to work out the home, with the exception of Kabul area.

Figure 3: Perceptions and violence

Source: Survey of Afghan People (2014) and violence data (2014) from Professor Mike Callen.

Figure 4 plots the Global Peace Index score across the x-axis and the respective LFP rate in

150 countries. The larger the score on the x-axis, the more dangerous/insecure the country is.


18

Figure 4: Security and labor force participation across countries

Source: WDI (2015) and Global Peace Index (2015). FLFP appears to have a negative correlation whereas there is no apparent correlation with male

labor force participation (MLFP) and security. This also reveals that security issues may

disproportionately affect women. The negative association, along with the regional differences

in Afghanistan, appears to indicate that security is a critical barrier that prevents women from

entering the labor market.

ii. Education/Skills

Overall, education levels are extremely low in Afghanistan. In 2015, women, on average,

had 2.5 years of education. According to the latest survey, lack of education and professional

courses were reported as the biggest challenge facing Afghan women. As shown in the

predictive model, education is an important factor that can determine if a female is in the labor

force or not. Direct evidence reveals that years of education is statistically different between

women in and not in the labor force. On average, women in the labor force have 5.5 years of

education, while women not in the labor force have 1.9 years. In addition, the literacy rate of

women in the labor force is 12% and 8% for women who are not in the labor force.55 It seems

that while there is a difference, overall female education levels are too low to draw any serious

conclusions about whether education is a key underlying barrier for women.

Figure 5 reveals that FLFP increases over time while women in the labor force’s years of

education decreases, especially during the period from 2010 to 2014. In the past decade, years of

55 These differences are statistically different from zero at 95% significance level.


19

education for women in the labor force has declined from 7.5 years in 2006 to 5.5 years in 2015.

At the same time, FLFP increased from 14.5% to 16.4%56 and the education gap between women

in and not in the labor force narrowed from 5.3 years to 3.5 years. The decline in the years of

education is largely due to decline in education in high school and increases in no schooling.

There is only a slight increase in higher education (see Figure 6). This implies that women with

less education can bypass the education constraint and enter the labor force, which also reveals

information about the demand for labor and the structure of the economy.

Figure 5: Time trend of female years of education and FLFP rate

Source: WDI (2105) and Survey of Afghan People (2015).

Figure 6: Education completion level time trend for women in labor force

Source: WDI (2105) and Survey of Afghan People (2015).

562014 WDI estimate.


20

Figure 7: Women’s education and FLFP across countries

Source: WDI (2015).

In addition, as the U-shaped hypothesis suggests, with extremely low levels of education, we

would expect high FLFP due to poverty and economic status – in Afghanistan, this does not

seem to be the case. In fact, as Figure 7 indicates, when looking at data from 103 of the world’s

poorest countries, average years of female education and FLFP do not appear to have a

meaningful relationship.57 Visually, there is high FLFP in countries with very low education.

While education is extremely important, it may not be the most critical underlying barrier

that prohibits women from participating in economic activities.

iii. Caregiving In the Afghan context, women are the primary caregivers. According to 2012 NRVA, 70%

of married women between the ages of 15 and 49 have children under the age of 5.58 The average

number of children under 5 per woman is 1.3. Without support in the household, childrearing is

likely to restrict women’s mobility preventing her from working. It is plausible to believe that

women in the labor market have fewer children. However, in Afghanistan, the difference in the

average number of young children between women in and not in the labor force is not significant

(see Table 1). Although literature suggests that burden from childcare is an important factor in

women’s decision to enter the labor market, it is not a key underlying barrier in the Afghan

context since it appears as through women are able to overcome this barrier.

57 “World Inequality Database on Education.” DME WIDE. UNESCO, 2015. <http://www.education-inequalities.org/>. 58 Women who identified themselves as married, divorced and widow.


21

Table 1: Average characteristics of women by labor force status

LF Status Age of first marriage

% of women with kids, age < 5

# of children, age < 5

# of children, age < 14

In Labor Force 17.9 70.7% 1.4 2.9 Not in Labor Force 17.6 70.5% 1.3 2.8

iv. Customs/Norms

Deep-seated cultural norms are important in influencing women’s decision to enter the labor

force, as well as men’s support of women entering. Along with anecdotal evidence, we rely

heavily on the Survey of Afghan People data to investigate women’s marriage and fertility

decisions, perceptions towards work and education, and differences among ethnic groups (that

stem from difference in cultural practices). We want to understand if longstanding norms

actually restrict mobility and preclude women from entering the labor force, as anecdotal

evidence suggests.

Literature suggests that early marriage and high fertility restricts women’s mobility.

Afghanistan has the highest fertility rate in South Asia of 5.1%. Women tend to get married

around the age of 17-18. There is also no practical difference between women in the labor force

and not in the labor force (see Table 1). Although the fertility rate is high and marriage ate is

quite low, evidence does not seem to indicate that these factors are the key barriers preventing

women from working.

Afghanistan is a multiethnic society divided into a wide variety of ethno linguistic groups.

The dominant groups are Pashtun, Tajik, Hazara, and Uzbek; other groups including Aimak,

Turkmen, Baloch, Pashai, Nuristani, Gujjar, and many others. While analyzing the provincial

differences in FLFP (Figure 5A), it was immediately apparent that ethnic norms influence

women’s decision to work. As Figure 8 shows, the level of women’s participation varies

significantly by ethnic group: 40% in Turkmen to 10% in Pashtun.59 We assume that men and

women within an ethnic group share similar cultural norms. Women in minority groups

participate in the economy more than those who identify as Pashtun, Hazara, Uzbek and Tajik.

To take a closer look, Table 3A reveals that being in the Pashtun ethnic group is negatively and

mostly correlated with FLFP. This is consistent with the fact that Pashtuns are more

59 We focus our analysis to the top 7 ethnic groups; however, the minority ethnic groups do tend to have higher FLFP. This could be due to cultural norms or the mere fact that some of these groups are marginalized or are extremely poverty.


22

conservative in their gender role attitudes – Pashtun women face more restrictions and are less

mobile.

Figure 8: FLFP by ethnicity and monthly income

Source: Survey of Afghan People (2015)

It can also be argued that income is a confounding factor as minority groups tend to be

marginalized and poorer than the dominant ethnic groups, especially the Pashtuns (see Figure 8).

As the literature suggests, women enter the labor force due to poverty; however, Figure 8 shows

that there is no obvious correlation between monthly HH income and FLFP. The number of

women in the labor force is consistent across monthly income groups, with the exception of the

one income group. Evidence and anecdotal information indicates that ethnic norms do persist and play a role in women’s decision to enter the labor market.

Mobility is another factor that restricts women. “Being under men’s control” and “cannot

leave home” are reported in the top ten problems facing Afghan women. While the majority of

women (75%) state that they should be allowed to work outside, less than half of men (40%)

hold that belief. As Figure 9 shows, this perception also varies significantly between women in

and not in the labor force. On average, women in the labor force are more associated with

believing they should work outside the home.

Both men and women’s attitudes toward women’s work affects women’s decision to work.

Roughly 23% of Afghans believe that women face a lack of job opportunities. This suggests that

women may prefer to work if jobs were available. Data also suggests that unemployment and

lack of opportunities concern women who are in the labor market more. In addition, women


23

might feel restricted to certain type of jobs. 90% of Afghans believe that female-only schools

and hospitals/clinics are appropriate for women – very few support women in nontraditional jobs.

Due to lack of data, there also seems to be little information about the work Afghan women

culturally engage in, for example, home-based handicraft work. Afghan women also report

“forced marriage” as one the top 5 problems they face. Women not in the labor force are more

concerned with this issue than those in the labor force. This could imply that lack of voice or

authority on marriage decisions does limit women’s choices.

Figure 9: Women’s perception on women’s mobility and on marriage age60

Source: Survey of Afghan People (2015).

Women’s own perceived internalized gender role also prohibits them from entering the labor

force. As Figure 9 shows, on average, women think the best age to marry is roughly 19 – 20.

The data on this indicator is tightly concentrated on the mean, which reveals that strong norms

around women’s role in the household persist. Women in the labor force report, on average, a

slightly higher ideal marriage age compared to women not in the labor force; however, both are

on the younger side. In addition, on average, men report an even younger ideal age for women

to marry – this difference is statistically significant and reveals that Afghan men have stronger

perceived gender roles for women. Overall, societal norms, for both men and women, appear to

be a significant barrier preventing women from entering the labor market.

v. Religion

Literature suggests that religion plays a role in restricting women’s decisions.61 Roughly 80%

of Afghans practice Sunni Islam while the rest are Shias – Islam is the dominant religion. In the

60 The differences in both graphs are significance at 99% significance level.


24

Afghan context, there is no convincing evidence to connect Islam to FLFP. Using the question,

“should religious leaders be involved in politics?” as a proxy for religious views, the predictive

model shows that having non-secular views is strongly correlated with FLFP. Figure 10 further

reveals that a larger percentage of men, who support women working outside home, state that

politics and religion should mix (non-secular) compared to the men who do not support women

working outside the home. Similarly, Figure 10 reveals that women in the labor force are more

likely, on average, to believe that politics and religion should mix when compared to women

who are not in the labor force.

Figure 10: Differences between people who feel that politics and religion should mix62

Source: Survey of Afghan People (2015).

Even when examining other countries, FLFP is quite high in both Bangladesh and Indonesia

– other majority Muslim countries. While this evidence may not be the most robust, it does not

seem that more open-minded men and women are less religious or vice versa; thus, religion does

not appear to be a underlying barrier limiting women’s participation in the economy.

vi. Legal Framework In every society, laws are written to protect citizens; in fact, in most countries, there are

specific laws to protect women. As mentioned, the Constitution of Afghanistan now promises

equality between men and women. It permits women to work outside the home, to engage in

political activity, and nominate a certain number of female candidates.63 The Women, Business,

61 Lehrer, E.l. “The Effects of Religion on the Labor Supply of Married Women.” Social Science Research 24.3 (1995). 62 The differences in both graphs are significance at 99% significance level. 63 Ghaisy, Richard, Jiayi Zhou, and Henrik Hallgren. 2015.


25

and the Law indicators reveal that Afghanistan addresses 24 of the 38 key legal areas with

respect to gender. In South Asia, the average number of laws addressed is 35 out of 38. Figure

11 reveals that South Asian countries with higher FLFP have addressed more non-discriminatory

laws.

Figure 11: Number of non-discriminatory and supportive laws and FLFP

Source: WDI (2015) and Women, Business, and the Law (2015)

The first order issue for Afghanistan is simply to have laws to protect women - the GoA can

substantially improve the legal and institutional environment for supportive laws towards

women. The second order issue is extremely critical and that is the distinction between de jure

and de facto laws. Many of the laws in the books appear to be only de jure and implementation

has been problematic. While many non-discriminatory laws do exist, the next step will be to

assess if these laws are in fact practiced by the legal system. In Afghanistan, personal law64 is

recognized as a valid source of law under the constitution, and in certain circumstances,

customary law can take precedence over constitutional provisions on nondiscrimination, which

indicates that traditional practices can influence the legal system despite protective measures

being in place for women. The legal framework of Afghanistan is affected by the lack of

political support and the prevailing norms in the society; thus, while the law is a large constraint,

it is not the key underlying barrier limiting women’s participation in the economy. Before

delving into an analysis of policy options, it’s relevant to understand Afghanistan’s

implementation capabilities.

64 Personal law refers to non-customary legal systems that stem from tradition or doctrinal texts, which are sometimes uncodied.


26

d. State Capabilities

Afghanistan, with international support, is in the process of building its capabilities. It is

essential that policy options are appropriate to the Afghan context to avoid the traps of premature

loadbearing and isomorphic mimicry.65 The focus of Afghan state building has been on

establishing formal government institutions, such as MoWA, usually working from the top down

and from the capital to the periphery. This approach views Afghanistan, which has been without

a coherent central government for more than 30 years, as a country with very little significant

capabilities. However, Afghanistan has a long history of strong informal ruling through

community jirgas and shuras.66 The provision of public goods has taken place in the absence of

an effective centralized system of government.67

In order to embrace Afghan ownership in the policymaking process, comprehensive policy

options on FLFP should be proposed after fully examining Afghanistan’s implementation

capabilities. In the Afghan context, any feasible intervention, especially on contentious women’s

issues, has to be internally driven, simple to implement, cost-effective, have minimal political

pushback, be somewhat centralized, and involve local informal levels of government.

