100234 THE COST OF THE GENDER GAP IN AGRICULTURAL PRODUCTIVITY in Malawi, Tanzania, and Uganda UNDP-UNEP POVERTY-ENVIRONMENT INITIATIVE Empowered lives. Resilient nations. THE COST OF THE GENDER GAP IN AGRICULTURAL PRODUCTIVITY in Malawi, Tanzania, and Uganda This report is a joint product of UN Women, the United Nations Development Programme–United Nations Environment Programme Poverty-Environment Initiative (UNDP-UNEP PEI) Africa, and the World Bank. The collaboration was led by UN Women, Eastern and Southern Africa Regional Office (ESARO). First published in October 2015. © 2015 UN Women, UNDP, UNEP, and the World Bank Group. The views expressed in this publication are those of the authors and do not necessarily reflect the views of UN Women, UNDP, UNEP, and the World Bank Group. The designation of geographical entities in this report, and the presentation of the material herein, does not imply the expression of any opinion whatsoever on the part of the publisher or the participating organizations concerning the legal status of any country, territory, or area, or of its authorities, or concerning the delimitation of its frontiers or boundaries. While reasonable efforts have been made to ensure that the contents of this publication are factually correct and properly referenced, UN Women, UNDP, UNEP, and the World Bank Group do not accept responsibility for the accuracy or completeness of the contents and shall not be liable for any loss or damage that may be occasioned directly or indirectly through the use of, or reliance on, the contents of this publication, including its translation into languages other than English. This publication may be reproduced in whole or in part and in any form for educational or nonprofit purposes without special permission from the copyright holder provided acknowledgment of the source is made. No use of this publication may be made for resale or for any other commercial purpose whatsoever without prior permission in writing from UN Women, UNDP, UNEP, and the World Bank Group. Cover photo: PaolikPhotos/Shutterstock Design and editing: Nita Congress Contents v Foreword vii Acknowledgments 1 Introduction: the gender gap in agricultural productivity 5 Three takeaways on the gender gap in agricultural productivity 7 Measuring the cost of the gender gap in agricultural productivity 13 Costing the factors that contribute to the gender gap in agricultural productivity 17 Finding the biggest bang for the buck: cost-effective policy interventions 21 Moving from recommendations to implementation 23 Appendix A:  Methodology for quantifying the cost of the gender gap in agriculture 25 Appendix B:  Methodology for costing the factors of production contributing to the gender gap in agriculture 27 References THE COST OF THE GENDER GAP IN AGRICULTURAL PRODUCTIVITY IN MALAWI, TANZANIA, AND UGANDA  / iii Foreword W omen form a large proportion of the Recognizing the need for more specific evidence of agricultural labor force in Sub-Saharan the economic gains from closing the gender gap, Africa and thus play a vital role in UN Women, the joint United Nations Development ensuring family nutrition and food security. In Programme–United Nations Environment Programme Eastern and Southern Africa, agriculture continues Poverty-Environment Initiative, and the World Bank to be a key engine for local and regional economies, collaborated on this study which measures the size represents a critical source of income and ensures of the gender gap in monetary terms. food security and nutrition. However, as has been widely documented, gender-based inequalities in The report provides a unique quantification of access to and control of productive and financial the costs in terms of lost growth opportunities resources inhibit agricultural productivity and reduce and an estimate of what societies, economies, food security. A new study measuring the economic and communities would gain if the gender gap in costs of the gender gap in agricultural productivity in agriculture is addressed. The findings of this report three African countries—Malawi, the United Republic are striking, and send a strong signal to policy makers of Tanzania (hereafter Tanzania), and Uganda— in Africa as well as development partners that closing provides further evidence that reducing the gender the gender gap is smart economics. Consider this: gap plays a significant role in poverty reduction and closing the gender gap in agricultural productivity improved nutritional outcomes. could potentially lift as many as 238,000 people out of poverty in Malawi, 80,000 people in Tanzania, and While there is mounting evidence on the link 119,000 people in Uganda. between promoting women’s equality and economic empowerment and other development outcomes, The Sustainable Development Goals (SDGs) offer such as sustainable agricultural and economic growth, a historic opportunity to shift from development gender issues are being inadequately reflected in in silos to a more integrated approach. This work agricultural policy strategies and programs. At the provides evidence and policy recommendations that same time, a changing climate means that there is can support the achievement of the SDGs—which a shrinking window of opportunity for action, and include a specific goal on achieving gender equality it is imperative that climate-smart approaches to and empowering all women and girls—as well as the agriculture help close the gender gap and promote objectives of the Comprehensive Africa Agricultural women’s empowerment, economic development, Development Programme (CAADP). The report and societal resilience to shocks. also provides guidance on the factors that must be targeted in order to close the gender gap and improve opportunities for women farmers. It concludes with THE COST OF THE GENDER GAP IN AGRICULTURAL PRODUCTIVITY IN MALAWI, TANZANIA, AND UGANDA  / v a set of general policy recommendations on how It is our hope that the report will be used by policy women’s empowerment, agriculture productivity, and makers and practitioners to propose and implement economic growth can be addressed in an integrated gender-sensitive—and environmentally sustainable— manner in order to achieve the SDGs at the national agriculture-related policies and programs. level. Yannick Glemarec Magdy Martinez-Soliman Deputy Executive Director Assistant Administrator and Director Policy and Programme Bureau Bureau for Policy and Programme Support UN Women United Nations Development Programme Caren Grown Ibrahim Thiaw Senior Director Deputy Executive Director Gender Group United Nations Environment Programme World Bank vi / THE COST OF THE GENDER GAP IN AGRICULTURAL PRODUCTIVITY IN MALAWI, TANZANIA, AND UGANDA Acknowledgments T his report is a joint production of UN Women, Comments were provided by Hodan Addou, Country the United Nations Development Programme– Representative, UN Women Uganda; Anna Collins- United Nations Environment Programme Falk, Country Representative, UN Women Tanzania; Poverty-Environment Initiative (UNDP-UNEP PEI) Olof Drakenberg, University of Gothenburg and Africa, and the World Bank. The collaboration was led PEI Technical Advisory Group member; Simone by UN Women, Eastern and Southern Africa Regional Ellis Oluoch-Olunya, Deputy Regional Director, UN Office (ESARO). Women ESARO; Alice Harding-Shackelford, Country Representative, UN Women Malawi; Ruth Hill, World The team of authors who contributed to the Bank; Themba Kalua, Regional Coordinator, UN preparation of this report was led by Niklas Buehren, Women ESARO; Linda Kasseva, UNEP Gender Unit; World Bank Africa Gender Innovation Lab (GIL), Gebrehiwot Kebedew, International Economist, and comprised of Markus Goldstein, World Bank UNDP Malawi; Isabell Kempf, Co-Director, UNDP- Africa GIL; Kajal Gulati, UN Women consultant; UNEP PEI; Talip Kilic, World Bank; James Mbata, Daniel Kirkwood, World Bank Africa GIL; Vanya Technical Adviser to the government of Malawi, Slavchevska, UN Women consultant; David Smith, UNDP Malawi; Edfas Mkandawire, UN Women, UNDP-UNEP PEI Africa;  Asa Torkelsson, UN Women; Malawi; Damaris Mungai, Programme Officer, UNEP and Moa Westman, UNDP-UNEP PEI Africa. Support Regional Office for Africa; Paulo Nunes, Ecosystem was provided by Zewdu Ayalew Abro, World Bank Services Economics Unit, UNEP; Lesley Reader, UN Africa GIL; Flavia Ciribello, UN Women; and Rachel Women Tanzania; and Victor Tsang, Programme Coleman, World Bank Africa GIL. Officer, UNEP Gender Unit. THE COST OF THE GENDER GAP IN AGRICULTURAL PRODUCTIVITY IN MALAWI, TANZANIA, AND UGANDA  / vii Introduction: the gender gap in agricultural productivity W  Agricultural omen comprise a large proportion of the agricultural labor force in Sub-Saharan Africa, ranging from 30 to 80 percent productivity = the gross (FAO 2011).1 Yet women farmers are consistently found to be less productive than male farmers. The value of output (in local currency) gender gap in agricultural productivity—measured by the value of agricultural produce per unit of cultivated produced per hectare of land land—ranges from 4 to 25 percent, depending on the country and the crop (World Bank and ONE 2014).2 fact that the gender gap persists suggests that the underlying constraints are still inadequately tackled This gap exists because women frequently have in agricultural policy strategies and programs. Low unequal access to key agricultural inputs such as land, agricultural productivity tends to reduce per hectare labor, knowledge, fertilizer, and improved seeds. The 3 yields and leads to more intense farming—resulting in overcultivation, soil erosion, and land degradation. These in turn fur ther undermine agricultural 1  Using individual-disaggregated, plot-level labor input productivity and environmental sustainability. The data from nationally representative household surveys, evidence presented in this report addresses this Palacios-Lopez, Christiaensen, and Kilic (2015) report the female share of agricultural labor for Malawi, Tanzania, and situation and offers guidance to policy makers on Uganda to be 52, 52, and 56 percent, respectively. how to increase agricultural productivity and national 2  Additional information on gender-disaggregated economic growth, strengthen food security, and productivity estimates for these countries can be found in support poverty reduction across Sub-Saharan Africa. Akresh (2005); FAO (2011); Gilbert, Sakala, and Benson (2002); Goldstein and Udry (2008); Hill and Vigneri (2014); Peterman, Behrman, and Quisumbing (2014); Tiruneh et al. (2001); and Udry (1996). that plots owned or managed by women are less likely Sheahan and Barrett (2014) report for their sample of six 3  to receive modern agricultural inputs and receive lesser Sub-Saharan countries that female-headed households amounts when applied. However, the sex of the plot statistically significantly apply, use, and own less modern manager or owner appears to be a lesser determinant of agricultural inputs compared to male-headed ones; and input use in Tanzania and Uganda compared to Malawi. 1 In this report, we estimate the monetary value of profile of female farmers in these countries. We then the gender gap in agricultural productivity in look at what the size of this gap means relative to Malawi, Tanzania, and Uganda. Box 1.1 presents a gross domestic product (GDP) and poverty reduction. BOX 1.1 Who is a woman farmer? I n the three countries profiled in this report, female farm managers are found to have lower levels of education and a smaller average family size, and to be less wealthy compared to all other plot managers. two years’ less schooling), and live in households with about 1.5 fewer household members. This latter statistic may be partly explained by the fact that 67 percent of the sole female managers are widowed, divorced, or separated compared to only 9 percent in the sample In Malawi, women farmers are older by over five years, on average, and of other plot managers. While women managers cultivate about also have lower levels of education as compared to male managers. 0.6 hectares of land on average, all other managers cultivate more Only 25 percent of sole female managers are married monogamously, than 1 hectare—a difference that is statistically significant. as compared to 87 percent of male managers. Seventy percent of them are widowed, divorced, or separated, compared to only 3 percent In Uganda, women managers cultivate plots that are on average among male managers. about 0.23 hectares smaller than those managed by males. They average about two years’ less schooling than male managers. And In Tanzania, women farm managers are about four years older than only 50 percent of the female managers are married, compared to all other managers, have a lower educational attainment (roughly 90 percent of the male managers. Malawi Tanzania Uganda 87% 67% 90% 50% 25% Widowed, divorced, 9% Married monogamously separated Married 1.0 70% 0.6   Hectares of land cultivated Less education Widowed, divorced, 3%   separated Smaller plots   Less education Less education 1½ fewer household members 5 years 4 years older older 2 / THE COST OF THE GENDER GAP IN AGRICULTURAL PRODUCTIVITY IN MALAWI, TANZANIA, AND UGANDA Annual cost We estimate that the gender BOX 1.2 of gender gap gap amounts to Linking the gender gap in $100 million in agricultural productivity to poverty Malawi, $105 million and food security  $100 million  in Tanzania, and $6 7 m ill i o n Uganda per year.4 in I n addition to impacts on overall national income, closing the gender agricultural productivity gap could Malawi Throughout this reduce poverty and improve nutrition: directly, because many poor people work in agriculture; and indirectly, study, we express because higher agricultural output may increase income monetary values for people employed in sectors linked to agriculture. At  $105 million  in terms of 2010 prices. These the same time, higher agricultural output can also lead to lower food prices. The combined impact of increasing the incomes and agricultural productivity of the poor Tanzania estimates can and lowering food prices could help improve nutrition help policy makers by enabling poor people to purchase more and better under s tand t he food, and by increasing their access to food from their scale of the gains own production.  $67 million  that could be made f rom des igning Also, although no attempt is made to quantitatively Uganda better policies to capture these in this report, there would likely be impacts on women’s empowerment and time use (Chung 2012; improve women’s Ruel, Alderman and the Maternal and Child Nutrition ability to use Study Group, 2013). The factor by which higher growth agriculture to lift themselves and their families out reduces poverty has been estimated in economywide of poverty and to contribute to economic growth models developed for Malawi, Tanzania, and Uganda. We use this poverty-agricultural growth conversion factor (box 1.2). It is important to stress that these potential (also referred to as elasticity) to compute the potential gains to do not come without cost. Closing the gender reduction in poverty and malnutrition from narrowing the gap will require changing existing or designing new gender gap (Benin et al. 2008; Pauw and Thurlow, 2011). policies, which may require additional resources. We then go beyond these figures to estimate the costs associated with gender gaps in access to boxes 3.1 and 3.2 for more details on the methodology individual agricultural inputs. This evidence can and data used for the costing exercise presented in help policy makers decide where efforts are most this report.) needed. For example, understanding that 97 percent of the gender gap in Tanzania is due to women’s lower Finally, we review existing evidence from impact access to male labor can help decision makers— evaluations and other sources on the effectiveness especially national governments, international of specific policies and interventions in closing organizations, and development partners—better the gender gap in access to different agricultural focus their efforts. For policy makers from countries inputs. Such evidence is essential for helping policy not covered in this study, our analysis could be makers think about how they can put the lessons replicated using data from their own country. (See from this analysis into practice. Unfortunately, existing knowledge of effective—let alone cost- efficient—policy instruments to resolve hurdles faced All dollars referred to in this report are U.S. dollars. 4  by women farmers is still nascent. For policies to work, THE COST OF THE GENDER GAP IN AGRICULTURAL PRODUCTIVITY IN MALAWI, TANZANIA, AND UGANDA  / 3 it is crucial to recognize that both men and women ▲▲ Making current agricultural policies more gender may face different constraints that hinder them from responsive. Such policies may include tweaking improving their agricultural practices and that it may existing policies, such as agricultural extension be necessary to rethink, innovate, and pilot in order to services, to purposely include both women and adequately address women-specific constraints and men. to document what works and what does not. The list of possible policy approaches can be split into two ▲▲ Designing new agricultural policies that are gender main groups: targeted. Policy makers can design agricultural policies that focus specifically on the needs of women farmers, for example, by promoting time- or labor-saving and sustainable technologies. 4 / THE COST OF THE GENDER GAP IN AGRICULTURAL PRODUCTIVITY IN MALAWI, TANZANIA, AND UGANDA Three takeaways on the gender gap in agricultural productivity T hree key policy lessons emerge from the additional investments from governments, their evidence presented in the remainder of this magnitude is sufficiently large to justify significant report. attention. The gender gap in agricultural The potential economic gains from productivity is large reducing the gender gap translate into significant poverty reduction and Even with the conservative assumptions used in this improved nutritional outcomes study, we find that there are large gains to be achieved if policy makers address the gender gap effectively. Increasing GDP by closing the gender gap in Annual crop output could increase by 2.1 percent in agricultural productivity has the potential to lift as Tanzania, 2.8 percent in Uganda, and 7.3 percent in many as 238,000 people out of poverty in Malawi, Malawi. Taking into account the contribution of crops approximately 80,000 people in Tanzania and to total agricultural output, the size of the agricultural 119,000 people in Uganda. In Tanzania, for example, sector in the overall economy, and spillover effects this gain also translates into a 0.7 percent reduction of higher agricultural output to other sectors of the in the incidence of undernourishment, which implies economy, we estimate the potential gross gains to that roughly 80,000 people would be lifted out of GDP to be $100 million in Malawi (or 1.85 percent of malnourishment per year. However, closing the GDP), $105 million in Tanzania (0.46 percent of GDP), gender gap could result in additional improvements and $67 million in Uganda (0.42 percent of GDP).1 as these estimates do not capture all the likely While achieving these gains would in itself require agriculture-nutrition linkages and other spillover effects. For example, increased income in women’s hands has implications for the intergenerational 1 The key empirical step we take to translate the estimated gross gains from closing the gap in agricultural yields between male and female farmers into gains of aggregate value addition (GDP) is to assume that the fraction of proportionally with the gains in total gross crop production. agricultural GDP associated with crop production would rise For more detail, see box 3.1. 5 transmission of hunger and malnutrition, as women women’s lower access to farm labor is one of the most tend to spend more of their income on children’s important constraints contributing to the gender health and education (Ruel, Alderman, and the gap in Malawi and Tanzania. Closing the gap in the Maternal and Child Nutrition Study Group 2013; quantity of male labor used could yield gross gains Smith et al. 2003). of over $45 million in Malawi and over $100 million in Tanzania. In Uganda, one priority should be improving To ensure the biggest “bang for the women’s access to agricultural machinery and other buck,” governments should identify production technologies, which has the potential and focus on the most costly to increase GDP by over $11 million. However, our constraints to women’s productivity knowledge of what works is far from complete. Further research should therefore be undertaken to This report helps lay the groundwork for deeper look at the relative impacts of specific policies and analysis as to where to invest for the most effective interventions as well as their cost-efficiency in order and cost-effective policies. Our analysis finds that to quantify their net benefits. Three takeaways   1  The gender gap in agricultural productivity is large. 2     Reducing it may reduce poverty and improve nutrition.        3  Reduce the gap by focusing on most costly constraints. 