Policy Research Working Paper 10672 Female Headship and Poverty in the Arab Region Analysis of Trends and Dynamics Based on a New Typology Shireen AlAzzawi Hai-Anh Dang Vladimir Hlasny Kseniya Abanokova Jere Behrman Development Economics Development Data Group January 2024 Policy Research Working Paper 10672 Abstract Various challenges are thought to render female-headed and dynamics. Most types of FHHs are less poor than non– households (FHHs) vulnerable to poverty in the Arab FHHs on average, but FHHs with a major share of female region. Yet, previous studies have had mixed results and adults are generally poorer. FHHs are more likely to escape the absence of household panel survey data hinders analysis poverty than households on average, but FHHs without of poverty dynamics. This paper addresses these challenges children are the most likely to do so. While more children by proposing a novel typology of FHHs and analyzes are generally associated with more poverty for FHHs, there synthetic panels constructed from 20 rounds of repeated is heterogeneity across countries in addition to heterogene- cross-sectional surveys spanning the past two decades from ity across measures of FHHs. The findings provide useful the Arab Republic of Egypt, Iraq, Jordan, Mauritania, the inputs for social protection and employment programs West Bank and Gaza, and Tunisia. The paper finds that the aiming at reducing gender inequalities and poverty in the definition of FHHs matters for measuring poverty levels Arab region. This paper is a product of the Development Data Group, Development Economics. It is part of a larger effort by the World Bank to provide open access to its research and make a contribution to development policy discussions around the world. Policy Research Working Papers are also posted on the Web at http://www.worldbank.org/prwp. The authors may be contacted at hdang@worldbank.org. The Policy Research Working Paper Series disseminates the findings of work in progress to encourage the exchange of ideas about development issues. An objective of the series is to get the findings out quickly, even if the presentations are less than fully polished. The papers carry the names of the authors and should be cited accordingly. The findings, interpretations, and conclusions expressed in this paper are entirely those of the authors. They do not necessarily represent the views of the International Bank for Reconstruction and Development/World Bank and its affiliated organizations, or those of the Executive Directors of the World Bank or the governments they represent. Produced by the Research Support Team Female Headship and Poverty in the Arab Region: Analysis of Trends and Dynamics Based on a New Typology Shireen AlAzzawi, Hai-Anh Dang, Vladimir Hlasny, Kseniya Abanokova, and Jere Behrman* Keywords: poverty, feminization, female-headedness typology, synthetic panels, Arab region, household surveys JEL Codes: I3, J16, N35, O1 * AlAzzawi (salazzawi@scu.edu) is lecturer, Santa Clara University, Santa Clara CA, USA; Dang (hdang@worldbank.org; corresponding author) is senior economist, Living Standards Measurement Unit, Development Data Group, World Bank, USA and is also affiliated with IZA, Indiana University, and London School of Economics and Political Science; Hlasny (vhlasny@gmail.com) is economic affairs officer, UN ESCWA, Beirut, Lebanon; Abanokova (kabanokova@worldbank.org) economist, Living Standards Measurement Unit, Development Data Group, World Bank, USA; Behrman (jbehrman@econ.upenn.edu) is WR Kenan Jr Professor of Economics & Sociology, University of Pennsylvania, USA. We would like to thank Kathleen Beegle, Nazmul Chaudhury, Ebad Ebadi, Daniel Halim, Emily Hannum, Dean Jolliffe, Michael Kevane, Mahdi Majbouri, and participants at the American Economic Association meeting (New Orleans) and the Central Banks of the Middle East and North Africa's Annual Conference on Development Economics (Rabat) for helpful comments on an earlier draft. We would also like to thank the Chief Economist’s Office for the Middle East and North Africa (MNACE) for guidance and financial support and the UK Foreign Commonwealth and Development Office (FCDO) for additional funding assistance through the Data and Evidence for Tackling Extreme Poverty (DEEP) Research Program. 