First and foremost, the program has to be internally driven. Past examples suggest that

outside interventions to affect change will be short-lived once the donors and implementers leave

or funds are cut. In order to be sustainable, programs have to be motivated by the GoA itself. In

2002, the MoWA was created with U.S. support to improve women’s rights by gathering

statistics, promoting public awareness of relevant laws, and protecting women from domestic

abuse, among other things. Gender issues are multidimensional and require cross-sector

collaboration – a comprehensive implementation plan needs to involve internal intra-ministerial

collaboration through working groups.

Given that bureaucratic capabilities are remain weak, the central government has to maintain

control over program implementation to ensure that goals and targets are met. This means

programs with many levels of delegation should be avoided. This, however, does not mean that

65 See Table 5A in the Appendix for Capability Trap definitions. 66 A jirga is a meeting held to resolve disputes which have an ad-hoc membership. A shura is a council or group of people who discuss particular issues. They can be held on a number of topics to include governance, security, and development. Historically, women have been excluded from these meetings. 67 Unruh, Jon Darrel and Rhodri Williams. Land and Post-conflict Peacebuilding. London: Earthscan, 2013.


27

local leaders will be neglected – programs have to receive buy-in from stakeholders across the

spectrum.

Afghanistan is not financially self-sufficient. It largely depends on international community,

especially the U.S. who has played a large role in contributing to programs for women’s issues.

In order to bring a long-term and sustainable progress, the program has to be simple and

financially appropriate – especially if the international community decides to exit in the future.

Examining cost effectiveness in the implementation of a program will be essential.

Women’s issues are extremely sensitive and there is a threat of political pushback from

strong stakeholders. Figure 1A maps various stakeholders by their influence, particularly if they

oppose or support programs targeted to increase FLFP. We can see that the international

community is the leading activist in this issue, whereas the Taliban is a leading opponent. The

rest of the stakeholders fall somewhere in between, with most in the opponent category. Figure

7A assesses stakeholders by their interest and power. Opposition groups seem to have medium

interest in our issue, but very high power. The international community has very high interest

but limited power with respect to competing local actors. Compared to the opposition groups

and Islamic leaders/scholars, the GoA may not have enough power to carry out a major reform

on women’s issues, even if it has the interest and motivation to do so. In fact, President Ghani

once met with religious leaders and mullahs to garner support for adding reserved seats for

female judges in the national courts. However, he received tremendous pushback and was

unable to execute the reform.68 In order to ensure a sustainable program, the various

stakeholders’ positions, relative power, and interests should be taken into consideration to create

a sound policy that receives minimal political pushback.

Afghanistan is a complex country and will require the Problem Driven Iterative Adaptation

(PDIA) approach where we continuously pilot, test, and validate our theory of change. Given the

capabilities, proposed options, while addressing the underlying barriers to FLFP, have to also be

practicable in the Afghan context.

V. Policy Options

Currently, the NEEP highlights six separate programs or policy options. The goal of the

NEEP is to target poor women in both rural and urban areas, produce significant and measurable 68 Mohib, Lael. “FLFP in Afghanistan.” Personal interview. 20 Jan. 2016.


28

results, create simple and cost-effective management arrangements that build on existing

capacities, and, ultimately, to sustainable economic development.69 Using these criteria, the six

programs vary in budget, scope, and coverage. Table 6A and Figure 3A present the six policy

options and the underlying assumptions they make to address low FLFP. While many of these

assumptions are closely scrutinized in our framework, we do not examine some of the aspects of

the NEEP, such as access to credit and the demand for labor.

Through our framework, we find a few significant and central insights with respect to FLFP

that will shape our policy recommendations:

• Security and violence deeply affect FLFP. High violence not only diminishes women’s willingness to leave their home, but it also raises concern in men, which further restricts women’s mobility. Security is the key underlying barrier limiting women’s participation in the economy. The level of security varies by region and is closely related to the regional economics and political factors. Therefore, a sound policy aimed to address FLFP requires local intervention that is specific to the regional context.

• Lack of protection from the legal and institutional system discourages women to engage in the economy. Low political support for women’s equality and the prevailing norms has resulted in a weak legal framework; thus, while the law is a large barrier, it is not the key underlying barrier limiting women’s participation in the economy.

• Education is an indicator that predicts FLFP well. Women also self-report that lack of education, illiteracy and professional trainings are the biggest challenge facing Afghan women. A supporting environment for women’s education is also instrumental, especially support from men toward equal education opportunities.

• Information matters, and the channel in which women receive information matters. Watching television, receiving information from the community shura, and having personal mobile phone are all positively correlated to women being in labor force. Obtaining information from the outside world provides exposure and can update women’s prior beliefs.

• Local leadership roles can improve outcomes for women. Women’s role as a leader in a Community Development Council (CDC) is highly accepted by women and men. Having a female leader in a village can raise female aspirations and has shown to increase years of female schooling.70 In the Afghan context, men’s perceptions on issues matter – their support is critical. Men’s support for women to work in co-ed schools is an important indicator that predicts FLFP. By having women in a more authoritative/leader role (i.e. a teacher), perceptions on traditional gender roles may be changing.

69 Afghanistan. Ministry of Labor, Social Affairs, Martyrs and Disabled. Draft of National Action Plan for Women's Economic Empowerment in Afghanistan 1395-1399. Ver. 2. Kabul: 2015. 70 Beaman, Lori, Esther Duflo, Rohini Pande, and Petia Topalova. 2012.


29

• Afghanistan is a weak state, but has a long history of strong informal ruling through

community jirgas and shuras. These traditional authorities have the strength and resources to exercise local control and provide services in many parts of the country. However, given that Afghanistan consists of many ethnic groups, achieving consensus is administratively difficult. Therefore, programs that seek to be effective have to take place from both top down and bottom up. Policy recommendations should be designed as a centralized intervention that involves community level engagement and decision-making.

Taking into account the NEEP criteria and building on the literature, past interventions, and our

findings, we assess 3 policy options: (1) local-level female quotas, (2) increase access to

information through television, and (3) business trainings. Table 7A provides a visual matrix of

the 3 options. We consider option (1) and (2) as they are technically correct and address the

underlying barrier as identified by the analysis. We consider option (3) as it represents the status

quo, is a large part of the NEEP, and a part of USAID’s Promote program.

a. Option 1 – Local-level female quotas

As suggested in the literature, female leaders can raise female aspirations and allow

communities to make different decisions, such as increase years of female schooling. This

channel affects the norms, which is a key underlying barrier suggested by the diagnostic.

Option 1 is to use shura-level quotas for women at the community level. The central

government will have to work with local religious leaders to initiate a plan to promote female

leadership in local villages to enable women to engage in substantive discussions. The

percentage of reserved seats can vary based on local sociopolitical conditions.

Gender quotas, as a top-down policy instrument, have been common in recent years. India is

a successful example where female leadership diminished bias against women and their ability to

lead, indirectly changed girls’ aspiration and parents’ aspiration for girls, and narrowed gender

gaps between boys and girls in school attainment.71 In Afghanistan, gender quotas in village

CDCs improved outcomes such as increased mobility and income generation, but given the

duration of the study, deeper gender roles remained unchanged.72 Gender quotas are also used at

the national level throughout the Middle East and North Africa in countries like Sudan, Turkey,

Kuwait, Iran, Jordan, etc.

71 Ibid. 2012. 72 Beath, Andrew, Fotini Christia, and Ruben Enikolopov. 2013.


30

According to Article 83 of the 2004 Afghan Constitution, 68 of the 249 total seats (27%) in

the Lower House (Wolesi Jirga) are reserved for women, comprising of at least two women for

each of the 34 provinces of the country.73 Article 84 also ensures that 50% of the individuals

appointed by the President in the Upper House (Meshrano Jirga) must be women. In the current

system, representation for women is 28% in the Lower House and 27% in the Upper House.74 In

each provincial council, at least 20% of the seats should be allocated to female candidates with

the most votes. At the constitutional court-level, women have no representation as justices. At

the community level, there are no laws that reserve seats for women in Shura.

Based on our analysis, deep-seated cultural norms are a key barrier preventing women from

entering the labor force. Changing perceptions toward gender roles in the family, village,

community, and society are extremely important to influence FLFP. Evidence suggests that

obtaining information from community shura differentiates women in the labor force from

women who are not. This could imply that women who receive information from shuras are

more engaged in the community decision-making process. The downside of this option is that

overall only 22% of women obtain information from shuras. Hypothetically, the impact of a

woman shura quota on changing female aspirations might be small due to the fact that only 22%

of women use this information or large due to snowballing effect or inspiring more women to

seek information given that it's a local female leader.

In addition, women’s leadership role as members of a CDC is widely accepted by the

majority of women and men. Given that men’s perceptions are extremely important in

Afghanistan, acceptance from men will provide local political support. By having women in a

more authoritative/leader role (i.e. a teacher), perceptions on traditional gender roles may be

changing.

There are also many unobserved underlying factors associated with different regions and

ethnic norms. With 77% of women living in rural and remote areas, such geographic variation is

extremely important. Therefore, while localized gender quotas can empower women through

changing aspirations and reducing men’s bias against women, there is still a large difference of

social beliefs, rituals, and norms across villages to consider.

73 Electoral Law of Afghanistan, 2010. 74 “Database of Quotas for Women.” Quota Database. International IDEA, Stockholm University and Inter-Parliamentary Union, 2015. <http://www.quotaproject.org/>.


31

At the community level, the shura is a council that resolves disputes and takes collective

decisions about important social issues and that carries a religious subtext.75 The buy-in from

religious scholars is critical. The unsuccessful attempt by President Ghani to reserve seats for

female judges in the national courts provides us an insightful lesson. While this option may be

technically correct, there is likely to be grave political pushback and challenges in the

administrative capabilities to monitor.

b. Option 2 – Increase access to information through television

As indicated in the diagnostic and predictive model, norms are an underlying barrier

affecting both men and women, and can be linked to women’s decision to enter the labor force.

Changing norms is quite difficult – behavioral economists have studied several interventions to

understand the channels through which people change their beliefs and practices. One of the

channels that has seen success is television: Option 2 is to use television as medium to affect

behavior change. As shown, FLFP is highly correlated with various modes of information sharing. One of the

most positively correlated indicators is watching television, along with using a personal mobile

phone. While it may be that women in the labor force have better access to these items due to

income effects, the reverse effect cannot be easily ruled out. Television and mobile phones do

provide access to information and exposure to the outside world. From the study in India, large

changes were observed in attitudes towards women with little to no targeted appeals. Cable

television is thought to be an effective form of persuasion because people emulate what they

perceive to be desirable behaviors and attitudes.76

In Afghanistan, the radio program “New Home, New Life”, implemented in 1994 by BBC

World Service, has been quite effective at delivering messages on social themes such as forced

marriages, blood feuds, landmines, immunization, opium addiction.77 Even through the shifting

political landscape and the Taliban occupation, the radio program delivered quality programming

with more than 70% of the population tuning in at some point.78 In fact, reports suggest that most

75 Unruh, Jon Darrel and Rhodri Williams. 2013. 76 Jensen, Robert. 2012. 77 New Home, New Life Final Report. USAID, June 2007. <http://pdf.usaid.gov/pdf_docs/Pdack342.pdf>. 78 “New Home, New Life Reflects Real Life.” BBC News. BBC <http://www.bbc.co.uk/mediaaction/where-we-work/asia/afghanistan/education>.


32

of the Taliban soldiers were also glued to the series.79 There is also evidence that the educational

messages were hitting home: After a long-running storyline about landmines, listeners to the

program were found to be statistically less likely to be killed by a landmine than non-listeners.80

Using a similar method to broadcast messages about women’s issues could have powerful effects,

especially if both young boys and girls are accessing the program.

Access to television and the popularity of various shows, especially from India, appear to be

increasing in Afghanistan. Since 2015, 80% have access to radio and roughly 60% have access

to television. In fact, television ownership is increasing as is obtaining information from

television. Between 2014 and 2015, there was an 8% increase in television ownership and in

2015, 56% of Afghans reported watching television. Among the television networks, Tolo TV is

the most frequently watched, capturing around 50% of the television viewers. Partnering with a

television network, like Tolo TV, to expand signals, the GoA can use an effective television

show that aims to touch upon women’s issues through a continuous storyline – similar to the

BBC program. Therefore, television can be served as a channel to address norms.

From an implementation perspective, a television program would be relatively simple and

inexpensive. While the GoA may have to invest in infrastructure for cable signal in rural areas,

the returns for such a program could be worth the initial investment.81 Since television is

spreading in Afghanistan, this may lead to changes in attitudes towards women even without

targeted interventions. In fact, despite the high poverty that prevents television ownership, we

may still see widespread effects as evidence from India suggests that people often watch

television with friends, neighbors, and relatives.

Furthermore, the program would be centrally managed and disseminated through television

networks and would minimize the need for a bureaucratic structure. In order to ensure political

supportability, locals would need to be involved for developing content that is tailored to the

Afghan context – religious leaders, mullahs, council members, teachers, and others will engage

in focus groups to help design the appropriate content for such a program. The GoA may not be

able to explicitly release the exact goals of the program, but by taking into account what the

people enjoy watching, they can develop a popular program with subtle thought-provoking

79 Brockes, Emma. “A Long Way from Ambridge.” The Guardian. Guardian News and Media, 23 Oct. 2001. <http://www.theguardian.com/world/2001/oct/23/afghanistan.terrorism3>. 80 Ibid. 81 We recommend such an analysis to be completed to understand the long-term impacts of a TV program.


33

messages. While there may not be concrete short-term gains, except for data on viewership and

access to information, the long-term gains should not be overlooked.

With appropriate messaging, an educational television or radio program can provide

information and exposure to the outside world that can gradually shift norms and standards

around traditional gender roles.

c. Option 3 – Business training program The GoA currently uses CDCs to provide communities with basic services in health,

education, water, etc. The CDCs already have a mandated mechanism for women’s participation

and in many regions, with donor funding, they have also established community-based

institutions, such as women’s business groups and lending groups.82 The GoA plans to partner

with these institutions to develop women’s business training centers that can provide technical

assistance in financial literacy, small business management, opening bank accounts, and other

necessary support in a safe female-only space approved by the GoA and village leaders. Option 3 is to provide business training to women.

While education appeared in the predictive model, the diagnostics does not directly point to

education/skills being a key barrier for women to engage in the labor market due to the fact that

women could bypass the barrier. Consistent with literature, women with more years of education

are more likely to be in the labor force; moreover, the effects were stronger when looking at

women who had attained university-level education. While business training provides relevant

skills that women desire, promoting and enhancing formal education may be more relevant.83

But, education might not be the most relevant barrier in the Afghan context due to its interaction

with persisting underlying norms – these very norms might also prevent girls from receiving an

education. If the norms component can be relaxed, it might be easier for women to access

education and business training may not be necessary. Therefore, education may be suboptimal

without addressing norms first. The current issue is that while business training is empowering,

it conflicts with certain pressures women feel – there will be a selection bias with respect to the

types of women accessing these programs.

82 Afghanistan. Ministry of Labor, Social Affairs, Martyrs and Disabled. Draft of National Action Plan for Women's Economic Empowerment in Afghanistan 1395-1399. Ver. 2. Kabul: 2015. 83 In the Survey of Afghan People, lack of professional courses and skills were cited to be one of the top problems facing Afghan women.


34

Moreover, from research in India, it was found that business training was not as useful for

women who were subject to extreme restrictions since they had too little agency to easily change

their aspirations or activities.84 For the Afghan context, this implies that business training may

not uniformly benefit women, and it might be difficult to gauge whom the program may benefit.

Business trainings, while perhaps political feasible after stakeholder buy-in, may be

administratively arduous for the GoA to administer. In order to reach the remote corners of the

country, the GoA would require a decentralized model with ownership at local levels.

Depending on the type of training and from the USAID examples, there could also be a large

cost burden on the GoA to implement a training program to target one million women in rural

and urban areas.

VI. Recommendations Our analysis provides great insights into the channels at which FLFP can be affected by

policy measures. While there are many approaches, we believe that focusing on a few key issues

will be in the best interest of the GoA. Our recommendations, while comprehensive, are the first

step in what should be an ongoing and iterative learning process.

a. Security Constraints In the Afghan context, security is a challenging issue and requires bold leadership from

several actors in the international arena. Our analysis leads us to believe that security tensions

arising from acts of violence is the key underlying barrier that prevents women from entering the

public space. Our main recommendations to the GoA does not delve into security as they are

beyond the scope of this analysis; however, we feel this is a first order concern that will need to

be addressed at a regional level to ensure the empowerment of not only Afghan women, but

Afghan people in general.

We recommend the GoA dedicate resources to meet with relevant actors and stakeholders

both internally and externally, and design a new regional strategy to alleviate the escalating

security tensions. This is by no means a simple task, but the approach to ensuring security needs

to be revisited. However, given the tense security situation, the GoA can work to encourage and

promote home-based women’s work, such as handicrafts and textiles. While this does not solve

84 Field, Erica, Seema Jayachandran, and Rohini Pande. 2010.


35

the mobility issue, it allows women to work and sheds light on an important industry in

Afghanistan.

b. Policy Intervention: Increase access to information through television Given the tradeoffs with respect to technical correctness, administrative feasibility, and

political supportability, we recommend Option 2, to increase access to information through

television.85 This channel directly addresses norms, which is the underlying barrier affecting

both men and women, and can be linked to women’s participation in the economy. FLFP is also

highly correlated with various modes of information sharing, and watching television is one of

the most positively correlated indicators. 72% women in the labor force obtain information via

television, whereas about 60% not in the labor force use television. The difference is statistically

significant. Cable television is thought to be an effective form of persuasion because people

emulate what they perceive to be desirable behaviors and attitudes; therefore, increased access to

information through television allows us to effectively target the key barriers that prevent women

entering the labor force.

With respect to administrative feasibility, a television program would be simple and cost-

effective. In Afghanistan, roughly 50% of the population has access to a television. Although

the up-front infrastructure cost is high, the long-term benefit would potentially be profound. In

addition, this centralized program would be effective in reaching women from rural and urban

areas with a low variable cost. As the stakeholder map suggests, the international community

has relatively high power and high interest to support a targeted women’s program. With a

sound design, international stakeholders can be engaged to provide financial assistance.

From a political perspective, this program would be disseminated like any other television

program, without the explicit knowledge of it being a women’s empowerment program. While

religious scholars or opposition groups may try to prevent women from watching television,

overtime, as televisions become widespread, it will be difficult for opposition groups to ban

television viewership. Also, the goal of the program would be to start a discourse, and watching

the show is not required for this to occur. Furthermore, the program will be centrally managed

and disseminated through local partner networks. This would also minimize the political push

back from religious scholars, who have the strength to exercise local control. The centralized

85 See Table 7A for analysis of the policy options.


36

structure of the policy also limits the negative influence from provincial governments and threats

from the Taliban.

c. Data Quality To precisely measure the magnitude of challenges facing women, policymakers need access

to appropriate and timely data. There are commendable data collection efforts in Afghanistan,

but currently, datasets are missing a few critical elements with respect to FLFP. We recommend

the GoA to revisit its national surveys to:

• Include questions to capture the activity of all the women in the household and not only

the married women.

• Probe women who self-identify as housewives to see if they participate in the informal

labor market through agriculture, domestic labor, handicraft work, etc.

• Capture the sector of employment.

• Directly ask labor force involvement questions instead of asking “In the last week, did

you do any work for pay, for profit, or for family gain?”

Section E of the Appendix provides a complete survey instrument that effectively captures the

relevant information needed to understand the complete picture of FLFP in Afghanistan.

d. Implementation Plan

Keeping in mind the larger goal of increasing FLFP, we propose using the PDIA approach

when implementing the program. Our analysis has already started the process and we have

completed the first two steps and parts of step four and five of the implementation plan. Figure

8A provides a visual on the following plan:

1. Gain authorization—Who in the GoA can reliably authorize work to be done on this issue?

The Office of the First Lady has expressed interest and commitment to women’s issues,

including issues around FLFP.

2. Nominate problem—What is the larger visible problem?

The problem that has been nominated by the authorizer and stakeholders is low female labor

force participation in Afghanistan.

3. Form team—Who are the key actors that need to be involved to work on this issue?

There needs to be a dedicated actor from multiple ministries to create a cross-ministerial team,

including MoWA, Communication, Economy, Agriculture, Labor, and others. These actors

should all be invested in improving the lives of women. The team should also consult with


37

regional, provincial, and local actors, such as shura leaders, mullahs, women’s NGOs, and

international actors to understand their viewpoints, but these members do not need to be on the

core team that takes decisions.

4. Construct the problem—What is the actual problem that affects the larger visible problem?

Based on our initial hypothesis, the problem lies within cultural norms that perpetuate strong

traditional gender roles and prevent women from entering the labor force. The other problem is

the security situation; however, for the purposes of this exercise, we will focus on norms.

5. Deconstruct the problem – What are the appropriate entry points?

We need to first examine the incentives of the opposition groups and stakeholders involved.

Given the strong opposition to women’s rights in certain parts of the country, we know that the

intervention has to be politically supportable and centralized. Using past experience and

evidence, we deconstruct the problem and find an entry point to be introducing Tolo TV signals

to more remote parts of the country, partnering with a television program to add relevant content,

and broadcasting a television program tailored for the Afghan audience to address women’s

issues. The main entry point is creating a partnership with an existing show or program to

deliver appropriate content on women’s issues.86 To develop content, the core team will have to

meet with stakeholders and people across the country to understand how best to design the

show/messages.

6. Take action steps—How do you enter? What steps are needed to make progress?

Using support from international organizations, the GoA can make the initial infrastructure

investments. This can be seen as a sunk cost. The GoA can then meet with relevant stakeholders

for content design and keep local leaders, opposition groups, and the international community

engaged. Using the partnership with a television program and/or channel, the GoA can introduce

signals to more remote parts of the country and develop ways to incentivize viewership. Before

beginning the program, the GoA can survey a random sample of the population and conduct

focus groups to get a baseline sense of perceptions and attitudes surrounding gender roles and

women’s issues.

86 Given that Indian television is popular, instead of reinventing the wheel, we can create a partnership with an Indian network to integrate content in one of their shows to air in Afghanistan.


38

7. Reflect on lessons—How are you capturing key learnings? How will you reflect?

Through this process, data should be collected – both qualitative and quantitative – data on

stakeholders involved, number of televisions, viewership statistics, program ratings, etc. Once

the program airs, the GoA would meet with the relevant stakeholders to make sure things are

progressing smoothly and concerns are addressed. Using village focus groups, the GoA can

quickly assess how the program is doing and determine any quick wins that can further motivate

the team.

8. Communicate quick wins – How will you keep key stakeholders engaged?

Donors and stakeholders often need to know that progress is being made. Given that norms

change very slowly, initial successes and quick wins will have to be based on viewership.

9. Iteration – Is the problem solved?

If viewership remains low or if there is backlash to the program, we will have to reconsider our

initial assessment/hypothesis. We can proceed in 3 ways: (1) Return to 1. Perhaps we do not

have the correct authorization to work on this problem. We need to revisit what we have the

authority to do and how we can get the authorization needed. (2) Return to 4. Perhaps we have

not constructed the problem well. The diagnostics may have misled us to the incorrect

underlying barrier. (3) Return to 5. Perhaps the problem is constructed correctly, but we have to

deconstruct it again to find a better entry point. While cultural norms may be the underlying

barrier, perhaps using a television intervention is not the correct channel.

If the program is successful and viewership is strong, then we should collect metrics on

perceptions and beliefs to see if things have changed since the baseline. If things have not

changed, then we return to 5. Perhaps we have not deconstructed the problem correctly and need

to find a better entry point. If things have changed positively, ensure the show can sustain itself

and keep monitoring and evaluating progress.

Essentially, the successful implementation of this program requires four elements: (1) the

GoA must have a dedicated and diverse team who bring their own expertise; (2) the GoA must

build and maintain relationships with key stakeholders; (3) the GoA must ensure ongoing and

continuous monitoring through feedback loops by collecting qualitative and quantitative data;

and (4) the GoA must reflect after each critical step and be prepared to iterate.


39

To summarize, based on the technical analysis, norms is one of the key underlying barriers

preventing women from entering the labor market.87 Our initial theory of change, in the Afghan

context, is that there is a lack of access to information and exposure to the outside world in a

culturally relatable manner. Addressing norms will require changing behavior and we

recommend the GoA to use a television program as a first step for long-term impact. This

process will be take time and it may require adaptation and iteration.

