Policy Research Working Paper 10809 Explaining Gender Differences in Economic Outcomes in Burkina Faso Aletheia Donald TM Tonmoy Islam Anja Robakowski Africa Region A verified reproducibility package for this paper is Gender Innovation Lab available at http://reproducibility.worldbank.org, June 2024 click here for direct access. Policy Research Working Paper 10809 Abstract Gender equality is central to economic development. This the business revenue gender gap. The wage gap increased paper examines gender gaps in Burkina Faso and find that most in more rural regions. Results from decomposition women’s labor force participation is 10 percentage points analysis show that women’s lack of capital and male workers, lower than men’s in 2019, while their wage earnings are lack of control over income and lower economic benefits 82 percent lower, business revenues are 61 percent lower, from marriage—along with lower levels of skills and farm- and value of agricultural production is 61 percent lower. ing inputs—have the largest associations with the gaps. The Nationally, gender gaps in labor force participation, busi- paper reviews evidence-based policy options for tackling ness revenues and crop sales are unchanged when compared the identified gaps for each sector, which include provid- to 2014 but increased significantly for wage earnings and ing vocational skills to women, improving their access to (to a lesser extent) for harvest value. The gender gap in capital, increasing the effectiveness of agricultural extension labor force participation increased in urban areas, while the services and expanding the provision of childcare services northern part of Burkina Faso witnessed large increases in and gender norms interventions. This paper is a product of the Gender Innovation Lab, Africa Rgion. It is part of a larger effort by the World Bank to provide open access to its research and make a contribution to development policy discussions around the world. Policy Research Working Papers are also posted on the Web at http://www.worldbank.org/prwp. The authors may be contacted at adonald@worldbank.org. A verified reproducibility package for this paper is available at http://reproducibility.worldbank. org, click here for direct access. RESEA CY LI R CH PO TRANSPARENT ANALYSIS S W R R E O KI P NG PA The Policy Research Working Paper Series disseminates the findings of work in progress to encourage the exchange of ideas about development issues. An objective of the series is to get the findings out quickly, even if the presentations are less than fully polished. The papers carry the names of the authors and should be cited accordingly. The findings, interpretations, and conclusions expressed in this paper are entirely those of the authors. They do not necessarily represent the views of the International Bank for Reconstruction and Development/World Bank and its affiliated organizations, or those of the Executive Directors of the World Bank or the governments they represent. Produced by the Research Support Team Explaining Gender Differences in Economic Outcomes in Burkina Faso Aletheia Donald, TM Tonmoy Islam and Anja Robakowski* Keywords: gender, productivity, Burkina Faso JEL: J31, J7, O12 * Donald: World Bank, adonald@worldbank.org; Islam: Elon University, tislam@elon.edu; Robakowski: World Bank, arobakow@worldbank.org. This paper is a product of the Gender Innovation Lab (Office of the Chief Economist, Africa Region) and of the Macroeconomics, Trade and Investment (MTI) Global Practice, and was prepared as a background paper for the 2023 Burkina Faso Country Economic Memorandum. We thank Kodzovi Senu Abalo, Markus Goldstein, Eugenie Maiga, Kofi Nouve, Daniel Pajank, Michael O’Sullivan and Christophe Rockmore for helpful comments. The findings, interpretations, and conclusions expressed in this paper are entirely those of the authors. They do not necessarily represent the views of the International Bank for Reconstruction and Development/World Bank and its affiliated organizations, or those of the Executive Directors of the World Bank or the governments they represent. 1. Introduction Achieving gender equality and the empowerment of women and girls by 2030 has been enshrined as a global goal by development actors including the United Nations. 1 To work toward this objective, policymakers will need to increase the participation of women in the economy and society. For gender- related policies and programs to be designed and implemented effectively, it is necessary to equip policy makers and project teams with appropriate knowledge and instruments. Moreover, it is important to identify the relative size and drivers of gender gaps across the economy in order to determine how best to boost economic growth through efforts to achieve gender parity. This approach is valuable for a country like Burkina Faso, where over 40 percent of the population lives below the poverty line (World Development Indicators, 2022). Burkina Faso ranks 144th out of 157 countries in terms of its human capital, and 184th out of 191 countries in terms of its Gender Development Index 2. Previous research from Burkina Faso has found evidence of discrimination in favor of boys within the household (Haddad and Reardon, 1993), widespread male control of household assets (van den Bold et al., 2015) and limited decision-making power for women (Kieran et al., 2015). In this paper, we provide a systematic analysis of gender gaps across Burkina Faso’s main economic sectors; examine how they vary over time, between age groups and across regions; identify the primary drivers of these gaps; and provide evidence-based policy recommendations. Using nationally representative household data from 2019, we find that women’s participation in paid work is 20 percentage points lower than that of men, and 10 percentage points lower when looking at both paid and unpaid work. Women in Burkina Faso earn 82 percent less in monthly wage earnings compared to men, while their business revenues are on average 61 percent lower. Within agriculture, women’s harvest values and crop sales are both approximately 60 percent lower than that of men’s. We further find that between 2014 and 2019, the overall gender gap in wage earnings increased, along with a small increase in the harvest value gender gap—but stayed the same for crop sales, business revenues and overall labor force participation. This means that Burkina Faso has not witnessed progress in lifting this enormous structural barrier to economic growth in the country. We run a series of decompositions to understand and explain the causes of the current gender gaps, which allow us to identify how much of the gender gap is driven by differences in endowments (differences in the 1 Goal 5: https://sdgs.un.org/goals/goal5. 2 Human capital is measured through the Human Capital Index, which captures the amount of human capital that a child born today can expect to attain by age 18. The Gender Development Index was accessed at: https://hdr.undp.org/gender-development-index#/indicies/GDI. 2 average levels of characteristics or resources that women have or use relative to men) versus differences in returns (differences in the benefit that women draw from these resources relative to men). We find that gender differences in labor force participation are primarily driven by differences in returns to worker characteristics, such as the number of young children in the household, and notably not due to characteristics such as differential educational attainment. A large portion of the gender wage earnings gap is due to occupational segregation across sectors, as well as men having higher-skilled jobs and more senior-level positions relative to women. Women also are more likely to work as unpaid workers in a family business or farm, widening the gender gap in wage earnings. For business revenues, much of the gender gap is driven by women-run businesses lacking electricity, as well as the fact that men tend not to work in businesses run by their female household members. In the agriculture sector, the gender gap is largely driven by women’s concentration in lower-value crops, as well as their lower adoption rates of farming technology (machinery, fertilizer and pesticides) compared to men. Across sectors, the most important drivers of gender gaps are women’s lack of capital and male workers, lack of control over income and lower economic returns to marriage. We conduct a thorough literature review of available impact evaluation evidence to identify what works to address these gaps. First, women’s skills can effectively be enhanced through financial assistance to attend trainings (Alfonsi et al., 2020; Bossuroy et al., 2022), apprenticeships (Hardy et al. 2019) and job fairs (Abebe et al. 2017). Second, productive utilization of capital/cash transfers has been strengthened by providing training and skills development (Blattman et al., 2016, Handa et al., 2018). Furthermore, teaching entrepreneurship skills has also been shown to improve incomes among entrepreneurs (Alibhai et al. 2019; Campos et al. 2017). If capital provision is combined with financial literacy, then the effects may be even higher, as studies show that providing financial education to female entrepreneurs helps to increases their income (Blattman et al., 2016; Blattman et al., 2020). To overcome occupational segregation, providing women with information on earnings for male-dominated trades is also emerging as a promising intervention (Gassier et al. 2022). In terms of women’s lower adoption rates of farming technology, providing information and subsidies has been shown to be effective (Carter et al, 2021; Fishman et al, 2019), as well as easing female farmers’ access to credit (Ndegwa et al, 2022). Improving extension services by sending video messages to female farmers about the proper way to grow crops, and providing tailored advice, has also been shown to lead to improvements in agricultural practices (Arouna et al., 2021; Vandevelde et al., 2021). Providing female farmers with insured loans and crop insurance has also been shown to lead women to invest more in agricultural activities (Mishra et al., 2021; Stoeffler et al., 2022). 3 To strengthen women’s lack of control over income and general economic agency, starting at an early age can yield dividends later in life. Studies have shown that teaching girls about their rights in school makes them more intolerant of domestic violence (Friedman et al., 2016), and that adolescent girls’ clubs can not only increase their later employment, but also reduce the likelihood of adolescent pregnancy and early marriage (Bandiera et al., 2020). Among adult couples, men can be taught to take up more of household responsibilities and inform them about gender equality (Doyle et al, 2018), which could free up time for women to do other economic activities. Finally, expanding access to childcare services can help women to work in economic activities (Ajayi et al. 2022, Bjorvatn et al. 2022, Donald and Vaillant 2022). The remainder of the paper proceeds as follows. Section 2 presents background data and summary statistics on the Burkinabè context and covers our estimation strategy. Our results on the magnitude of the gender gaps and their drivers are covered in Section 3. Section 4 provides additional data and policy recommendations, while Section 5 concludes. 2. Setting and Empirical Strategy 2.1 Context and Data Burkina Faso is a largely rural country, with about 70 percent of the population living in rural areas (World Bank 2020). Labor force participation (across both paid and unpaid work) is high, increasing from 70 to 80 percent between 2014 and 2019. 3 However, this does not translate to low poverty levels. In 2021, Burkina Faso had a GDP per capita of less than $800 in constant 2015 dollars, and the expected years of schooling, according to the Human Capital Index data, for both boys and girls in Burkina Faso is seven years. Moreover, the country is undergoing political turmoil and a series of military coups 4, which could have a severe economic impact. Although the overall educational attainment remains low, the gender gap in both gross primary and secondary school enrollment has reversed in recent years. Enrollment for boys and girls in primary school is approximately equal (92.2 percent for boys vs. 93 percent for girls), and enrollment for girls in secondary school is slightly higher at 39.4 percent for boys vs. 41.8 percent for girls (UNESCO 2020). These gaps are smaller than the Sub-Saharan African average of a 4-5 pp gender gap in favor of boys for both primary and 3 Changes in labor dynamics have been largest in the North, where labor force participation grew from 67 to 87 percent over the same period. 4 “Burkina Faso soldiers announce military takeover, Kaboré ‘suspended’” France24. https://www.france24.com/en/live-news/20220124-burkina-faso-president-kabore-held-by-mutinying-soldiers- sources-tell-france-24; “The old junta leader makes way for the new in Burkina Faso's second coup of the year” NPR. https://www.npr.org/2022/10/02/1126478594/burkina-faso-coup. 4 secondary school (UNESCO, 2020). Burkina’s adolescent fertility rate is also lower than that of peer countries such as Mali, Niger and Uganda, but still high at 100 births per 1,000 adolescent girls. Moreover, Burkina Faso has reduced its maternal mortality rate from 516 to 320 per 100,000 live births between 2000 and 2017, in line with the overall reduction in Sub-Saharan Africa of 39 percent over this period (WHO 2019). Despite this progress, stark gender disparities in Burkina Faso persist. As of July 2022, merely 16.9% of seats in parliament were held by women (Parline 2023). Only 9.9% of adult women have reached at least a secondary level of education, compared to 15.7% of their male counterparts (UNESCO 2020). Women of reproductive age (15-49 years) often face barriers with respect to their sexual and reproductive health and rights; only 28.1% of women in 2020 had their need for family planning satisfied with modern methods, e.g., condoms or the pill (UNDP 2021). Women are customarily expected to be submissive to their husbands (Nikièma et al., 2008; Kieran et al., 2015), and while a high proportion of women work in agriculture, most of them are engaged in subsistence farming (Weber, 2018), often on their husband’s plots. Although the law allows for women to own property, the prevalence of local norms ensures that few women own land (Kieran et al., 2015). These gender disparities could widen further as the county is facing armed conflicts with Islamist groups, who reportedly now control about 40 percent of the country, and there have been armed conflicts in 10 of the 13 regions in Burkina Faso. 5 How do gender disparities manifest in terms of labor force participation and earnings? Our primary source of data to investigate this question is the “Enquête Harmonisée sur les Conditions de Vie des Ménages 2018-2019” (EHCVM). The EHCVM is a nationally representative survey carried out in each country of the West African Economic and Monetary Union (WAEMU). The Burkina Faso survey was conducted by the National Institute of Statistics and Demography (INSD), the national statistical agency of Burkina Faso. The survey used a stratified two-stage sampling, with 585 enumeration areas randomly selected in the first step, and 12 households randomly drawn from each enumeration area in the second step. EHCVM surveyed around 7,000 households. The survey asked a rich variety of questions about different individual characteristics, such as income, educational attainment, and sector of employment. Figure 1 shows male and female employment by sector in the EHCVM data, depicted as a Venn diagram. Individuals can have more than one occupation—for example, someone may be both a farmer and a business manager, or a farmer and a paid employee. In each of the 7,000 households covered by the survey, 5 “Burkina Faso – Events of 2022” Human Rights Watch. https://www.hrw.org/world-report/2023/country- chapters/burkina- faso#:~:text=Burkina%20Faso's%20human%20rights%20situation,result%20of%20two%20military%20coups. 