6 / THE COST OF THE GENDER GAP IN AGRICULTURAL PRODUCTIVITY IN MALAWI, TANZANIA, AND UGANDA Measuring the cost of the gender gap in agricultural productivity T his section presents estimates of the foregone account. We do this by calculating the conditional income (total GDP and agricultural GDP) gender gap, which is estimated conditional on plot and poverty reduction potential that result area and agro-climatic conditions. Figure 3.1 shows from the gender gap in agricultural productivity in that the unconditional gender gap ranges from 13 Malawi, Tanzania, and Uganda. Box 3.1 outlines the to 28 percent for the three countries studied. The methodology, which is presented in more detail in conditional gender gap is even more substantial, appendix A. ranging from 28 to 31 percent. These findings echo earlier analysis (World Bank and ONE 2014). In order to make these estimates, we compute the unconditional and conditional values of the gender gap in agricultural productivity. In this report, FIGURE 3.1 agricultural productivity is defined as the value of Unconditional and conditional gender gap in output per hectare. The difference in this measure agricultural productivity in percentages between male and female farmers constitutes the Unconditional gap Conditional gap unconditional gender gap, as described in box 3.1. But this unconditional gender gap does not take into account the fact that, on average, women work on 31% 30% 28% 28% smaller plots than men. Generally, farmers are more productive on smaller plots; one reason postulated for this is that they are able to use physical and labor 16% resources to cultivate their plots more intensely (see, 13% for example, Carletto, Gourlay, and Winters 2013 for robust evidence on this inverse relationship for several African countries). But despite cultivating smaller plots relative to men, women are still less productive; this implies that the gender gap would be Malawi Tanzania Uganda even larger if we take the smaller size of their plots into 7 BOX 3.1 Methodology: Measuring the economic costs of the gender gap in agricultural productivity 1 Traditionally, agricultural productivity is measured based on household-level analysis. In contrast, we here look at the plot level and identify the plot manager, measuring the in two scenarios. In the first scenario, we assume that there is no difference between male and female agricultural output— that is, there is no gender gap and agricultural productivity difference in productivity between plots cultivated by women of women’s plots is equal to plots cultivated by men. In the and men. We convert the agricultural output produced by second scenario, we use the actual productivity estimates women and men farmers on their plots into monetary values obtained in the first step to calculate the value of output by multiplying the output obtained per unit of land with obtained from female plots in the presence of the gender gap. the median unit and crop-specific price in the respective The difference between the no gender gap scenario and the enumeration area (or at a higher level of aggregation if gender gap scenario gives the additional output value from needed). In Tanzania and Uganda, output was measured in closing the gender gap in productivity. kilograms; in Malawi, a wide range of measurement units was employed. We then aggregate the total value of all crops per unit of land associated with each gender. The difference in this value between women’s and men’s plots constitutes the 3 The final step includes computing the size of the gap relative to agricultural GDP. To do this, we first need to know what fraction of agricultural GDP comes from crop unconditional gender gap in agricultural productivity. This production. Second, we need to know the share of agricultural is the first step in estimating the national income that is GDP in overall GDP. Because growth in the agricultural sector foregone because of the gender agricultural productivity gap. influences other sectors of the economy, the cost of the gender gap is likely higher than just the foregone agricultural 2 As a next step, we calculate the fraction of land cultivated by women and men, after accounting for the fact that women cultivate smaller plots than men. For example, in GDP. To take this into account, we use an estimate of the multiplier between agricultural sector growth and the rest of the economy obtained from other studies in Malawi, Tanzania, women cultivate 20 percent of the cultivated Tanzania, and Uganda (Benin et al. 2008; Mabiso, Pauw, plots but because their plots are on average smaller by and Benin 2012; Pauw and Thurlow 2011). 0.47 hectares, women manage only about 13 percent of the land. Combining this fraction with the estimated A more technical description of the methodology is given gender gap in agricultural productivity, we compute the in appendix A. percentage difference between harvested value of output Because agriculture also employs more than two- soil erosion, and land degradation—which in turn thirds of the population in these countries—including further undermine agricultural productivity and some of the poorest citizens—increasing agricultural environmental sustainability. production can make a significant contribution to reducing poverty. Moreover, improvements in the We treat the plot of land, with identification of the agricultural sector may have considerable spillover gender of the plot manager or decision maker, as effects for other sectors of the economy. Therefore, the unit of analysis. This identification was made the analysis presented here extends to outcomes possible by the data structure of the Living Standards related to poverty and nutrition; we here define the Measurement Study—Integrated Sur veys on poor as those living on less than $1.25 a day. Note Agriculture; see box 3.2 for more detail. Using this that low agricultural productivity can also lead to gender-disaggregated, plot-level data allows us to more intense farming, resulting in overcultivation, capture differences in agricultural productivity even 8 / THE COST OF THE GENDER GAP IN AGRICULTURAL PRODUCTIVITY IN MALAWI, TANZANIA, AND UGANDA The motivation for plot-level analysis rests on the BOX 3.2 assumption that female farmers face a different (and Data used for costing the gender possibly larger) set of constraints relative to male gap in agricultural productivity farmers which may hinder them from accessing inputs and output markets to similar degrees or at F or our analysis, we use data from the World Bank’s Living Standards Measurement Study—Integrated Surveys on Agriculture (http://go.worldbank.org/ the same prices. If households were to act as a single unit that allocates resources such that overall welfare BCLXW38HY0). Specifically, the analysis presented is maximized, these market imperfections might not here uses data from Malawi’s third Integrated Household matter as much. If, however, we consider a collective Survey collected in 2010/11, the second wave of the household model in which individual preferences Tanzanian National Panel Survey collected in 2010/11, matter, it becomes imperative to conduct analysis at and the 2011/12 wave of the Uganda National Panel the plot level with identification of the plot manager. Survey. The economic literature (such as Duflo and Udry These surveys are nationally representative and 2004) provides a multitude of examples suggesting link welfare, agriculture, and income. The data are that the collective household model may indeed be disaggregated at the plot level and contain information the more appropriate approximation of reality; these on which member of the household makes agricultural include evidence on the importance of the gender decisions about each of the plots cultivated by the of the recipient of cash transfers to household-level household. There are some differences across these data outcomes. sets in terms of assigning responsibility for each plot cultivated by the household. The Malawi questionnaire allows only one person to be listed as the decision maker, To express the gender gap in agricultural productivity while the Tanzanian and Ugandan data allow for multiple in monetary terms and to put these numbers into decision makers. Plots can be managed by women only, perspective relative to each country’s GDP, this study by men only, or by women and men jointly. In this study, maintains an additional set of assumptions. One key we only consider the difference in crop output obtained assumption is the absence of general equilibrium between women-only managed plots and all other plots. effects. For example, in the calculations presented, In Malawi, for example, women make decisions on about 26 percent of all agricultural plots; 76 percent of these increased productivity of women farmers affects plots are also actually owned by them, suggesting a neither male farmers’ productivity nor agricultural strong relationship between ownership and decision- prices. While there are good reasons to believe that making power, but there is no one-to-one correspondence general equilibrium effects such as these exist, the between plot management and land ownership or direction of these effects can go either way. For household headship. instance, while standard economic theory would predict lower prices when increased supply of All other macrolevel statistics, such as agricultural GDP and national GDP, are obtained from the 2015 World agricultural produce meets unchanged demand Development Indicators. in a closed economy, household nutrition may benefit from both the price and the income effect of increased agricultural productivity. among women and men who belong to the same household but cultivate different plots. The main advantage of this level of analysis is that it explicitly measures the productivity of women farmers who are frequently neglected in analytical work that only looks at the gender of the household head. THE COST OF THE GENDER GAP IN AGRICULTURAL PRODUCTIVITY IN MALAWI, TANZANIA, AND UGANDA  / 9 Malawi In Malawi, closing the unconditional gender gap equates to a 2.2 percent reduction in the poverty In Malawi, the unconditional gender gap is estimated headcount, which is equivalent to more than 238,000 to be 28 percent. The costs of this unconditional people being lifted out of poverty each year.2 gender gap equate to Given the small difference between the unconditional ▲▲ 7.3 percent of current crop production; or and conditional gender gaps in Malawi, the corresponding figures for closing the conditional ▲▲ 6.1 percent of agricultural GDP (or about gender gap are very similar: a 2.4 percent reduction $90 million);1 or in poverty, with nearly 264,000 people lifted out of poverty every year. ▲▲ 1.85 percent of total GDP (or approximately $100 million), including the multiplier effects of Tanzania benefits to other sectors in the economy. In Tanzania, the unconditional gender gap of If we base the estimates on the conditional gender 16 percent represents gap of 31 percent instead, then the costs of the gap equate to ▲▲ 2.1  percent of current agricultural output; or ▲▲ 8.1 percent of current crop output, or ▲▲ 1.5 percent of agricultural GDP (or over $85 million);3 or ▲▲ 6.7 percent of agricultural GDP, or ▲▲ 2.1 percent of total GDP (or about $110 million). The poverty growth elasticity used in the all-sector growth 2  scenario is −0.76; that used in the agriculture-led growth scenario is −1.19 (Dorosh and Thurlow 2014). Crop output accounts for 83 percent of agricultural GDP 1  Crop output makes up approximately 70 percent of 3  in Malawi. agricultural GDP in Tanzania. Closing  $100 million $90 million increase in increase in total GDP the agricultural GDP gender 4 $ 7.3% increase in 238,000 28% people lifted 4 $ crop production out of poverty gap in  Malawi  10 / THE COST OF THE GENDER GAP IN AGRICULTURAL PRODUCTIVITY IN MALAWI, TANZANIA, AND UGANDA Closing  $105 million $85 million increase in increase in total GDP the agricultural GDP gender 4 $ 2% increase in 80,000 people 16% 4 $ crop production lifted out gap of poverty; 80,000 in more people  adequately  Tanzania nourished ▲▲ 0.46 percent of total GDP (or about $105 million), ▲▲ 2.7 percent of agricultural GDP, or including the multiplier effects of benefits to other sectors in the economy.4 ▲▲ 0.86 percent of total GDP (or about $196 million). The above estimates only take into account the Building on poverty-growth elasticities derived from benefits to women who manage their own plots. an economywide general equilibrium approach However, gendered agricultural policies may also reported in Dorosh and Thurlow (2014), we calculate benefit women farmers who jointly manage their plots the potential benefits of closing the gender gap in with men—which is the case for about 54 percent terms of poverty reduction and nutrition. The gross of the plots. Even if only about one-fifth of these gains from closing the unconditional gender gap jointly managed plots also realize higher productivity, in agricultural productivity translate into an annual then the costs of the gender gap would equate to 0.41 percent reduction in the poverty headcount, 3.8 percent of the current level of crop production, which is equivalent to nearly 80,000 people being 2.7 percent of agricultural GDP, and 0.84 percent of lifted out of poverty every year.5 total GDP. Closing the gender gap could lead to a 0.72 percent If we consider the conditional gender gap of reduction in the incidence of undernourishment, 30 percent, the estimated costs are almost twice as which equates to more than an additional 80,000 high, equating to people being adequately nourished every year.6 ▲▲ 3.9 percent of crop output, or 5  These estimates are based on poverty-growth elasticity for Tanzania experienced in all sectors of the economy, which is −0.49. 6 Undernourishment is defined as the share of the population For Tanzania, this multiplier effect is estimated to be 4  consuming less than 2,550 kilocalaries available per male around 1.23. adult equivalent (Pauw and Thurlow 2011). THE COST OF THE GENDER GAP IN AGRICULTURAL PRODUCTIVITY IN MALAWI, TANZANIA, AND UGANDA  / 11 Uganda If we use the conditional gender gap of 28 percent, then the costs of the gap equate to10 The unconditional gender gap in agricultural productivity in Uganda is 13 percent.7 The costs of ▲▲ 6.1 percent of current crop production, or this unconditional gender gap equate to ▲▲ 3.6 percent of agricultural GDP (or about ▲▲ 2.8 percent of current crop output; or $126 million), or ▲▲ 1.6 percent of agricultural GDP (or about ▲▲ 0.9 percent of total GDP (or about $145 million).11 $58 million);8 or Combing the gross gains in GDP with the poverty- ▲▲ 0.42 percent of total GDP (or nearly $67 million), growth elasticities reported by Dorosh and Thurlow including the multiplier effects of benefits to other (2014), we estimate that potential poverty reduction sectors in the economy.9 benefits of closing the unconditional gender gap equate to a 0.9 percent reduction in poverty in Uganda, or approximately 119,000 people being lifted out of poverty. 7 In Uganda, 27 percent of plots and 20 percent of all Basing the estimates on the conditional gender cultivated land is under the sole management of women; the remaining 73 percent of plots and 80 percent of all gap, closing the gender gap would be equivalent to cultivated land is managed either jointly by women and a 2.0 percent reduction in poverty or nearly 260,000 men or solely by men. people being lifted out of poverty. Crop production accounts for 59 percent of agricultural 8  GDP in Uganda. We use a multiplier of 1.11 as the benefits of raising 9  In Uganda, women cultivate plots that are on average 10 agricultural production also include spillovers to other 23 percent smaller than those of male farmers. sectors in the economy. We also assume that closing the gender gap influences all agricultural sectors equally in Spillover effects and economywide linkages are taken 11  Uganda. into account in estimating this GDP benefit. Closing  $67 million $58 million increase in increase in total GDP the agricultural GDP gender 4 $ 2.8% increase in 119,000 13% 4 $ crop production people lifted gap out of poverty in  Uganda  12 / THE COST OF THE GENDER GAP IN AGRICULTURAL PRODUCTIVITY IN MALAWI, TANZANIA, AND UGANDA Costing the factors that contribute to the gender gap in agricultural productivity I n this section, we provide the results from decomposition analysis to identify which BOX 4.1 constraints to women’s agricultural productivity Quantifying the benefits from contribute most to the gender agricultural productivity narrowing the gender gap in gap. This information can help governments prioritize agricultural productivity by specific those policies that are likely to have the biggest determinants impact on closing the gap in agricultural yields. The decomposition analysis (see box 4.1) extracts the P lots managed by women may be less productive than those managed by men due to observable factors, such as differences in experience and education, importance of specific determinants of agricultural land quality, quantity of agricultural inputs used, and productivity in terms of potential gross gains in the choice of crops grown. However, an agricultural productivity difference could persist even when women GDP. 1 We study several determinants, including and men have similar observable characteristics and use manager characteristics, household demographics, the same quantity of inputs, as women may derive lower household wealth, plot characteristics, crop choice, returns from using these inputs. The Oaxaca-Blinder use of fertilizer, farming techniques, and labor inputs. decomposition approach (Blinder 1973; Oaxaca 1973) The policy recommendations are framed by the has been widely used in other areas of the economics choice of these determinants. Table 4.1 provides an literature such as in studies analyzing the wage gap overview of those determinants that stand out in between male and female workers (see, for example, World Bank 2011). This decomposition method can also terms of impact potential. The choice of determinants be employed to determine how much of the gender gap was based on data availability. It is recognized that arises from the different levels of agricultural inputs a number of these determinants are proximate used by women and men and how much arises from the lower returns that women obtain from using these inputs. For more detail on the Oaxaca-Blinder decomposition method, see appendix B. 1  This analysis builds on earlier work by Kilic, Palacios- Lopez, and Goldstein (2015), Slavchevska (2015), and Ali et al. (2015). Similar analysis is carried out for Ethiopia, Niger, and Nigeria by Aguilar et al. (2015), Backiny-Yetna and McGee (2015), and Oseni et al. (2015), respectively. 13 TABLE 4.1 Relationship of selected determinants to the gender gap in agricultural productivity and to GDP MALAWI TANZANIA UGANDA % of In terms of % of % of In terms of % of % of In terms of % of  Determinant gap GDP ($) GDP gap GDP ($) GDP gap GDP ($) GDP Qty of male family labor per household 45.19 45,110,180 0.84 97.34 102,180,543 0.45 n.a.  n.a.  n.a.  High-value crops 28.43 28,378,296 0.53 3.00 3,153,441 0.01 13.29 8,872,253 0.06 Agricultural implements 17.76 17,722,900 0.33 8.18 8,591,710 0.04 9.02 6,021,846 0.04 Pesticide use 0.97 964,601 0.02 12.03 12,630,384 0.06 4.45 2,973,106 0.02 Inorganic fertilizer use 5.32 5,313,775 0.10 6.39 6,707,789 0.03 3.04 2,026,367 0.01 Wealth index 3.29 3,288,461 0.06 −0.10 −106,908 0.00 n.a.  n.a.  n.a.  Note: n.a. = not available. Statistically significant factors are marked in bold. GDP values are 2010 dollars. Percentages may not sum to 100 because a number of determinants can be negative. Only a selection of those that reduce the gender gap are shown here, and together, they may overcompensate. causes that can be linked to ultimate causes of the in Tanzania and 45 percent in Malawi. For example, gender gap. A key challenge for future research will 97 percent of the gender gap is related to unequal be to understand which of these may be at play by access to male family labor in Tanzania. This could focusing on these ultimate determinants. Note that, potentially be linked to a number of other factors except where stated, the interventions discussed in including the segregation of tasks, rural women’s this section have not yet been rigorously evaluated. limited voice and agency, their lack of access to finance to hire male labor and invest in machinery, Women and men farmers have very and limited time-saving infrastructure. One key different levels of access to male reason that women farm managers have less access family labor to male family labor is that the majority of them are widowed, separated, or divorced: this is the case for A large part of the gender gap can be attributed to 67 percent of sole female plot managers in Tanzania. differential access to male family labor in Tanzania In fact, it is quite possible that these women became and Malawi. Equalizing the access to male family labor plot managers entirely because of their head-of- would reduce the estimated gender gap 97 percent household status. These high rates of widowhood, separation, and divorce mean that women have fewer people in the household to draw on for farm labor. In Access to male family labor Tanzania, the households of female farm managers have an average of one fewer person than all other makes up much of the gender gap in Tanzania and Malawi households. + $102 97% million In Malawi, women try to compensate for their lower Closing the gap = use of male family labor with more intense use of Tanzania potential gains in female labor, including themselves and, occasionally, national income + $45 their children. However, these additional inputs are 45% million not sufficient to offset the lost productivity brought Malawi about by lower use of male family labor. Closing the 14 / THE COST OF THE GENDER GAP IN AGRICULTURAL PRODUCTIVITY IN MALAWI, TANZANIA, AND UGANDA gap in the use of male family labor alone would lead to gains equal to 0.80 percent of national income in BOX 4.2 Malawi ($45 million in 2010 prices) and 0.45 percent Role of market access in gender in Tanzania ($102 million in 2010 prices). agricultural productivity gap POLICY IMPLICATIONS Designing policies that directly reduce inequality in I n Uganda, we examine whether selling more than 50 percent of crop output in the market also contributes to the gender gap in agricultural productivity. access to male farm labor can take two avenues. One We find that this factor alone explains 40 percent of option is to tackle constraints related to women’s the unconditional gender productivity gap in Uganda, access to household male labor. Another option is to equating to $26 million (in 2011 prices) of potential gross think about policies that help women farmers access gains in GDP. A multitude of factors could explain why women sell less of their output to the market. First, they substitutes for household labor, such as hired workers may face imperfect markets and thus prioritize their and labor-saving technologies. families’ food security by cultivating crops for home consumption. Moreover, because women tend to cultivate Women and men farm different crops smaller plots, they may not have enough produce left to sell to the market after fulfilling their families’ Women farmers are less likely to grow cash or export consumption needs (World Bank 2008). Second, women crops that men sell to the market for higher incomes. may lack access to various inputs, such as intensive labor, required for cultivating marketable crops (World In Malawi, the primary cash crop, tobacco, is only Bank 2008). Third, social norms may dictate which type planted on 3 percent of women’s plots compared to of crops women cultivate. 10 percent of men’s plots. Overall, there is a 28 percent gender gap between women and men in the fraction of land devoted to export crops in Malawi. Closing these gender gaps in the cultivation of cash crops Women are disadvantaged in has the potential of raising GDP over $28 million accessing agricultural machinery and in Malawi, $3 million in Tanzania, and $8 million in production technologies Uganda. In all three countries profiled, women’s access POLICY IMPLICATIONS to agricultural implements and machinery is Several complementary policies can play a significant significantly lower than that of men.2 Differences role. First, improving women’s control over marketed in the use of implements and machinery explain output so they can better take charge of the income 18 percent of the gender gap in Malawi, 8 percent in they earn has the potential to shift the underlying Tanzania, and 9 percent in Uganda. In Malawi, women conditions in which women farmers operate (Hill own fewer agricultural implements and machinery, and Vigneri 2014). Second, strengthening women’s groups and networks so that women can sell in larger For Malawi, the agricultural implements index is based 2  quantities can help women reach markets and sell on ownership of a hand hoe, slasher, ax, sprayer, panga their produce at a lower cost. Box 4.2 further explores knife, sickle, treadle pup, watering can, ox cart, ox plow, the issue of market access. tractor, ridger, cultivator, generator, motorized pump, grain mill or other implements, chicken house, livestock kraal, poultry kraal, storage house, granary, barn, and/or pigsty. For Tanzania, the agricultural implements index includes ownership of carts, animal carts, wheelbarrow, livestock, donkeys, hoes, spraying machine, water pump, reapers, tractor, tractor trailer, plow, harvesting machine, hand miller, coffee pulper, and/or fertilizer distributor. THE COST OF THE GENDER GAP IN AGRICULTURAL PRODUCTIVITY IN MALAWI, TANZANIA, AND UGANDA  / 15 such as weighing machines, spraying pumps, panga et al. 2014). Therefore, it will be critical to carefully knives, axes, and irrigation equipment. In Tanzania, understand women’s machinery needs and, where ownership of livestock, spraying machines, water appropriate, challenge the existing social norms (for pumps, and plows is lower among women. example, via female-led tractor provision services). If the binding constraint for women farmers stems In addition to this gender mechanization gap, from the heavy upfront investment typically required women also use lower levels of advanced agricultural to purchase machinery, providing them with rental technologies, such as pesticide and inorganic or leasing options is a policy alternative. In addition fertilizer. About 12 percent of the gap, or $13 million, to promoting efficient and targeted use of pesticide in potential gross gains in Tanzania can be accounted and fertilizer use through voucher programs, small for by the gender difference in pesticide use. Lower nudges such as timeliness of delivery and smaller use of inorganic fertilizer by women is equivalent packages of fertilizer and seed more appropriately to potential gross gains in GDP of over $2 million sized for women’s smaller plots can potentially have in Uganda. In Tanzania, average use of inorganic a huge impact (Duflo, Kremer, and Robinson 2009).3 fertilizer continues to be low, irrespective of whether it is a female- or male-managed plot, so the difference Along with these short-term policy shifts, broader between women and men is not significant. policy changes such as reforming land rights in favor of women have the potential to increase women’s Overall, the gender gap in the use of agricultural long-term agricultural investments, even if unrelated mechanization and technology equates to national to machinery or fertilizer (Ali, Deininger, and Goldstein income of approximately $24 million in Malawi, 2014). This may further help address the problem $11 million in Uganda, and $21 million in Tanzania. of small fragmented farms that affect smallholder farmers of both sexes and that can render the use of POLICY IMPLICATIONS agricultural machinery uneconomical. Cash vouchers or in-kind transfers may help women increase their use of machinery. However, women are unlikely to purchase and operate heavy units of 3  Efficient and targeted use of pesticides should be machinery if it is inadequate for their needs or if it restricted to legal chemicals applied with caution to reduce is deemed culturally or socially inappropriate (Njuki risks to farmers and ecosystems. 16 / THE COST OF THE GENDER GAP IN AGRICULTURAL PRODUCTIVITY IN MALAWI, TANZANIA, AND UGANDA Finding the biggest bang for the buck: cost-effective policy interventions N ow that we know the costs of the gender Naturally, the relative cost-benefit ratio of various gap in agricultural productivity in Malawi, interventions should also be weighed against other Tanzania, and Uganda, as well as the factors factors, such as ease of implementation and cultural that contribute the most to this gap, it is critical to and social context.  identify the most cost-effective policies. There may be many policy options available, but clearly these will be of little practical use to policy makers if their Identify where the implementation is more costly than the value of the benefits they are able to achieve. By identifying some benefit of closing the of the policies that may have the highest benefit- to-cost ratio, we hope to provide a useful starting gender gap outweighs the cost of point for further analysis that could offer practical the respective policy option guidance for policy makers who need to work out how to respond to the gender gap while making best use of limited resources. Policy priority 1: Narrow the gender productivity gap due to lack of access This report has highlighted that lack of access to labor to adequate labor, crop choice, and low use of machinery and nonlabor technological inputs First, it is possible to increase women’s labor contribute to the majority of the gender gap across productivity by enabling them to adopt labor-saving the three countries studied. In the remainder of this technologies on farm or by freeing up their time by section, we outline some policy priorities that could adoption of labor-saving technologies at home such address these constraints; these policy options are as the use of energy-efficient and environmentally summarized in table 5.1. The next step for policy friendly improved cooking stoves. These stoves are makers is to engage in this cost-benefit analysis to widely regarded as a means to reduce the amount identify where the benefit of closing the gender gap of time required for fetching firewood and thus outweighs the cost of the respective policy option. increase the time available for productive work— 17 TABLE 5.1 Summary of potential policy options for addressing the gender gap in agricultural productivity Potential GDP gains to: Policy priority Policy instrument Malawi Tanzania Uganda Research priority LABOR POLICIES Enhance women’s use of Cash vouchers/discounts on purchase ▼▼Effective substitutes technologies that save for male labor time on farm and off farm Doorstep delivery and training ▼▼Constraints that prevent women from High High Low Agricultural cash vouchers hiring labor Improve access to hired male and female labor ▼▼Women’s needs Queue-jumping incentives for labor to provide and preferences for services to women farmers first technology adoption CROP CHOICE POLICIES Improve access to markets Encourage formation of groups so women can ▼▼Understand what and agricultural groups have access to markets makes women adopt certain crops Encourage crops Provide crops with nutritional benefits and ▼▼Examine the that match women’s related training constraints women preferences Medium Low Medium face in cultivating high-value crops and determine if those Market crops as women’s Promote crops so they are considered to be are different from crops both women’s and men’s crops constraints farmers generally face NONLABOR INPUT POLICIES Package fertilizer in small amounts Innovative delivery mechanisms such as free ▼▼Understand women’s Improve access and delivery machinery and incentivize use of inorganic equipment needs and organic fertilizer and Information and communication–based ▼▼Examine constraints pesticide nudges such as mobile phone reminders about that particularly using inputs influence women’s Medium Low Medium use of nonlabor Cash and in-kind transfers for input purchase inputs Agricultural cash or discounts on purchase ▼▼Understand mechanisms and Expand the use of Where direct use is inappropriate, encourage policy packages that culturally appropriate machinery custom-hiring markets might particularly machinery for women work for women Machine-use training tied to women’s working schedule 18 / THE COST OF THE GENDER GAP IN AGRICULTURAL PRODUCTIVITY IN MALAWI, TANZANIA, AND UGANDA with the additional potential benefits of reducing Lack of labor availability, as established by this deforestation rates and respiratory diseases. report, may prevent women from undertaking the cultivation of crops that demand heavy time burdens. Second, policies could focus on enabling women’s Women are more time constrained than men in many access to hired labor. Prevalent cultural norms may countries due to their domestic care and child-rearing prevent women from hiring male labor, especially if responsibilities. Moreover, differences in agricultural specific agricultural tasks are performed by women know-how, risk-taking ability, and preference toward and men separately (Fafchamps 2001). Hence, ensuring household food security constitute some policies involving both women and men, such as other reasons for women’s preference for food crops. awareness and sensitization campaigns, may be In fact, evidence from cocoa production in Ghana and needed to reform existing structures. coffee production in Uganda—both of which are cash crops for these countries—suggests that differences Little direct evidence exists of policies that directly in productivity between women and men disappear help remove the labor shortages that women face. when women have the same access to productive Potential policy options include cash vouchers for inputs and sell their produce in the same way as men. hiring labor, price discounts on the purchase of Yet the reality in which women operate seldom gives labor-saving machinery, and doorstep delivery of them the same access to markets and methods of machinery and training (Duflo, Kremer, and Robinson cultivation (Hill and Vigneri 2014). Accessing markets 2009; Seidenfeld et al. 2014). However, the lack of may also be problematic for women due to cultural evidence on the effectiveness of these programs norms and women’s lower access to transport, means that it will be necessary to experiment with both of which restrict their mobility. In light of these and rigorously evaluate pilot programs that determine cultural and social circumstances, engaging men in which policy innovations work and which do not. promoting gender equality and challenging these social norms with women will be key to progress. Policy priority 2: Enable women  farmers to move into cultivation of high-value cash crops Engage men in promoting gender A key finding of this report is that women are less likely to grow cash crops and that this plays a significant equality and challenging social role in the gender productivity gap. Too often, women may shy away from growing higher-value crops due norms together with women to labor or cash shortages, especially if growing cash crops is culturally seen as a male activity (Hill There are a number of policy options for either and Vigneri 2014). Because women cultivate smaller enabling women to raise their productivity for the plots, they may not be inclined to cultivate cash crops they already grow or for incentivizing them crops because of the need to scale up (Fafchamps to shift into more profitable crops. Strengthening 1992). Moreover, because women seldom own land female farmer groups may allow women to not only and/or have weak land tenure rights, they may be scale up investments but also access markets by less motivated to make investments in cash crop reducing unit costs. Such interventions can also cultivation (Goldstein and Udry 2008; Morrison, Raju, allow women to address labor shortages by receiving and Sinha 2007). help from others in the group (Hill and Vigneri 2014). Understanding what women want in terms of crop cultivation is also crucial, especially if they prefer THE COST OF THE GENDER GAP IN AGRICULTURAL PRODUCTIVITY IN MALAWI, TANZANIA, AND UGANDA  / 19 growing crops that embody certain characteristics vouchers for fertilizer may still lead to low uptake. For such as nutritional value. In such cases, a gradual 1 instance, in Mozambique, an experimental evaluation introduction of cash crops may be required. Another of the fertilizer subsidy program found that farmers’ option is to promote certain crops as women’s uptake of fertilizer was relatively low and had varied crops, although the policy design for this intervention impact on farmers’ incomes and yields. Relatively low is somewhat complicated given the shifting cultural uptake may not only be related to lack of credit but norms with women taking up roles that traditionally could be because farmers do not know how to use fell within a man’s domain (Saito et al. 1994). the technology effectively (Carter, Laajaj, and Yang 2013). Third, we know that fertilizer subsidies may not Policy priority 3: Improve women be as successful in improving uptake as small, time- farmers’ access to and use of nonlabor limited discounts, such as free delivery of fertilizer inputs in agricultural production right after the harvest season (Duflo, Kremer, and Robinson 2009). Several policies, both gendered and not, have been implemented and evaluated to improve the uptake Despite the diversity of evaluation work done in this of technologies such as fertilizer, pesticide, and area, much remains to be learned about what works. improved seed varieties (Peterman, Behrman, and For example, information and communication–based Quisumbing 2014), as well as other rural technologies nudges in the form of mobile phone reminders could that can save time and increase farm productivity. be tried. Similarly, packaging fertilizer and improved When it comes to policies aimed at improving the seed varieties in smaller quantities (to prevent adoption rate of such technologies, we know more spoilage and to make it more convenient for women to about what does not work than about what does use on their smaller plots) could be tried to see if such work, especially for sustained, long-term use of these an intervention would improve uptake of fertilizer. nonlabor inputs. First, we know that even if fertilizer is Moreover, farmer training on sustainable input use given free to women farmers, it may not necessarily could be scheduled at a time that fits with women’s improve farm profit since it increases spending on household and farm schedules to see if it improves other complementary inputs as well—which has huge learning about technologies. Needless to say, these consequences for the sustainability of such a policy policies should be formulated and evaluated with (Beaman et al. 2013). Second, we know that giving careful consideration of the larger constraints within which women farmers operate, such as low access to land and credit, insecure land tenure, limited access to appropriate technologies, mobility-related 1 See, for example, evidence on women’s preference for growing orange flesh sweet potato in Uganda in challenges that may prevent women’s ability to reach Quisumbing et al. (2014). input dealers and markets, and risk-taking abilities. 20 / THE COST OF THE GENDER GAP IN AGRICULTURAL PRODUCTIVITY IN MALAWI, TANZANIA, AND UGANDA Moving from recommendations to implementation T his report has highlighted the importance So what are some characteristics of good and cost- of fully including women in the agricultural effective policies that narrow the gender gap in development process. The three countries agriculture? profiled in this report are meant as an illustration  of the kinds of cost-benefit analyses that donors, governments, and international organizations We know very little about should be undertaking before investing their scarce resources in various policies. Yet we know very little what exactly works in about what exactly works in narrowing the gender narrowing the gender gap and how gap and how much it costs. much it costs. The next stage needs to be to find cost-effective solutions through combining the implementation of Good policies work on improving innovative pilot interventions with careful evaluation. choices Because the gender gap in agriculture operates within the broader context of the bigger gender gap If the aim of development policy is to ensure that in society, it is important that policy makers, donors, women become more productive, then policy makers and development partners carefully consider their should carefully consider if women are operating understanding of which key problems women out of choice or constraints. Since there is a thin line face, why particular policies would work, and what between the two, agricultural gender policy should operational challenges they may face when trying be cognizant of how women farmers make their to actually implement policies. Because the gender agricultural decisions. Various policy instruments gap is deeply cultural and societal, it is imperative affect women’s constraints and choices differently. that policy makers use a combination of economic and behavioral shifts to narrow the gender gap in agriculture. 21 Good policies are built upon refined them on their smaller plots. Carefully refining and and redefined problems redefining policy scope is critical to maximizing benefits from closing the gender gap. Investing in carefully diagnosing and refining the scope of problems can significantly reduce Good policies may have to shift implementation costs and ensure that policies cultural norms are cost-effective. Lessons can be learned from experiments and research in other development Government agencies, donors, and development efforts. For example, lack of access to clean water practitioners work within embedded social and was diagnosed as one of the factors leading to a cultural norms. Attacking the problem of the gender high number of cases of diarrhea among children in gap in agricultural productivity first begins with rural Kenya. One intervention implemented was to shifting the mindset through which policy is framed cover water springs at the source in order to avoid and implemented. It requires making it acceptable contamination. Yet the intervention only moderately for women to cultivate cash crops and agricultural helped improve the quality of water at home (Ahuja, machinery. It means that it must be acceptable Kremer, and Zwane 2010). Closer diagnosis revealed for women to hire male labor and that men find it that the problem was in fact the contamination of acceptable to work for a woman. Tools that may the water at home. be particularly useful here are behavioral policy instruments such as identity cues and framing, Similarly, it is quite possible that the gender gap in microincentives, and reminders. Policy makers, agricultural productivity is not caused by a lack of donors, and international agencies must reassess the access to fertilizer per se, but to a lack of fertilizer realities under which they frame agricultural policies. marketed in small quantities so women can use 22 / THE COST OF THE GENDER GAP IN AGRICULTURAL PRODUCTIVITY IN MALAWI, TANZANIA, AND UGANDA APPENDIX A Methodology for quantifying the cost of the gender gap in agriculture We estimate agricultural productivity in terms of gross value per hectare, and A is the total arable land,2 value of output (in local currency) per unit of land (in which can be obtained from the World Bank’s World hectares). We obtain the quantity produced of each Development Indicators (http://data.worldbank.org/ crop on each plot and multiply total crop quantity by products/wdi). We express the mean harvest value the median crop sale value per appropriate unit in the per hectare on female plots (female productivity, Yf ) respective enumeration area. If village-level unit sale in terms of the mean harvest value per hectare on prices are not available for some crops, we use the male-managed plots (male productivity, Ym ) using prices available for the next higher level geographical the estimate of the gender gap—say 28 percent in area. Next, we add the values of output of all the crops Malawi—in the following manner: grown on the plot and divide the aggregate value of Yf = 0.72Ym (A.2) output by the plot size in order to obtain the gross value of output per hectare.1 The difference in these Total harvested value obtained from women’s and values of output per hectare obtained on male- and men’s cultivated land at the national level is expressed female-managed plots constitutes the unconditional as below. gender gap in agricultural productivity. Q = YfPA + Ym(1−P)A (A.3) Based on the identified gender gap in agricultural Here P represents the proportion of land controlled productivity and the estimate of the share of land by female managers based on the fraction of plots under women’s control, we can monetize the controlled by women. This fraction is based on the gender gap in terms of potential gains in agricultural average area of their plots relative to the average production and total economic output. To do this, the area of men’s plots. In Malawi, for example, women following formula to estimate the total quantity of plant 26 percent of all plots, but because theirs are, output obtained by women and men at the national level is useful (FAO 2011): Q = Y *A (A.1) 2 Arable land includes land under temporary crops (double- cropped areas are counted once), temporary meadows for Here Q is the total harvested output (in local currency mowing or pasture, land under market or kitchen gardens, units for the year of the survey), Y is the mean harvest and land temporarily fallow. For more information, see the World Development Indicators table notes (available at http://data.un.org/_Docs/WDI%20definitions.pdf). Since arable land includes plots that are temporarily fallow, it may 1 Ideally, plot size data measured by global positioning be useful to adjust the estimate by obtaining an estimate system (GPS) should be utilized, but GPS-measured of fallow land from the microlevel surveys and subtract area data are usable only in the case of Malawi. GPS- that fraction from the total arable land to better estimate measured data were collected for about 80 percent of the cultivated land. Often, farmers’ reports of fallowing are plots in Tanzania; thus, using GPS data would drop about rather low. For example, in the Malawi data only around 20 percent (1,312) of the plots from the analysis. 1 percent of the 18,990 plots are listed as fallow. THE COST OF THE GENDER GAP IN AGRICULTURAL PRODUCTIVITY IN MALAWI, TANZANIA, AND UGANDA  / 23 on average, 0.046 hectares smaller than men’s plots, crop production forms 83 percent of total agricultural the proportion of area under women’s control is about GDP. This 7.3 percent higher crop output translates 24 percent. to a 6.06 percent higher agricultural GDP, which is roughly around $89.9 million (in 2010 prices). Substituting equation A.2 into equation A.3 gives the total harvested value, Q, in the presence of the Because of the many economywide spillover effects identified gender gap in agricultural productivity. We between the agricultural sector and the rest of the term this scenario the baseline. We can also obtain economy, total GDP is expected to be higher by more the potential harvest value, Q*, under the scenario than the $89.9 million. We need an estimate of the of no gender gap in agricultural productivity, that is, multiplier between the agricultural sector and the when Yf = Ym . rest of the economy. Here we draw on economywide models for each country. For instance, the multiplier The additional output from closing the gender gap for Malawi is about 1.11, implying that each additional in agricultural productivity, as a proportion of the dollar generated in the agricultural sector leads to an baseline harvest value, is expressed as follows. additional $0.11 in benefits in the non-agricultural sector (Benin et al. 2008). Consequently, the Δ = (Q*−Q)/Q (A.4) $89.9 million higher agricultural GDP in Malawi due In Malawi’s case, closing the unconditional gender to closing the agricultural gender productivity gap gap will lead to an increase of total crop harvest of results in a total benefit of $99.8 million added to total 7.3 percent. GDP. Overall, total GDP will be higher by 1.85 percent if the gender gap in agricultural productivity is closed. To link the increase to agricultural GDP and total GDP, we need a few more pieces of information. First, we need to know what fraction of agricultural GDP comes from crop production.3 For example, in Malawi 3 Agricultural GDP includes forestry, hunting, finishing, livestock, and crop production (again, see the World Development Indicators table notes available at http:// data.un.org/_Docs/WDI%20definitions.pdf). We separate the specific contribution of each subsector of agriculture to total agricultural GDP from the national account statistics of each country. For Malawi, the national accounts report a combined figure for crop production, livestock production, and hunting. A 2005 FAO country brief reports from agricultural GDP to obtain an estimate of crop GDP, that livestock constitutes 9.9 percent of agricultural which is about 83 percent of agricultural GDP. Using the GDP. We therefore assume that livestock production as Malawi Social Accounting Matrix for 2004, Benin et al. a percentage of agricultural GDP has remained largely (2008) estimate that crop GDP is close to 86 percent of unchanged since those earlier studies and is about agricultural GDP. The difference between the two sources is 10 percent. In 2010, fisheries and forestry constituted about small and is driven by the different estimates of the size of 7 percent of agricultural GDP in Malawi. We subtract the the livestock sector. As is our practice throughout the paper, contributions of livestock, fisheries, and forest production we report results with the more conservative estimates. 24 / THE COST OF THE GENDER GAP IN AGRICULTURAL PRODUCTIVITY IN MALAWI, TANZANIA, AND UGANDA APPENDIX B Methodology for costing the factors of production contributing to the gender gap in agricultural productivity Plots managed by women farmers may be less managed by a woman, and 0 otherwise. M is a productive due to observable factors including vector of explanatory variables pertaining to other inequalities in manager attributes such as experience characteristics of the plot manager; X is a set of plot- and education, plot characteristics, agricultural level characteristics including size and quality; I is a technology and input use, and crop choice. A gender vector of plot-level controls for nonlabor input use; gap may persist even after accounting for these L is a set of plot-level controls for labor inputs; C is a factors. For example, after controlling for manager vector of indicator variables accounting for whether characteristics, plot characteristics and size, input the primary crop cultivated on the plot is a cash crop;2 use, and geographical features, the gender gap in ε is an error term. The term λh captures community Malawi decreases to 0.02 percent and is no longer and geographical characteristics. statistically significant at any level. The portion of the gap that cannot be explained by observable factors To closely compare results across the three countries, may be associated with differences in the returns we define the variables in a similar way and control associated with using these factors of production on for the same set of variables wherever possible and women’s plots as compared to men’s. To determine meaningful. By doing this, we develop a comparable exactly how much of the gap is due to levels of framework for analysis and discussion so that inputs used and how much is because of returns to differences in outcomes across countries are not those inputs, we employ an Oaxaca-Blinder-type linked to differences in definitions of the variables or decomposition. The central piece in the Oaxaca- set of variables included, but instead to differences Blinder decomposition approach is the following in the levels and coefficients associated with those production function. variables. This is not always possible because survey questions may not be structured in the same way. For 1n(Yih) = c0 + αFih + Mihγ + Xihδ + 1n(Iih)η + example, the agricultural implement indexes that we 1n(Lih)θ + Cihθ + λh + εih (B.1) construct for each country are based on different Here i denotes the plot planted by a member of agricultural assets. In some countries, we account household h; Y is the value of agricultural output for livestock and oxen power; in others, we do not. per unit of land (hectare); F equals 1 if the plot is 1 This difference in definition may perhaps capture 1 There are two agricultural seasons in all countries. But for production during the long rainy season. If women’s we estimate the gender gap using data for the long access to productive resources is even more limited when rainy season or the main season only, as the majority total resources are generally low, then the gender gap in of households cultivate land then. We assume that the productivity may be even larger. gender gap will be similar in the shorter rainy season. Such an assumption may not hold perfectly, but it is quite 2 The primary crop is often identified by the respondent of likely that farmers reserve most of their limited resources the survey. THE COST OF THE GENDER GAP IN AGRICULTURAL PRODUCTIVITY IN MALAWI, TANZANIA, AND UGANDA  / 25 the varying significance of agricultural implements R = Q + U (B.5) in explaining the gender agricultural productivity gap where in different countries. Q = (E(Xm ) − E(Xf ))'β* The Oaxaca-Blinder decomposition attempts to explain how much of the mean outcome difference gives the proportion of the gender productivity gap between two groups (female- and male-managed that results from group differences in the predictors plots) are accounted for by group differences in the (referred to in the literature as level effect); and predictors. The aggregate decomposition follows U = (E(Xm )'(βm − β*) + E(Xf )'(β* − βf)) from the linear model specified below. Yl = Xl' βl + εl, E(εl ) (B.2) is the residual or unexplained part that results from unequal returns to the predictors (structural effect) where l ∈ (f, m) and stands for female-managed (Aguilar et al. 2015; Blinder 1973; Fortin, Lemieux, and plots (f )or male-managed plots (m),3 X is a vector Firpo 2011; Jann 2008; Oaxaca 1973). of predictors (and a constant term), and β is a vector of slope coefficients including the intercept. We can The nondiscriminatory vector of coefficients β* can write the gap as be estimated in a number of ways (Fortin, Lemieux, R = E(Ym ) − E(Yf ) = E(Xm )'βm − E(Xf )'βf (B.3) and Firpo 2011; Jann 2008). Here β* is estimated from a pooled regression over all plots, with a where E(εl ) = 0. dummy variable identifying group membership (plots managed by a woman versus plots managed by a Using algebraic manipulations, the expression man as suggested in Jann 2008 and Fortin, Lemieux, in equation B.3 can be rewritten into a part of the and Firpo 2011). differential due to differences in the levels of the predictors and a part due to differences in the The primary focus from the decomposition results coefficients associated with the predictors. The is on the contribution of differences in the levels latter part is often referred to as the discrimination of factors of production to the gender agricultural component, especially if it is linked to an immutable productivity gap. The main goal is to estimate how characteristic such as gender (Fortin, Lemieux, much additional output could be obtained from and Firpo 2011). We assume that there is some closing the gender gap in accessing the various nondiscriminatory coefficient vector β* through which factors of production that contribute most to the the difference in the predictors is weighted so that gender productivity gap. For example, if differences R = (E(Xm ) − E(Xf ))'β* + (E(Xm )'(βm − β*) in fertilizer use explain a significant fraction of the + E(Xf )'(β* − βf)) (B.4) gender gap in agricultural productivity, then we discuss how much of the benefits associated with The expression in equation B.4 provides a twofold closing the gender gap in productivity could be decomposition, achieved by closing the gender gap in access to fertilizer. While equitable access to production factors such as land, physical inputs, machines, and livestock Under male-managed plots, we also include jointly 3  may have benefits beyond increasing agricultural managed plots, wherever joint management data is productivity, the approach taken here only focuses available such as in Tanzania and Uganda. on the benefits obtained from improved agricultural productivity by equalizing access to these factors. 26 / THE COST OF THE GENDER GAP IN AGRICULTURAL PRODUCTIVITY IN MALAWI, TANZANIA, AND UGANDA References Aguilar, A., E. Carranza, M. Goldstein, T. Kilic, and Blinder, A. S. 1973. “Wage Discrimination: Reduced G. Oseni. 2015. “Decomposition of Gender Form and Structural Estimates.” Journal of Human Dif ferentials in Agricultural Productivity in Resources 8 (4): 436–55. doi: 10.2307/144855. Ethiopia.” Agricultural Economics 46.3: 311–34. Carletto, C., S. Gourlay, and P. Winters. 2013. Ahuja, A., M. Kremer, and A. P. Zwane. 2010. “From Guesstimates to GPStimates: Land Area “Providing Safe Water: Evidence from Randomized Measurement and Implications for Agricultural Evaluations.” Annual Review of Resource Economics Analysis.” Policy Research Working Paper 6550, 2 (1): 237–56. World Bank, Washington, DC. Akresh, R. 2005. “Understanding Pareto Inefficient Carter, M. R., R. Laajaj, and D. Yang. 2013. “The Intrahousehold Allocations.” IZA Discussion Impact of Voucher Coupons on the Uptake of Papers 1858, IZA, Bonn. Fertilizer and Improved Seeds: Evidence from a Randomized Trial in Mozambique.” American Ali, D. A., D. Bowen, K. Deininger, and M. F. Duponchel. Journal of Agricultural Economics 95 (5): 1345–51. 2015. “Investigating the Gender Gap in Agricultural Productivity: Evidence from Uganda.” Policy Chung, K. 2012. “An Introduction to Nutrition- Research Working Paper 7262, World Bank, Agriculture Linkages.” MINAG/DE Research Washington, DC. Report 72E, Directorate of Economics, Ministry of Agriculture, Maputo, Mozambique. Ali, D. A., K. Deininger, and M. Goldstein. 2014. “Environmental and Gender Impacts of Land Dorosh, P., and J. Thurlow. 2014. “Beyond Agriculture Tenure Regularization in Africa: Pilot Evidence versus Nonagriculture: Decomposing Sectoral from Rwanda.” Journal of Development Economics Grow th–Pover ty Linkages in Five African 110: 262–75. Countries.” IFPRI Discussion Paper 1391, International Food Policy Research Institute, Backiny-Yetna, P., and K. McGee. 2015. “Gender Washington, DC. Differentials in Agricultural Productivity in Niger.” Policy Research Working Paper 7199, World Bank, Duflo, E., M. Kremer, and J. Robinson. 2009. “Nudging Washington, DC. Farmers to Use Fertilizer: Theory and Experimental Evidence from Kenya.” NBER Working Paper No. Beaman, L. A., D. Karlan, B. Thuysbaert, and C. R. Udry. 15131, National Bureau of Economic Research, 2013. “Profitability of Fertilizer: Experimental Cambridge, MA. Evidence from Female Rice Farmers in Mali.” NBER Working Paper No. 18778, National Bureau Duflo, E., and C. Udry. 2004. “Intrahousehold of Economic Research, Cambridge, MA. Resource Allocation in Côte d’Ivoire: Social Norms, Separate Accounts and Consumption Choices.” Benin, S., J. Thurlow, X. Diao, C. McCool, and NBER Working Paper 10498, National Bureau of F. Simtowe. 2008. “Agricultural Growth and Economic Research, Cambridge, MA. Investment Options for Poverty Reduction in Malawi.” IFPRI Discussion Paper 74, International Fafchamps, M. 1992. “Cash Crop Production, Food Food Policy Research Institute, Washington, DC. Price Volatility, and Rural Market Integration in the Third World.” American Journal of Agricultural Economics 74 (1): 90–99. THE COST OF THE GENDER GAP IN AGRICULTURAL PRODUCTIVITY IN MALAWI, TANZANIA, AND UGANDA  / 27 —. 2001. “Intrahousehold Access to Land and Morrison, A. R., D. Raju, and N. Sinha. 2007. “Gender Sources of Inefficiency: Theory and Concepts.” Equality, Poverty and Economic Growth.” Policy Access to Land, Rural Poverty and Public Action, Research Working Paper 4349, World Bank, edited by A. de Janvry, G. Gordillo, E. Sadoulet, Washington, DC. and J.-P. Platteau, 68–96. Oxford, UK: Oxford University Press. Njuki, J., E. Waithanji, B. F. Sakwa, J. Kariuki, E. Mukewa, and J. Ngige. 2014. “A Qualitative Assessment of FAO (Food and Agriculture Organization of the United Gender and Irrigation Technology in Kenya and Nations). 2005. “Livestock Sector Brief: Malawi.” Tanzania.” Gender, Technology and Development FAO, Rome. 18 (3): 303–40. —. 2011. The State of Food and Agriculture 2010–11. Oaxaca, R. 1973. “Male-Female Wage Differentials Women in Agriculture: Closing the Gender Gap for in Urban Labor Markets.” International Economic Development. Rome: FAO. Review 14: 693–709. For tin, N., T. Lemieux, and S. Firpo. 2011. Oseni, G., P. Corral, M. Goldstein, and P. Winters. 2015. “Decomposition Methods in Economics.” In “Explaining Gender Differentials in Agricultural Handbook of Labor Economics, 4th ed., edited by Production in Nigeria.” Agricultural Economics 46.3: D. Card and O. Ashenfelter, 1–102. Amsterdam: 285–310. Elsevier (North Holland). Palacios-Lopez, A., L. Christiaensen, and T. Kilic. 2015. Gilbert, R. A., W. D. Sakala, and T. D. Benson. 2002. “How Much of the Labor in African Agriculture Is “Gender Analysis of a Nationwide Cropping Provided by Women?” Policy Research Working System Trial Survey in Malawi.” African Studies Paper WPS 7282, World Bank, Washington, DC. Quarterly 6 (1): 223–43. Pauw, K., and J. Thurlow. 2011. “Agricultural Growth, Goldstein, M., and C. Udry. 2008. “The Profits of Poverty, and Nutrition in Tanzania.” Food Policy Power: Land Rights and Agricultural Investment 36 (6): 795–804. in Ghana.” Journal of Political Economy 116 (6): 981–1022. Peterman, A., J. A. Behrman, and A. R. Quisumbing. 2014. A Review of Empirical Evidence on Gender Hill, R. V., and M. Vigneri. 2014. “Mainstreaming Gender Dif ferences in Nonland Agricultural Inputs, Sensitivity in Cash Crop Market Supply Chains.” Technology, and Services in Developing Countries. Gender in Agriculture: Closing the Knowledge Gap, New York: Springer. edited by A. R. Quisumbing, R. Meinzen-Dick, T. L. Raney, A. Croppenstedt, J. A. Behrman, and Quisumbing, A. R., and D. Rubin, C. Manfre, A. Peterman, 315-341. New York: Springer. E. Waithanji, M. van den Bold, D. Olney, and R. Meinzen-Dick. 2014. “Closing the Gender Asset Jann, B. 2008. “The Blinder-Oaxaca Decomposition Gap: Learning from Value Chain Development in for Linear Regression Models.” Stata Journal 8 (4): Africa and Asia.” IFPRI Discussion Paper 1321, 453–79. International Food Policy Research Institute, Washington, DC. Kilic, T., A. Palacios-Lopez, and M. Goldstein. 2015. “Caught in a Productivity Trap: A Distributional Ruel, M. T., H. Alderman, and the Maternal and Child Perspective on Gender Differences in Malawian Nutrition Study Group. 2013. “Nutrition-Sensitive Agriculture.” World Development 70: 416–63. Interventions and Programmes: How Can They Help to Accelerate Progress in Improving Maternal Mabiso, A., K. Pauw, and S. Benin. 2012. “Agricultural and Child Nutrition?” Lancet 382 (9891): 536–51. Growth and Poverty Reduction in Kenya: Technical Analysis for the Agricultural Sectoral Development Saito, K. A., H. Mekonnen, and D. Spurling. 1994. Strategy (ASDS)—Medium Term Investment Plan “Raising the Productivity of Women Farmers in (MTIP).” ReSAKSS Working Paper 35, International Sub-Saharan Africa.” Discussion Paper Series 230, Food Policy Research Institute, Washington, DC. World Bank, Washington, DC. 28 / THE COST OF THE GENDER GAP IN AGRICULTURAL PRODUCTIVITY IN MALAWI, TANZANIA, AND UGANDA Seidenfeld, D., S. Handa, G. Tembo, S. Michelo, Tiruneh, A., T. Tesfaye, W. Mwangi, and H. Verkuijl. C. Harland Scott, and L. Prencipe. 2014. “The 2001. “Gender Dif ferentials in Agricultural Impact of an Unconditional Cash Transfer on Food Production and Decision-Making among Security and Nutrition: The Zambia Child Grant Smallholders in Ada, Lume, and Gimbichu Woredas Programme.” Institute of Development Studies, of the Central Highlands of Ethiopia.” International Brighton, UK. Maize and Wheat Improvement Center (CIMMYT) and Ethiopian Agricultural Research Organization Sheahan, M., and C. B. Barrett. 2014. “Understanding (EARO). the Agricultural Input Landscape in Sub-Saharan Africa: Recent Plot, Household, and Community- Udry, C. 1996. “Gender, Agricultural Production, and Level Evidence.” Policy Research Working Paper the Theory of the Household.” Journal of Political 7014, World Bank, Washington, DC. Economy 104 (5): 1010–46. Slavchevska, V. 2015. “Gender Dif ferences in World Bank. 2008. Gender in Agriculture Sourcebook. Agricultural Productivity: The Case of Tanzania.” Washington, DC: World Bank. Agricultural Economics 46.3: 335–55. —. 2011. World Development Report 2012: Gender Smith, L. C., U. Ramakrishnan, A. Ndiaye, L. Haddad, Equality and Development. Washington, DC: World and R. Martorell. 2003. “The Importance of Bank. Women’s Status for Child Nutrition in Developing Countries.” Research Report 131, International World Bank and ONE. 2014. Levelling the Field: Food Policy Research Institute, Washington, DC. Improving Opportunities for Women Farmers in Africa. Washington, DC: World Bank. THE COST OF THE GENDER GAP IN AGRICULTURAL PRODUCTIVITY IN MALAWI, TANZANIA, AND UGANDA  / 29 UN WOMEN, grounded in the vision of equality as enshrined in the Charter of the United Nations, works for the elimination of discrimination against women and girls; the empowerment of women; and the achievement of equality between women and men as partners and beneficiaries of development, human rights, humanitarian action, and peace and security. Placing women’s rights at the center of all its efforts, UN Women leads and coordinates United Nations system efforts to ensure that commitments on gender equality and gender mainstreaming translate into action throughout the world. It provides strong and coherent leadership in support of Member States’ priorities and efforts, building effective partnership with civil society and other relevant actors. The WORLD BANK GROUP (WBG) is a vital source of financial and technical assistance to developing countries around the world. The formation of a unified WBG has created a unique partnership working to end extreme poverty within a generation and boost shared prosperity. The World Bank Africa Region Gender Innovation Lab’s (GIL’s) analytical work underpins World Bank financing and helps inform the development communities’ investments. Through conducting impact evaluations, the GIL supports the design of innovative, scalable interventions to address gender inequality across Africa, with the goal of enabling project teams and policy makers to advocate for better gender integration using evidence. The POVERTY-ENVIRONMENT INITIATIVE (PEI) of the United Nations Development Programme (UNDP) and the United Nations Environment Programme (UNEP) supports country-led efforts to mainstream poverty-environment linkages into national development planning and budgeting. PEI provides financial and technical assistance to government partners to set up institutional and capacity-strengthening programs and carry out activities to address the particular poverty-environment context. PEI is funded by the governments of Norway, Spain, Sweden, the United Kingdom, and the European Union and with core funding of UNDP and UNEP.