VII. Concluding Remarks

There have been clear gains in women’s empowerment in Afghanistan since the end of the

Taliban era; however, progress has been highly variable. The Afghan Constitution promises

women equal legal status to men, the right to go to school, access to jobs and hold political

office. In reality, the statistics are quite dismal. Currently, female labor force participation rate

in Afghanistan is around 16%. Limited participation of half the eligible working population in

the economy has deep implications for the country’s wellbeing and for overall growth and

productivity.

We believe that greater female participation in the economy is a major opportunity for the

country’s reconstruction, but it will require addressing obstacles at all levels, including through

shifts in social customs, greater accommodation of women in the market, and inclusive policy

that addresses the key barriers that women face to join the labor force. Security is a critical

barrier that disproportionately affects women - this issue needs to be addressed by the GoA in

tandem with a program that aims to affect FLFP. Prevailing norms is another underlying barrier

that prevents women from entering the labor force, but we believe that a targeted intervention,

such as increasing access to information, can help to lessen this barrier.

In sum, while progress has been made, there is a long way to go for women in Afghanistan.

We know from literature that gender equality matters for development. When women make

progress, all of society makes progress. We hope the GoA can start a long-term dialogue to

empower not only women, but all of Afghanistan.

87 For the purposes of this analysis, we will not be addressing the other underlying barrier, security, since it is beyond our scope.


40

References

Afghanistan. Ministry of Labor, Social Affairs, Martyrs and Disabled. Draft of National Action

Plan for Women's Economic Empowerment in Afghanistan 1395-1399. Ver. 2. Kabul:

2015.

Bahri, J. “Western Gender Policies in Afghanistan: Failing Women and Provoking Men.”

Gender, Technology and Development 18, no. 2 (2014): 163–85.

Beaman, Lori, Esther Duflo, Rohini Pande, and Petia Topalova. “Female Leadership Raises

Aspirations and Educational Attainment for Girls: A Policy Experiment in India.”

Science 335, no. 6068 (February 3, 2012): 582–86.

Beath, Andrew, Fotini Christia, and Ruben Enikolopov. “Empowering Women through

Development Aid: Evidence from a Field Experiment in Afghanistan.” American

Political Science Review 107, no. 03 (August 2013): 540–57.

Besamusca, Janna, Kea Tijdens, Maarten Keune, and Stephanie Steinmetz. “Working Women

Worldwide. Age Effects in Female Labor Force Participation in 117 Countries.” World

Development 74 (October 2015): 123–41.

Cazes, S., and S. Verick. The Labour Markets of Emerging Economies: Has Growth Translated

into More and Better Jobs? Geneva and Basingstoke, UK: ILO and Palgrave Macmillan,

2013. �

“Database of Quotas for Women.” Quota Database. International IDEA, Stockholm University

and Inter-Parliamentary Union, 2015. <http://www.quotaproject.org/>.

Dixon, R.B. Women in Agriculture: Counting the Labor Force in Developing Countries.

Population and Development Review 8 (1982): 539–566.

Duflo, Esther. “Women Empowerment and Economic Development.” Journal of Economic

Literature 50, no. 4 (December 1, 2012): 1051–79.

Field, Erica, and Kate Vyborny. “Women's Mobility in Pakistan.” CERP. Center for Economic

Research in Pakistan, July 2014. <http://cerp.org.pk/womens-mobility/>.

Field, Erica, Seema Jayachandran, and Rohini Pande. “Do Traditional Institutions Constrain

Female Entrepreneurship? A Field Experiment on Business Training in India.” The

American Economic Review 100, no. 2 (May 1, 2010): 125–29.

Fogli, Alessandra, and Laura Veldkamp. “Nature or Nurture? Learning and the Geography of

Female Labor Force Participation.” Econometrica 79, no. 4 (July 1, 2011): 1103–38.


41

Gaddis, Isis, and Stephan Klasen. “Economic Development, Structural Change, and Women’s

Labor Force Participation: A reexamination of the feminization U hypothesis.” Journal of

Population Economics J Popul Econ 27.3 (2013): 639-81.

“Gender.” Afghanistan. USAID, 26 Jan. 2016. <https://www.usaid.gov/afghanistan/gender-

participant-training>.

Ghaisy, Richard, Jiayi Zhou, and Henrik Hallgren. Afghanistan's Private Sector: Status and

Ways Forward. Rep. Stockholm: SIPRI, 2015.

Ghani, Ashraf. “Twitter Feed.” Twitter. 8 Nov. 2014.

Ghani, Ejaz, William R. Kerr, and Stephen D. O’Connell. “Political Reservations and Women’s

Entrepreneurship in India.” Journal of Development Economics 108 (May 2014): 138–53.

Goldin, Claudia. “The U-Shaped Female Labor Force Function in Economic Development and

Economic History.” Working Paper. National Bureau of Economic Research, April 1994.

Haghighat-Sordellini, Elhum. “Determinants of Female Labor Force Participation: A Focus on

Muslim Countries.” International Review of Sociology 19, no. 1 (March 1, 2009): 103–

25.

Hausmann, Ricardo, Dani Rodrik, and Andrés Velasco. “Growth Diagnostics.” The Washington

Consensus Reconsidered (2008): 324-55.

Jensen, Robert. “Do Labor Market Opportunities Affect Young Women’s Work and Family

Decisions? Experimental Evidence from India.” The Quarterly Journal of Economics

127, no. 2 (May 1, 2012): 753–92.

Jensen, Robert, and Emily Oster. “The Power of TV: Cable Television and Women’s Status in

India.” The Quarterly Journal of Economics 124, no. 3 (August 1, 2009): 1057–94.

Lavy, Victor, and Alexander Zablotsky. “Women’s Schooling and Fertility under Low Female

Labor Force Participation: Evidence from Mobility Restrictions in Israel.” Journal of

Public Economics 124 (April 2015): 105–21.

Lehrer, E.l. “The Effects of Religion on the Labor Supply of Married Women.” Social Science

Research 24.3 (1995): 281-301.

Loh, Katherine. “Does the Ghani Administration Mean a Greater Role for Women in the Afghan

Economy?” In Asia. The Asia Foundation, 19 Nov. 2014. <http://asiafoundation.org/in-

asia/2014/11/19/does-the-ghani-administration-mean-a-greater-role-for-women-in-the-

afghan-economy/>.


42

Lundberg, Shelly J., Robert A. Pollak, and Terence J. Wales. “Do Husbands and Wives Pool

Their Resources? Evidence from the United Kingdom Child Benefit.” The Journal of

Human Resources 32.3 (1997): 463.

Mehmood, Bilal, Sama Ahmad, and Muhammad Imran. “What Derives Female Labor Force

Participation in Muslim Countries? A Generalized Method of Moments Inference.”

Pakistan Journal of Commerce and Social Sciences 9.1 (2015): 120-30.

Mohib, Lael. “FLFP in Afghanistan.” Personal interview. 20 Jan. 2016.

New Home, New Life Final Report. USAID, June 2007.

<http://pdf.usaid.gov/pdf_docs/Pdack342.pdf>.

“New Home, New Life Reflects Real Life.” BBC News. BBC

<http://www.bbc.co.uk/mediaaction/where-we-work/asia/afghanistan/education>.

Parto, S.R. and Mihran, R. Understanding Gender in Agricultural Production: An Annotated

Bibliography for the Case of Afghanistan. Afghanistan Public Policy Research

Organization, Kabul, Afghanistan (2010).

Pritchett, Lant, Michael Woolcock, and Matt Andrews. “Capability Traps? The Mechanism of

Persistent Implementation Failure.” Center for Global Development Working Paper 234

(December 2010).

Tavva, Srinivas, Malika Abdelali, Aden Aw-Hassan, Barbara Rischkowsky, Markos Tibbo, and

Javed Rizvi. “Gender Roles in Agriculture: The Case of Afghanistan.” Indian Journal of

Gender Studies 20.1 (2013): 111-34.

Unruh, Jon Darrel., and Rhodri Williams. Land and Post-conflict Peacebuilding. London:

Earthscan, 2013.

Verick, Sher. “Female Labor Force Participation in Developing Countries.” IZA World of Labor

Izawol (2014).

“World Inequality Database on Education.” DME WIDE. UNESCO, 2015.

<http://www.education-inequalities.org/>.

World Bank. Afghanistan: Women's Role in Afghanistan's Future - Taking Stock of

Achievements and Continued Challenges. Rep. no. AC34474. Washington DC: World

Bank, 2013.

World Bank. World Development Report 2012: Gender Equality and Development. Washington,

DC: World Bank, 2012.


43

Appendix

Section A. Figures and Tables Figure 1A. Spectrum of Influence

Source: Authors’ own work Figure 2A. Fishbone Diagram for FLFP

Source: Authors’ own work

Des

ai a

nd L

i A

naly

zing

FL

FP in

Afg

hani

stan

44

Figu

re 3

A: N

EEP

Pol

icy

Map

ping

Incr

ease

wom

en’s

acc

ess

to a

g in

puts

, ext

ensi

on

serv

ices

, and

mar

kets

Impr

ove

the

qual

ity

of g

ende

r st

atis

tics

Pro

mot

e cr

eati

ve

indu

stri

es f

or m

arke

ts

Rem

ove

the

lega

l bar

rier

s to

wom

en’s

eco

nom

ic

part

icip

atio

n

Pro

mot

e fi

nanc

ial

incl

usio

n

•En

surin

g gov

ernm

ent’s

lega

l obli

gatio

ns:

birth

certif

icate

s, e-

tazk

eera

, etc.

•

Rem

oving

regu

lator

y obs

tacle

s •

Upda

ting

socia

l pra

ctice

s: inh

erita

nce l

aw,

dowr

ies, c

ontro

l ove

r HH

expe

nditu

res

•Im

prov

e acc

ess t

o fina

nce

•Fi

nanc

ial lit

erac

y and

small

bu

sines

s man

agem

ent t

raini

ng

•Ed

ucat

ion on

finan

cial li

tera

cy,

small

bus

iness

man

agem

ent,

open

ing a

ccou

nts,

etc.

•In

crea

se w

omen

’s bu

sines

s and

sa

vings

/lend

ing g

roup

s.

Intr

oduc

e bu

sine

ss-

trai

ning

cen

ters

thro

ugh

CD

Cs

•Ra

ise q

uality

of p

rodu

cts

•So

urce

food

supp

lies f

or p

ublic

se

rvice

s fro

m lo

cal s

uppli

ers

•Pr

ovide

exte

nsion

serv

ices f

or

wom

en

•Im

prov

e mar

ketin

g an

d inc

reas

e inf

orm

ation

/impr

ove

weak

value

chain

s •

Impr

ove q

uality

of p

rodu

cts

Prom

otionof

wom

en’swork

Education/Skills

•Do

es se

curit

y and

vio

lence

affe

ct wo

men

’s wi

llingn

ess t

o wo

rk?

Security

Marriage/Fertility

Perceptions

•W

hen

do w

omen

ge

t mar

ried?

•

How

man

y chil

dren

do

they

hav

e?

•Ar

e wo

men

the

prim

ary c

areg

ivers

?

Credit/Banking

Laws

•Ca

n wo

men

acc

ess

cred

it/fina

nce?

•

Wha

t is th

e leg

al sta

tus o

f wom

en?

•Do

es th

is pr

even

t wo

men

from

en

terin

g?

•Do

wom

en w

ant t

o wo

rk?

•Do

they

und

ersta

nd th

eir

retu

rns t

o lab

or?

•Do

fam

ilies s

uppo

rt th

is?

•Ar

e wo

men

edu

cate

d?

•Do

they

hav

e th

e rig

ht

skills

?

Institutional

Setting

•Do

jobs

exis

t?

•Ar

e th

ey re

levan

t for

wom

en

given

their

per

cept

ions?

•

Do w

omen

hav

e ac

cess

to

thes

e job

s?

•Do

jobs

requ

ire

docu

men

tatio

n? A

re

thes

e re

adily

ac

cess

ible?

•

Do b

usine

sses

face

hir

ing la

ws th

at p

reve

nt

from

ent

ering

?

Organize/

Advocate

•Ca

n wo

men

adv

ocat

e fo

r the

mse

lves?