1. Introduction Are female-headed households (FHHs) in the Arab region more likely to be poor, and increasingly so? Social and cultural barriers often hinder women’s economic participation in the region and several recent studies find that women are at an increasing disadvantage compared to men in labor markets (Amara and Jemmali, 2018; AlAzzawi and Hlasny, 2022). The COVID-19 pandemic further deepened gender inequality in many countries (Dang and Nguyen, 2021; Alon et al., 2022). Yet, few studies have investigated the topic of poverty feminization in the region. These are important policy questions since many countries in the region specifically target FHHs based on the premise that they are more vulnerable to poverty and probably increasingly vulnerable to poverty, particularly during the pandemic. 1 Furthermore, the Arab region is home to countries of different income levels with diverse social and cultural circumstances, resulting in different context-specific factors contributing to poverty feminization. That poverty is more prevalent among women than men is widely assumed, and various explanations are offered for it. These include lower school enrollment rates and less work experience (Grant and Behrman, 2010), limited access to income-generating assets such as land (Deere and Leon, 2003), credit and other financial services (Demirguc-Kunt et al., 2013), physical and social capital, and technology (World Bank, 2011; Klasen et al., 2015), and market discrimination (Buvinic and Gupta, 1997). There is, however, far less agreement on the existence of “feminization of poverty” affecting FHHs (Chant, 2010; Duflo, 2012; Klasen et al., 2015; Bradshaw et al., 2017). Buvinic and Gupta (1997) observe that of 65 studies covering Africa, Asia, Latin America and the Caribbean, 38 1 These include the Arab Republic of Egypt’s largest poverty-targeting cash transfer program, Takaful, and other subsidy programs in Jordan, Lebanon, and Tunisia (NAF, 2020; Nasri, 2020; ESCWA, 2021; World Bank, 2022). 2 studies find that FHHs were overrepresented among the poor, 15 others found that poverty was associated with certain FHH types, and the remaining eight studies show no such relationship. While Quisumbing et al. (2001) and Medeiros and Costa (2008) find FHHs to be consistently poorer in only 10 developing countries in Africa, Asia and Latin America, Chant (2003) fails to obtain a similar finding in studies for the three continents. More recently, Milazzo and van de Walle (2017) even find that despite a growing population share of FHHs in Africa, FHHs saw faster poverty reduction than male-headed households (MHHs). Furthermore, one particular challenge in understanding the current literature on poverty feminization is variations in how FHHs are defined (see Appendix A, Table A.1; we return to this discussion in the next section). Several authors investigated the gender dimension of poverty in the Arab region before the onset of the Arab Spring uprisings in 2011 (Nassar, 1997; Datt et al., 2001; El-Laithy, 2001). More recent post Arab-Spring studies examined poverty dynamics for the whole population and different population groups (e.g., Dang and Ianchovichina, 2018). Yet, these studies do not investigate the gender prism; just a few studies explicitly examine poverty feminization related to FHHs but only on a single-country basis (AlAzzawi, 2018; Amara and Jemmali, 2018; AbdelLatif et al., 2019). Furthermore, these few studies stopped short of examining poverty dynamics for FHHs due to the lack of panel data. 2 These knowledge gaps prevent efficient and cost-effective policy interventions, since policies that address chronic poverty could be quite different from those that tackle transient poverty. 3 2 Mixed results exist regarding static poverty across countries. For example, comparison between MHHs and FHHs by self-reported headship revealed that for the Arab Republic of Egypt, FHHs are less likely to be poor (AlAzzawi, 2018; AbdelLatif et al., 2019) while the opposite result holds for Tunisia (Amara and Jemmali, 2018). 3 For example, while social safety-net programs better address transient poverty (e.g., as they help prevent the non- poor but vulnerable households from falling into poverty), longer-term investments in human capital and infrastructure can tackle chronic poverty. See, e.g., Barret (2005) and Ravallion (2016) for further discussion on different policy interventions regarding chronic poverty versus transitory poverty. 3 Our study makes several new contributions to the literature. First, we propose and evaluate a novel typology of FHHs consisting of four main types and several sub-types, which are based on self-reported responses, demographic characteristics, and socio-economic characteristics. This approach allows us to employ more nuanced headship definitions that reach beyond the traditional identification solely based on the household head’s gender to better include other aspects of household female composition. Our proposed typology also calls for more attention to the important role of children in defining FHHs, since FHHs with children could show remarkably different, static and dynamic, poverty outcomes from those of FHHs without any children or those of non-FHHs. Second, we study the trends in the FHH poverty–gender nexus and poverty dynamics, for six countries across the Arab region—namely the Arab Republic of Egypt, Iraq, Jordan, Mauritania, the West Bank and Gaza (Palestine), and Tunisia—for which little knowledge exists regarding poverty by FHH status. Third, despite the absence of actual panel data, we construct synthetic panels that allow us to examine FHH poverty dynamics for these countries. By conducting analyses on both poverty incidence and dynamics, we contribute to a better understanding of the dynamic economic well- being of FHHs over time. In fact, to our knowledge, we offer the first multi-country study that investigates to what extent FHH poverty exists across a number of countries in the Arab region, and whether FHHs, according to a variety of household types, are more likely to enter or escape poverty over time, using recent survey data. We also make a new data contribution by carefully assembling and harmonizing relevant, up-to-date surveys from multiple sources in a region that is well recognized for limited data access. We find that the shares of FHHs widely vary, ranging from 10 percent to more than 40 percent depending on the countries and definitions. Compared with non-FHHs, most types of FHHs 4 (including self-reported, potential, and most-educated-female-adult FHHs) are 1 percent to 4 percent less likely to be poor while majority-female-adult FHHs are 3 percent to 5 percent more likely to be poor. We also find considerable mobility in and out of poverty over the past decade, with the average poor FHH having between 21 and 54 percent chance of escaping poverty, depending on the country. Yet, country heterogeneity exists, with Iraq, Jordan, and Mauritania having upward mobility rates of between 41 and 54 percent, and Egypt, the West Bank and Gaza and Tunisia having upward mobility rates between 21 and 31 percent. More children are generally associated with more poverty and lower chances of escaping poverty for FHHs. The upward mobility rates out of poverty for FHHs without children, FHHs with children, and non-FHHs across all countries are respectively 42 percent, 37 percent, and 36 percent. The corresponding figures for downward movement into poverty for these FHH types are respectively 14 percent, 17 percent, and 19 percent. Our results on mobility are robust to different definitions of FHHs, alternative estimation models, and sample sizes. The rest of the paper is organized as follows. Section 2 discusses the various definitions of female headship in the existing literature before proposing our new typology of female-headed households (Section 2.1) and introduces the synthetic panel method that allows us to assess FHH poverty feminization dynamically without actual panel data (Section 2.2). Section 3 reviews the available data (Section 3.1), welfare aggregates and standardization measures (Section 3.2), and presents descriptive statistics (Section 3.3). Section 4 reports the main results for cross-sectional poverty (Section 4.1) and poverty dynamics (Section 4.2), and Section 5 concludes with key findings and policy implications. Appendixes A and B present additional estimation results and 5 descriptive statistics, Appendix C discusses the synthetic-panel method, and Appendix D provides further analyses with equivalence scales. 2. Analytical Framework 2.1. Typology of female-headed households Households vary in their compositions and socioeconomic characteristics. In the countries in our sample, the majority of households are comprised of a married couple with one or two income earners, with or without children. Single-head households vary broadly: from widowed retirees who may have already worked for many years and are now living with older children who might be supporting them, to a middle-aged mother who got divorced or lost her husband and is struggling to meet ends by joining the labor market for the first time. Among this group, the presence of another adult male, whether an earner or not, is yet another confounding factor, as well as the presence of children. The heterogeneous nature of FHHs and the need to separately study different FHH types have been discussed extensively in the literature on poverty feminization (e.g. Kabeer, 1997; Quisumbing, et al., 2001; Klasen et al., 2015; