THE COST OF GENDER INEQUALITY GENDER INEQUALITY, HUMAN CAPITAL WEALTH, AND DEVELOPMENT OUTCOMES IN UGANDA QUENTIN WODON AND ADENIKE ONAGORUWA AUGUST 2019 THE COST OF GENDER INEQUALITY GENDER INEQUALITY, HUMAN CAPITAL WEALTH, AND DEVELOPMENT OUTCOMES IN UGANDA QUENTIN WODON AND ADENIKE ONAGORUWA BACKGROUND TO THIS SERIES Reducing gender inequality makes economic sense In addition, many girls are married or have children apart from being the right thing to do. Achieving before the age of 18, before they may be physically gender equality and empowering all women and girls and emotionally ready to become wives and mothers. is the fifth sustainable development goal and is a top Women and girls also face higher risks of gender- priority for governments. Countries can achieve this based violence in their homes, at work, and in public goal if they take appropriate steps. This note is part spaces. Their voice and agency is often lower than of a series that aims to measure the economic cost of that of males, whether this is within the household, at gender inequality globally and regionally by examining work, or in national institutions. This also affects their the impacts of gender inequality in a wide range of areas children. For example, children of young and poorly and the costs associated with those impacts. Given educated mothers often face higher risks of dying by that gender inequality affects individuals throughout age five, being malnourished, and doing poorly in school. their life, economic costs are measured in terms of Fundamentally, gender inequality disempowers women losses in human capital wealth, as opposed to annual and girls in ways that deprive them of their basic human losses in income or economic growth. The notes also rights. aim to provide a synthesis of the available evidence on successful programs and policies that contribute This lack of opportunities for girls and women entails to gender equality in multiple areas and achieve the large economic costs not only for them, but also for Sustainable Development Goals (SDGs). their households and countries. Achieving gender equality would have dramatic benefits for women and In many countries, girls’ average educational attainment girls’ welfare and agency. This, in turn, would greatly remains lower than boys and adult women are less benefit their households and communities, and help literate than men. Apart from these gender gaps in countries reach their full development potential. It would educational attainment, discrimination and social norms reduce fertility in countries with high population growth, shape the terms of female labor force participation. as well as reduce under-five mortality and stunting, Women are less likely than men to join the labor force thereby contributing to ushering the demographic and to work for pay. When they do, they are more transition and the associated benefits from the likely to work part-time, in the informal sector, or in demographic dividend. occupations that have lower pay. These disadvantages translate into substantial gender gaps in earnings, which in turn decrease women’s bargaining power and voice. 1 | THE COST OF GENDER INEQUALITY: GENDER INEQUALITY, HUMAN CAPITAL WEALTH, AND DEVELOPMENT OUTCOMES IN UGANDA | AUGUST 2019 KEY RESULTS as knowledge of HIV-AIDS and birth registrations. This note has two objectives. The first is to estimate • In many of those areas, associated costs are high. This potential losses in national wealth due to inequality in is for example the case for impacts on fertility and earnings between men and women in Uganda. The second population growth which affect standards of living. is to document the impact of gender inequality in other areas such as fertility and population growth, indicators of • Two main factors lead women to have less earnings and child health, and measures of women’s agency. Overall, the thereby lower human capital wealth than men: lower analysis suggests that the development impact of gender labor force participation rates and fewer hours worked inequality is large. Key findings from the note are as follows: in the labor market, and lower pay. These factors keep many women in a productivity trap due in part to social • A country’s wealth is the basis for its long-term norms relegating them to unpaid care and informal work. development and the creation of income (Gross Domestic Product). The three main components • To increase human capital wealth, investments of wealth are produced capital, natural capital, and throughout the life cycle are needed, from early human capital. In Uganda, human capital – the childhood development and learning in schools to building value of the future earnings of the labor force, job-relevant skills that employers demand, encouraging accounts for 50 percent of the country’s wealth. entrepreneurship and innovation, and ensuring that women have the same access to opportunities and • Inequality in human capital wealth between men resources as men. Policies are also needed to tackle other and women remains substantial. In Uganda women aspects of gender inequality, such as child marriage. account for 39 percent of human capital wealth versus 61 percent for men. These estimates are similar to those observed for the sub-Saharan African region. INTRODUCTION • Women’s human capital could increase from US$102 billion to US$163 billion if women were earning as much This note has two main objectives. The first is to estimate as men, yielding a gain in wealth of up to US$61 billion. potential losses in national wealth due to inequality in earnings between men and women in Uganda. The second • Per capita, gender equality could increase wealth by is to document the impact of gender inequality in selected up to US$1,619, an increase of 11.8 percent versus the other domains, including fertility and population growth, base value of total wealth per capita in the country. health outcomes for young children, and measures of These estimates are not meant to be precise, but they women’s agency. give an order of magnitude of potential benefits. Consider first potential losses in national wealth due to • Gender inequality also has large impacts in other gender inequality. There is a substantial literature on the areas. Achieving gender equality could reduce impact of gender inequality on economic growth and the fertility rate by 11 percent from its base performance. By focusing on wealth, the approach used value. This could lead to a substantial reduction for measurement in this note is different. Wealth is the in population growth and thereby an increase in assets base that enables countries to produce income human capital wealth per capita and well-being. (Gross Domestic Product or GDP). A country’s wealth includes various types of capital. Produced capital comes • Gender equality could also lead to a reduction of under- from investments in assets such as factories, equipment, five mortality and stunting by respectively 13 percent or infrastructure. Natural capital includes assets such as and five percent. It could increase women’s decision- agricultural land and other renewable and non-renewable making ability within the household by more than a natural resources. However, the largest component of fourth. There could also be benefits towards reducing countries’ wealth often resides in their people. If gender intimate partner violence or improving indicators such equality in earnings were achieved, countries could increase AUGUST 2019 | THE COST OF GENDER INEQUALITY: GENDER INEQUALITY, HUMAN CAPITAL WEALTH, AND DEVELOPMENT OUTCOMES IN UGANDA | 2 their human capital wealth, and thereby their total wealth capital (this sums up to more than 100 percent due to substantially. This could enable them to strengthen the negative net foreign assets). The share of human capital sustainability of their development path, and thereby future in total wealth increased over time, while the share of growth in GDP per capita. The note provides estimates of natural capital decreased. In per capita terms, total wealth potential losses in national wealth in Uganda due to gender stood at US$13,731 in 2014 versus US$9,010 in 1995. inequality in earnings. This represents an annual growth rate of two percent, with the difference versus the growth in total wealth due Consider next the impact of gender inequality in other to population growth. Human capital wealth stood at areas. Gender inequality has implications not only for US$6,889 per person in 2014 versus US$3,026 in 1995. women’s earnings, but also for a wide range of other areas which for simplicity can be considered as pertaining to Most of the country’s natural capital consists of crop women’s roles as wives and mothers “at home”, as opposed land. There was a decrease in recent years in the value to their role “at work”. In practice, the spheres of home of crop land per capita that led to an overall decrease and work are not always easy to delineate, and they are not in total wealth per capita between 2010 and 2014. This independent of each other. But for expository purpose, decrease due essentially to a drop in the production of these impacts “at home” are considered separately. plantains according to data from the Food and Agricultural Specifically, three main types of impacts are considered: Organization. As production of plantains fell, this lead to impacts on (i) women’s total fertility towards the end of a drop in the valuation of cropland since that valuation is their reproductive age and thereby population growth, (ii) based on the discounted value of the sale of agricultural under-five mortality and stunting, and (iii) women’s agency. production in the future, with future expected production The note is structured as follows. The first two sections based on existing patterns production and expected gains provide estimates of national wealth for Uganda and losses in productivity. With the exception of the recent decrease from gender inequality in earnings. The next section in the value of agricultural land, which could be reversed in documents the impact of gender inequality in other areas. the future, virtually all categories of assets saw an increase The last section provides a few general guidelines on policy in value per capita over time. options to achieve gender equality. A brief conclusion follows. BASELINE ESTIMATES OF UGANDA’S NATIONAL WEALTH This section presents baseline estimates of human capital and total wealth from Lange et al. (2018) for Uganda. Tables 1 and 2 provide estimates in absolute value and per capita terms. All estimates are in constant US dollars of 2014. As mentioned earlier, total wealth includes natural capital, produced capital, human capital, and net foreign assets. Uganda’s wealth stood at US$519 billion in 2014. This represented a large increase in real terms of 182 percent over 20 years (annual growth rate of 5.3 percent per year). Human capital wealth reached US$260 billion in 2014, an even larger increase of 321 percent since 1995 (average annual growth rate of 7.5 percent). Human capital accounted in 2014 for half of total wealth, versus 38 percent for natural capital and 14 percent for produced 3 | THE COST OF GENDER INEQUALITY: GENDER INEQUALITY, HUMAN CAPITAL WEALTH, AND DEVELOPMENT OUTCOMES IN UGANDA | AUGUST 2019 Table 1: Baseline Estimates of Uganda’s Total Wealth (US$, millions) Millions, constant 2014 USD 1995 2000 2005 2010 2014 Total wealth 183,930 211,616 349,216 463,389 518,839 Produced capital (including urban land) 20,492 33,362 49,839 46,543 70,722 Natural capital 105,668 99,343 166,346 212,559 199,092 Forests, timber resources 649 622 815 975 1,027 Forests, non-timber resources 795 714 633 508 408 Protected areas 9,576 9,012 15,837 23,048 25,343 Land 94,648 88,995 148,913 187,729 172,284 Cropland 75,220 71,190 120,094 147,312 130,299 Pastureland 19,428 17,805 28,819 40,417 41,985 Sub-soil assets - - 148 299 30 Human capital 61,777 83,670 141,016 210,382 260,304 Net foreign assets -4,007 -4,758 -7,985 -6,096 -11,279 Population (millions) 20.4 23.8 28.0 33.1 37.8 Share of total wealth Produced capital (including urban land) 11% 16% 14% 10% 14% Natural capital 57% 47% 48% 46% 38% Human capital 34% 40% 40% 45% 50% Net foreign assets -2% -2% -2% -1% -2% Source: Lange et al. (2018). The profile of wealth of Uganda by broad asset categories as a whole due to drops in wealth in a few large countries is not unlike that of a typical sub-Saharan African country. such as the Dominican Republic of Congo and Nigeria. While on average wealth per capita in sub-Saharan Africa Still, there is scope for improving Uganda’s performance is at twice the level observed in Uganda, the shares of in building its wealth. At some point in the future, once wealth by broad type of asset are similar, with 16 percent exploitation starts, Uganda will benefit from wealth related for produced capital, 36 percent for natural capital, and 50 to oil production. But more fundamentally, as is the case percent for human capital in the region. Uganda however for other countries transitioning from low to middle income has seen its wealth increase substantially over time, while countries, human capital should increase further and this has not been the case in per capita terms for the region gender equality has a role to play. Table 2: Baseline Estimates of Uganda’s Per Capita Wealth (US$) Per Capita, constant 2014 USD 1995 2000 2005 2010 2014 Total wealth 9,010 8,907 12,453 13,979 13,732 Produced capital (including urban land) 1,004 1,404 1,777 1,404 1,872 Natural capital 5,177 4,182 5,932 6,412 5,269 Forests, timber resources 32 26 29 29 27 Forests, non-timber resources 39 30 23 15 11 Protected areas 469 379 565 695 671 Land 4,637 3,746 5,310 5,663 4,560 Cropland 3,685 2,997 4,283 4,444 3,449 Pastureland 952 749 1,028 1,219 1,111 Sub-soil assets - - 5 9 1 Human capital 3,026 3,522 5,029 6,346 6,889 Net foreign assets -196 -200 -285 -184 -299 Source: Lange et al. (2018). AUGUST 2019 | THE COST OF GENDER INEQUALITY: GENDER INEQUALITY, HUMAN CAPITAL WEALTH, AND DEVELOPMENT OUTCOMES IN UGANDA | 4 LOSSES IN HUMAN Researchers looking at the impact of gender inequality on development have typically focused on annual measures CAPITAL WEALTH DUE TO of income or growth in income (e.g. Elborgh-Woytek et al., 2013; Cuberes and Teigner, 2015; McKinsey Global GENDER INEQUALITY Institute, 2015). These analyses focus on losses in Gross Domestic Product (GDP) from inequality between women Gender inequality is meant to represent differences in a and men in labor markets. This focus on income is natural wide range of development outcomes between men and since GDP is the standard measure according to which the women, as explained in Box 1. Many of these differences economic performance of countries is measured today. result from social norms and opportunity gaps between Yet GDP growth is a short-term measure of performance, men and women. Closing these gaps would help improve which may be misleading about the health of an economy development outcomes for women (and in some cases men because it does not reflect whether a country is investing depending on the country). When gaps are closed between in the assets base that will sustain its long-term growth. For men and women, we consider that gender equality is example, a country could deplete its natural capital base achieved. While other approaches could be used to analyze or fail to invest in its people and still be able generate high gender inequality and define what gender equality means rates of GDP growth in the short run, although probably or implies, this approach has the merit of being simple. not in the long-run. BOX 1: WHAT IS MEANT BY GENDER INEQUALITY AND ACHIEVING GENDER INEQUALITY Gender inequality takes many forms. In this section, the focus is on differences in labor market outcomes for men and women due to differences in labor force participation as well as earnings when working. Some of the differences in earnings may themselves be related to other differences including in educational attainment between men and women. But as noted in subsequent sections, gender inequality also takes many other forms. Child marriage affects mostly girls and is therefore a form of gender inequality. Women also tend to have less decision-making ability in their household than men. The impact of these and other forms of gender inequality are considered in this note, and some costs associated with impacts are measure. We consider that gender equality is achieved in this note when gap between men and women are closed. In the case of earnings, this means that women earn as much as men. In the case of early childbearing or child marriage, we consider that gender equality is achieved when both are eliminated. More details on what is meant by gender equality are provided when discussing specific indicators and the simulations that are undertaken to model gender equality. In this note, following Wodon (2018, see also Wodon Estimations of human capital wealth based on the future and de la Brière, 2018), we rely on a different approach earnings of the labor force are done separately for men to measure the losses in earnings that result from gender and women. Hence, we can estimate losses in human inequality or, equivalently, the gains associated with gender capital wealth due to gender inequality in a simple way equality in labor markets. Instead of measuring losses from (see appendix 1 on the methodology). In 2014, women inequality as annual flows (the GDP approach), we focus on accounted for 39 percent of human capital wealth losses in human capital (the wealth approach). The rationale in Uganda versus 61 percent for men. These are also for this focus on wealth is discussed in Box 2. In practice, essentially the proportions observed for sub-Saharan Africa this is done by measuring lifetime losses in earnings. on average, and the values are similar to existing estimates More precisely, human capital wealth is defined as the of gender shares in GDP for the region in other studies. present value of the future earnings of today’s labor force, How large are the losses in wealth resulting from gender considering individuals aged 15 and above. inequality in Uganda? As shown in Table 3, women’s human 5 | THE COST OF GENDER INEQUALITY: GENDER INEQUALITY, HUMAN CAPITAL WEALTH, AND DEVELOPMENT OUTCOMES IN UGANDA | AUGUST 2019 BOX 2: USING HUMAN CAPITAL WEALTH DATA TO MEASURE THE IMPACT OF GENDER INEQUALITY When considering the impact of gender inequality on earnings, at least three arguments justify using a wealth (stock) approach as opposed to a GDP (flow) approach to measure losses in earnings due to gender inequality. First, using a flow approach does not reveal the full magnitude of the losses in earnings faced by women throughout their working life. Estimates of losses from gender inequality in labor markets based on human capital wealth are substantially larger than those based on GDP simply because wealth is larger than GDP. The full magnitude of the losses from gender inequality appears only when considering human capital wealth or women’s earnings over their lifetime. Second, a flow approach tends to emphasize losses for individuals at the peak of their earnings, since they account for a larger share of the labor earnings in GDP. Again, it seems more appropriate to look at individuals’ lifetime earnings to better reflect expected losses from gender inequality. This should give a higher weight to younger individuals than with the flow approach. Third, a wealth approach is forward-looking as it emphasizes sustainability. GDP, or more precisely the consumption component of GDP, is essentially is the annual return or income that a country reaps from its wealth, the assets base that it uses for production. By focusing on wealth, countries can complement GDP measures and focus on long-term sustainable investments. We rely in this note on measures of wealth developed by the World Bank in the Changing Wealth of Nations study (Lange et al., 2018). Building on previous reports (World Bank, 2006 and 2011), the new wealth estimates cover the period 1995 to 2014. They includes not only estimates of produced capital and natural capital, as did previous reports, but also estimates of human capital following the approach suggested by Jorgensen and Fraumeni (1992a, 1992b). The estimations of human capital are based on household survey data. They represent a significant improvement over past estimates where total wealth included a large unexplained residual called intangible capital. This residual, it turns out, consists for the most part of human capital. By measuring the shares of human capital wealth associated to men and women at the country level, the methodology enables us to estimate lifetime earnings losses due to gender inequality. Source: Wodon (2018); Wodon and de la Brière (2018). capital could increase from US$102 billion to US$163 but also children), gender inequality is leading to a loss in billion under gender equality (see Box 3 on the limitations wealth of US$1,619 per person. These losses are large for of these measures). This represents a loss in wealth of up a low-income country such as Uganda. They underscore to US$61 billion due to gender inequality. The estimated the benefits that could be reaped from achieving gender increase in human capital wealth from the base is 23.5 equality. Over time, the total wealth lost due to gender percent in 2014, and for total wealth (including natural and inequality increases from US$8 billion in 1995 to US$61 produced capital as well as net foreign assets), the increase billion in 2014. This increase comes from population in wealth is at 11.8 percent. Global estimates as well as growth, as well as higher standards of living. But other estimates for the sub-Saharan Africa region are of a similar factors that affect human capital wealth also play a role, order of magnitude in percentage terms from the base. On including factors that affect the share of labor earnings in a per capita basis (including not only the adult population GDP over time. AUGUST 2019 | THE COST OF GENDER INEQUALITY: GENDER INEQUALITY, HUMAN CAPITAL WEALTH, AND DEVELOPMENT OUTCOMES IN UGANDA | 6 Table 3: Loss in Wealth from Gender Inequality in Uganda (US$, total and per capita) 1995 2000 2005 2010 2014 Wealth, millions, constant 2014 USD Baseline gender shares of human capital Men’s share of human capital 55.5% 55.9% 56.5% 60.9% 60.9% Women’s share of human capital 44.5% 44.1% 43.5% 39.1% 39.1% Human capital wealth by gender Human capital, men 34,267 46,803 79,613 128,215 158,640 Human capital, women 27,510 36,867 61,403 82,167 101,664 Loss from gender inequality Counterfactual human capital, women 35,705 48,992 82,929 132,218 162,850 Increase in human capital 8,195 12,125 21,526 50,051 61,186 Loss as share of baseline human capital 13.27% 14.49% 15.26% 23.79% 23.51% Loss as share of baseline total wealth 4.46% 5.73% 6.16% 10.80% 11.79% Per capita wealth, constant 2014 USD Baseline wealth Human capital per capita, men 1,679 1,970 2,839 3,868 4,199 Human capital per capita, women 1,348 1,552 2,190 2,479 2,691 Loss from gender inequality Loss in human capital per capita 401 510 768 1,510 1,619 Source: Based on Wodon (2018); see also Wodon and de la Brière (2018). BOX 3 LIMITATIONS OF THE METHOD USED TO COMPUTE LOSSES IN HUMAN CAPITAL WEALTH The estimation of the losses in human capital wealth provided in this note simply assumes that women could work and earn as much as men. The estimation does not consider potential effects on men of rising earnings and hours worked for women. We do not account for the fact that men’s earnings may decrease if women become better educated and have access to the same jobs as men (in part thanks to reductions in occupational segregation). We also assume that women can allocate more time to labor market work without a negative impact on men’s working hours, thus not considering the possibility of men having to allocate more time to household chores or unpaid care. Women tend to do most of the domestic work, especially in developing countries. As women work more hours in paid employment, they may have less time for unpaid domestic work, which could affect the number of hours that men may be able to spend in paid employment, depending on options for elderly, child, or other care services available to households. Many other effects could be at work as women catch up with men in earnings. Here, for simplicity, we only compute how much more human capital Uganda could gain if women had the same lifetime earnings profile as men without any decrease in men’s earnings. In that sense, the estimate could be an upper bound of losses from gender inequality since we do not factor in potential general equilibrium effects. However, higher earnings for women could also lead to more economic activity with positive multiplier effects on the economy and wages. Furthermore, if systems for the provision of care to family members were expanded, a substantial share of the time now allocated to unpaid care could become paid care work. The literature also suggests that as countries develop and women join the labor market or work longer hours, this may primarily reduce free time and time spent on domestic chores. Overall, especially through multiplier effects, unleashing women’s earnings potential could generate larger earnings and human capital gains for both men and women. We also do not account for intergenerational benefits through better education, health, and employment opportunities for their children. Source: Wodon (2018); Wodon and de la Brière (2018). 7 | THE COST OF GENDER INEQUALITY: GENDER INEQUALITY, HUMAN CAPITAL WEALTH, AND DEVELOPMENT OUTCOMES IN UGANDA | AUGUST 2019 SELECTED OTHER are expected to have over their lifetime2. This focus stems from our interest in looking at the impact of gender IMPACTS OF GENDER inequality on population growth, which has implications for the ability of Uganda to reap the benefits from the INEQUALITY demographic dividend (Box 4). Since we consider the number of children that women have towards the end of Gender inequality has implications not only for women’s their reproductive life, we account for desired fertility and earnings, but also for a wide range of other areas pertaining substitution effects in the timing of birth when simulating to women’s roles as wives and mothers “at home”, as opposed the effect of gender equality on fertility. to their role “at work”1. “At home”, gender inequality leads among others to child marriage, early childbearing, and low The impact of gender inequality on total fertility is educational attainment for girls. This, in turn, leads to higher estimated using regression analysis by simulating changes fertility, and thereby higher population growth. Furthermore, in the characteristics of women that are related to when girls are married or have children before the age of 18, gender inequality and affect total fertility3. Results are they may not be physically and emotionally ready to become provided in Table 4 using data from both the 2011 and wives and mothers. As a result, children of young and poorly 2016 Demographic and Health Surveys (DHS) to test for educated mothers often face higher risks of dying by age robustness of the findings. Under gender equality, total five, being malnourished, and doing poorly in school. Gender fertility could be reduced with the 2016 DHS from 6.73 equality at home also contributes to women facing higher children per women towards the end of their reproductive risks of intimate partner violence. Their voice and agency are life to 5.94 children per women. This is a reduction on limited within the household, as well as at work or in national average of 0.79 child per women or 11.7 percent. The institutions. Fundamentally, gender inequality disempowers effects are of the same order of magnitude with the 2011 women in ways that deprive them of their basic human DHS. The largest share (more than three fourths) of the rights. A number of these impacts are considered in this reduction comes from the impact of child marriage on section following the framework outlined in Appendix 2.’ total fertility. When women marry early, given low access to modern contraception method in Uganda, this leads IMPACT ON TOTAL FERTILITY them to have children earlier and more children over their lifetime. Ending child marriage by itself could reduce total The factors leading to fertility are complex. For this study, fertility by about nine percent. Clearly, gender equality we consider the impact of gender inequality on total could help speed up Uganda’s demographic transition. fertility defined as the number of live births that women 1 In practice, the spheres of home and work are not always easy to delineate, and they are not independent of each other. For example, the burden of domestic work has implications for the ability to engage in so-called “productive” work. Still, it is useful for exposition to consider both spheres sequentially. 2 Total fertility is defined here as the number of live births that a woman has over her lifetime. This definition is used for econometric work aiming to measure the marginal impact of child marriage on fertility. By contrast traditional “total fertility rates” are population-level statistics. Our definition of “total fertility” is thus similar, but not exactly the same as “total fertility rates” traditionally measured. The econometric analysis is conducted for women ages 35-49 for sample size considerations. More details on the methodology are available in Onagoruwa and Wodon (2018a). 3 Six assumptions are used for simulating gender equality based on the variables included in the regression analysis: (1) child marriage is eliminated; (2) women are assumed to have the same education as men; (3) gender inequality is assumed through higher earnings to lift households who are in the poorest quintile to the second quintile of well-being, and households in the second quintile to the third; (4) the spousal age gap (the difference in age between the wife and her husband/partner) is reduced to five to nine years for women that have a spousal age gap of more than 10 years; (5) women are assumed to be involved in most decisions made in the household; and (6) women do not accept wife beating. The impact of gender inequality on total fertility is obtained by comparing predicted fertility under current conditions with predicted fertility under gender equality. AUGUST 2019 | THE COST OF GENDER INEQUALITY: GENDER INEQUALITY, HUMAN CAPITAL WEALTH, AND DEVELOPMENT OUTCOMES IN UGANDA | 8 Table 4: Impact of Ending Gender Inequality on Total Fertility (Number of Live Births) Expected value under Expected value under Absolute Percentage gender inequality gender equality Difference change (%) Total fertility 2011 7.12 6.42 0.70 9.83 2016 6.73 5.94 0.79 11.74 Average effect 6.93 6.18 0.75 10.79 Source: Authors. Estimates conducted with the 2011 and 2016 DHS surveys. BOX 4: THE DEMOGRAPHIC DIVIDEND While different definitions of the demographic dividend have been proposed in the literature, the term is commonly associated with the improvements in standards of living and accelerated economic growth that can result when a developing country achieves a population structure that is favorable in terms of economic growth thanks to a reduction in birth (and death) rates that is followed after a short period by rapid fertility decline. As a result, the share of the population of working age individuals increases sharply for a period of time, which tends to generate faster economic growth. In addition, with lower dependency ratios, households are better able to support themselves and invest among others in education, nutrition, and health (or human capital broadly conceived). These investments in turn may lead younger generations to be better educated and more productive in adulthood. This demographic and human capital transition may help reduce poverty rates dramatically. Achieving gender equality should help reduce population growth and improve education in countries where fertility rates remain high, thereby helping to usher in the demographic dividend. IMPACT ON UNDER-FIVE MORTALITY The results are in Table 5. The estimated effects are larger AND STUNTING with the 2011 DHS than with the 2016 DHS. It makes sense to consider the average effect for both years as representing By weakening conditions for early childhood development, the more reliable estimates. Under gender equality, the gender inequality may have negative impacts on young predicted rate of under-five mortality could be reduced children. Early childhood is critical for a child’s development. by an average of 0.77 percentage point or 13 percent of Poor conditions early in life affect brain development and the baseline value when considering estimates from both capabilities, with lasting consequences in adulthood. Children surveys. The predicted under-five stunting rate could be born of very young mothers tend to have higher risks of reduced by 1.7 percentage point or seven percent of the under-five malnutrition and mortality than children born average baseline values in the two surveys. These reductions of older mothers. Part of the reason is that some young are far from negligible, but they are not as large as some may mothers may simply not yet be ready to give birth. When have thought. This is in large part because while the marginal mothers are poorly nourished, this may put their children impact of an early childbirth (being born of a mother at higher risk of intrauterine growth restriction. These and younger than 18, considered as one dimension of gender other effects have implications for children as they grow up inequality) on the risk of under-five mortality and stunting is and in adulthood. For example, research suggests a loss in relatively large, only a small share of children are born from productivity in adulthood for stunted as children. mothers younger than 18. In other words, even large effects at the margin do not imply major shifts nationally. While As for fertility, the analysis of the impact of gender gender equality could make a difference, it could not end inequality on under-five mortality and stunting is conducted under-five mortality and stunting. using regression analysis and simulation techniques4. 4 In a similar way to the analysis for fertility, assumptions are used for simulating gender equality (: (1) early childbearing is eliminated; (2) women are assumed to have the same education as men; (3) gender inequality is assumed through higher earnings for women to lift households who are in the poorest quintile to the second quintile of well-being, and households in the second quintile to the third; (4) the spousal age gap (the difference in age between the wife and her husband/partner) is reduced to five to nine years for women that have a spousal age gap of more than 10 years; (5) women are assumed to be involved in most decisions made in the household; (6) women do not accept wife beating; and (7) women do not have problems getting permission to access medical help for themselves. The impact of gender inequality on the rates of under-five malnutrition and stunting is obtained by comparing predicted rates under current conditions with the rates predicted under gender equality. 9 | THE COST OF GENDER INEQUALITY: GENDER INEQUALITY, HUMAN CAPITAL WEALTH, AND DEVELOPMENT OUTCOMES IN UGANDA | AUGUST 2019 Table 5: Impact of Ending Gender Inequality on Under-five Mortality and Stunting (%) Expected value under Expected value under Absolute Percentage gender inequality gender equality Difference change (%) Under-five mortality 2011 6.49 5.31 1.18 18.18 2016 4.76 4.41 0.35 7.35 Average effect 5.63 4.86 0.77 12.77 Under-five stunting 2011 33.81 31.03 2.78 8.22 2016 28.05 27.47 0.58 2.07 Average effect 30.93 29.25 1.68 5.15 Source: Authors. Estimates conducted with the 2011 and 2016 DHS surveys. IMPACT ON AGENCY AND DECISION-MAKING ABILITY Gender inequality is associated with losses in decision-making, thereby limiting women’s aspirations, as well as their voice and agency (Klugman et al., 2014). To assess the impact of gender inequality on decision-making ability, an index is constructed to take a value between zero and 1005. The impact of gender inequality on the index is obtained by comparing predicted values under current conditions with values predicted under gender equality6. The impact of gender inequality is again substantial, as shown on Table 6. Under gender equality, on average for the two DHS surveys the index of decision-making ability for women could increase from 70.7 (out of a maximum value of 100) to 89.4. This represents a gain on average of 18.7 percentage points or 26.8 percent of the base value of the index, which is rather large. Table 6: Impact of Gender Inequality on Women’s Decision-making Ability (Index) Expected value under Expected value under Absolute Percentage gender inequality gender equality Difference change (%) Decision-making 2011 66.37 87.94 21.57 32.50 2016 75.01 90.85 15.84 21.12 Average effect 70.69 89.40 18.71 26.81 Source: Authors. Estimates conducted with the 2011 and 2016 DHS surveys. 5 The variables included in the index are of four types. First, women currently married are asked in the surveys about who makes decisions in the household in four areas: health care, household purchases, visits to friends and relatives, and the use of the husband’s earnings. For each question, women may typically respond according to four modalities: they alone make decisions, they make decisions with the husband/partner, the husband makes decisions alone, or another person makes the decisions (or the husband has no earnings for the question pertaining to use of earnings). Second, women are also asked if they can refuse to have sex with their husband and if they can request their husband to use a condom when having sex. In addition, women respond to four different circumstances assessing if a husband is justified in beating their wife in those instances: if the wife goes out without telling her husband, if she neglects her children, if she argues with her husband, or if she refuses to have sex with him. Finally, women are asked whether getting their husband’s permission to get medical help for themselves is a major problem or not. While an alternative approach could be used to consider different types of decision-making separately, the results are not very different when doing so. The benefit of an overall index is that it provides a single summary measure of decision-making ability as well as the impact of child marriage on that measure. For more detailed work on decision-making, it is however recommended to also consider different types of decision-making separately. 6 The approach is similar as to what was done for total fertility and child health. The impact of gender equality on decision-making ability for women is based on the following: (1) child marriage is eliminated; (2) women are assumed to have the same education as men; (3) gender inequality is assumed through higher earnings for women to lift households who are in the poorest quintile to the second quintile of well-being, and households in the second quintile to the third; (4) the spousal age gap is reduced to five to nine years for women that have a spousal age gap of more than 10 years; and (5) several variables related to women’s decision-making ability in the village or area where a specific woman lives are assumed to be improved. AUGUST 2019 | THE COST OF GENDER INEQUALITY: GENDER INEQUALITY, HUMAN CAPITAL WEALTH, AND DEVELOPMENT OUTCOMES IN UGANDA | 10 OTHER IMPACTS While child marriage does not affect the index in a statistically significant way, our analysis suggest that Beyond the impacts estimated above, gender inequality is educational attainment does, and so does gender also associated with other negative effects for women. A inequality. This matters given that young women few of those effects can be illustrated in the case of child are disproportionately impacted (HIV prevalence marriage as a key component of gender inequality based on is almost four times higher among young women analysis for Uganda (World Bank, 2017). than men). Increased agency and education can help reduce risk of infection, as for example women are • Intimate Partner Violence: The available data on better able to negotiate condom use and protect intimate partner violence for Uganda suggests themselves from rape and other forms sexual violence. that levels of intimate partner violence (IPV) are high in comparison with other countries. There is • Birth registration: Gender inequality could also affect evidence both in Uganda and in other countries that birth registration for children. In some countries, gender inequality is associated with higher risks of when mothers have children below the minimum IPV. For example, multiple studies have suggested legal age for marriage, this could lead to lower birth that child marriage may increase risks of IPV. In registration rates if women are fearful that having turn, the health implications of these impacts can a child at a young age suggests that marriage took be serious, as can be their cost implications for place before the minimum legal age. More generally, women and households. In Uganda, analysis suggests there is clear evidence that birth registration is related that child marriage contributes substantially to among others to the mother’s educational attainment, IPV. The impact of gender inequality, including its which itself is affected by gender inequality. implications for educational attainment, is larger. • Knowledge of HIV/AIDS: Gender inequality may also have an impact on women’s knowledge about HIV/ AIDS through its impact on education. Knowledge of HIV/AIDS is measured through an index that accounts for responses to a wide range of questions. 11 | THE COST OF GENDER INEQUALITY: GENDER INEQUALITY, HUMAN CAPITAL WEALTH, AND DEVELOPMENT OUTCOMES IN UGANDA | AUGUST 2019 ECONOMIC IMPLICATIONS smaller than the impact from lost earnings estimated in Table 2, over time the impact would increase and could These various impacts have implications for standards of become large. living as well as the provision of public services. Consider the contribution of gender inequality to higher fertility Lower population growth from ending gender inequality and population growth. Simulations using demographic could also reduce the cost for the government of providing projection tools suggest that in Uganda, ending child basic services to the population. For example, as total marriage could reduce population growth by 0.17 fertility is reduced, the size of the cohorts of new students percentage points under current conditions. When ending entering primary school would be reduced, leading to gender inequality as opposed to only ending child marriage, savings in service provision. The same applies to other the impacts could be 25 to 30 percent higher according services in multiple areas such as healthcare or basic to estimations based on total fertility rates. Ending gender infrastructure. Savings from a reduction in population inequality could therefore lead to a reduction in population growth could be used for investments in improving the growth of about 0.22 percentage point. This reduction quality of the services provided – or pay for the cost of would be cumulative from year to year, since every year the policies aiming to achieve gender equality (such as universal annual rate of population growth would be lower than the secondary education for girls). rate under business as usual condition. Finally, gender inequality also has implications for future As a result, Uganda’s population could be several generations. The analysis above suggests that gender percentage points smaller by 2030 if gender equality were inequality leads to an increase in the rates of under-five achieved today. This would generate – all other things being mortality and stunting. In the recent study on the impacts equal, higher levels of wealth per capita since total wealth of child marriage in Uganda, valuations of these losses were would not necessarily be affected at least in the short and provided. The valuations could be even larger for achieving medium term by the reduction in fertility rates, while gender equality. the population growth rate and thereby the population size would decrease. A higher level of wealth per capita would in turn be beneficial for future standards of living. In economic terms, while the impact of gender equality from lower total fertility and population growth could be initially AUGUST 2019 | THE COST OF GENDER INEQUALITY: GENDER INEQUALITY, HUMAN CAPITAL WEALTH, AND DEVELOPMENT OUTCOMES IN UGANDA | 12 CONCLUSION The impacts of gender inequality on a wide range of development outcomes are large and the economic case for investing in girls and women is strong. Losses in human capital wealth in Uganda due to gender inequality are estimated at up to US$61 billion. From its base value in 2014, Uganda’s national wealth could increase by 11.8 percent under gender equality in earnings. On a per capita basis, this could generate a gain of up to US$1,619 per person in wealth. Gender inequality also has large impacts in other areas, including fertility and population growth, under-five mortality and stunting, and women’s decision-making ability within the household. In many of those areas, associated costs are likely to be high as well. This is for example the case for impacts on fertility and population growth which affect standards of living (GDP per capita as well as wealth per capita). In separate publications at the World Bank, policy options to tackle gender inequality have been documented. For example, to increase women’s earnings, investments throughout the life cycle are needed, starting with early childhood development and learning in schools, and continuing with improved job opportunities in adulthood. As noted in Wodon and de la Brière (2018), successful interventions can be implemented to address time use constraints, facilitate access to productive assets, and solving market and institutional failures that penalize women. Interventions need to be tailored in terms of age (young women face specific barriers and opportunities), poverty (very poor women need more than a single intervention) and type of workforce participation among women (considering wage workers, entrepreneurs and farmers). But smart delivery and implementation can lead to positive impacts. Addressing constraints often requires incentives and nudges, but what is also needed is to take on women’s subordinate position in the family and the traditional division of labor for household chores and care. While more work would be needed to adjust policy options to the specific context of Uganda, the good news is that achieving greater gender equality in labor markets and other areas would generate substantial economic gains for countries apart from a better life for women. 13 | THE COST OF GENDER INEQUALITY: GENDER INEQUALITY, HUMAN CAPITAL WEALTH, AND DEVELOPMENT OUTCOMES IN UGANDA | AUGUST 2019 APPENDIX 1: (including both the employed and the self-employed) in GDP as available in the Penn World Tables. Second and METHODOLOGY FOR separately, the estimations also rely on two variables obtained from data compiled by the United Nations Population HUMAN CAPITAL WEALTH Division: (1) population data by age and sex (so that the data in the household surveys can be better calibrated); ESTIMATES and (2) mortality rates by age and gender (so that the expected years of work can be adjusted, accounting for the Human capital wealth is defined as the discounted value of fact that some workers will die before age 65). Again, we future earnings for a country’s labor force. In practice, we adjust data from the surveys to population estimates from estimate how likely it is that various types of individuals will the United Nations to ensure that estimates are adequate. be working, and how much they will earn when working. For individuals in the 15-to-24 age group, the probability of By “various types” of individuals, we mean individuals remaining in school is also considered. categorized by age, sex, and level of education. Essentially, we use household surveys to construct a dataset that Given the estimation of human capital wealth based on captures (1) the probability that individuals are working Mincerian wage regressions, the measure accounts not depending on their age, sex, and years of education; and only for the number of years of schooling completed by (2) their likely earnings when working, again, by age, sex workers, but also for the earning gains associated with and years of schooling. This is done separately for men and schooling (which implicitly factors in the quality of learning in women, and results in estimates of human capital wealth by school), whether individuals work (labor force participation), gender. Typically, women earn significantly less than men. and for how many years they work (accounting for health conditions through life expectancy). Estimations of human Estimates of the likelihood of working for individuals are capital wealth are done separately for men and women. based on observed values in household and labor force This means that once we have estimates of human capital surveys. Estimates of expected earnings are based on wealth by gender, we can estimate losses in human capital Mincerian wage regressions. The regressions are used to wealth due to gender inequality in a very simple way. If we compute expected earnings throughout individuals’ working denote a country’s human capital wealth as measured from life, considering their sex, education level, and assumed the expected future earnings of women and men as HCM experience (computed based on age and the number of years and HCW, respectively, and the adult population of men and of education completed). Expected earnings are computed women by POPM and POPW, the earnings per adult men and for all individuals in the surveys from age 15 to age 65, noting women can be defined as hcM=HCM/POPM and hcW=HCW/ that some individuals may go to school beyond age 15. The POPW. Under gender equality, interpreted as ensuring analysis also considers the life expectancy of the labor force. that adult men and women have the same future expected In countries with high life expectancy, workers are expected earnings, human capital for women would increase from hcW to work until age 65, but in other countries they may not be to hcM. Therefore, the loss in human capital wealth from able to. For simplicity, when estimating the present value of gender inequality is measured as (hcM-hcW)×POPW. Details future earnings, the same discount factor for future earnings are provided in Wodon (2018). is applied to all countries. The household surveys used for the computation of the earnings profiles—as well as the probability of working—are nationally representative. The surveys are in most cases of good quality, but they may still generate estimates that are not consistent with either the system of national accounts or population data for the countries. Therefore, two adjustments are made. First, to ensure consistency of the earnings profiles from the surveys with published data from national accounts, earnings estimates from the surveys are adjusted to reflect the share of labor earnings AUGUST 2019 | THE COST OF GENDER INEQUALITY: GENDER INEQUALITY, HUMAN CAPITAL WEALTH, AND DEVELOPMENT OUTCOMES IN UGANDA | 14 APPENDIX 2: CONCEPTUAL Once impacts in various domains are estimated, costs can be measured for selected impacts. As shown in Figure 1, FRAMEWORK FOR estimates of the monetary benefits from ending gender inequality can be computed among others in terms of (i) ANALYZING THE IMPACTS Higher growth in GDP per capita and lesser budgetary needs for service provision as a result of lower population OF GENDER INEQUALITY growth; (ii) Higher labor earnings as a result of better health and less stunting in childhood; (iii) Higher labor earnings This note is part of a series that aims to measure the for women in adulthood (the focus of this note); and (iv) economic cost of gender inequality by looking at the Benefits associated with children’s lives saved. This list is by impacts of gender inequality and the associated costs no means exhaustive, but it includes some of the largest in multiple domains. The series also aims to provide a expected benefits. synthesis of the available evidence on successful programs and policies that have been shown to contribute to gender Finally, the benefits from gender equality at the levels equality in multiple areas. of individuals and households have broader implications. By raising standards of living (among others through The framework for the analysis of the impacts and costs of higher GDP per capita with lower population growth and gender inequality builds on recent work on the economic higher earnings for women), gender equality will reduce impacts of child marriage, low educational attainment poverty. Since girls and women from lower socio-economic for girls, and human capital wealth at the World Bank. backgrounds are the most affected by gender inequality, Conceptually, five potential domains of impacts of gender promoting gender equality will also contribute to shared inequality are considered, as shown in Figure 1: (1) fertility prosperity. and population growth; (2) health and nutrition; (3) child marriage and educational attainment; (4) labor force participation and earnings; and (5) agency, including decision-making and the risk of gender-based violence. FIGURE A1: CONCEPTUAL FRAMEWORK FOR MEASURING THE COST OF GENDER INEQUALITY Associated Losses/Gains Development 5 Domains of “Impacts” Outcomes Welfare Fertility and population growth Gains GENDER INEQUALITY World Bank twin goals: Health, nutrition and violence Reduction in Earning extreme poverty Gains and shared Educational attainment Complex direct and prosperity (growth and child marriage indirect “impacts” for the bottom 40 percent) Budget Labor, Earnings & Productivity Savings Decision-making and violence Other Benefits Source: Wodon and de la Brière (2018). 15 | THE COST OF GENDER INEQUALITY: GENDER INEQUALITY, HUMAN CAPITAL WEALTH, AND DEVELOPMENT OUTCOMES IN UGANDA | AUGUST 2019 AUGUST 2019 | THE COST OF GENDER INEQUALITY: GENDER INEQUALITY, HUMAN CAPITAL WEALTH, AND DEVELOPMENT OUTCOMES IN UGANDA | 16 REFERENCES Cuberes, D. and M. Teignier. 2015. How Costly Are Labor Gender Gaps? Estimates for the Balkans and Turkey. World Bank Policy Research Working Paper 7319. Washington, DC: The World Bank. Elborgh-Woytek, K., M. Newiak, K. Kochhar, S. Fabrizio, K. Kpodar, P. Wingender, B. Clements, and G. Schwartz. 2013. Women, Work and the Economy: Macroeconomic Gains from Gender Equity. IMF Staff Discussion Note. Washington, DC: International Monetary Fund. Jorgensen, D.W. and B.M. Fraumeni. 1992a. The Output of Education Sector, in Z. Griliches (ed.). Output Measurement in the Service Sectors. Chicago, IL: University of Chicago Press. Jorgensen, D.W. and B.M. Fraumeni. 1992b. Investment in Education and US Economic Growth. Scandinavian Journal of Economics, 94(Supplement): 51–70. Lange, G. M., Q. Wodon, and K. Carey. 2018. The Changing Wealth of Nations 2018: Sustainability into the 21st Century. Washington: The World Bank. McKinsey Global Institute. 2015. The Power of Parity: How Advancing Women’s Equality Can Add $12 Trillion to Global Growth. London: McKinsey Global Institute. Wodon, Q. 2018. What Is the Cost of Gender Inequality in Lost Earnings? Global Estimates Based on the Changing Wealth of Nations, Mimeo, Washington, DC: The World Bank. World Bank. 2006. Where Is the Wealth of Nations? Measuring Capital for the 21st Century, Washington, DC: The World Bank. World Bank. 2010. Stepping Up Skills for More Jobs and Higher Productivity. Washington, DC: The World Bank. World Bank. 2011. The Changing Wealth of Nations? Measuring Sustainable Development in the New Millennium, Washington, DC: The World Bank. World Bank. 2012. World Development Report 2012: Gender Equality and Development. Washington, DC: The World Bank. World Bank. 2017. Accelerating Uganda’s Development: Educating Girls and Ending Child Marriage and Early Childbearing. Washington, DC: The World Bank. 17 | THE COST OF GENDER INEQUALITY: GENDER INEQUALITY, HUMAN CAPITAL WEALTH, AND DEVELOPMENT OUTCOMES IN UGANDA | AUGUST 2019 asdasd Recommended citation for this note: Wodon, Q. and A. Onagoruwa. 2019. Gender Inequality, Human Capital, and Development Outcomes in Uganda. The Cost of Gender Inequality Notes Series. Washington, DC: The World Bank. This note was prepared by World Bank staff. It relies on a framework designed for work on the economic impact of gender inequality globally. Funding for the work was provided by the Canadian Government, the Children’s Investment Fund Foundation, and the Global Partnership for Education. Findings, interpretations, and conclusions expressed in this note are entirely those of the author(s) and should not be attributed in any manner to the World Bank, to its affiliated organizations or to members of its Board of Executive Directors or the countries they represent. The World Bank does not guarantee the accuracy of the data included in this work. Information and illustrations contained in this note may be freely reproduced, published or otherwise used for noncommercial purposes without permission from the World Bank. However, the World Bank requests that the original study be cited as the source. © 2019 The World Bank, Washington, DC 20433. AUGUST 2019 | THE COST OF GENDER INEQUALITY: GENDER INEQUALITY, HUMAN CAPITAL WEALTH, AND DEVELOPMENT OUTCOMES IN UGANDA | 18 THE COST OF GENDER INEQUALITY: GENDER INEQUALITY, HUMAN CAPITAL WEALTH, AND DEVELOPMENT OUTCOMES IN UGANDA