5 every adult in the household 6 is asked questions about their labor force participation. Any individual aged 15 and above who reports receiving a salary is counted as a wage worker, while any individual aged 15+ who reports working without pay for a family business or farm is counted as an unpaid worker. If an individual reports solely or jointly managing a plot in the household, they are classified as a farmer, while if they solely or jointly manage a business or small-scale enterprise, they are classified as an entrepreneur. Figure 1: Venn Diagram of Men’s and Women’s Employment Figure 1 shows that most women are unpaid workers, while most men are farmers. Among women who receive compensation for their work, entrepreneurship is the most common profession. To compare changes over time, we also use data from the 2014 (LSMS) Enquête Multisectorielle Continue. The 2014 LSMS is a nationally representative survey conducted by the National Institute of Statistics and Demography (INSD), with support from the Swedish Cooperation. The data was collected in four rounds from January to December 2014. Identically to the EHCVM survey, stratified two-stage sampling was employed. A total of 905 enumeration areas were selected in the first stage, and 12 households were randomly selected from each EA in the second stage. In total, 10,860 households were selected for the survey. 7 The LFP, earnings and agriculture output data in both the 2014 LSMS and 2018/2019 EHCVM were collected using very similar survey questions. The primary difference between the two surveys is that the 2014 LSMS survey did not have any specific questions asking if a person worked without pay, so we focus only on overall LFP (including paid and unpaid workers, and workers looking for a job in the last four weeks) for comparisons over time. 6 Defined as being 15 years old or above for the purposes of this paper. 7 https://microdata.worldbank.org/index.php/catalog/2538/study-description. 6 2.2 Econometric Specification Gender inequality has been shown to matter for economic growth (Hsieh et al, 2019, Esteve-Volart, 2009). One of the main reasons why gender inequality can hinder long run economic growth is because discrimination can lead to talent being misallocated from productive to non-productive sectors (Santos Silva and Klasen, 2021). For example, in Hsieh et al (2019), individuals want to work in occupations that match their innate abilities. However, because of discrimination, women and minorities may be able to work in sectors which match their ability only at a reduced wage. Moreover, discrimination during educational attainment can make the cost of education much higher for women and minorities (Hsieh et al, 2019). Gender inequality can also hamper growth by resulting in less effective and representative institutions, and lower investment in the next generation. Based on these theoretical models, we examine gender differences across labor force participation rates, wage earnings, business revenues, value of harvest and crop sales for the working age population (15-64 years old) among the Burkinabè to understand the sources behind the gaps in labor force participation and earnings. We measure paid labor force participation by looking at the share of wage workers by gender, and overall labor force participation by looking at the share of wage workers and unpaid workers by gender. For both categories, we also consider any adult who was actively looking for work in the four weeks prior to the survey. Monthly wage earnings are calculated as the sum of monthly income earned by all wage workers and unpaid workers in their primary job, along with any allowances (food, transportation, etc.) that they received. Entrepreneurship revenues are calculated at the business-level. We define our main outcome as the total gross revenue earned by each enterprise reported by the household. We then examine differences for businesses solely or jointly managed by a woman vs. businesses solely or jointly managed by a man. We only look at businesses that report being active, and do not examine businesses that have closed or been dormant for extended periods of time. In addition to wanting to keep outcomes consistent across sectors, we focus on revenues rather than profits because not all the labor costs were included in the survey. For agriculture, we first look at the value of crop sales for each plot owned or farmed by the household and examine differences by gender of the plot manager. The harvest value, on the other hand, is imputed as it is not given directly in the dataset. 8 8 Using the value of the crop sold and the quantity of crop sold, we estimate the price per unit of each crop grown. We then find the average price of each crop grown, and then multiply that by the total harvest of each farmer. This gives us an estimate of the harvest value grown by each farmer. 7 We first run simple regressions to calculate the raw gender gaps. For the LFP regression, where the dependent variable is a dummy variable, we use a logistic regression as follows: 1 ( ,, = 1) = (1) 1+exp (−,, +,, ) Where the dependent variable equals 1 if the individual is in the labor force, and zero otherwise. The variable ,, = 1 if the person i, working in sector j={wage employment, entrepreneurship or agricultural production } and company/business sector/plot k is female, and 0 if the person is male. The coefficient gives us the extent of the differences in outcomes between men and women. 9 For regressions of the gender gap in earnings, we run the following specification: ,, = + ,, + ,, (2) The variable ,, is the dependent variable measuring earnings of person i of gender G = {man, woman} working in sector j={wage employment, entrepreneurship or agricultural production} and company k. For ,, , we use an inverse hyperbolic sine transformation. 10 We then estimate equation (1) and (2) separately for urban and rural areas, as well as by region and age group (individuals aged 15-29 vs. 30-64). The urban/rural classification was taken from the dataset: the survey coded whether the respondent lived in an urban or rural area. We use robust standard errors. 11 Next, we use Kitagawa-Oaxaca-Blinder (KOB) decompositions to understand the factors underlying the gender gaps in the 2019 data. The method calls for running two separate regressions, one for men and one for women: ,,, = + ,,, + ,,, (3) 9 In this and in all subsequent regressions, we use the analytic weights of Stata as weights, which allows us to run regressions on smaller subsamples and obtain the standard errors. Analytic weights scale the regression by the square root of the weight, and this method allows us to get an unbiased estimate of the variance of the residual (for more information, see “What is the Effect of Specifying aweights in regress?”). We did a test where we used the full sample and ran the regressions using survey weight command and analytic weights respectively. The coefficients were identical, and the standard errors were very similar in almost all the cases. 10 The one exception is wage earnings, where due to most of the data consisting of zeroes, we instead use untransformed values for wage earnings, winsorized at the 1% and 99% levels in both years. 11 Although not reported, we also ran the KOB decomposition using clustered standard errors at the enumeration area level. While the clustered standard errors were calculated in the regressions using the full sample, the clustered standard errors were not produced when regressions on subsamples were run (by age or by urban/rural classification). However, we saw that for the full sample, the clustered standard errors and the robust standard errors were virtually identical. 8 where ,,, is the same as described above. ,,, includes different individual-level characteristics, and characteristics of the sector of employment j and organization k where the person works. After the regression, the following equation is calculated: ����� ����� , − ����� ����� , = + , + ������ ����� , − ( + , + , ) (4) where, ����� ����� , , ������ , , , are the average values of the dependent variable and regressors for gender G (M for Men, and W for Women). Since ����� , = 0, we can remove them from the equation. This gives: , = ����� , − ����� ����� ����� ������ , = − + , − , (5) Adding and subtracting ������ , from the right-hand side of the equation and rearranging gives us: , − ����� ����� ����� ������ ������ , = − + (, − , ) + ( − ), (6) The equation above has decomposed the average difference between men and women (����� , − , ) into ����� two parts: the part due to differences in average endowments between men and women (represented by ����� − + ( ������ , − , )), and the part due to the differences in the returns to the endowments (( − ������ ), ). If the difference in the outcome is primarily due to endowments, it would mean that women have, on average, less of a ‘productive’ characteristic or input, or more of an ‘unproductive’ characteristic or input. If on the other hand, the difference in the outcome variable is primarily driven by the coefficients, then the endowments of men and women do not yield equal returns. This would imply that there are differences in returns to worker characteristics for men and women because of discrimination, social norms, market failures or other institutional constraints. 3. Results We present results for labor force participation, wage earnings, business revenues, and farm output in turn. 3.1 Labor Force Participation We estimate econometric specification (1). The first column of Table 1 Panel A shows that the gender gap in paid labor force participation is 20 percentage points, while the second column shows that the gender gap in overall labor force participation (pooling paid and unpaid workers together) is 10 percentage points. 9 Focusing on overall labor force participation, Table 2 shows that the gap has slightly dropped over time, though the change is not significant. The gender gap fell in the northern region, but rose in urban areas,12 and in central and southern regions. The gender gap in LFP is now largest in the Center of the country (which includes the Capital Region), followed by the West, and is 13 percentage points larger in urban areas compared to rural areas (20 vs. 7 percentage points). Gender differences are only slightly different across age groups, with gender differences being marginally larger among the 30-64 group (though this group experienced a significant decline in the LFP gap across years). Next, we turn to the decomposition analysis results in Table 3. Columns (1), (2) and (3) consider paid labor force participation, while columns (4), (5), and (6) show results for overall labor force participation. Levels of education and household wealth are not associated with the gender gap in labor force participation. Rather, the gender gap in both types of labor force participation is entirely explained by differences in returns to worker characteristics (i.e., by discrimination, social norms, or institutional constraints). Further, women benefit less than men in terms of their labor force participation from increased age (i.e., labor market experience), from living in an urban area and from having children in the household. They also face lower returns to being married. This may be because they are expected to work less and adopt or intensify their role as caregivers upon marriage—even given the same number of children—or generally have less control over whether they can work once they are married. We delve deeper to understand the differences in LFP. Appendix Table 2 looks at the KOB decomposition by age. Columns (1), (2) and (3) show the decomposition results for those aged 15-29, confirming that here too, gender gaps are entirely driven by differences in returns. However, for adults aged 30-64—shown in columns (4), (5) and (6)—we note that 2.1 percentage points of the gap is driven by differences in endowments, in particular lower ownership of a vehicle and financial account. We also look at LFP by urban/rural, and the results are illustrated in Appendix Table 3. In rural areas, differences in endowments are not correlated with the gender gap, but in urban areas, 3.1 percentage points of the overall 20 percentage point gender gap can be explained by differences in endowments. Similar to the results for older adults, we see that urban women’s lower financial account and vehicle ownership 12 One caveat is that areas denoted as urban/rural have changed for some enumeration areas in Burkina Faso over time. In the last LSMS done in 2014, the average urbanicity in the provinces was about 13 percent, and it rose slightly to 15 percent by 2019. Appendix Table 1 shows the proportion of urban population by province in 2014 and 2019. However, there are some provinces that faced a larger increase in urban population than other provinces. Both 2014 and 2019 surveys used a two-stage sampling method, but the enumeration areas picked were different in each of those samples, which does not allow us to distinguish whether the changes in LFP or earnings in urban and rural areas across time were due to more people moving to urban areas, or due to rural areas transforming to urban areas. 10 increases the gender gap. We also see that women’s age—which is slightly lower than men’s on average— contributes to women’s lower labor force participation in urban areas. Taken together, we find that the overall gender gap in labor force participation in Burkina Faso is driven by differences in returns, not differences in endowments. While we do see that differences in endowments make up a small part of the gender gap among adults aged 30+ and in urban areas, tackling social norms and easing women’s care work through childcare or other time-saving technologies may be more effective in decreasing gender inequality in the likelihood of working, compared to policy solutions like increasing educational attainment among girls. 3.2 Wage Earnings We estimate specification (2) to compute gender gaps in wage earnings, and present the results in column (3) of Table 1 Panel A. We find an 82% gap in wage earnings. This is driven by a remarkable difference in the likelihood that women are not paid at all for their work: 91% of female wage workers are unpaid, versus 66% of male workers. When we restrict our results to non-zero wage earnings, we find a 30% gender gap, shown in Column (4). However, since women’s unpaid wage work is economically important, we focus on the overall wage earnings definition in the remainder of the analysis. Table 2 shows that that the gender gap has in wage earnings has increased significantly between 2014 and 2019. This increase is particularly prominent for younger adults, where the gap was 54% in 2014 but 76% in 2019. The gender gaps in wage earnings have increased in the North and West and most dramatically in the South—but decreased in the Central region, where the country’s capital is located. This matches findings by urban/rural, which show that the widening of the gap is coming exclusively from rural areas. Table 4 shows the KOB decomposition of the 82% gender gap monthly wage earnings. Contrary to labor force participation, Table 4 shows that over half of the gap is due to differences in endowments. The single largest driver of the wage earnings gender gap is occupational segregation across sectors, with 86% of wage-earning women being employees in the lower-productivity agriculture sector, compared to only 57% of wage-earning men (Table 1 Panel B). Working in agriculture is highly correlated with being an unpaid employee: almost all the female employees in agriculture are not paid any monetary amount for their work. Second, we see that a relative lack of seniority and skills among women explain around a third of the overall gender gap in wage earnings. Table 1 Panel B shows descriptive statistics related to these differences. 10% of men are managers, versus 3% of women (and 7% of men are skilled, versus 2% of women). Though not as large in magnitude, we also find that women’s lower education rates–14.7% of male employees have secondary education or more, versus only 4.1% of female employees—and lower financial account 11 ownership—39.2% of male employees have one, versus only 11.1% of female employees—explain a sizeable portion of the gender gap in wage earnings. In terms of the latter, while lower financial account ownership can be a consequence of lower wage earnings, empirical literature has also established it as a cause (Carranza et al. 2022). The other element increasing the gender gap is the value of the coefficients. Women’s lower economic returns to being married alone explain over a quarter over the gender gap. As discussed above, this may be because women are expected to adopt or intensify their role as caregivers and housekeepers upon marriage or generally have decreased economic control due to social norms. Concurrently, men may obtain a higher economic gain from marriage as they are more able to work when there is an additional person in their household taking on the care of work and chores. We also find that women have lower economic returns to being a manager, and from working in the lower-productivity agriculture sector—where they are often unpaid workers on their husbands’ fields. Turning to Appendix Table 4, we find that among the 15–29-year-old age group, about 45 percent of the gender wage gap is due to the endowments, but for the 30-64 age group, it is 69 percent. Moreover, there are some crucial differences in the drivers of the gender gap in wage earnings across age groups. Differences in skills and in education are significantly related to the gender gap only for older adults, while lower financial account ownership is only significantly related to larger gender gaps for younger adults. We also note that adults aged 30 to 64 have significantly lower returns to having young children in the household. Lower returns to marriage are present for both. Equalizing these returns could reduce the gender gap by 30 percent among the 15-29 age group, and about 28 percent of the wage gap among 30–64-year-olds. In Appendix Table 5, we see that gender gaps in wage earnings are driven by both differences in endowments and in returns in urban and rural areas, with the latter explaining more of the gender gap in wage earnings in both. The drivers of the gender gap summarized above are generally present in both urban and rural areas, though lower financial account ownership is correlated with a higher gender gap only in rural areas, and women’s lower completion of secondary education increases the gender gap more in urban areas compared to rural areas. In sum, we find a large gender wage gap in Burkina Faso—and much of the gap can be explained by women’s unpaid work in agriculture, as well as the fact that women work in low-skilled jobs and are further down the seniority ladder. Addressing both horizontal and vertical occupational segregation is the central priority for reducing gender gaps in wage earnings in Burkina Faso. 12 3.3 Business Revenues We next turn to business revenues. Table 1 Panel A shows that the gender gap in business revenues is 61%, while Table 2 shows that the size is virtually unchanged since 2014. Breakdowns by age group show that the business revenue gender gap has increased for younger adults and decreased for older adults, though neither of these changes is statistically significant. We likewise see a small non-significant increase in the business revenue gender gap in rural areas (however, there is a particularly large and significant increase in the North). The gap in business revenues did fall significantly in the Western region. Next, we examine the correlates of the gender gap in business revenues in Table 5. We find that differences in business revenues between female and male entrepreneurs are mostly driven by differences in entrepreneur and business characteristics: they explain approximately 64% of the gap, and differences in returns are not significant. Among these differences in characteristics, lower electrification of female-run businesses emerges as an important factor. As shown in Table 1 Panel B, 14% of male-run businesses have electricity, versus 2.9% of women-run businesses. This difference can explain around 9.4% of the gender gap in business revenues. The gender gap in business revenues is also correlated with lower financial-account and phone ownership. 63% of male entrepreneurs have a financial account, versus 27% of female entrepreneurs, as shown in Table 1 Panel B. This lower account ownership among female entrepreneurs alone can explain a quarter of the gender gap in business revenues. However, the most important driver of the gender gap in business revenues is female entrepreneurs’ low use of male household labor, which limits their business revenues. Over the course of a year, male entrepreneurs use 254 days of male household labor, while female entrepreneurs only use 6 days of male household labor (Table 1 Panel B). This factor alone explains 76% of the overall gender gap in business revenues. We subsequently decompose business revenues by age group in Appendix Table 6. Interestingly, almost all the gender business revenue gap for the younger age group is attributable to differences in returns, but for the older age group, the gap is mainly due to the differences in endowments. While the correlates of the gender gap mirror those of the overall sample for the older cohort (lack of electricity, financial accounts, and male household labor), lower revenues for female entrepreneurs aged 15-29 are correlated with lower returns to marriage, lower returns to account ownership and secondary education, lower business phone ownership and less male household labor. Appendix Table 7 shows the drivers of the gender gap in urban vs. rural areas. We note that gender gaps in urban areas are completely driven by gender differences in entrepreneur and business characteristics, while 13 in rural areas both differences in characteristics and returns to those characteristics matter equally. While the factors identified as driving the gender gap hold across both urban and rural areas, we also find that female businesses’ lower likelihood of maintaining accounting records helps increase the gap in rural areas, while women’s concentration in the clothing sector increases the gender gap in urban areas. As shown in Table 1 Panel B, 8.2% of male-run businesses in the country keep business records, as opposed to only 3% of female-run businesses. Female-run businesses are also slightly more likely to be in the clothing sector (28.1% for women versus 25.4% for men). In sum, we find large gender gaps in business revenues—and that three-quarters of the gender gap in business revenues can be explained through women’s lower use of male household labor in their businesses. This result echoes what we found for wage earnings, were women were less likely to work as managers. We also find that a lack of access to utilities, such as electricity, and access to finance matter. 3.4 Agricultural Production Lastly, we show the results for agricultural production. Column (5) of Table 1 Panel A shows a 61% gender gap in the imputed value of agricultural harvests, while Column (6) shows a 62% gender gap in the value of crop sales. Table 2 shows that the gender gap in harvest value has increased by a small but significant 5 percentage points between 2014 and 2019, while the gender gap in crop sales is statistically unchanged. The increase in the agricultural production gender gap holds across both age groups but is proportionately larger for younger adults, with the increase in the gap being most pronounced in the West of the country. For crop sales, we instead notice a significant decline in the gender gap among younger adults, as well as in the North of the country. We next decompose the factors underlying gender gaps in both harvest value and crop sales in Table 6. Both gender gaps can be explained by a mix of gender differences in endowments and differences in returns, though differences in endowments are more important for crop sales. The largest driver of the gender gap in agricultural harvest value is female farmers’ concentration in lower-value crops, as opposed to cash crops like cotton or cashews. This variable alone explains 1 percent of the overall gender gap. The next largest factor is the fact that men in the household tend not to work on female-managed plots: this explains an additional 10 percent of the gender gap. This is similarly echoed by the fact that women face much lower returns to their agricultural production from a higher number of adults in the household. Unlike for business revenues, we also find that women’s lower use of male hired (non-household) labor widens the gender gap for both agricultural production and crop sales. 14 Further, we see that women’s lower use of fertilizers, pesticides and tools in farming increase the agricultural gender gap. These variables are the largest driver of the gender gap in the value of crop sales. Indeed, 47% of male-managed plots use fertilizer versus 26% of female-managed plots, 54% of male- managed plots use pesticides versus 31% of female-managed plots, and 59% of male-managed plots use agricultural tools/machinery versus 33% of female-managed plots (Table 1 Panel B). Reaching parity in the use of fertilizers, pesticides and agricultural equipment between men and women farmers could help to reduce the gap in output value by 0.19 points, or about 18 percent of the raw gender gap in harvest value. It would also reduce the gender gap in crop sales by about 0.3 points, or about 30 percent of the raw gender gap. Moreover, we also find that female plot managers and less likely to control (i.e., be able to make decisions over) crop revenue and face lower returns from this revenue, which further increases the gender gap. We next look at the decomposition by age group. Appendix Table 8 shows results for harvest value. Women’s lower likelihood of farming an export crop is similarly important for younger and older adults. The lower number of days that household and non-household men work on female-managed plots and the lower ownership of agricultural equipment are more important in explaining the gender gap in crop value for younger adults, while the lower use of fertilizer, pesticides and control over income are more important among older adults. These results are similar for gender gaps in crop sales (shown in Appendix Table 9), though the result on lower labor provision by household men is flipped, mattering only for older adults. Overall, we find that the reason why women plot managers have lower harvest value and crop sales compared to their male counterparts is their concentration in food crops and lower use of both labor and non-labor inputs. Increasing fertilizer, pesticides and farming equipment use among women could help to substantially increase their revenues. Equipping women with the means to grow higher-value crops, like those that are exported (such as cotton), could further help to increase farm revenues. 4. Discussion and Recommendations The analysis above showed that the most important drivers of gender gaps across sectors are women’s lack of capital and male workers, lack of control over income and lower economic benefits from marriage. We also found sector-specific drivers, highlighting the importance of conducting this kind of analysis for policy prioritization. Rounding out our findings from the decomposition analysis, additional survey data provide further context and information on challenges that hamper gender equality. While we have discussed above that women in 15 Burkina Faso shoulder virtually all the care work in households, a recent survey 13 provides quantitative precision: 96% of women reported that they do most or all the cooking in their households, while 80% reported that they do most or all the childcare. In addition, women’s agency in decisions on spending money is limited: only 66% of surveyed women report being able to fully decide how to spend their own income, and 60% indicate not fully participating in decisions on how to spend household money even on major household purchases. Women’s economic participation is also restricted by their limited ability to fully control their health and reproductive lives. For example, while Burkina Faso has succeeded in reducing maternal mortality more than some peer countries (Mali, Niger, Uganda) and the mortality ratio per 100,000 live births (320) is lower than the regional average (425), it remains well above the global mean (160), and mortality reduction in Burkina Faso has progressed at a slower pace than in aspirational peer countries (Ethiopia, Rwanda). Nationally representative data shows that in 75% of households, the husband is the main decisionmaker over the wife’s health, while female genital mutilation is highly pervasive across income groups—ranging from 77% of women having been mutilated in the first income quintile, to 69% in the fifth income quintile (DHS 2010). This is despite female genital mutilation having been legally banned since 1996, showing the importance of addressing customary norms beyond legal changes. The UN’s most recent data from 2015 shows that the situation has not improved, with 68% of women aged 15-29 having suffered genital mutilation (UN 2019). Lastly, while the gender gap in both gross primary and secondary school enrollment has reversed in recent years, more than half of Burkinabè women aged 15-24 are illiterate, and girls have lower competency in math, especially in late primary school (PASEC 2015). These gender gaps in harmonized test scores have not decreased since 2017. In conjunction with prevailing social norms and women’s lower access to market information, these challenges contribute to the fact that women in Burkina Faso concentrate in lower-paying sectors, as employees, as entrepreneurs and as farmers. Our calculations from the 2018/2019 EHCVM show a large degree of occupational segregation across sectors, with male-dominated industries being more lucrative. For example, the average monthly earnings in construction are 1,390,000 CFA, while the average earnings in the food sector are 262,000 CFA. Women also make up most family workers and are underrepresented both in middle and senior management. In agriculture, virtually all export crops are grown by men. Three main policy priorities emerge from our findings. First, increasing women’s skills and engagement in higher-value sectors (in entrepreneurship, but also in agriculture). Second, increasing women’s access to 13 This survey was conducted as part of the endline for the World Bank’s ‘Travaux à Haute Intensité de Main d’œuvre’ (THIMO) project in Burkina Faso. 16 productive inputs such as business capital, agricultural inputs and machinery—but most importantly, labor. Third, increasing women’s physical security and household agency. To formulate recommendations for each of these priorities, we conducted an in-depth literature review. We selected studies for inclusion as follows: (1) the paper must draw the conclusions from a randomized controlled trial (RCT) that has been conducted in the last 10 years, (2) the paper has to be focused on the West Africa region, and if none is found from the West Africa region, then other regions in Africa may be considered. However, if other regions are considered, then multiple RCTs must be available. And finally, (3) the RCT must be preferably conducted with women participants. But if there are no studies found that specifically target women, then multiple RCTs from the region must be available to make the policy recommendations. 4.1 Increase Women’s Skills and Engagement in Higher-Value Sectors Skills Livelihood skills and vocational training can help women and girls start successful businesses and take advantage of economic opportunities to transition into the workforce. Evidence from Burkina Faso regarding vocational and entrepreneurial skills training is thin and more research is needed—though the ongoing Sahel Women's Empowerment and Demographic Dividend (SWEDD) impact evaluation will yield valuable lessons. Adolescent empowerment programs for girls, generally consisting of a combination of safe spaces, life skills training, and vocational skills training, have been found effective in other settings to increase young women’s labor force participation and economic outcomes (Buehren et al 2017, Bandiera et al. 2020). For example, an impact evaluation of an intervention in Liberia, which provided classroom- based technical and life skills training coupled with support to enter the workforce for adolescent girls, produced strong evidence of increased employment by 47%, increased earnings by 80%, increased access to and control of financial resources, and improved self-confidence (Adoho et al. 2014). Livelihood skills training can also be effective for women following instability, such as in post-conflict situations. Moreover, income generation and livelihood opportunities are often considered essential in discouraging remilitarization during the reintegration process of ex-combatants. Experimental evidence from Maniema, DRC shows that an income generating and budget management activity, as part of gender-targeted programming to support female combatants’ socio-economic reintegration, produced positive results and increased beneficiaries’ asset ownership (World Bank 2013). Lastly, evidence shows that outside of formal education and more traditional technical and business skills training, psychology-based, mindset-oriented training can improve entrepreneurs’ business outcomes (e.g., Alibhai et al. 2019). An impact evaluation in Togo compared the impacts of a business training program and a personal initiative training on 17 microentrepreneurs. Women who received the personal initiative training increased profits by 40%, while those who received the business training increased earnings by only 5% (Campos et al. 2017). Representation in Management As we have seen in our analysis above, women are not well-represented in management positions in Burkina Faso. Though not thoroughly tested through RCTs, the effectiveness of gender quotas has been assessed through a range of credible studies. While some studies have shown that quotas are not required for empowering women (Van der Windt et al, 2018), others have shown some promising results. For example, quotas have been shown to increase the participation of women in leadership (O’Brien and Rickne, 2016, Okedele, 2021). There is also evidence that showcasing women as role models can increase interest in leadership among younger girls (Porter and Serra, 2020, Breda et al, 2021). More generally, piloting and testing different ways to change gender norms regarding female leadership—including role-model campaigns showcasing female manager success stories—will be important in making progress on this front. Overcoming Occupational Segregation One way to reduce occupational segregation is through skills-based training and apprenticeships for women, as data indicate that women working in male-dominated sectors tend to have benefitted from apprenticeships in those sectors. Employers and public sector agencies have several strategies at their disposal to increase women’s success in apprenticeship programs, including awareness-raising sessions, adapting language in advertising and descriptions in apprenticeships, and profiling real-life examples of women in these jobs to ensure they avoid gender stereotypes and attract a diverse set of candidates (Schomer and Hammond 2020). More broadly, certain skills-based trainings have been found to be helpful in helping women break into male-dominated sectors. For example, Croke, Goldstein, and Holla (2018) found that an information and communications technology (ICT) training, which gave access to 85 hours of classroom- based training spread across 10 weeks, resulted in university graduates being 26 percent more likely to work in the ICT sector. Providing a monetary incentive to apply to a job can also increase the quality of the applicants (Abebe et al, 2021), and may encourage women to apply to jobs where occupational segregation is high. Organizing job fairs can help firms and potential employees learn about what types of potential employees and jobs are available and help them in their search for a formal job (Abebe et al, 2017). In addition, providing information about earnings and opportunities across different sectors can increase women’s participation in lucrative sectors (Del Carpio and Guadalupe, 2022; Wiswall and Zafar, 2015), with a study in the Republic of Congo showing positive results (Gassier et al, 2022). 18 Involving men in this effort has been shown to be crucial. Engaging couples in joint planning around male- dominated activities, such as export-crop farming, can lead to increased female engagement in those activities (Donald et al., 2022). Moreover, providing male mentors to women is a promising avenue: having a male role model is significantly associated with women crossing over into male-dominated sectors in pooled data from 97 countries (World Bank 2022). 4.2 Increase Women’s Access to Productive Inputs Financial Inclusion and Business Capital Increasing women’s financial inclusion and capital to run their firms and farms is an important policy priority to close gender gaps in Burkina Faso, though it is worth noting that the country is already ahead in terms of financial inclusion relative to its neighboring countries. 14 One way to do this is by encouraging the formation and expansion of Village Savings and Loan Associations, which have been shown to be effective across contexts (Bass et al. 2014). More broadly, providing women with secure savings mechanisms (including through digital/mobile technology) can help women in generating revenue and saving for working capital for future investment (Bastian et al., 2018). A study in Côte d’Ivoire, for example, found that offering a private commitment savings account to female agro-processing workers led to an 11% increase in their earnings, primarily by increasing their attendance at work (Carranza et al, 2022). Furthermore, making it easier for women to obtain loans can increase women’s liquidity and investment. Blattman et al (2016) showed that providing microloans along with business skills training helped to increase the incomes of ultra-poor women. Since women may have lower creditworthiness using traditional scoring mechanisms (due to lower assets, profitability, and tenure), experimenting with new methods to assess creditworthiness is a promising avenue to increase women’s access to capital. For example, using psychometric credit scoring as a substitute for loan collateral was found to have positive impacts on women’s credit, and survival of their firms during the COVID-19 pandemic and conflict in Ethiopia (Alibhai et al., 2022). Bossuroy et al. (2022) also find that providing group savings promotion in the context of a multi-faceted ‘graduation’ program increased profits in women-owned businesses, along with their psycho-social well-being. Offering cash-flow based loans to procure machineries and other productive inputs, such as hired labor, is a further policy priority worth testing. 14 https://www.worldbank.org/en/publication/globalfindex/Report. 19 Lastly, ensuring that women’s ownership of national ID cards—often required to access different financial products—is at least equal to that of men is important to make progress on this priority. Exploring equal civil registration by gender as a precursor to ID cards will be an important first step. Use and Returns to Labor Our findings consistently showed that women managers had fewer men (both from within and from outside of the households) working as employees, and that this widened the gender gap across outcomes. Providing women with cash transfers to help them hire labor is a recommended policy option (Daidone et al., 2015), while providing households with labor-saving technology can help ease some of women’s time constraints. A recent RCT that provided oxen to farmers in Côte d’Ivoire increased the use of different complementary inputs and reduced the farm labor of women of the household (Brudevold-Newman et al., 2023). Providing subsidies to rent agricultural equipment can also increase the use of tools in agricultural production (Caunedo and Kala, 2022) and help free up women’s time. Lastly, one of the reasons for women managers’ low use of male labor may be prevailing gender norms in Burkinabè society that make men reluctant to work under a female manager. One reason could be that men assume that a female manager would be not as productive as a male manager, resulting in lower earnings, and so men employees may not want to work under women managers. This hypothesis merits further investigation, as well as the piloting of gender norms trainings to increase men’s willingness to work for women. Use of Modern Agricultural Inputs Women farmers tend to use lower levels of fertilizers, pesticides and farming equipment. An absence of these inputs is associated with a lower level of farm revenue on female-controlled farms. One way to tackle this issue is by providing subsidies for technology adoption—which can not only lead to an increase in harvest values, but also a sustained use of that technology even after the subsidy is not provided (Carter et al, 2021, Fishman et al, 2019). Accessing credit has also allowed farmers to use more chemical fertilizers and higher-yielding seeds (Ndegwa et al, 2022), while the availability of training facilities helped to increase horticultural income among Kenyan farmers (Shimizutani et al, 2021). Agricultural extension services can also be used to inform farmers about the proper use of inputs. Sending video messages to farmers about the proper way to grow crops led to better seed selection and inputs (Vandevelde et al., 2021). Arouna et al (2021) has shown that providing advice to the farmers one-on-one can increase farm yield and profit. Maertens et al. (2021) has demonstrated that showing farmers effective plot cultivation over a season can increase technology adoption by farmers. Lastly, female farmers can be more likely to adopt a technology from female extension agents (Kondylis et al., 2016)—while the joint 20 couples’ training in Côte d’Ivoire mentioned above (Donald et al., 2022) led to a higher use of agricultural inputs. Returns to Land One of the main drawbacks faced by women is the low returns to their farmland. One credible way to tackle this issue is through large-scale land regularization and certification programs. In Benin, formalizing land ownership led to more investment on agricultural lands, with women leaving land fallow for longer (so that the soil regains nutrients) and spending less on land security (Goldstein, et al., 2018). In Malawi, an erosion (strengthening) of women’s land rights was found to reduce (increase) investment on land (Deininger et al., 2021). An emerging policy option to increase women’s returns to land is through conditionally subsidized land registration. An ongoing experiment in rural Uganda conducted by the Gender Innovation Lab offered freehold land titles to rural households. Simply requiring the presence of the wife during their program interaction or making the benefits of titling land in both spouses’ names salient to the husbands can shift demand away from solo titling (i.e., in men’s names only) by up to 83%, half of which gets channeled towards joint titling (i.e., in both spouses’ names). Making the land title offer conditional on joint titling was even more powerful: fully shifting demand away from solo titling and towards joint titling. Lastly, strengthening women’s property rights through civil marriage is a frontier policy option, currently being tested in Côte d’Ivoire (Donald et al. 2020). 4.3 Increase Women’s Physical Security and Household Agency Knowledge of Rights and Laws A lack of knowledge and understanding of the law can hamper the productive capabilities of women. Teaching girls about civics in school makes them more intolerant of domestic violence (Friedman et al., 2016). Informing girls about empowerment not only increased their propensity of being self-employed, but also reduced the likelihood of adolescent pregnancy and early marriage (Bandiera et al., 2020). Clear understanding of individual rights and legal responsibilities has been shown to decrease bribe payments and instances of GBV (Croke et al. 2020). Reforming remaining gender-regressive laws (mobility, factory jobs, and work in hazardous jobs) that do not provide equal treatment for women—and disseminating information on these changes to the broader population—may also help in strengthening women’s agency in Burkina Faso. 21 Reproductive Agency Increasing reproductive agency through the provision of life skills training to girls has been shown to be an effective strategy. Adolescent girls participating in the empowerment programs mentioned above improved decision-making ability and control of their bodies, reduced rates of childbirth, cohabitation, and having nonconsensual sex, and were more likely to engage in income-generating activities (Bandiera et al. 2019; Bandiera et al. 2020). Encouraging school enrollment, delayed marriage and delayed birth through unconditional and conditional cash transfers have also been shown to be effective. One study from Malawi found that both conditional cash transfers (CCTs) and unconditional cash transfers (UCTs) resulted in fewer girls dropping out of school, though UCTs resulted in significantly fewer early marriages and early births than either the CCT or the control group (Baird et al. 2011). Care and Domestic Work In Burkina Faso, the fertility rate remains high, and the bulk of the childcare and domestic work falls on the shoulders of mothers. One approach to tackling this issue is by focusing on intra-household dynamics. Specific couples’ interventions, such as Engaging Men Through Accountable Practice (EMAP) in the DRC, and Bandebereho in Rwanda, have been shown to increase men’s involvement in household and childcare activities (Doyle et al. 2018; Vaillant et al. 2020). Another approach is to improve access to childcare services, either by subsidizing provision to existing centers, or by opening centers where previously they did not exist. Recent evidence from Burkina Faso shows that mobile childcare centers for women participating in a public works project improved both their employment and financial outcomes, while child development scores also increased (Ajayi et al. 2022). In Uganda, subsidizing childcare led to a 44 percent increase in household income, driven by an increase in mothers’ business revenues and fathers’ wage earnings. Lastly, a recent impact evaluation in the Democratic Republic of the Congo, which tested the opening of new childcare centers in rural areas, found that both women and their husbands increased their engagement in commercial activities, leading to gains in agricultural productivity, household income and women’s subjective well-being (Donald and Vaillant 2022). 5. Conclusion Existing research and data on Burkina Faso document high gender inequality in the country. However, gender gaps in economic outcomes and the root causes behind them have not been sufficiently explored. This paper investigates the underlying reasons for gender gaps in earnings in Burkina Faso, using recent household survey data to understand their change over time and the factors associated with them. 22 We find a 10-percentage point gender gap in labor force participation for paid work, and a 20-percentage point gender gap across all work. This gender gap is slightly lower among younger adults compared to older adults and is concentrated in urban areas. Importantly, this gender gap is not linked to educational attainment. Rather, it is driven by women facing lower economic benefit from marriage and young children in the household—indicating that social norms regarding women’s role primary function as a caregiver are what underlie this gap. For wage earnings, the magnitude of the estimated gap depends primarily on whether women’s unpaid work is considered. If we restrict to paid work, women earn 30% less than men—a sizeable amount, but still the smallest gender gap across all outcomes we examine. If, however, we include women’s unpaid work, the gender gap rises to a striking 82%: the largest among all gender gaps. These gaps have increased over time, particularly for young adults, and are largest in the agriculture sector due to women’s role as contributing family workers on household land. This again reinforces how one of the major gender inequalities is that women are more likely than men not to be paid at all for their work. Unlike for labor force participation, differences in characteristics between men and women explain most of the gender gap in wage earnings. In particular, a large portion of the gender wage gap is due to men’s holding higher-skilled jobs and working in managerial positions relative to women. Women also tend to work more often than men as unpaid workers in a family business or farm, and this increases the gender wage gap. Increasing women’s skills and providing them with information on wages and opportunities in more lucrative sectors can help lower the gender wage gap. The gender gap in business revenues is 61%, with no significant differences across time, age groups or urban vs. rural areas. The most important driver of the gender gap in business revenues is female entrepreneurs’ low use of male household labor, with male entrepreneurs using 42 times more male labor a year compared to female entrepreneurs. The gender gap in business revenues can also be explained by differences in financial account and phone ownership, as well as differences in utilities such as use of electricity. Helping women with the means to substitute for this lack of male household labor with external hired labor, shifting norms around men working for women and husbands working as contributing family workers for their wives, and providing women with secure and private ways to save could help lower the gender gap in business revenues. In the agriculture sector, we estimate the gender gap to be 61% for harvest values. This gap is not driven by gender differences in the likelihood of selling crops, since the gender gap in the value of crop sales is virtually the same at 62%. The gap in harvest value has increased by 5 percentage points since 2014 and is similar across age groups. In both cases, much of the gender gap is being driven by women’s low use of machinery, fertilizer and pesticides, as well as the fact that they are less likely to grow higher-value export 23 crops. Just like in the business sector, we also find that men tend to work much less frequently under women. Providing women with the financial means and the information to increase their use of these inputs, and encouraging men to increase their labor provision to women-run farms and firms, can help close gender gaps in Burkina Faso. One overall finding from our analysis is that gender gaps in economic outcomes have either stayed the same or worsened between 2014 and 2019. This calls for urgent action, and underlines that women need targeted support to help reduce gender earnings gaps, but also that engaging men must be part of the answer. Training women to acquire new skills—including socio-emotional skills and entrepreneurial orientation—, providing them with information on earnings across sectors, and engaging men to change norms and behaviors around gendered occupations can bolster women’s inclusion in higher-value sectors. Promoting cash transfers, secure savings mechanisms and exploring new methods to evaluate creditworthiness, while also expanding the use of digital technology for agricultural extension services and providing women with subsidized labor, fertilizer and machinery can help close the gender gap in productive inputs. Lastly, increasing the number of available childcare centers, encouraging school enrollment, delayed marriage and delayed birth through cash transfers, reforming remaining gender-regressive laws, and increasing men's participation in household and childcare responsibilities through media messaging and couples’ trainings can help increase women’s agency and physical security. A multi-pronged policy approach that tackles the variety of root causes underlying the gender earnings gap is recommended and should be part of Burkina Faso’s path toward increased growth and shared prosperity. Our analysis points to a menu of policy options for boosting growth through gender parity in Burkina Faso, but it also highlights areas where there is a need for more rigorous testing—and comparison—of interventions that work to close gender gaps. More high-quality data collection and research is acutely needed to move the needle for women’s economic empowerment in Burkina Faso. 24 6. Bibliography Abebe, G. T., Caria, S., Fafchamps, M., Falco, P., Franklin, S., Quinn, S., & Shilpi, F. (2017). Job fairs: Matching firms and workers in a field experiment in Ethiopia. World Bank Policy Research Working Paper, (8092). Adoho, F., Chakravarty, S., Korkoyah, D. T., Lundberg, M. K., & Tasneem, A. (2014). The impact of an adolescent girls employment pro gram: The EPAG project in Liberia. World Bank Policy Research Working Paper, (6832). Ajayi, K. F., Dao, A., & Koussoube, M. E. J. (2022). The Effects of Childcare on Women and Children: Evidence from a Randomized Evaluation inBurkina Faso (No. 10239). The World Bank. Aker, J. C., & Ksoll, C. (2016). Can mobile phones improve agricultural outcomes? Evidence from a randomized experiment in Niger. Food Policy, 60, 44-51. Alfonsi, L., Bandiera, O., Bassi, V., Burgess, R., Rasul, I., Sulaiman, M., & Vitali, A. (2020). Tackling youth unemployment: Evidence from a labor market experiment in Uganda. Econometrica, 88(6), 2369- 2414. Alibhai, S., Buehren, N., Frese, M., Goldstein, M., Papineni, S., & Wolf, K. (2019). Full esteem ahead? Mindset-oriented business training in Ethiopia. Mindset-Oriented Business Training in Ethiopia (June 17, 2019). World Bank Policy Research Working Paper, (8892). Alibhai, S., Cassidy, R., Goldstein, M., & Papineni, S. (2022). Evening the Credit Score? Impact of Psychometric Loan Appraisal for Women Entrepreneurs. Arouna, A., Michler, J. D., Yergo, W. G., & Saito, K. (2021). One size fits all? Experimental evidence on the digital delivery of personalized extension advice in Nigeria. American Journal of Agricultural Economics, 103(2), 596-619. Bandiera, O., Buehren, N., Goldstein, M., Rasul, I., & Smurra, A. (2019). Empowering Adolescent Girls in a Crisis Context (No. 32115). The World Bank Group. Bandiera, O., Buehren, N., Burgess, R., Goldstein, M., Gulesci, S., Rasul, I., & Sulaiman, M. (2020). Women’s empowerment in action: evidence from a randomized control trial in Africa. American Economic Journal: Applied Economics, 12(1), 210-59. Baird, S., McIntosh, C., & Özler, B. (2011). Cash or condition? Evidence from a cash transfer experiment. The Quarterly journal of economics, 126(4), 1709-1753. Bass, J., Bolton, P., Murray, S., Cole, G., Robinette, K., Poulton, C., ... & Annan, J. (2014). Study of effectiveness of a social-economic intervention for sexual violence survivors in Eastern DRC. Bastian, G., Bianchi, I., Goldstein, M., & Montalvao, J. (2018). Short-term impacts of improved access to mobile savings, with and without business training: Experimental evidence from Tanzania. Documents de travail, 478. Berry, J., Karlan, D., & Pradhan, M. (2018). The impact of financial education for youth in Ghana. World Development, 102, 71-89. 25 Bjorvatn, K., Ferris, D., Gulesci, S., Nasgowitz, A., Somville, V., & Vandewalle, L. (2022). Childcare, labor supply, and business development: Experimental evidence from Uganda (No. 17243). CEPR Discussion Papers. Blattman, Christopher, Eric P. Green, Julian Jamison, M. Christian Lehmann, and Jeannie Annan. 2016. "The Returns to Microenterprise Support among the Ultrapoor: A Field Experiment in Postwar Uganda." American Economic Journal: Applied Economics, 8 (2): 35-64. Blattman, C., Fiala, N., & Martinez, S. (2020). The Long-Term Impacts of Grants on Poverty: Nine-Year Evidence from Uganda's Youth Opportunities Program. American Economic Review: Insights, 2(3), 287- 304. Blinder A. S. 1973. Wage discrimination: Reduced form and structural estimates. Journal of Human Resources 8: 436–455. Breda, T., Grenet, J., Monnet, M., & Van Effenterre, C. (2021). Do female role models reduce the gender gap in science? Evidence from French high schools. Brudevold-Newman, A., Donald, A., & Rouanet, L. (2023). Steered Away from the Fields: Short-Term Impacts of Oxen on Agricultural Production and Intra-household Labor Supply. Bossuroy, T., Goldstein, M., Karimou, B., Karlan, D., Kazianga, H., Parienté, W., ... & Wright, K. A. (2022). Tackling psychosocial and capital constraints to alleviate poverty. Nature, 605(7909), 291-297. Buehren, N., Chakravarty, S., Goldstein, M., Slavchevska, V., & Sulaiman, M. (2017). Adolescent Girls’ Empowerment in Conflict-Affected Settings: Experimental Evidence from South Sudan. Working Paper. Agricultural Production and Intra-Household Labor Supply Campos, F., Frese, M., Goldstein, M., Iacovone, L., Johnson, H. C., McKenzie, D., & Mensmann, M. (2017). Teaching personal initiative beats traditional training in boosting small business in West Africa. Science, 357(6357), 1287-1290. Carranza, E., Donald, A., Grosset, F., & Kaur, S. (2022). The Social Tax: Redistributive Pressure and Labor Supply (No. w30438). National Bureau of Economic Research. Carter, Michael, Rachid Laajaj, and Dean Yang. 2021. "Subsidies and the African Green Revolution: Direct Effects and Social Network Spillovers of Randomized Input Subsidies in Mozambique." American Economic Journal: Applied Economics, 13 (2): 206-29. Caunedo, J., & Kala, N. (2021). Mechanizing agriculture (No. w29061). National Bureau of Economic Research. Chegere, M. J., Eggert, H., & Söderbom, M. (2022). The Effects of Storage Technology and Training on Postharvest Losses, Practices, and Sales: Evidence from Small-Scale Farms in Tanzania. Economic Development and Cultural Change, 70(2), 000-000. Clark, S., Paul, M., Aryeetey, R., & Marquis, G. (2018). An assets-based approach to promoting girls' financial literacy, savings, and education. Journal of adolescence, 68, 94-104. Croke, K., Goldstein, M., & Holla, A. (2018). Can Job Training Decrease Women's Self-Defeating Biases? Experimental Evidence from Nigeria (No. 30495). The World Bank Group. 26 Croke, K., Garcia Mora, M. E., Goldstein, M., Mensah, E. R., & O'Sullivan, M. (2020). Up Before Dawn: Experimental Evidence from a Cross-Border Trader Training at the Democratic Republic of Congo? Rwanda Border. Rwanda Border (January 27, 2020). World Bank Policy Research Working Paper, (9123). Daidone, S., Davis, B., Dewbre, J., González-Flores, M., Handa, S., Seidenfeld, D., & Tembo, G. (2014). Zambia’s Child Grant Programme: 24-month impact report on productive activities and labour allocation. Food and Agriculture Organization of the United Nations, Rome. De Hoop, J., Groppo, V., & Handa, S. (2020). Cash transfers, microentrepreneurial activity, and child work: Evidence from Malawi and Zambia. The World Bank Economic Review, 34(3), 670-697. Del Carpio, L., & Guadalupe, M. (2022). More women in tech? Evidence from a field experiment addressing social identity. Management Science, 68(5), 3196-3218. Deininger, K., Xia, F., Kilic, T., & Moylan, H. (2021). Investment impacts of gendered land rights in customary tenure systems: Substantive and methodological insights from Malawi. World Development, 147, 105654. DHS (2010). Institut National de la Statistique et de la Demographie (INSD) and ICF International. 2012. 2010 Enquête Démographique et de Santé et à Indicateurs Multiples du Burkina Faso. Calverton, MD: ICF International. Donald, A., & Vaillant, J. (2022). Experimental Evidence on Rural Childcare Provision. Donald, A., Goldstein, M., & Rouanet, L. (2022). Two Heads Are Better Than One. Donald, A., Goldstein, M., Hartman, A., La Ferrara, E., O'Sullivan, M., & Stickler, M. (2020). What’s Mine is Yours (No. 34081). The World Bank Group. Doyle, K., Levtov, R. G., Barker, G., Bastian, G. G., Bingenheimer, J. B., Kazimbaya, S., ... & Shattuck, D. (2018). Gender-transformative Bandebereho couples’ intervention to promote male engagement in reproductive and maternal health and violence prevention in Rwanda: Findings from a randomized controlled trial. PloS one, 13(4), e0192756. Dupas, Pascaline, Dean Karlan, Jonathan Robinson, and Diego Ubfal. 2018. "Banking the Unbanked? Evidence from Three Countries." American Economic Journal: Applied Economics, 10 (2): 257-97. Esteve-Volart, B. (2009). Gender discrimination and growth: theory and evidence from India. Mimeo: York University. Fishman, R., Smith, S. C., Bobić, V., & Sulaiman, M. (2019). Can Agricultural Extension and Input Support Be Discontinued? Evidence from a Randomized Phaseout in Uganda. The Review of Economics and Statistics, 1-45. Galor, O., & Weil, D. N. (1999). From Malthusian stagnation to modern growth. American Economic Review, 89(2), 150-154. Gassier, M., Rouanet, L. M., & Traore, L. (2022). Addressing Gender-Based Segregation through Information: Evidence from a Randomized Experiment in theRepublic of Congo (No. 9934). The World Bank. 27 Goldstein, M., Houngbedji, K., Kondylis, F., O'Sullivan, M., & Selod, H. (2018). Formalization without certification? Experimental evidence on property rights and investment. Journal of Development Economics, 132, 57-74. Haddad, L., & Reardon, T. (1993). Gender bias in the allocation of resources within households in Burkina Faso: a disaggregated outlay equivalent analysis. The Journal of Development Studies, 29(2), 260-276. Handa, S., Natali, L., Seidenfeld, D., Tembo, G., Davis, B., & Zambia Cash Transfer Evaluation Study Team. (2018). Can unconditional cash transfers raise long-term living standards? Evidence from Zambia. Journal of Development Economics, 133, 42-65. Hardy, M., Mbiti, I. M., Mccasland, J. L., & Salcher, I. (2019). The apprenticeship-to-work transition: experimental evidence from Ghana. World bank policy research working paper, (8851). Hsieh, C. T., Hurst, E., Jones, C. I., & Klenow, P. J. (2019). The allocation of talent and us economic growth. Econometrica, 87(5), 1439-1474. Kieran, C., Gray, B., & Gash, M. (2018). Understanding Gender Norms in Rural Burkina Faso: A Qualitative Assessment. Grameen Foundation, Washington, DC. Kitagawa E. M. 1955. Components of a difference between two rates. Journal of the American Statistical Association 50: 1168–1194 Kondylis, F., Mueller, V., Sheriff, G., & Zhu, S. (2016). Do female instructors reduce gender bias in diffusion of sustainable land management techniques? Experimental evidence from Mozambique. World Development, 78, 436-449. Maertens, A., Michelson, H., & Nourani, V. (2021). How do farmers learn from extension services? Evidence from Malawi. American Journal of Agricultural Economics, 103(2), 569-595. McKay, S. (1998). The effects of armed conflict on girls and women. Peace and Conflict, 4(4), 381-392. Mishra, K., Gallenstein, R. A., Sam, A. G., Miranda, M. J., Toledo, P., & Mulangu, F. (2021). Do index- insured loans spur technology adoption? Evidence from a randomized control trial in northern Ghana. Am. J. Agric. Econ, 103, 923-943. Natali, L., Handa, S., Peterman, A., Seidenfeld, D., & Tembo, G. (2018). Does money buy happiness? Evidence from an unconditional cash transfer in Zambia. SSM-Population Health, 4, 225-235. Ndegwa, M. K., Shee, A., Ward, P. S., Liu, Y., Turvey, C. G., & You, L. The Impact of Risk Contingent Credit and Traditional Credit on Smallholders’ Agricultural Investment, Productivity and Welfare: Experimental Evidence from a Randomized Control Trial in Rural Kenya. Productivity and Welfare: Experimental Evidence from a Randomized Control Trial in Rural Kenya. Nikièma, B., Haddad, S., & Potvin, L. (2008). Women bargaining to seek healthcare: norms, domestic practices, and implications in rural Burkina Faso. World Development, 36(4), 608-624. O’brien, D. Z., & Rickne, J. (2016). Gender quotas and women's political leadership. American Political Science Review, 110(1), 112-126. Oaxaca R. 1973. Male–female wage differentials in urban labor markets. International Economic Review 14: 693–709. 28 Okedele, A. (2021). Women, quotas, and affirmative action policies in Africa. The Palgrave Handbook of African Women's Studies, 449-463. Omotilewa, O. J., Ricker‐Gilbert, J., & Ainembabazi, J. H. (2019). Subsidies for agricultural technology adoption: Evidence from a randomized experiment with improved grain storage bags in Uganda. American Journal of Agricultural Economics, 101(3), 753-772. PASEC. (2015). Education system performance in francophone sub-Saharan Africa: Competencies and learning factors in primary education PASEC 2014. Dakar: CONFEMEN “Parline – global data on national parliaments.” Inter-Parliamentary Union. 2023. https://data.ipu.org/ Porter, C., & Serra, D. (2020). Gender differences in the choice of major: The importance of female role models. American Economic Journal: Applied Economics, 12(3), 226-254. Santos Silva, M., & Klasen, S. (2021). Gender inequality as a barrier to economic growth: a review of the theoretical literature. Review of Economics of the Household, 19, 581-614. Schomer, I., & Hammond, A. (2020). Stepping Up Women's STEM Careers in Infrastructure. Shimizutani, S., Taguchi, S., Yamada, E., & Yamada, H. (2021). The Impact of “Grow to Sell” Agricultural Extension on Smallholder Horticulture Farmers: Evidence from a Market-Oriented Approach in Kenya. Available at SSRN 3969797. Stoeffler, Q., Carter, M., Guirkinger, C., & Gelade, W. (2022). The spillover impact of index insurance on agricultural investment by cotton farmers in burkina faso. The World Bank Economic Review, 36(1), 114- 140. “UNESCO Institute for Statistics” United Nations Educational, Scientific and Cultural Organization. 2020. http://data.uis.unesco.org/Index.aspx UN (2019). Mutilations Génitales Féminines & Mariage d’Enfants. https://www.unicef.org/burkinafaso/media/2221/file/Rapport%20sur%20les%20mutilations%20g%C3% A9nitales%20f%C3%A9minines%20et%20le%20mariage%20des%20enfants.pdf United Nations Development Programme, Human Development Report (2021). "Gender Development Index." Retrieved from http://hdr.undp.org/en/content/gender-development-index-gdi.. Vaillant, J., Koussoubé, E., Roth, D., Pierotti, R., Hossain, M., & Falb, K. L. (2020). Engaging men to transform inequitable gender attitudes and prevent intimate partner violence: a cluster randomised controlled trial in North and South Kivu, Democratic Republic of Congo. BMJ Global Health, 5(5), e002223. Van den Bold, M., Dillon, A., Olney, D., Ouedraogo, M., Pedehombga, A., & Quisumbing, A. (2015). Can integrated agriculture-nutrition programmes change gender norms on land and asset ownership? Evidence from Burkina Faso. The journal of development studies, 51(9), 1155-1174. Van der Windt, P., Humphreys, M., & de la Sierra, R. S. (2018). Gender quotas in development programming: Null results from a field experiment in Congo. Journal of Development Economics, 133, 326-345. Vandevelde, S., Van Campenhout, B., & Walukano, W. (2021). Accounting for spillovers in assessing the effectiveness of video messages to improve potato seed quality: evidence from Uganda. The Journal of Agricultural Education and Extension, 27(4), 503-534. 29 Weber, Michael, 2018. "Burkina Faso Jobs Diagnostic: Overview and Suggestions for a Strategic Framework for Jobs," Jobs Group Papers, Notes, and Guides 30672458, The World Bank. Wiswall, M., & Zafar, B. (2015). How do college students respond to public information about earnings?. Journal of Human Capital, 9(2), 117-169. “World Development Indicators.” Washington, D.C. :The World Bank., n.d. World Bank. (2021). People's Perspectives on ID and Civil Registration in Rwanda. World Bank. World Bank (2013). “Reintegration Opportunities for Vulnerable Households of Ex-Combatants Demobilized through PNDDR : An Overview of the Socioeconomic Situation of Pilot Project Beneficiaries Two Years Following Project Closure - Maniema, Democratic Republic of Congo,” https://openknowledge.worldbank.org/ handle/10986/17859?locale-attribute=en. WHO 2019. Trends in Maternal Mortality: 2000 to 2017. Geneva, World Health Organization, 2019 30 7. Tables Table 1 Panel A: Gender Differences in LFP and Earnings in Burkina Faso (2019) (1) (2) (3) (4) (5) (6) (7) Wage Paid Wage Business Harvest Paid LFP All LFP Earnings Earnings Revenues Values Crop Sales Female -0.203*** -0.101*** -31,850.73*** -35,310.99*** -0.936*** -0.949*** -0.965*** (0.009) (0.008) (2,128.868) (4569.264) (0.124) (0.033) (0.049) 39,014.52*** 115,483.9*** 11.849*** 11.983*** 11.776*** Constant (2,056.267) (2475.621) (0.105) (0.015) (0.022) % gap 20 pp 10 pp 82% 30% 61% 61% 62% Observations 22,666 22,666 10,616 2,206 4,646 13,578 4,952 Note: Standard errors are in parenthesis. ***/**/* indicate statistical significance at the 1/5/10 percent level respectively. % gap values for business revenue, harvest value and crop sales are obtained with the exp(b)-1 transformation (Halvorsen and Palmquist (1980)), while that of LFP were obtained from the solving for the logit equation. Wage earnings was in CFAs, so the wage gap was calculated by dividing the female dummy coefficient with the constant. For the paid LFP and All LFP results, the marginal effects are reported, so they do not have a constant. Table 1 Panel B: Summary Statistics (2019) All Men Women VARIABLES mean sd mean sd mean sd Variables used in Labor Force Regression Decomposition Overall LFP 0.798 0.402 0.853 0.354 0.782 0.432 LFP (Paid Work Only) 0.615 0.487 0.726 0.446 0.523 0.500 Individual Lives in an Urban Area 0.310 0.463 0.325 0.468 0.298 0.457 Individual Age 32.93 13.32 33.04 13.67 32.84 13.02 Individual Age Squared 1,262 984.9 1,279 1,011 1,248 962.6 Number of Adults in HH 4.465 2.816 4.490 2.822 4.444 2.810 Number of Children Aged 0-5 in HH 1.154 1.328 1.071 1.254 1.223 1.384 Individual own a Financial Account 0.275 0.446 0.404 0.491 0.166 0.372 Individual is Married 0.661 0.473 0.588 0.492 0.723 0.448 Individual is Divorced/Married/Separated 0.048 0.214 0.014 0.116 0.077 0.267 Individual Completed Secondary Schooling 0.079 0.270 0.106 0.308 0.057 0.232 Individual owns a Vehicle 0.514 0.500 0.649 0.477 0.401 0.490 Wealth PCA - Based on Home Condition -0.156 1.596 -0.143 1.615 -0.167 1.581 Wealth PCA - Based on Appliances Ownership 0.008 2.033 0.030 2.019 -0.011 2.044 Has a Disability 0.0008 0.0275 0.0005 0.0227 0.0009 0.031 Is Female 0.545 0.498 Number of Observations 22,666 10,270 12,396 Variables used in Monthly Wage Decomposition Monthly Income (In CFA) 19,573 59,167 39,029 82,790 7,174 31134 Individual Lives in an Urban Area 0.237 0.425 0.368 0.482 0.154 0.361 Individual Age 29.95 12.06 26.48 10.71 32.16 12.35 31 Individual Age Squared 1,042 846.0 815.6 711.8 1,187 891.8 Number of Adults in HH 4.809 2.972 5.123 3.102 4.608 2.868 Number of Children Aged 0-5 in HH 1.206 1.393 0.873 1.187 1.418 1.472 Individual owns a Financial Account 0.220 0.415 0.392 0.488 0.111 0.314 Individual is Married 0.607 0.488 0.348 0.476 0.773 0.419 Individual is Divorced/Married/Separated 0.033 0.179 0.012 0.110 0.046 0.210 Individual Completed Secondary Schooling 0.082 0.275 0.147 0.354 0.041 0.199 Individual owns a Vehicle 0.434 0.496 0.496 0.500 0.394 0.489 Individual Works as Apprentice 0.030 0.171 0.049 0.216 0.0179 0.133 Individual Works as Manager 0.058 0.234 0.103 0.304 0.029 0.169 Individual Works as Skilled Worker 0.037 0.188 0.069 0.253 0.016 0.127 Individual Works as Family Worker 0.748 0.434 0.570 0.495 0.862 0.345 Individual Works in Agri. Sector 0.746 0.435 0.572 0.495 0.858 0.349 Individual Woks in Man. Sector 0.056 0.230 0.117 0.321 0.018 0.132 Individual Works in Service Sector 0.197 0.398 0.312 0.463 0.124 0.330 Is Female 0.611 0.488 Number of Observations 10,591 4,295 6,296 Variables used in Monthly Business Revenues Regression Decomposition Asinh(Revenue) 11.34 2.710 11.85 3.003 10.91 2.353 Individual Lives in an Urban Area 0.398 0.489 0.400 0.490 0.396 0.489 Individual Age 38.62 10.84 39.41 10.50 37.95 11.09 Individual Age Squared 1,609 882.7 1,663 865.3 1,563 894.8 Number of Adults in HH 3.921 2.657 3.862 2.572 3.970 2.726 Number of Children Aged 0-5 in HH 1.163 1.257 1.279 1.270 1.065 1.237 Individual owns a Financial Account 0.435 0.496 0.633 0.482 0.268 0.443 Individual is Married 0.867 0.340 0.885 0.319 0.851 0.356 Individual is Divorced/Married/Separated 0.069 0.253 0.0267 0.161 0.104 0.305 Individual Completed Secondary Schooling 0.036 0.188 0.0567 0.231 0.019 0.139 Individual owns a Vehicle 0.682 0.466 0.879 0.326 0.516 0.500 Business has Electricity 0.079 0.271 0.140 0.347 0.029 0.168 Business has Running Water 0.014 0.119 0.025 0.156 0.006 0.075 Business has a Phone 0.888 0.316 0.945 0.228 0.839 0.367 Business is Registered 0.033 0.180 0.065 0.247 0.006 0.081 Business Maintains Accounting Records 0.054 0.226 0.082 0.274 0.030 0.172 Business has Machines 0.132 0.339 0.196 0.397 0.078 0.269 Number of Men Employees 0.372 3.272 0.784 4.793 0.025 0.326 Number of Women Employees 0.170 7.256 0.283 10.72 0.075 0.415 Number of Days HH Female Member Worked 134.0 168.1 23.34 95.62 227.3 159.2 Number of Days HH Male Member Worked 119.6 171.8 254.2 170.1 6.025 42.38 Business is in Food Sector 0.387 0.487 0.243 0.429 0.509 0.500 Business is in Clothing Sector 0.269 0.443 0.254 0.436 0.281 0.450 Business is in Construction Sector 0.126 0.332 0.182 0.386 0.080 0.271 Business is in Service Sector 0.516 0.500 0.615 0.487 0.431 0.495 32 Is Female Business Manager 0.543 0.498 Number of Observations 4,646 2,077 2,569 Variables used in Agric. Output Regression Decomposition Asinh(Harvest Value) 11.79 1.601 11.98 1.638 11.05 1.189 Individual Age 41.58 11.71 42.34 11.47 38.70 12.16 Individual Age Squared 1,866 989.8 1,924 982.1 1,646 988.1 Number of Adults in HH 3.484 2.044 3.633 2.077 2.913 1.804 Number of Children Aged 0-5 in HH 1.461 1.346 1.529 1.327 1.203 1.386 Individual owns a Financial Account 0.323 0.468 0.378 0.485 0.113 0.317 Individual is Married 0.919 0.272 0.956 0.205 0.780 0.414 Individual is Divorced/Married/Separated 0.054 0.225 0.0122 0.110 0.211 0.408 Individual Completed Secondary Schooling 0.007 0.0846 0.008 0.091 0.003 0.052 Individual owns a Vehicle 0.818 0.386 0.899 0.301 0.510 0.500 Household Owns Mechanized Farming Equipment 0.373 0.484 0.410 0.492 0.233 0.423 Area Of Parcel of Land 2.129 15.79 2.248 16.21 1.674 14.04 Area of Parcel of Land Squared 253.7 7,320 267.9 7,721 199.8 5,531 Distance Of Land from Home 24.21 118.4 24.64 132.2 22.57 29.96 Crop Grown for Export 0.043 0.202 0.053 0.224 0.004 0.059 Farm Manager is the Owner 0.613 0.487 0.684 0.465 0.341 0.474 Manager Controls the Sale of Crops 0.323 0.468 0.334 0.472 0.280 0.449 Number Of Days HH Female Member Worked 67.69 90.43 69.60 92.45 60.42 81.89 Number Of Days HH Male Member Worked 64.67 80.25 72.72 84.91 34.02 48.21 Total Days Worked by Non-HH Men 2.219 8.240 2.390 8.749 1.568 5.865 Total Days Worked by Non-HH Women 1.293 4.837 1.309 5.025 1.234 4.044 All Harvest from Parcel Sold 0.099 0.298 0.107 0.309 0.0668 0.250 Parcel Of Land uses Fertilizer 0.323 0.468 0.349 0.477 0.227 0.419 Parcel Of Land uses Pesticides 0.309 0.462 0.338 0.473 0.199 0.399 Is Female Farmer 0.208 0.406 Number of Observations 13,577 10,693 2,884 Variables used in Agric. Output Sold Regression Decomposition Asinh(Output Sold) 11.59 1.397 11.78 1.378 10.82 1.197 Individual Age 41.01 11.56 41.63 11.44 38.43 11.71 Individual Age Squared 1,816 974.5 1,864 973.8 1,614 951.4 Number of Adults in HH 3.579 2.173 3.664 2.164 3.217 2.176 Number of Children Aged 0-5 in HH 1.509 1.329 1.548 1.302 1.343 1.426 Individual owns a Financial Account 0.342 0.474 0.402 0.490 0.0891 0.285 Individual is Married 0.928 0.258 0.955 0.207 0.816 0.387 Individual is Divorced/Married/Separated 0.043 0.203 0.0125 0.111 0.172 0.377 Individual Completed Secondary Schooling 0.005 0.071 0.006 0.077 0.001 0.036 Individual owns a Vehicle 0.848 0.359 0.925 0.263 0.521 0.500 Household Owns Mechanized Farming Equipment 0.536 0.499 0.585 0.493 0.332 0.471 Area Of Plot of Land 2.021 14.78 2.158 14.10 1.444 17.36 33 Area Of Plot of Land Squared 222.6 7,229 203.5 6,876 303.0 222.6 Distance of Land from Home 27.74 138.0 28.05 152.7 26.40 31.52 Crop Grown for Export 0.115 0.319 0.140 0.347 0.009 0.093 Farm Manager is the Owner 0.607 0.488 0.669 0.471 0.346 0.476 Manager Controls the Sale of Crop 0.884 0.321 0.896 0.305 0.830 0.376 Number of Days HH Female Member Worked 68.28 90.89 69.13 92.24 64.72 84.93 Number of Days HH Male Member Worked 63.45 79.77 70.47 84.07 33.97 48.21 Total Days Worked by Non-HH Men 2.180 7.008 2.417 7.585 1.183 3.534 Total Days Worked by Non-HH Women 1.747 6.206 1.862 6.655 1.261 3.738 All Harvest from Plot Sold 0.270 0.444 0.287 0.452 0.198 0.399 Plot of Land uses Fertilizer 0.429 0.495 0.470 0.499 0.256 0.437 Plot of Land uses Pesticides 0.493 0.500 0.537 0.499 0.308 0.462 Is Female Farmer 0.192 0.394 Number of Observations 4,952 3,987 965 34 Table 2: Comparison of the Gender Gaps in Different Sectors Between 2014 and 2019 Gap in 2014 Gap in 2019 Significance of 2014 vs. 2019 Gap LFP 0.12** 0.10** By Age Age 15-29 0.11** 0.09** Age 30-64 0.13** 0.11** ** By Urbanicity Urban 0.12** 0.20** ** Rural 0.12** 0.07** * By Region North 0.23** 0.07** ** Central 0.07** 0.16** ** South 0.03** 0.06** ** West 0.14** 0.12** Monthly Wages 68%** 82%** ** By Age Age 15-29 54%** 76%** ** Age 30-64 82%** 88%** ** By Urbanicity Urban 64%** 49%** ** Rural 58%** 97%** ** By Region North 68%** 94%** ** Central 73%** 62%** ** South 36%** 87%** ** West 75%** 83%** * By Age Business Revenues 62%** 61%** By Age Age 15-29 37%** 65%** Age 30-64 69%** 59%** By Urbanicity Urban 49%** 50%** Rural 62%** 66%** By Region North 11% 74%** ** Central 56%** 39% South 60%** 71%** West 84%** 62%** ** Harvest Value 56%** 61%** ** By Age Age 15-29 48%** 58%** * 35 Age 30-64 59%** 61%** By Region North 37%** 25%** ** Central 44%** 46%** South 57%** 59%** West 38%** 76%** ** Crop Sales 63%** 62%** By Age Age 15-29 70%** 58%** ** Age 30-64 61%** 62%** By Region North 47%** 10% ** Central 51%** 33%** South 59%** 57%** West 70%** 75%** The values were computed using the 2018/2019 EHCVM and the 2014 LSMS of Burkina Faso. * indicate significance at 10 percent, ** at 5 percent. 36 Table 3: Kitagawa-Oaxaca-Blinder Decomposition of Labor Force Participation (1) (2) (3) (4) (5) (6) VARIABLES Paid Work Paid+Unpaid Work Overall Endow. Coeff. Overall Endow. Coeff. Male 0.726*** 0.853*** (0.005) (0.004) Female 0.523*** 0.752*** (0.006) (0.005) Raw Difference 0.203*** 0.101*** (0.008) (0.007) Difference due to -0.009** 0.001 Endowments (0.004) (0.003) Difference due to 0.212*** 0.101*** Coefficients (0.009) (0.007) Individual Lives in an -0.001 0.020*** -0.000 0.013** Urban Area (0.001) (0.007) (0.001) (0.005) Individual Age -0.006 0.750*** 0.000 0.285*** (0.005) (0.140) (0.001) (0.103) Individual Age Squared 0.011 -0.317*** -0.001 -0.157*** (0.009) (0.069) (0.002) (0.049) Number of Adults in HH 0.001 -0.034*** -0.000 0.009 (0.001) (0.012) (0.000) (0.008) Number of Children Aged 0-5 in -0.001 0.012* 0.000 0.012** HH (0.001) (0.007) (0.000) (0.006) Individual owns a Financial Account -0.030 -0.009 0.001 0.008 (0.025) (0.007) (0.005) (0.006) Individual is Married 0.013 0.038** -0.000 0.026** (0.012) (0.017) (0.001) (0.013) Individual is Divorced/Widowed/ 0.006 0.001 -0.000 0.001 Separated (0.006) (0.001) (0.000) (0.001) Individual Completed Secondary 0.004 -0.006** -0.000 -0.005*** Schooling (0.004) (0.003) (0.001) (0.002) Individual owns a Vehicle (Bicycle, -0.006 0.004 0.000 -0.002 Motorcycle, Car) (0.005) (0.010) (0.002) (0.008) Wealth PCA - Based on Home Condition 0.000 -0.000 -0.000 -0.000 (0.000) (0.000) (0.000) (0.000) Wealth PCA - Based on Appliances -0.000 0.000 -0.000 -0.000 Ownership (0.000) (0.001) (0.000) (0.001) Individual has a Disability -0.000 0.000** 0.000 0.000 (0.000) (0.000) (0.000) (0.000) Constant -0.247*** -0.089* (0.065) (0.045) Observations 22,666 22,666 22,666 22,666 22,666 22,666 Robust standard errors in parentheses. *** p<0.01, ** p<0.05, * p<0.1 37 Table 4: Kitagawa-Oaxaca-Blinder Decomposition of Monthly Wage Earnings (1) (2) (3) VARIABLES Overall Endow. Coeff. Male 39,029.396*** (1,582.821) Female 7,173.695*** (475.450) Raw Difference 31,855.701*** (1,652.687) Difference due to 17,597.556*** Endowments (1,386.519) Difference due to 14,258.144*** Coefficients (1,533.097) Individual Lives in an Urban Area 178.834 -2,129.334* (541.726) (1,116.533) Individual Age -261.938 6,386.937 (2,037.440) (26,288.082) Individual Age Squared -687.070 2,807.738 (1,926.942) (13,785.755) Number of Adults in HH -259.638*** -4,575.236** (98.879) (2,163.628) Number of Children Aged 0-5 in -185.325 1,246.540 HH (136.780) (797.302) Individual owns a Financial Account 2,421.506*** 557.291 (583.067) (752.266) Individual is Married -4,268.291*** 9,056.051*** (914.900) (2,378.079) Individual is Divorced/Widowed/ -248.119** -44.469 Separated (115.665) (283.117) Individual Completed Secondary 1,976.755*** -489.806 Schooling (575.774) (776.826) Individual owns a Vehicle (Bicycle, -15.031 -293.616 Motorcycle, Car) (107.220) (1,204.066) Individual Works as Apprentice -619.218*** -542.991** (176.251) (241.723) Individual Works as Manager 8,500.897*** 1,520.036* (845.543) (815.221) Individual Works as Skilled Worker 2,460.303*** -1,041.654*** (391.392) (363.568) Individual Works in the Agriculture Sector 9,344.021*** -37,670.460*** (1,393.355) (5,400.111) Individual Works in the Service Sector -740.131 -3,799.391*** (866.043) (1,129.125) Constant 43,270.510*** (14,906.302) Observations 10,591 10,591 10,591 The reference dummies are the family worker and the manufacturing sector. Robust standard errors in parentheses. *** p<0.01, ** p<0.05, * p<0.1 38 Table 5: The Kitagawa-Oaxaca-Blinder Decomposition of Business Revenues (1) (2) (3) VARIABLES Overall Endow. Coeff. Male 11.849*** (0.105) Female 10.913*** (0.061) Raw Difference 0.936*** (0.121) Difference due to 0.598*** Endowments (0.179) Difference due to 0.338 Coefficients (0.221) Individual Lives in Urban Area 0.002 -0.169* (0.008) (0.089) Individual Age 0.102** -4.747* (0.049) (2.500) Individual Age Squared -0.096** 2.619** (0.044) (1.245) Number of Adults in HH 0.003 -0.074 (0.003) (0.200) Number of Children Aged 0-5 in HH -0.012 0.251* (0.013) (0.135) Individual owns a Financial Account 0.235*** 0.190* (0.040) (0.098) Individual is Married 0.001 0.410 (0.008) (0.384) Individual is Divorced/Married/Separated -0.009 0.011 (0.021) (0.033) Individual Completed Secondary Schooling -0.013 -0.011 (0.023) (0.029) Individual has a Vehicle (bicycle, 0.049 -0.015 motorcycle, car) (0.040) (0.217) Business has Electricity 0.088*** 0.050** (0.025) (0.021) Business has Running Water -0.003 -0.006 (0.027) (0.018) Business has a Phone 0.043** -0.293 (0.019) (0.409) Business Maintains Accounting Records 0.006 0.021 (0.048) (0.019) Business is Registered -0.028 -0.068* (0.022) (0.037) Business has Machines 0.018 0.001 (0.022) (0.041) Number of Men Employees 0.037 -0.002 (0.087) (0.013) Number of Women Employees -0.001 -0.036*** (0.010) (0.009) Number of Days HH Male Member Worked 0.715*** 0.034 (0.142) (0.052) Number of Days HH Female Member -0.446*** -0.016 Worked (0.071) (0.070) Business is in Food Sector -0.065** -0.152* (0.031) (0.091) 39 Business is in Clothing Sector 0.013* -0.120 (0.008) (0.082) Business is in Construction Sector -0.042* -0.053 (0.023) (0.049) Constant 2.512** (1.248) Observations 4,646 4,646 4,646 The reference dummy is the service sector. Robust standard errors in parentheses. *** p<0.01, ** p<0.05, * p<0.1 40 Table 6: The Kitagawa-Oaxaca-Blinder Decomposition of Harvest Value and Crop Sales (1) (2) (3) (4) (5) (6) Harvest Value Crop Sales VARIABLES Overall Endow. Coeff. Overall Endow. Coeff. Male 11.981*** 11.778*** (0.019) (0.027) Female 11.047*** 10.816*** (0.026) (0.046) Raw Difference 0.934*** 0.962*** (0.032) (0.054) Difference due to 0.526*** 0.587*** Endowments (0.031) (0.049) Difference due to 0.408*** 0.375*** Coefficients (0.040) (0.064) Individual Age 0.015 -0.415 0.077* -1.471 (0.030) (0.671) (0.042) (1.067) Individual Age Squared -0.018 0.069 -0.087** 0.763 (0.027) (0.341) (0.039) (0.543) Number of Adults in HH 0.022*** 0.148*** 0.024*** 0.099 (0.006) (0.050) (0.006) (0.081) Number of Children Aged 0-5 0.019*** -0.032 0.004 0.068 In HH (0.004) (0.030) (0.003) (0.048) Individual owns a Financial 0.010 0.000 0.032** -0.067*** Account (0.008) (0.011) (0.014) (0.015) Individual is Married 0.007 0.356* 0.031* 0.836*** (0.016) (0.204) (0.017) (0.226) Individual is Divorced/Married -0.008 0.090* -0.002 0.159*** /Separated (0.022) (0.050) (0.024) (0.045) Individual Completed -0.000 -0.003* 0.002* -0.001 Secondary Schooling (0.001) (0.002) (0.001) (0.001) Individual has a Vehicle 0.011 0.050 -0.035 -0.137 (bicycle, motorcycle, car) (0.016) (0.053) (0.025) (0.087) Household owns Mechanized 0.052*** 0.029 0.127*** 0.014 Agriculture Equipment (0.007) (0.022) (0.014) (0.046) Area of Plot of Land 0.003 0.006 0.013 0.056* (0.003) (0.012) (0.013) (0.029) Area of Plot of Land Squared -0.001 -0.002 0.004 -0.021 (0.002) (0.003) (0.012) (0.019) Land Distance from House 0.000 -0.100*** -0.000 -0.073** (0.000) (0.016) (0.000) (0.028) Crop Grown for Export 0.161*** -0.003 0.106*** -0.005 (0.009) (0.003) (0.011) (0.005) Manager of Farm is the Owner -0.003 0.019 -0.019 -0.018 (0.010) (0.028) (0.015) (0.046) 41 Manager Controls Income from -0.003* 0.668*** 0.012** 0.719*** Plot (0.002) (0.126) (0.005) (0.109) Total Days Worked by HH Men 0.098*** 0.006 0.045*** 0.010 (0.012) (0.023) (0.012) (0.039) Total Days Worked by HH -0.001 -0.140*** -0.001 -0.046 Women (0.002) (0.023) (0.001) (0.035) Total Days Worked by Men 0.013*** -0.003 0.020*** 0.008 (non-HH) (0.003) (0.006) (0.004) (0.015) Total Days Worked by Women 0.000 0.022** 0.002 0.042*** (non-HH) (0.001) (0.009) (0.002) (0.014) All Harvest Sold 0.006*** 0.010 0.039*** 0.005 (0.002) (0.008) (0.008) (0.028) Plot of Land uses 0.073*** 0.004 0.127*** -0.057 Urea/Phosphates/NPK/DAP (0.007) (0.018) (0.014) (0.036) Plot of Land uses 0.069*** 0.036** 0.066*** 0.135*** Pesticides/Fungicides/Herbicides (0.007) (0.018) (0.012) (0.043) Constant -0.408 -0.644 (0.393) (0.516) Observations 13,577 13,577 13,577 4,952 4,952 4,952 Robust standard errors in parentheses, *** p<0.01, ** p<0.05, * p<0.1 42 8. Appendix Tables Appendix Table 1: Urbanicity by Province in Burkina Faso in 2014 and 2019 Province ID Province 2014 2019 1 Bam 0.051 0.090 2 Bazega 0.085 0.070 3 Bougouriba 0.153 0.163 4 Boulgou 0.210 0.142 5 Boulkiemde 0.130 0.253 6 Comoe 0.287 0.233 7 Ganzourgou 0.053 0.065 8 Gnagna 0.029 0.024 9 Gourma 0.109 0.210 10 Houet 0.536 0.685 11 Kadiogo 0.808 0.900 12 Kenedougou 0.115 0.083 13 Kossi 0.047 0.089 14 Kouritenga 0.120 0.306 15 Mouhoun 0.139 0.134 16 Nahouri 0.167 0.150 17 Namentenga 0.026 0.040 18 Oubritenga 0.043 0.121 19 Oudalan 0.016 0.091 20 Passore 0.056 0.083 21 Poni 0.099 0.085 22 Sanguie 0.070 0.059 23 Sanmatenga 0.080 0.096 24 Seno 0.051 0.064 25 Sissili 0.123 0.083 26 Soum 0.055 0.083 27 Sourou 0.055 0.050 28 Tapoa 0.018 0.016 29 Yatenga 0.107 0.140 30 Zoundweogo 0.080 0.087 31 Bale 0.062 0.082 32 Banwa 0.055 0.036 33 Loba 0.079 0.059 34 Komandjoari 0.239 0.160 35 Kompienga 0.089 0.115 36 Koulpelogo 0.110 0.039 37 Kourweogo 0.084 0.059 38 Leraba 0.064 0.064 39 Loroum 0.105 0.080 40 Nayala 0.055 0.143 41 Noumbiel 0.128 0.436 42 Tuy 0.084 0.338 43 Yagha 0.030 0.008 44 Ziro 0.061 0.021 45 Zondoma 0.137 0.099 Average Across Provinces 0.220 0.257 43 Appendix Table 2: Kitagawa-Oaxaca-Blinder Decomposition of Labor Force Participation (Paid+Unpaid Work) by Age Group (1) (2) (3) (4) (5) (6) VARIABLES Age 15-29 Age 30-64 Overall Endow. Coeff. Overall Endow. Coeff. Male 0.752*** 0.938*** (0.008) (0.005) Female 0.667*** 0.828*** (0.008) (0.007) Raw Difference 0.085*** 0.110*** (0.011) (0.008) Difference due to -0.012*** 0.021*** Endowments (0.005) (0.004) Difference due to 0.097*** 0.089*** Coefficients (0.011) (0.010) Individual Lives in an -0.000*** 0.016** -0.002*** 0.021** Urban Area (0.000) (0.008) (0.000) (0.009) Individual Age -0.015 -0.422 0.006*** -0.060 (0.012) (0.537) (0.001) (0.429) Individual Age Squared -0.004 0.289 -0.008*** -0.032 (0.012) (0.277) (0.001) (0.207) Number of Adults in HH 0.002 0.007 0.000 0.015 (0.001) (0.015) (0.000) (0.011) Number of Children Aged 0- -0.003 0.015* -0.000*** 0.004 5 in HH (0.002) (0.009) (0.000) (0.008) Individual owns a Financial 0.015*** 0.007 0.014*** 0.001 Account (0.003) (0.006) (0.003) (0.010) Individual is Married -0.007 0.005 0.002** 0.083*** (0.005) (0.011) (0.001) (0.031) Individual is Divorced/Widowed/ 0.001 -0.000 -0.003** 0.006* Separated (0.001) (0.000) (0.002) (0.003) Individual Completed Secondary -0.008*** -0.010*** 0.002** -0.004 Schooling (0.001) (0.003) (0.001) (0.003) Individual owns a Vehicle 0.002* 0.011 0.012*** 0.004 (Bicycle, Motorcycle, Car) (0.001) (0.008) (0.003) (0.016) Wealth PCA - Based on 0.004*** -0.000 -0.002*** 0.000 Home Condition (0.001) (0.000) (0.001) (0.000) Wealth PCA - Based on 0.001*** -0.001 0.000 0.000 Appliances Ownership (0.000) (0.001) (0.000) (0.001) Individual has a Disability 0.000 -0.000 0.000** 0.000 (0.000) (0.000) (0.000) (0.000) Constant 0.179 0.050 (0.259) (0.220) Observations 10,711 10,711 10,711 11,941 11,941 11,941 Robust standard errors in parentheses. *** p<0.01, ** p<0.05, * p<0.1. 44 Appendix Table 3: Kitagawa-Oaxaca-Blinder Decomposition of Labor Force Participation (Paid+Unpaid Work) by Urban/Rural (1) (2) (3) (4) (5) (6) VARIABLES Urban Rural Overall Endow. Coeff. Overall Endow. Coeff. Male 0.781*** 0.888*** (0.008) (0.005) Female 0.585*** 0.822*** (0.010) (0.006) Raw Difference 0.196*** 0.065*** (0.013) (0.008) Difference due to 0.031*** -0.005 Endowments (0.005) (0.004) Difference due to 0.165*** 0.070*** Coefficients (0.014) (0.008) Individual Age 0.039*** 0.836*** 0.001 -0.086 (0.004) (0.158) (0.002) (0.127) Individual Age Squared -0.028*** -0.476*** 0.005 0.036 (0.003) (0.074) (0.008) (0.060) Number of Adults in HH 0.000 0.001 -0.000 0.024** (0.000) (0.013) (0.000) (0.011) Number of Children Aged 0- -0.001* 0.003 -0.001 0.015** 5 in HH (0.000) (0.007) (0.001) (0.007) Individual owns a Financial 0.019*** -0.005 -0.016 0.008** Account (0.003) (0.013) (0.024) (0.004) Individual is Married -0.001 0.084*** 0.010 0.026 (0.001) (0.019) (0.017) (0.018) Individual is Divorced/Widowed/ -0.001 0.005*** 0.002 0.001 Separated (0.002) (0.002) (0.004) (0.001) Individual Completed Secondary -0.009*** -0.024*** 0.002 -0.000 Schooling (0.001) (0.006) (0.003) (0.001) Individual owns a Vehicle 0.009*** -0.003 -0.008 0.007 (Bicycle, Motorcycle, Car) (0.002) (0.011) (0.013) (0.010) Wealth PCA - Based on 0.002*** -0.018 0.000 -0.008 Home Condition (0.001) (0.012) (0.001) (0.007) Wealth PCA - Based on 0.001*** 0.018 0.000 0.006 Appliances Ownership (0.000) (0.017) (0.000) (0.007) Individual has a Disability -0.000*** -0.000 -0.000 0.000 (0.000) (0.000) (0.000) (0.000) Constant -0.256*** 0.041 (0.071) (0.057) Observations 9,861 9,861 9,861 12,799 12,799 12,799 Robust standard errors in parentheses. *** p<0.01, ** p<0.05, * p<0.1. 45 Appendix Table 4: Kitagawa-Oaxaca-Blinder Decomposition of Monthly Wages by Age Group (1) (2) (3) (4) (5) (6) Age 15-29 Age 30-64 VARIABLES Overall Endow. Coeff. Overall Endow. Coeff. Male 13,959.73*** 92,453.0*** (1,159.740) (3,711.685) Female 3,371.458*** 10,763.5*** (347.883) (865.377) Raw Difference 10,588.27*** 81,689.5*** (1,210.793) (3,811.231) Difference due to 4,739.494*** 56,368.1*** Endowments (1,027.677) (3,224.216) Difference due to 5,848.782*** 25,321.4*** Coefficients (1,136.809) (3,943.294) Individual Lives in -365.272 -2,404.8*** 2,780.598 -691.542 an Urban Area (278.347) (933.410) (1,794.628) (2,977.77) Individual Age 2,777.583 -77,927.442 -8,371.664*** 228,211.60 (1,907.21) (65,164.895) (2,897.903) (146,997.0) Individual Age -3,255.5* 47,030.543 7,318.324*** -102,887.3 Squared (1,963.13) (35,337.165) (2,820.320) (73,552.45) Number of Adults -185.198 -4,073.02** 125.107 -661.317 in HH (126.330) (1,946.924) (97.186) (4,343.308) Number of Children 259.154** -714.440 -935.887*** 4,700.16** Aged 0-5 in HH (126.034) (607.336) (293.134) (2,126.934) Individual owns a 1,360*** 805.748 3,254.135 2,226.623 Financial Account (384.069) (519.399) (2,054.753) (2,517.225) Individual is -1,869.8* 3,724.135** -1,642.309** 18,491.6** Married (1,059.8) (1,816.016) (720.831) (7,741.180) Individual is Divorced/ 21.284 -134.503 -654.631** 802.787 Widowed/Separated (27.340) (82.746) (329.292) (880.664) Individual Completed 120.321 -562.159 9,997.520*** -744.407 Secondary Schooling (199.42) (418.813) (2,558.136) (2,306.7) Individual owns a Vehicle 48.063 965.395 -462.986 -2,301.716 (Bicycle, Motorcycle, Car) (41.558) (700.151) (594.454) (4,923.57) Individual Works as -244.35* -577.69* -1,140.15*** -306.23 Apprentice (147.09) (323.146) (351.292) (248.697) Individual Works as 2,079*** 38.334 20,410.16*** 4,092.51* Manager (483.99) (381.446) (2,631.810) (2,248.34) Individual Works as 234.571 47.497 6,249.461*** -2,626*** Skilled Worker (148.42) (295.877) (1,147.301) (863.518) Individual Works in 4,162*** -33,286*** 20,635.49*** -35,64*** the Agriculture Sector (932.4) (7,088.141) (4,125.394) (9,079.82) Individual Works in -403.127 -2,110.82* -1,195.082 -5,370** the Service Sector (343.14) (1,085.986) (3,212.446) (2,582.52) Constant 75,028.*** -81,972.7 (27,512.4) (74,970.4) Observations 5,903 5,903 5,903 4,688 4,688 4,688 The reference dummies are family worker and manufacturing sector. Robust standard errors in parentheses. *** p<0.01, ** p<0.05, * p<0.1 46 Appendix Table 5: Kitagawa-Oaxaca-Blinder Decomposition of Monthly Wages by Urban/Rural (1) (2) (3) (4) (5) (6) Urban Rural Overall Endow. Coeff. Overall Endow. Coeff. Male 88,236.180*** 10,414.785*** (3,193.346) (928.905) Female 44,692.701*** 350.511*** (2,089.437) (62.999) Raw Difference 43,543.479*** 10,064.274*** (3,816.176) (931.039) Difference due to 17,187.349*** 4,441.201*** Endowments (1,527.404) (941.417) Difference due to 26,356.130*** 5,623.073*** Coefficients (3,999.809) (871.652) Individual Age 963.665 75,679.626 -662.908 11,162.860 (588.082) (70,318.410) (1,356.438) (11,173.964) Individual Age Squared -537.293 -28,527.808 346.444 -4,485.751 (579.477) (37,623.201) (1,036.316) (4,230.380) Number of Adults 205.309 78.882 -275.362*** -4,720.178*** in HH (174.758) (5,013.285) (83.966) (1,579.305) Number of Children Aged -311.249 4,286.668* 42.065 -222.173 0-5 in HH (286.585) (2,238.269) (86.906) (793.724) Individual owns a Financial 1,012.753 113.283 1,681.028*** 405.325 Account (665.705) (3,840.605) (436.804) (318.736) Individual is -1,100.427** 10,633.930** -2,386.262** 4,621.336** Married (539.423) (5,074.315) (823.402) (1,851.284) Individual is Divorced/ -451.522 -705.460 -103.181* 70.545 Widowed/Separated (508.104) (1,029.446) (54.110) (86.881) Individual Completed 2,298.444*** -670.720 309.421** -173.724 Secondary Schooling (775.995) (3,566.892) (127.560) (143.203) Individual owns a Vehicle 13.361 550.209 6.169 299.741 (Bicycle, Motorcycle, Car) (612.445) (4,578.642) (12.716) (779.649) Individual Works as -480.883*** -955.571 -62.550** -224.075*** Apprentice (184.701) (1,025.306) (28.634) (76.100) Individual Works as 7,813.785*** 7,187.670** 1,286.452*** -3.543 Manager (656.712) (3,484.212) (236.033) (182.858) Individual Works as 3,280.616*** -3,452.283** 173.668** -160.536* Skilled Worker (535.828) (1,536.530) (83.856) (91.850) Individual Works in the 4,485.839*** -6,475.619*** 4,987.933*** -52,861.344** Agriculture Sector (743.844) (1,591.811) (1,090.634) (11,054.282) Individual Works in the -5.047 -16,744.753** -901.715 -208.370 Service Sector (105.677) (5,065.529) (690.681) (607.989) Constant -14,641.925 52,122.962*** (31,856.671) (16,174.822) Observations 3,790 3,790 3,790 6,801 6,801 6,801 Robust standard errors in parentheses. *** p<0.01, ** p<0.05, * p<0.1. 47 Appendix Table 6: The Kitagawa-Oaxaca-Blinder Decomposition of Business Revenues by Age Group (1) (2) (3) (4) (5) (6) Age 15-29 Age 30-64 VARIABLES Overall Endow. Coeff. Overall Endow. Coeff. Male 11.717*** 11.878*** (0.208) (0.119) Female 10.672*** 10.995*** (0.122) (0.068) Raw Difference 1.045*** 0.883*** (0.241) (0.137) Difference due to 0.001 0.668*** Endowments (0.432) (0.186) Difference due to 1.044** 0.215 Coefficients (0.512) (0.234) Individual Lives in Urban 0.007 -0.190 -0.001 -0.168* Area Dummy (0.020) (0.176) (0.008) (0.100) Individual Age 0.086 -26.923 -0.002 2.069 (0.186) (17.916) (0.009) (4.894) Individual Age Squared -0.075 13.366 -0.000 -0.430 (0.194) (9.801) (0.004) (2.359) Number of Adults in HH -0.051 0.007 -0.002 -0.209 (0.032) (0.465) (0.006) (0.212) Number of Children Aged 0-5 -0.015 0.293 -0.036 0.277* In HH (0.059) (0.245) (0.030) (0.153) Individual owns a Financial 0.113 0.479* 0.265*** 0.128 Account (0.075) (0.246) (0.046) (0.107) Individual is Married -0.030 1.035** 0.006 0.468 (0.073) (0.469) (0.026) (0.475) Individual is Divorced/Married -0.001 0.003 -0.019 0.030 /Separated (0.007) (0.012) (0.031) (0.047) Individual Completed 0.033 0.094** -0.019 -0.042 Secondary Schooling (0.033) (0.038) (0.023) (0.037) Individual has a Vehicle -0.071 0.532 0.077* -0.191 (bicycle, motorcycle, car) (0.077) (0.397) (0.045) (0.257) Business has Electricity 0.045 0.027 0.105*** 0.051** (0.027) (0.040) (0.030) (0.022) Business has Running Water -0.003 0.001 -0.000 0.001 (0.027) (0.013) (0.027) (0.027) Business has a Phone 0.108** -0.823 0.024 -0.084 (0.051) (0.601) (0.020) (0.477) Business Maintains 0.016 -0.037* -0.003 0.030 Accounting Records (0.020) (0.020) (0.055) (0.022) Business is registered 0.003 0.023 -0.045 -0.073* 48 (0.007) (0.024) (0.030) (0.040) Business operates machines -0.025 -0.008 0.036 0.005 (0.029) (0.094) (0.029) (0.045) Number of Men Employees -0.083 -0.037 0.106 0.017 (0.158) (0.026) (0.087) (0.018) Number of Women Employees 0.003 0.036 -0.010 -0.048*** (0.015) (0.044) (0.017) (0.011) Number of Days HH Male 0.824** -0.124 0.678*** 0.066 Member Worked (0.356) (0.147) (0.146) (0.054) Number of Days HH Female -0.696*** -0.023 -0.406*** 0.012 Member Worked (0.190) (0.098) (0.073) (0.078) Business is in Food Sector -0.082 -0.192 -0.061* -0.152 (0.055) (0.194) (0.036) (0.101) Business is in Clothing -0.010 -0.106 0.011 -0.109 Sector (0.015) (0.147) (0.010) (0.090) Business is in Construction -0.094* 0.070 -0.033 -0.098** Sector (0.055) (0.127) (0.023) (0.045) Constant 13.541* -1.333 (8.028) (2.540) Observations 989 989 989 3,657 3,657 3,657 Robust standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1, The reference dummy is the service sector 49 Appendix Table 7: The Kitagawa-Oaxaca-Blinder Decomposition of Business Revenues by Urban/Rural (1) (2) (3) (4) (5) (6) Urban Rural VARIABLES Overall Endow. Coeff. Overall Endow. Coeff. Male 12.380*** 11.494*** (0.182) (0.116) Female 11.681*** 10.410*** (0.071) (0.087) Raw Difference 0.699*** 1.085*** (0.195) (0.145) Difference due to 0.892*** 0.570*** Endowments (0.335) (0.187) Difference due to -0.193 0.515** Coefficients (0.383) (0.251) Individual Age 0.037 -4.282 0.129 -6.695* (0.046) (3.367) (0.080) (3.628) Individual Age Squared -0.033 2.153 -0.130* 3.619** (0.036) (1.744) (0.074) (1.778) Number of Adults in HH 0.009 -0.527 -0.000 0.316 (0.010) (0.327) (0.002) (0.248) Number of Children Aged 0-5 -0.003 0.207 -0.021 0.255 In HH (0.011) (0.166) (0.020) (0.188) Individual owns a Financial 0.118*** 0.189 0.305*** 0.062 Account (0.045) (0.189) (0.063) (0.079) Individual is Married -0.001 0.378 0.010 0.918 (0.007) (0.401) (0.017) (0.723) Individual is Divorced/Married 0.016 0.014 -0.044 0.042 /Separated (0.026) (0.044) (0.035) (0.052) Individual Completed -0.013 -0.009 0.001 0.002 Secondary Schooling (0.052) (0.065) (0.001) (0.002) Individual has a Vehicle 0.063 -0.033 0.048 -0.038 (bicycle, motorcycle, car) (0.043) (0.239) (0.060) (0.299) Business has Electricity 0.224*** 0.098** 0.024** 0.006 (0.061) (0.047) (0.011) (0.009) Business has Running Water -0.001 0.005 0.004 -0.001 (0.061) (0.043) (0.003) (0.001) Business has a Phone 0.007 -0.810 0.049 -0.264 (0.006) (0.521) (0.033) (0.481) Business Maintains -0.217** -0.271*** 0.010* 0.012 Accounting Records (0.091) (0.098) (0.005) (0.013) Business is registered 0.135 0.126** 0.004 0.003 (0.119) (0.059) (0.004) (0.004) 50 Business operates machines 0.003 -0.032 0.036 0.004 (0.029) (0.110) (0.029) (0.029) Number of Men Employees 0.152 -0.005 -0.007 0.006 (0.160) (0.030) (0.081) (0.009) Number of Women Employees -0.022 -0.071*** -0.012** 0.001 (0.038) (0.021) (0.005) (0.004) Number of Days HH Male 0.772** 0.076 0.653*** -0.061 Member Worked (0.310) (0.090) (0.124) (0.050) Number of Days HH Female -0.318** 0.018 -0.365*** -0.197** Member Worked (0.133) (0.078) (0.073) (0.088) Business is in Food Sector -0.063** 0.181 -0.073 -0.469*** (0.032) (0.111) (0.053) (0.145) Business is in Clothing 0.057** -0.101 -0.011 -0.144 Sector (0.026) (0.145) (0.010) (0.091) Business is in Construction -0.030 0.024 -0.040 -0.071 Sector (0.039) (0.082) (0.026) (0.059) Constant 2.479 3.209** (1.830) (1.591) Observations 2,472 2,472 2,472 2,174 2,174 2,174 Robust standard errors in parentheses, *** p<0.01, ** p<0.05, * p<0.1, The reference dummy is the service sector 51 Appendix Table 8: The Kitagawa-Oaxaca-Blinder Decomposition of Harvest Value by Age (1) (2) (3) (4) (5) (6) Age 15-29 Age 30-64 VARIABLES Overall Endow. Coeff. Overall Endow. Coeff. Male 11.862*** 12.001*** (0.051) (0.021) Female 11.001*** 11.064*** (0.053) (0.029) Raw Difference 0.861*** 0.937*** (0.074) (0.035) Difference due to3 0.482*** 0.545*** Endowments (0.082) (0.035) Difference due to 0.379*** 0.392*** Coefficients (0.097) (0.045) Individual Age 0.060 10.521* -0.030* -1.855 (0.086) (5.851) (0.018) (1.414) Individual Age Squared -0.047 -5.210* 0.027 0.765 (0.079) (3.107) (0.017) (0.699) Number of Adults in HH 0.016 0.132 0.020*** 0.145*** (0.015) (0.123) (0.005) (0.056) Number of Children Aged 0-5 -0.054** -0.115 0.032*** 0.001 In HH (0.025) (0.108) (0.007) (0.029) Individual owns a Financial -0.006 0.001 0.014 -0.002 Account (0.019) (0.032) (0.009) (0.011) Individual is Married 0.009 0.181 0.070* 0.357* (0.014) (0.288) (0.041) (0.215) Individual is Divorced/Married -0.005 0.014 -0.062 0.123 /Separated (0.012) (0.012) (0.044) (0.075) Individual Completed -0.006* -0.004 0.000 -0.002 Secondary Schooling (0.003) (0.004) (0.001) (0.002) Individual has a Vehicle 0.035 0.270*** 0.003 -0.018 (bicycle, motorcycle, car) (0.027) (0.086) (0.017) (0.061) Household owns Mechanized 0.050*** 0.066 0.049*** 0.013 Agriculture Equipment (0.017) (0.049) (0.007) (0.025) Area of Plot of Land 0.001 0.035 0.007 -0.010 (0.006) (0.110) (0.005) (0.012) Area of Plot of Land Squared -0.002 -0.014 -0.004 0.004 (0.006) (0.032) (0.003) (0.004) Land Distance from House 0.000 -0.143*** 0.000 -0.095*** (0.001) (0.048) (0.000) (0.018) Crop Grown for Export 0.133*** -0.011** 0.164*** -0.003 (0.021) (0.005) (0.010) (0.003) Manager of Farm is the Owner 0.014 0.023 -0.007 0.012 (0.016) (0.053) (0.011) (0.031) 52 Manager Controls Income from 0.007 0.624*** -0.002 0.699*** Plot (0.005) (0.233) (0.002) (0.148) Total Days Worked by HH Men 0.149*** 0.082 0.087*** -0.001 (0.055) (0.065) (0.011) (0.026) Total Days Worked by HH 0.001 -0.292*** -0.000 -0.137*** Women (0.010) (0.076) (0.001) (0.025) Total Days Worked by Men 0.035** -0.028 0.009*** -0.004 (non-HH) (0.014) (0.024) (0.002) (0.005) Total Days Worked by Women 0.003 0.060** 0.002* 0.014 (non-HH) (0.004) (0.030) (0.001) (0.009) All Harvest Sold 0.016* 0.035* 0.005** 0.004 (1=Yes) (0.009) (0.018) (0.002) (0.009) Plot of Land uses 0.020 -0.048 0.085*** 0.024 Urea/Phosphates/NPK/DAP (0.015) (0.040) (0.008) (0.019) Plot of Land uses 0.053*** -0.035 0.074*** 0.051** Pesticides/Fungicides/Herbicides (0.013) (0.040) (0.008) (0.020) Constant -5.762** 0.306 (2.775) (0.768) Observations 2,142 2,142 2,142 11,435 11,435 11,435 Robust standard errors in parentheses. *** p<0.01, ** p<0.05, * p<0.1. 53 Appendix Table 9: The Kitagawa-Oaxaca-Blinder Decomposition of Crop Sales by Age Group (1) (2) (3) (4) (5) (6) Age 15-29 Age 30-64 VARIABLES Overall Endow. Coeff. Overall Endow. Coeff. Male 11.632*** 11.803*** (0.065) (0.029) Female 10.757*** 10.837*** (0.092) (0.052) Raw Difference 0.875*** 0.966*** (0.113) (0.060) Difference due to3 0.298*** 0.629*** Endowments (0.106) (0.057) Difference due to 0.577*** 0.338*** Coefficients (0.146) (0.070) Individual Age 0.046 2.129 0.013 2.088 (0.104) (8.699) (0.025) (2.361) Individual Age Squared -0.032 -0.766 -0.023 -0.861 (0.098) (4.630) (0.027) (1.164) Number of Adults in HH 0.030** 0.209 0.018*** 0.075 (0.015) (0.145) (0.006) (0.094) Number of Children Aged 0-5 -0.122*** 0.406*** 0.001 0.043 In HH (0.034) (0.134) (0.008) (0.048) Individual owns a Financial 0.048 -0.085* 0.031* -0.052*** Account (0.031) (0.049) (0.016) (0.014) Individual is Married -0.010 0.766** 0.076** 0.355 (0.013) (0.316) (0.035) (0.243) Individual is Divorced/Married 0.015 0.046* -0.034 0.073 /Separated (0.019) (0.028) (0.040) (0.056) Individual Completed 0.000 -0.006 0.004*** 0.000 Secondary Schooling (0.004) (0.007) (0.001) (0.000) Individual has a Vehicle -0.101** -0.030 -0.013 -0.132 (bicycle, motorcycle, car) (0.048) (0.166) (0.030) (0.103) Household owns Mechanized 0.076*** -0.015 0.132*** 0.009 Agriculture Equipment (0.027) (0.087) (0.016) (0.051) Area of Plot of Land -0.007 0.226* 0.034 -0.087 (0.021) (0.132) (0.022) (0.063) Area of Plot of Land Squared 0.008 -0.088 -0.012 0.024 (0.028) (0.070) (0.009) (0.018) Land Distance from House 0.009 -0.103 -0.000 -0.086*** (0.009) (0.076) (0.000) (0.030) Crop Grown for Export 0.071*** -0.012 0.112*** -0.005 (0.023) (0.012) (0.013) (0.006) Manager of Farm is the Owner -0.002 0.044 -0.030* -0.051 (0.023) (0.087) (0.017) (0.050) 54 Manager Controls Income from -0.003 0.579*** 0.015** 0.745*** Plot (0.005) (0.190) (0.007) (0.131) Total Days Worked by HH Men 0.052 0.001 0.041*** 0.011 (0.033) (0.089) (0.012) (0.045) Total Days Worked by HH 0.005 -0.187 0.000 -0.018 Women (0.009) (0.127) (0.001) (0.042) Total Days Worked by Men 0.028** 0.014 0.019*** 0.009 (non-HH) (0.013) (0.068) (0.005) (0.015) Total Days Worked by Women -0.004 0.036 0.004 0.043*** (non-HH) (0.008) (0.044) (0.003) (0.015) All Harvest Sold 0.085*** 0.081 0.029*** -0.014 (1=Yes) (0.028) (0.063) (0.008) (0.029) Plot of Land uses 0.051* -0.185** 0.142*** -0.019 Urea/Phosphates/NPK/DAP (0.027) (0.083) (0.017) (0.037) Plot of Land uses 0.055** 0.049 0.069*** 0.167*** Pesticides/Fungicides/Herbicides (0.025) (0.086) (0.014) (0.051) Constant -2.532 -1.981 (4.054) (1.217) Observations 785 785 785 4,167 4,167 4,167 Robust standard errors in parentheses. *** p<0.01, ** p<0.05, * p<0.1. 55