Is th

ere

som

eone

adv

ocat

ing o

n th

eir b

ehalf

?

Econom

ic

Opportunities

•Ar

e th

eir a

venu

es fo

r wo

men

to m

arke

t the

ir wo

rk?

•Is

ther

e inf

orm

ation

of

wom

en’s

work

? •

Is th

is hig

h qu

ality

work

?

FemaleLaborD

iagnostics

Challengestoentering

thelabor

market

SupplyofLabor

Challenges

Dem

andfor

LaborChallenges

Norms

Institutional

Setting

Des

ai a

nd L

i A

naly

zing

FL

FP in

Afg

hani

stan

45

Figu

re 4

A: M

ap o

f Afg

hani

stan

So

urce

: http

://w

ww

.lib.

utex

as.e

du/m

aps/

afgh

anis

tan.

htm

l


46

Figure 5A. FLFP rates in Afghanistan by Province

Source: Survey of Afghan People 2014 Figure 6A. Number of Violence Incidences in Afghanistan by Province

Source: Incidence of Violence Data from Professor Michael Callen


47

Figure 7A. Stakeholder Analysis Green = Our side Red = Their Side Blue = Uncertain & Mixed

Source: Authors’ own work Figure 8A. PDIA Diagram

Source: Authors’ own work

Des

ai a

nd L

i A

naly

zing

FLF

P in

Afg

hani

stan

48

Tabl

e 1A

: Pro

gram

s to

Prom

ote

Wom

en’s

Rol

e in

the

Econ

omy

in A

fgha

nist

an

Prog

ram

s Sp

onso

rs/P

artn

ers

Stat

us

Prog

ram

Goa

ls W

omen

’s L

eade

rshi

p D

evel

opm

ent

(WLD

) Pro

gram

U

SAID

- Pr

omot

e O

ngoi

ng

Pers

onal

dev

elop

men

t and

lead

ersh

ip m

ento

ring

prog

ram

for 2

5,00

0 ed

ucat

ed g

irls a

ged

betw

een

18 a

nd 3

0.

Wom

en in

Gov

ernm

ent P

rogr

am

USA

ID -

Prom

ote

Ong

oing

Pr

ovid

e in

tern

ship

s with

var

ious

gov

ernm

enta

l org

aniz

atio

ns, a

t nat

iona

l and

loca

l lev

els,

with

the

goal

of

obta

inin

g pe

rman

ent e

mpl

oym

ent p

ositi

ons i

n th

e A

fgha

n ci

vil s

ervi

ce fo

r 3,0

00 u

nive

rsity

gra

duat

e w

omen

ag

ed b

etw

een

21 a

nd 3

0.

Wom

en’s

Rig

hts G

roup

s and

Co

aliti

ons (

Mus

harik

at) P

rogr

am

USA

ID -

Prom

ote

Ong

oing

W

omen

’s ri

ghts

focu

sed

orga

niza

tions

in e

ach

of th

e pr

ovin

ces w

ill p

artic

ipat

e in

gen

der-

rela

ted

issue

-ba

sed

coal

ition

s. A

min

imum

of 5

,000

act

ivist

s will

par

ticip

ate

in th

ese

coal

ition

s to

iden

tify

maj

or

prio

ritie

s to

adva

nce

wom

en’s

righ

ts th

roug

h an

d be

yond

the

trans

form

atio

n de

cade

. A

fgha

n W

omen

’s L

eade

rshi

p in

the

Econ

omy

(AW

LE) P

rogr

am

USA

ID -

Prom

ote

Ong

oing

H

elp

educ

ated

wom

en se

cure

em

ploy

men

t with

adv

ance

men

t pot

entia

l and

impr

ove

the

viab

ility

and

in

com

e gr

owth

of w

omen

-ow

ned

busi

ness

es.

Scho

lars

hip

Fund

for A

fgha

n W

omen

U

SAID

- Pr

omot

e O

ngoi

ng

Supp

ort w

omen

to u

nder

take

und

ergr

adua

te a

nd p

ost-g

radu

ate

stud

y at

uni

vers

ities

in A

fgha

nist

an, t

he

regi

on, a

nd in

tern

atio

nally

. A

dole

scen

t Girl

s Ini

tiativ

e -

Afg

hani

stan

Wor

ld B

ank

Aug

201

1 –

Dec

201

4 En

hanc

e th

e jo

b sk

ills o

f you

ng w

omen

and

faci

litat

e th

eir a

cces

s to

wag

e em

ploy

men

t

Nat

iona

l Sol

idar

ity P

rogr

am 3

W

orld

Ban

k Ju

ne 2

010

– Se

pt 2

015

Cont

inue

to b

uild

and

stre

ngth

en th

e ca

paci

ty o

f Com

mun

ity D

evel

opm

ent C

ounc

ils

Non

-For

mal

App

roac

h to

Tra

inin

g Ed

ucat

ion

and

Jobs

in A

fgha

nista

n Pr

ojec

t

Wor

ld B

ank

Ong

oing

In

crea

se th

e po

tent

ial f

or e

mpl

oym

ent a

nd h

ighe

r ear

ning

s of t

arge

ted

youn

g A

fgha

n w

omen

and

men

in

rura

l and

sem

i-urb

an a

reas

thro

ugh

non-

form

al sk

ills t

rain

ing

AF

Rur

al E

nter

prise

Dev

elop

men

t Pr

ogra

m

Wor

ld B

ank

Ong

oing

In

crea

sed

inco

me

and

susta

inab

le e

mpl

oym

ent o

ppor

tuni

ties f

or m

en a

nd w

omen

thro

ugh

supp

orte

d ru

ral

ente

rpris

es.

The

MoW

A In

itiat

ive

to S

uppo

rt Po

licy

and

Adv

ocac

y (M

ISPA

) U

SAID

Th

e A

sia

Foun

datio

n

Jan

2006

–

Jan

2011

St

reng

then

the

Min

istry

of W

omen

’s’ A

ffairs

’ (M

oWA

) cap

acity

to se

rve

as a

vis

ible

and

effe

ctiv

e ag

ent f

or

polic

y an

d ad

voca

cy.

Am

bass

ador

’s S

mal

l Gra

nts P

rogr

am

to S

uppo

rt G

ende

r Equ

ality

(ASG

P)

USA

ID

Crea

tive

Ass

ocia

tes

Inte

rnat

iona

l, In

c.

July

200

9 –

Nov

201

1 Pr

ovid

e in

stitu

tiona

l cap

acity

dev

elop

men

t and

adv

ocac

y su

ppor

t to

wom

en-f

ocus

ed A

fgha

n ci

vil-s

ocie

ty

orga

niza

tions

(CSO

s) v

ia g

rant

s in

orde

r to

impr

ove

gend

er e

qual

ity a

nd to

spec

ifica

lly h

elp

wom

en se

cure

op

portu

nitie

s and

adv

ocat

e fo

r the

mse

lves

. Fle

xibl

e gr

ants

add

ress

em

ergi

ng lo

cal A

fgha

n w

omen

’s g

roup

s’

need

s, w

hile

em

pow

erin

g th

em to

exp

and

thei

r act

iviti

es.

Emer

genc

y Ed

ucat

ion

Reh

abili

tatio

n &

Dev

elop

men

t Pro

ject

W

orld

Ban

k Ju

ne 2

002

– Ju

ly 2

006

Supp

ort t

he G

over

nmen

t's e

mer

genc

y ed

ucat

ion

prog

ram

, by

spec

ifica

lly h

elpi

ng in

incr

easi

ng a

cces

s to

educ

atio

n in

bot

h th

e fo

rmal

, and

non

-for

mal

edu

catio

n sy

stem

s for

und

er-s

erve

d gr

oups

, in

parti

cula

r w

omen

and

girl

s Re

info

rcin

g th

e ab

ility

of A

fgha

n w

omen

to im

prov

e th

eir l

ivin

g co

nditi

ons

EU C

omm

issio

n H

ealth

Net

TPO

Dec

201

0 –

Dec

201

3 En

cour

age

wom

en to

take

up

activ

ities

supp

ortin

g th

eir l

ivel

ihoo

ds. E

xplo

re d

efin

ition

s of f

amily

vio

lenc

e an

d ha

rmfu

l pra

ctic

es a

nd id

entif

y cu

ltura

lly a

ppro

pria

te so

lutio

ns. B

uild

cap

acity

of l

ocal

org

aniz

atio

ns a

nd

rele

vant

aut

horit

ies a

nd e

nabl

e th

em to

add

ress

thes

e pr

actic

es a

nd p

rote

ct v

ictim

s.

Rein

forc

e co

llabo

ratio

n w

ith th

e M

inis

try o

f Wom

en A

ffairs

. G

ende

r Equ

ality

Pro

ject

II (G

EP II

) U

ND

P Ja

n 20

13 –

D

ec 2

015

Supp

ort a

roun

d 30

0 w

omen

to st

art n

ew b

usin

esse

s eac

h ye

ar a

nd to

reac

h a

tota

l of 9

00 w

omen

by

2015

. Th

e pr

ojec

t will

als

o es

tabl

ish

4 w

omen

’s p

rodu

ctio

n ce

ntre

s; a

rrang

e ex

posu

re v

isits

for 2

50 w

omen

en

trepr

eneu

rs; i

ntro

duce

six

new

cle

an te

chno

logi

es a

nd su

ppor

t 30

func

tiona

l wom

en’s

co-

oper

ativ

es in

ta

rget

ed a

reas

.

Des

ai a

nd L

i A

naly

zing

FLF

P in

Afg

hani

stan

49

Tabl

e 2A

: Pro

gram

s to

Prom

ote

Wom

en’s

Rol

e in

Oth

er E

cono

mie

s C

ount

ry

Und

erly

ing

Barr

iers

In

terv

entio

n O

utco

mes

In

dia

Educ

atio

n/Sk

ills

Poor

self-

empl

oyed

wom

en w

ere

train

ed in

bas

ic

finan

cial

lite

racy

and

bus

ines

s ski

lls, a

nd

enco

urag

ed to

iden

tify

conc

rete

fina

ncia

l goa

ls.

(Fie

ld, J

ayac

hand

ran,

and

Pan

de; 2

010)

Am

ong

Hin

du w

omen

, tra

inin

g in

crea

sed

borr

owin

g an

d bu

sine

ss in

com

e fo

r tho

se fa

cing

mor

e re

stric

tions

. H

owev

er, M

uslim

wom

en fa

iled

to b

enef

it fr

om th

e tra

inin

g pr

ogra

m.

Indi

a N

orm

s (F

emal

e le

ader

ship

an

d so

cial

stat

us)

Stat

e-le

vel i

mpl

emen

tatio

n of

pol

itica

l re

serv

atio

ns fo

r wom

en.

(Gha

ni, K

err,

and

O’C

onne

ll; 2

014)

Polit

ical

rese

rvat

ions

insp

ired

mor

e w

omen

to o

pen

esta

blis

hmen

ts, a

nd th

ey d

id so

at a

smal

l est

ablis

hmen

t sc

ale

in in

dust

ries w

here

they

had

exp

erie

nce

and/

or th

e su

ppor

t net

wor

ks o

f oth

er w

omen

. In

dia

N

orm

s (A

spira

tions

/role

m

odel

)

A 1

993

law

rese

rved

lead

ersh

ip p

ositi

ons f

or

wom

en in

rand

omly

sele

cted

vill

age

coun

cils.

(B

eam

an, D

uflo

, Pan

de, a

nd T

opal

ova;

201

2)

Fem

ale

lead

ersh

ip in

fluen

ces a

dole

scen

t girl

s’ c

aree

r as

pira

tions

and

edu

catio

nal a

ttain

men

t.

Indi

a N

orm

s (A

ttitu

des a

nd

beha

vior

s)

Intro

duct

ion

of c

able

tele

visi

on o

n w

omen

’s st

atus

in

rura

l Ind

ia.

(Jen

sen

and

Ost

er; 2

009)

Sign

ifica

nt d

ecre

ases

in th

e re

porte

d ac

cept

abili

ty o

f do

mes

tic v

iole

nce

tow

ard

wom

en a

nd so

n pr

efer

ence

, as

wel

l as i

ncre

ases

in w

omen

’s a

uton

omy

and

decr

ease

s in

fert

ility

. In

dia

Job

oppo

rtuni

ties

3 ye

ars o

f rec

ruiti

ng se

rvic

es to

hel

p yo

ung

wom

en in

rura

l vill

ages

get

jobs

in th

e bu

sine

ss

proc

ess o

utso

urci

ng in

dust

ry (a

new

indu

stry

). (J

ense

n; 2

012)

You

ng w

omen

in tr

eatm

ent v

illag

es w

ere

sign

ifica

ntly

le

ss li

kely

to g

et m

arri

ed o

r hav

e ch

ildre

n du

ring

this

pe

riod,

cho

osin

g in

stea

d to

ent

er th

e la

bor m

arke

t or

obta

in m

ore

scho

olin

g or

pos

t sch

ool t

rain

ing.

B

angl

ades

h Ed

ucat

ion/

Skill

s;

Nor

ms (

Early

m

arria

ge)

Peer

-led

sess

ions

in sa

fe sp

aces

(spa

ces w

here

gi

rls c

an m

eet o

n a

regu

lar b

asis

) to

prov

ide

train

ing,

fina

ncia

l lite

racy

, and

coo

king

oil

ince

ntiv

es to

del

ay m

arria

ge.

-

Ong

oing

Saud

i A

rabi

a Ed

ucat

ion/

Skill

s;

Job

oppo

rtuni

ties

Prov

ide

firm

s fin

anci

al su

ppor

t to

hire

wom

en a

nd

train

wom

en fo

r fac

torie

s. -

Ong

oing

Saud

i A

rabi

a Se

curit

y

Subs

idiz

ing

trans

porta

tion

for w

orki

ng w

omen

to

help

them

reta

in jo

bs a

nd e

ncou

rage

une

mpl

oyed

w

omen

to se

ek e

mpl

oym

ent.

-

Ong

oing

Paki

stan

Se

curit

y;

Nor

ms (

Mob

ility

) In

trodu

cing

wom

en’s

-onl

y tra

nspo

rt an

d tra

nspo

rt vo

uche

rs.

-

Ong

oing


50

Table 3A: Linear Correlation to FLFP

Feather Pearson Correlation Rank

Education University level education 0.32 1

Years of education 0.28 2 No education -0.22 3 Geographic

Northwest region 0.11 15 Southwest region -0.11 18 Language

Speak Dari 0.16 5 District-level share of men who speak Dari 0.17 4 District-level share of men who speak English 0.11 17 District-level share of men who speak Pashto -0.12 12 Ethnic Group

Pashtun -0.15 6 Security

Province-level number of violence incidences -0.13 8 District-level share of men who report there were accidents caused by mines and explosive devices -0.13 9 District-level share of men who report there were physical injuries as a result of violence in the household -0.12 11 Information

Received information from community shuras 0.14 7 Watch TV 0.12 13 Use personal mobile phone (not jointly by the family) 0.12 14 Men's and Women's Attitudes towards Gender

Women should be allowed to work outside home 0.11 19

District-level share of men who think there unemployment is a problem for women 0.12 10

District-level share of men agree that women can work in co-ed schools 0.11 16 District-level share of men agree that women can work in a private company outside the home (factory, shop, business) with female employees only 0.10 20

Note: Pearson product-moment correlation coefficient method


51

Table 4A: Logistic Regression

Note: Weights applied. Robust standard errors in parentheses, *** p<0.01, ** p<0.05, * p<0.1.

(1) (2) (3) (4) Independent Variables Is the woman in labor force? No=0, Yes=1 Characteristics Years of education 0.0244*** 0.0266*** 0.0247*** 0.0247***

(0.00177) (0.00182) (0.00181) (0.00176) Married 0.0523*** 0.0448** 0.0509*** 0.0478***

(0.0183) (0.0181) (0.0181) (0.0182) District average years of men’s education

-0.00933*** (0.00207)

-0.00383 (0.00311)

-0.00782*** (0.0022)

Security and Violence Province-level number of violentincidences

-0.000031***(0.0000)

-0.0000637*(0.0000)

-0.0000285***(0.0000)

District-level share of men who think there were accidents caused by mines and explosive devices

-0.0419***(0.0161)

-0.0203(0.0261)

-0.0371**(0.0177)

Information Source Received information from community shuras

0.0427***(0.0143)

0.0134(0.0161)

0.0169(0.015)

0.0407***(0.0142)

Watch TV 0.0200* 0.0164 0.0147 0.018

(0.0117) (0.0129) (0.0122) (0.0118) Women’s Perception Women should be allowed to work outside home

0.0385***(0.0116)

0.0447***(0.0126)

0.0474***(0.012)

0.0386***(0.0115)

Politics and religion should mix at the local level

0.0406***(0.0119)

0.0192(0.012)

0.0281**(0.0121)

0.0412***(0.0118)

District–LevelMen’s Perception Men – Women should have equal opportunities like men in education

0.0697** (0.0311)

0.00583 (0.0385)

0.0427 (0.0313)

Men – it is acceptable for women to work in a co-ed schools

0.0278 (0.0281)

0.0519 (0.0349)

0.0655** (0.03)

District Fixed Effects YES Province Fixed Effects YES Ethnic Group Fixed Effects YES Constant 0.0363 0.205 0.261*** 0.378*

(0.03) (0.178) (0.0574) (0.195) Observations 4,739 4,739 4,739 4,739 R-squared 0.123 0.279 0.18 0.139


52

Table 5A: Capability Trap Definitions Term Definition Premature load bearing This occurs when unrealistic expectations

about the level and rate of improvement of capability lead to stresses and demands on systems that cause capability to weaken (if not collapse).

Isomorphic mimicry When you build institutions and processes in weak states that look like those found in functional states.

Source: Pritchett, Lant, Michael Woolcock, and Matt Andrews. “Capability Traps? The Mechanism of Persistent Implementation Failure.” Center for Global Development Working Paper 234 (December 2010). Table 6A: Policy Options in the NEEP Policy Options Underlying assumptions Removing legal barriers to women’s economic participation

Assumes laws, social practices, and regulatory obstacles is the constraint

Promoting financial inclusion Assumes access to credit and financial literacy is the constraint

Introducing business training centers Assumes education/skills and women’s collective organizing in a safe space is the constraint

Promoting creative industries for domestic/international markets

Assumes lack of marketing and lack of demand is the constraint

Increasing women’s access to agricultural inputs, extension services, and markets

Assumes education/skills and lack of demand is the constraint

Improving quality of gender statistics Assumes lack of proper data is the constraint in assessing FLFP issues

Source: Draft 2 of the NEEP

Des

ai a

nd L

i A

naly

zing

FLF

P in

Afg

hani

stan

53

Tabl

e 7A

: Ana

lysis

of P

olic

y O

ptio

ns


54

Section B. Data Sources

Dataset Year(s) Source Description

The National Risk and Vulnerability Assessment (NRVA)

2003; 2005; 2007; 2011

Central Statistics Organization of the Government of the Islamic Republic of Afghanistan

This survey is the only nation-wide survey available in Afghanistan. It captures problems that the population face throughout the year, especially in terms of labor, poverty and food security. In the 2011-2012 survey, it includes 20,828 households.

Women, Business and the Law 2016 World Bank

Data on 173 economies. It provides quantitative measures of laws and regulations that affect women’s economic opportunities in seven areas: accessing institutions, using property, getting a job, providing incentives to work, going to court, building credit and protecting women from violence.

Survey of the Afghan People 2004 – 2015 The Asia

Foundation

The longest-running public opinion poll in the country. 9,271 Afghan citizens from all 34 provinces were interviewed on issues central to the country's development. The survey includes questions on women's right.

Incidence of Violence Data 2008 – 2015 Professor Mike

Callen

It recorded the time and type of violent incidents in Afghanistan by GEO codes from 2008 to 2014.

Global Peace Index 2015 Institute for Economics and Peace

It measures the state of peace in 161 countries according to 23 indicators that gauge the absence of violence or the fear of violence.

World Development Indicators 2015 World Bank

Data on labor force participation rates, literacy rates, years of education, and GDP per capita for Afghanistan and other countries listed in the paper.


55

Section C. Diagnostics Framework We use the diagnostic tree approach and evaluate for evidence of the following:

1) Whether such a problem exists, and 2) Whether that problem appears to be a true barrier for women to enter the labor market.

Using the principles adapted from Hausmann’s Growth Diagnostics, we use the following pieces of evidence to determine the key underlying barriers women in Afghanistan face: 1) Direct evidence: Observations, surveys or impact evaluations that give descriptive evidence

of the scale of the problem. We mainly used Survey of Afghan People and NRVA, combined with others databases listed in Appendix Section B to find direct evidence regarding factors such as education, security, etc. We are looking at direct correlations, for example, if we argue that security is the key underlying barrier in the Afghan context, we would see a correlation between highly secure areas and female labor force participation.

2) Benchmarking: Although many elements might be imperfect in absolute terms, we would expect an underlying factor to be a problem relative to comparable countries (i.e. Muslim, South Asian, history of colonists, etc.). In Afghanistan, we know that women don't have much autonomy and have limited mobility for various reasons. It can be argued that Pakistan has similar constraints for women as Afghanistan. If in Pakistan, a female-only transport option shows increased mobility for women, then this would indicate that this transport/transport options is a real problem in Afghanistan. Or, if security is a key underlying barrier in Pakistan, then we should also find it to be a barrier in Afghanistan.

3) Changes leads to changes: If something were indeed an underlying barrier, we would expect to see changes in female labor force participation if it were relaxed. For example, if education were an underlying barrier with respect to FLFP, then a successful training or education program would be accompanied by an increase in the FLFP rate.

4) Bypassing the constraint: Is there evidence that women are attempting to bypass certain constraints that hold them back from better outcomes? If the variable of interest were indeed the key underlying constraint, women would not have entered the labor market without the barrier being addressed. For example, if education were the key underlying barrier, women with little or no education would find it difficult to enter the labor force. If we continue to see women with no education joining the labor force, education may not be the key barrier, although having more education may help with increasing the tendency to work outside home or securing a job.

5) “Camels and Hippos”: In a world where a particular constraint exists, what types of institutions or practices would we expect to see? Does that match reality? In the Afghan context, if education were the particular constraint, we would have seen institutions or agencies to provide education at different levels because there was a demand. If security were the particular constraint, we would have seen practices to ensure safety, for example, by restricting women’s mobility.


56

Section D. Predictive Model

Using the 2015 Survey of Afghan People, we use a binary dependent variable (Yi) model to

determine the factors (X1…Xn) influencing women’s decision to enter the labor force. Building

on the diagnostic, the predictive model validates much of the empirical findings. We first used

the Pearson Correlation method to select top 100 variables that are most correlated with female

labor force – the variable of interest – out of 681 variables. Table 3A provides a list of indicators

and the linear correlation to FLFP.

The most correlated indicators are put into a cross-validated unbalanced lasso-logistic

regression model that allows us to understand the relative importance of each indicator.

Continuous variables are standardized before putting into the model. This predictive model

gives us a list of relative importance coefficients for each variable, allowing for better

interpretation of the model and identification of those covariates most strongly associated with

the outcome.

Based on the results of the model, we select relevant indicators that are politically and

administratively feasible to influence in a real world setting and use a OLS model to find the

grouping of variables with the best R-squared.

The general equation used is:

!! = !! + !!!! +⋯+ !!!! + !! where, Yi, denotes female labor force participation and X1…Xn represents various determinant

factors leading to female participation in the labor force.

To test the robustness of the variables of interest: education, norms (e.g., perceptions on

whether should be allowed to work outside home), and sources of obtaining information (e.g.,

watching TV, getting information from community shura, personal mobile phone), a series of

OLS regressions are conducted. For each variables of interest, we first regress the dependent

variable of labor force participation (Y!) on the variable of interest only (!!). Next, in Table 8A

to 12A, we gradually add on more independent variables (!!…!!) to see if the size and the

significance of the coefficient on !! change. This is to proxy whether there is a causal link

between the variable of interest and the dependent variable of female labor force. If the variable

of interest is the underlying factor influencing female’s decision to enter the labor force, its

significance will remain despite the inclusion of other confounding factors. This approach

validates the empirical findings from the diagnostic and OLS predictive model (see Table 4A).

Des

ai a

nd L

i A

naly

zing

FLF

P in

Afg

hani

stan

57

Tabl

e 8A

: OLS

Reg

ress

ion

on F

emal

e La

bor

Forc

e Pa

rtic

ipat

ion:

Var

iabl

e of

Inte

rest

Nor

ms

VA

RIA

BLE

S (1

) (2

) (3

) (4

) (5

) (6

) (7

) (8

) (9

) (1

0)

Wom

en s

houl

d be

allo

wed

to w

ork

outs

ide

hom

e 0.

0778

***

(0.0

117)

0.

0631

***

(0.0

115)

0.

0777

***

(0.0

126)

0.

0760

***

(0.0

129)

0.

0901

***

(0.0

117)

0.

0527

***

(0.0

114)

0.

0511

***

(0.0

114)

0.

0488

***

(0.0

114)

0.

0394

***

(0.0

116)

0.

0390

***

(0.0

116)

Fi

xed

effe

cts

Et

hnic

ity

Prov

ince

D

istri

ct

Prov

ince

-leve

l num

ber o

f vio

lent

in

cide

nces

-0

.000

046*

**

(0.0

00)

-0.0

0004

6***

(0

.000

) -0

.000

042*

**

(0.0

00)

-0.0

0004

1***

(0

.000

) -0

.000

031*

**

(0.0

00)

-0.0

0003

1***

(0

.000

) Y

ears

of e

duca

tion

0.

0234

***

0.02

22**

* 0.

0236

***

0.02

41**

* 0.

0243

***

(0.0

0169

) (0

.001

72)

(0.0

0178

) (0

.001

79)

(-0.

0018

) R

ecei

ved

info

rmat

ion

from

co

mm

unity

shur

as

0.04

36**

* (0

.014

4)

0.04

34**

*

(0.0

143)

0.

0433

***

(0.0

143)

0.

0433

***

(0.0

143)

U

se p

erso

nal m

obile

pho

ne (n

ot

join

tly b

y th

e fa

mily

)

0.

0320

**

(0.0

133)

0.

0258

* (0

.013

3)

0.02

66**

(0

.013

3)

0.02

65**

(0

.013

3)

Mar

ried

0.

0506

***

0.04

91**

* 0.

0440

**

(0.0

184)

(0

.018

4)

(0.0

191)

Po

litic

s and

relig

ion

shou

ld m

ix a

t the

lo

cal l

evel

0.03

37**

* (0

.011

9)

0.03

90**

* (0

.011

9)

0.03

91**

* (0

.011

9)

Dis

trict

-leve

l sha

re o

f men

who

thin

k w

omen

shou

ld h

ave

equa

l op

portu

nitie

s lik

e m

en in

edu

catio

n

0.

0651

**

(0.0

311)

0.

0658

**

(0.0

311)

Dis

trict

-leve

l sha

re o

f men

who

thin

k th

ere

wer

e ac

cide

nts c

ause

d by

min

es

and

expl

osiv

e de

vice

s

-0

.045

4***

(0

.016

) -0

.045

1***

(0

.016

)

Dis

trict

ave

rage

yea

rs o

f men

’s

educ

atio

n

-0

.009

21**

* (0

.002

03)

-0.0

0931

***

(0.0

0204

) D

istri

ct-le

vel s

hare

of m

en w

ho th

ink

it is

acc

epta

ble

for w

omen

to w

ork

in

a co

-ed

scho

ols

0.03

08

(0.0

276)

0.

0291

(0

.028

)

Age

0.00

0462

(0

.000

5)

Hou

seho

ld si

ze

-0

.000

0232

(0

.001

48)

Con

stan

t 0.

102*

**

0.52

9**

0.03

26

0.28

3 0.

154*

**

0.12

1***

0.

102*

**

0.04

50**

0.

0458

0.

0356

(0.0

092)

(0

.244

) (0

.032

) (0

.202

) (.0

103)

(.0

102)

(.0

107)

(.0

199)

(.0

298)

(0

.033

)

Obs

erva

tions

4,

739

4,73

9 4,

739

4,73

9 4,

739

4,73

9 4,

739

4,73

9 4,

739

4,73

9 R

-squ

ared

0.

009

0.03

7 0.

103

0.20

2 0.

03

0.10

8 0.

112

0.11

6 0.

124

0.12

4 N

ote:

Wei

ghts

app

lied.

Rob

ust s

tand

ard

erro

rs in

par

enth

eses

, **

* p<

0.01

, **

p<0.

05, *

p<0

.1.

Des

ai a

nd L

i A

naly

zing

FLF

P in

Afg

hani

stan

58

Tabl

e 9A

: OLS

Reg

ress

ion

on F

emal

e La

bor

Forc

e Pa

rtic

ipat

ion:

Var

iabl

e of

Inte

rest

- Ed

ucat

ion

VA

RIA

BLE

S (1

) (2

) (3

) (4

) (5

) (6

) (7

) (8

) (9

) (1

0)

Yea

rs o

f edu

catio

n 0.

0242

***

0.02

40**

* 0.

0242

***

0.02

62**

* 0.

0242

***

0.02

34**

* 0.

0222

***

0.02

36**

* 0.

0241

***

0.02

43**

*

(0

.001

67)

(0.0

0168

) (0

.001

73)

(0.0

0171

) (0

.001

67)

(0.0

0169

) (0

.001

72)

(0.0

0178

) (0

.001

79)

(-0.

0018

) Fi

xed

effe

cts

Et

hnic

ity

Prov

ince

D

istri

ct

Prov

ince

-leve

l num

ber o

f vio

lent

in

cide

nces

-0

.000

043*

**

(0.0

000)

-0

.000

045*

**

(0.0

000)

-0

.000

042*

**

(0.0

000)

-0

.000

041*

**

(0.0

000)

-0

.000

041*

**

(0.0

000)

-0

.000

031*

**

(0.0

000)

W

omen

shou

ld b

e al

low

ed to

wor

k ou

tsid

e ho

me

0.

0527

***

(0.0

114)

0.

0511

***

(0.0

114)

0.

0488

***

(0.0

114)

0.

0394

***

(0.0

116)

0.

0390

***

(0.0

116)

R

ecei

ved

info

rmat

ion

from

co

mm

unity

shur

as

0.04

36**

* (0

.014

4)

0.04

34**

* (0

.014

3)

0.04

33**

* (0

.014

3)

0.04

33**

* (0

.014

3)

Use

per

sona

l mob

ile p

hone

(not

jo

intly

by

the

fam

ily)

0.03

20**

(0

.013

3)

0.02

58*

(0.0

133)

0.

0266

**

(0.0

133)

0.

0265

**

(0.0

133)

Mar

ried

0.

0506

***

(0.0

184)

0.

0491

***

(0.0

184)

0.

0440

**

(0.0

191)

Polit

ics a

nd re

ligio

n sh

ould

mix

at

the

loca

l lev

el

0.

0337

***

(0.0

119)

0.

0390

***

(0.0

119)

0.

0391

***

(0.0

119)

D

istri

ct-le

vel s

hare

of m

en w

ho

thin

k w

omen

shou

ld h

ave

equa

l op

portu

nitie

s lik

e m

en in

edu

catio

n

0.

0651

**

(0.0

311)

0.

0658

**

(0.0

312)

Dis

trict

-leve

l sha

re o

f men

who

th

ink

ther

e w

ere

acci

dent

s cau

sed

by

min

es a

nd e

xplo

sive

dev

ices

-0

.045

4***

(0

.016

) -0

.045

1***

(0

.016

)

Dis

trict

ave

rage

yea

rs o

f men

’s

educ

atio

n

-0

.009

21**

* (0

.002

03)

-0.0

0931

***

(0.0

0204

) D

istri

ct-le

vel s

hare

of m

en w

ho

thin

k it

is a

ccep

tabl

e fo

r wom

en to

w

ork

in a

co-

ed sc

hool

s

0.

0308

(0

.027

6)

0.02

91

(0.0

28)

Age

0.00

0462

(0

.000

493)

H

ouse

hold

size

-0.0

0002

32

(0.0

0148

) C

onst

ant

0.09

87**

* 0.

474*

* 0.

0412

0.

281*

0.

155*

**

0.12

1***

0.

102*

**

0.04

50**

0.

0458

0.

0356

(0

.005

72)

(0.1

92)

(0.0

275)

(0

.171

) (0

.008

49)

(0.0

102)

(0

.010

7)

(0.0

199)

(0

.029

8)

(0.0

33)

Obs

erva

tions

47

39

4739

47

39

4739

47

39

4739

47

39

4739

47

39

4739

R

-squ

ared

0.

086

0.11

3 0.

172

0.27

4 0.

105

0.10

8 0.

112

0.11

6 0.

124

0.12

4 N

ote:

Wei

ghts

app

lied.

Rob

ust s

tand

ard

erro

rs in

par

enth

eses

, **

* p<

0.01

, **

p<0.

05, *

p<0

.1.

Des

ai a

nd L

i A

naly

zing

FLF

P in

Afg

hani

stan

59

Tabl

e 10

A: O

LS R

egre

ssio

n on

Fem

ale

Labo

r Fo

rce

Part

icip

atio

n: V

aria

ble

of In

tere

st -

Wat

ch T

V

VA

RIA

BLE

S (1

) (2

) (3

) (4

) (5

) (6

) (7

) (8

) (9

) (1

0)

(11)

W

atch

TV

0.

0917

***

0.07

49**

* 0.

0605

***

0.05

53**

* 0.

0831

***

0.02

68**

0.

0201

* 0.

0211

* 0.

0204

* 0.

0200

* 0.

0199

*

(0.0

114)

(0

.011

8)

(0.0

125)

(0

.013

1)

(0.0

115)

(0

.011

1)

(0.0

112)

(0

.011

2)

(0.0

112)

(0

.011

7)

(0.0

117)

Fi

xed

effe

cts

Et

hnic

ity

Prov

ince

D

istri

ct

Pr

ovin

ce-le

vel n

umbe

r of

viol

ent i

ncid

ence

s

-0

.000

039*

**

(0.0

000)

-0

.000

042*

**

(0.0

000)

-0

.000

044*

**

(0.0

000)

-0

.000

04**

* (0

.000

0)

-0.0

0004

1***

(0

.000

0)

-0.0

0003

1***

(0

.000

0)

-0.0

0003

1***

(0

.000

0)

Yea

rs o

f edu

catio

n

0.02

34**

* 0.

0229

***

0.02

25**

* 0.

0238

***

0.02

44**

* 0.

0246

***

(0.0

017)

(.0

0172

) (.0

0172

) (.0

0177

) (.0

0177

) (.0

0179

) W

omen

shou

ld b

e al

low

ed to

w

ork

outs

ide

hom

e

0.

0493

***

(0.0

115)

0.

0466

***

(0.0

114)

0.

0474

***

(0.0

114)

0.

0385

***

(0.0

116)

0.

0382

***

(0.0

116)

Po

litic

s and

relig

ion

shou

ld

mix

at t

he lo

cal l

evel

0.03

50**

* (0

.011

9)

0.03

54**

* (0

.011

9)

0.04

06**

* (0

.011

9)

0.04

07**

* (0

.011

9)

Rec

eive

d in

form

atio

n fr

om

com

mun

ity sh

uras

0.04

29**

* (0

.014

3)

0.04

27**

* (0

.014

3)

0.04

27**

* (0

.014

3)

0.04

27**

* (0

.014

3)

Mar

ried

0.05

36**

* 0.

0523

***

0.04

73**

(0.0

183)

(0

.018

3)

(0.0

191)

D

istri

ct-le

vel s

hare

of m

en w

ho

thin

k th

ere

wer

e ac

cide

nts

caus

ed b

y m

ines

and

exp

losi

ve

devi

ces

-0

.041

9***

(0

.016

1)

-0.0

418*

* (0

.016

5)

Dis

trict

-leve

l sha

re o

f men

who

th

ink

wom

en sh

ould

hav

e eq

ual o

ppor

tuni

ties l

ike

men

in

educ

atio

n

0.

0697

**

(0.0

311)

0.

0704

**

(0.0

313)

Dis

trict

ave

rage

yea

rs o

f men

’s

educ

atio

n

-0.0

0933

***

(0.0

0207

) -0

.009

43**

* (0

.002

08)

Dis

trict

-leve

l sha

re o

f men

who

th

ink

it is

acc

epta

ble

for

wom

en to

wor

k in

a c

o-ed

sc

hool

s

0.

0278

(0

.028

1)

0.02

63

(0.0

284)

Age

0.

0004

57

(0

.000

493)

H

ouse

hold

size

0.

0000

26

(0

.001

48)

Con

stan

t 0.

109*

**

0.55

2**

0.05

70*

0.29

6 0.

166*

**

0.14

1***

0.

112*

**

0.08

56**

* 0.

0384

* 0.

0363

0.

0259

(0.0

072)

(0

.223

) (0

.031

3)

(0.2

01)

(0.0

101)

(0

.010

1)

(0.0

112)

(0

.012

3)

(0.0

206)

(0

.03)

(0

.033

4)

Obs

erva

tions

4,

739

4,73

9 4,

739

4,73

9 4,

739

4,73

9 4,

739

4,73

9 4,

739

4,73

9 4,

739

R-s

quar

ed

0.01

5 0.

041

0.10

1 0.

199

0.03

1 0.

106

0.10

9 0.

113

0.11

6 0.

123

0.12

4 N

ote:

Wei

ghts

app

lied.

Rob

ust s

tand

ard

erro

rs in

par

enth

eses

, **

* p<

0.01

, **

p<0.

05, *

p<0

.1.

Des

ai a

nd L

i A

naly

zing

FLF

P in

Afg

hani

stan

60

Tabl

e 11

A: O

LS R

egre

ssio

n on

Fem

ale

Labo

r Fo

rce

Part

icip

atio

n: V

aria

ble

of In

tere

st -

Info

rmat

ion

from

Com

mun

ity sh

uras

V

AR

IAB

LES

(1)

(2)

(3)

(4)

(5)

(6)

(7)

(8)

(9)

(10)

(1

1)

Rec

eive

d in

form

atio

n fr

om

com

mun

ity sh

uras

0.

0790

***

(0.0

148)

0.

0688

***

(0.0

149)

0.

0249

(0

.015

7)

0.03

42**

(0

.017

1)

0.06

13**

* (0

.014

9)

0.04

10**

* (0

.014

3)

0.04

25**

* (0

.014

4)

0.04

34**

* (0

.014

3)

0.04

34**

* (0

.014

3)

0.04

33**

* (0

.014

3)

0.04

33**

* (0

.014

3)

Fixe

d ef

fect

s

Ethn

icity

Pr

ovin

ce

Dis

trict

Prov

ince

-leve

l num

ber o

f vi

olen

t inc

iden

ces

-0.0

0003

9***

(0

.000

0)

-0.0

0004

1***

(0

.000

0)

-0.0

0004

13**

* (0

.000

0)

-0.0

0004

13**

* (0

.000

0)

-0.0

0004

13**

* (0

.000

0)

-0.0

0003

14**

* (0

.000

0)

-0.0

0003

12**

* (0

.000

0)

Yea

rs o

f edu

catio

n

0.02

39**

* 0.

0231

***

0.02

36**

* 0.

0236

***

0.02

41**

* 0.

0243

***

(0.0

0168

) (0

.001

7)

(0.0

0178

) (0

.001

78)

(0.0

0179

) (0

.001

8)

Wom

en sh

ould

be

allo

wed

to

wor

k ou

tsid

e ho

me

0.05

37**

* (0

.011

3)

0.04

88**

* (0

.011

4)

0.04

88**

* (0

.011

4)

0.03

94**

* (0

.011

6)

0.03

90**

* (0

.011

6)

Use

per

sona

l mob

ile p

hone

(n

ot jo

intly

by

the

fam

ily)

0.

0258

* (0

.013

3)

0.02

58*

(0.0

133)

0.

0266

**

(0.0

133)

0.

0265

**

(0.0

133)

M

arrie

d

0.05

06**

* 0.

0506

***

0.04

91**

* 0.

0440

**

(0.0

184)

(0

.018

4)

(0.0

184)

(0

.019

1)

Polit

ics a

nd re

ligio

n sh

ould

m

ix a

t the

loca

l lev

el

0.

0337

***

(0.0

119)

0.

0337

***

(0.0

119)

0.

0390

***

(0.0

119)

0.

0391

***

(0.0

119)

D

istri

ct-le

vel s

hare

of m

en

who

thin

k w

omen

shou

ld

have

equ

al o

ppor

tuni

ties l

ike

men

in e

duca

tion

0.

0651

**

(0.0

311)

0.

0658

**

(0.0

312)

Dis

trict

-leve

l sha

re o

f men

w

ho th

ink

ther

e w

ere

acci

dent

s cau

sed

by m

ines

an

d ex

plos

ive

devi

ces

-0

.045

4***

(0

.016

) -0

.045

1***

(0

.016

4)

Dis

trict

ave

rage

yea

rs o

f m

en’s

edu

catio

n

-0.0

0921

***

(0.0

0203

) -0

.009

31**

* (0

.002

04)

Dis

trict

-leve

l sha

re o

f men

w

ho th

ink

it is

acc

epta

ble

for

wom

en to

wor

k in

a c

o-ed

sc

hool

s

0.

0308

(0

.027

6)

0.02

91

(0.0

28)

Age

0.

0004

62

(0

.000

493)

H

ouse

hold

size

-0

.000

023

(0

.001

48)

Con

stan

t 0.

142*

**

0.54

4**

0.10

3***

0.

328*

0.

198*

**

0.14

4***

0.

108*

**

0.04

50**

0.

0450

**

0.04

58

0.03

56

(0

.006

53)

(0.2

37)

(0.0

296)

(0

.194

) (0

.009

04)

(0.0

0919

) (0

.010

4)

(0.0

199)

(0

.019

9)

(0.0

298)

(0

.033

) O

bser

vatio

ns

4,73

9 4,

739

4,73

9 4,

739

4,73

9 4,

739

4,73

9 4,

739

4,73

9 4,

739

4,73

9 R

-squ

ared

0.

008

0.03

8 0.

097

0.19

7 0.

023

0.10

7 0.

111

0.11

6 0.

116

0.12

4 0.

124

Not

e: W

eigh

ts a

pplie

d. R

obus

t sta

ndar

d er

rors

in p

aren

thes

es,

***

p<0.

01, *

* p<

0.05

, * p

<0.1

.

Des

ai a

nd L

i A

naly

zing

FLF

P in

Afg

hani

stan

61

Tabl

e 12

A: O

LS R

egre

ssio

n on

Fem

ale

Labo

r Fo

rce

Part

icip

atio

n: V

aria

ble

of In

tere

st -

Pers

onal

Mob

ile P

hone

V

AR

IAB

LES

1 2

3 4

5 6

7 8

9 10

11

Use

per

sona

l mob

ile p

hone

(not

jo

intly

by

the

fam

ily)

0.09

53**

* (0

.013

8)

0.08

63**

* (0

.013

8)

0.08

42**

* (0

.013

8)

0.08

53**

* (0

.014

4)

0.09

30**

* (0

.013

8)

0.03

47**

* (0

.013

1)

0.03

08**

(0

.013

2)

0.02

46*

(0.0

132)

0.

0228

* (0

.013

4)

0.02

41*

(0.0

135)

0.

0241

* (0

.013

5)

Fixe

d ef

fect

s

Ethn

icity

Pr

ovin

ce

Dis

trict

Prov

ince

-leve

l num

ber o

f vi

olen

ce in

cide

nces

-0

.000

0418

***

(0.0

000)

-0

.000

0418

***

(0.0

000)

-0

.000

0418

***

(0.0

000)

-0

.000

0418

***

(0.0

000)

-0

.000

0418

***

(0.0

000)

-0

.000

0418

***

(0.0

000)

-0

.000

0418

***

(0.0

000)

Yea

rs o

f edu

catio

n

0.02

33**

* 0.

0226

***

0.02

40**

* 0.

0232

***

0.02

39**

* 0.

0240

***

(0.0

017)

(0

.001

71)

(0.0

0177

) (0

.001

79)

(0.0

0179

) (0

.001

81)

Wom

en sh

ould

be

allo

wed

to

wor

k ou

tsid

e ho

me

0.05

01**

* (0

.011

5)

0.04

78**

* (0

.011

5)

0.04

62**

* (0

.011

5)

0.03

76**

* (0

.011

6)

0.03

73**

* (0

.011

6)

Polit

ics a

nd re

ligio

n sh

ould

mix

at

the

loca

l lev

el

0.

0336

***

(0.0

119)

0.

0341

***

(0.0

119)

0.

0394

***

(0.0

119)

0.

0396

***

(0.0

119)

M

arrie

d

0.05

09**

* 0.

0506

***

0.04

92**

* 0.

0442

**

(0.0

183)

(0

.018

4)

(0.0

184)

(0

.019

1)

Rec

eive

d in

form

atio

n fr

om

com

mun

ity sh

uras

0.

0434

***

(0.0

143)

0.

0435

***

(0.0

143)

0.

0435

***

(0.0

143)

W

atch

TV

0.

0168

0.

0168

0.

0167

(0.0

113)

(0

.011

8)

(0.0

119)

D

istri

ct-le

vel s

hare

of m

en w

ho

thin

k th

ere

wer

e ac

cide

nts c

ause

d by

min

es a

nd e

xplo

sive

dev

ices

-0.0

423*

**

(0.0

161)

-0

.042

3**

(0.0

161)

Dis

trict

-leve

l sha

re o

f men

who

th

ink

wom

en sh

ould

hav

e eq

ual

oppo

rtuni

ties l

ike

men

in

educ

atio

n

0.

0668

**

(0.0

313)

0.

0673

**

(0.0

313)

Dis

trict

ave

rage

yea

rs o

f men

’s

educ

atio

n

-0.0

0961

***

(0.0

0206

) -0

.009

70**

* (0

.002

06)

Dis

trict

-leve

l sha

re o

f men

who

th

ink

it is

acc

epta

ble

for w

omen

to

wor

k in

a c

o-ed

scho

ols

0.

0269

(0

.028

1)

0.02

54

(0.0

284)

Age

0.

0004

54

(0

.000

492)

Hou

seho

ld si

ze

0.00

0050

1

(0.0

0148

) C

onst

ant

0.13

0***

0.

511*

* 0.

0855

***

0.29

1 0.

186*

**

0.14

7***

0.

115*

**

0.05

78**

* 0.

0383

* 0.

0397

0.

0291

(0.0

0643

) (0

.241

) (0

.029

9)

(0.2

1)

(0.0

0906

) (0

.009

23)

(0.0

104)

(0

.019

7)

(0.0

206)

(0

.030

2)

(0.0

335)

Obs

erva

tions

4,

739

4,73

9 4,

739

4,73

9 4,

739

4,73

9 4,

739

4,73

9 4,

739

4,73

9 4,

739

R-s

quar

ed

0.01

4 0.

043

0.10

6 0.

205

0.03

2 0.

106

0.11

0.

114

0.11

7 0.

124

0.12

4 N

ote:

Wei

ghts

app

lied.

Rob

ust s

tand

ard

erro

rs in

par

enth

eses

, **

* p<

0.01

, **

p<0.

05, *

p<0

.1.


62

Section E. Questionnaire Module A: Labor

A1 Are you currently working for pay, including farm work or tending to livestock, or occasional work?

Yes 1 Skip to A4

No 2

A2 If no, are you currently actively looking for work?

Yes 1 Skip to A5

No 2

A3 What is the main reason you are not looking for work?

Student / pupil 1

Housewife 2 Skip to A6

Retired/ too old 3

Illness / injured 4

Handicapped 5

Military service 6

Have a job that starts in future 7 Skip to A5

Temporarily laid off 8

Do not want to work 9

No jobs available / discouraged 10

Other (specify) __________ 666

A4 What is your main occupation or work? Use Employment Code: ____________

A5 In what occupation or what sector will you work or are you looking for work? Use Employment Code: ____________

A6 If you're a housewife, do you contribute to household income, through agriculture, handicraft/textile work, or domestic labor?

Yes 1

No 2

Employment Codes:

Farmer (own land / tenant farmer) 1 Private Office – Executive/ Manager 9

Farm laborer (other's land) 2 Self employed professional 10

Laborer, domestic, unskilled worker 3 Small business owner 11 Informal sales/ business 4 School Teacher 12 Skilled worker/artisan 5 University Teacher 13

Government Office - Clerical worker 6 Military/ Police 14

Private Office - Clerical worker 7 Other (specify)_________ 666

Government Office – Executive/ Manager 8 Refused 888

ANALYZING FEMALE LABOR FORCE PARTICIPATION IN …

Documents

Transcript of ANALYZING FEMALE LABOR FORCE PARTICIPATION IN …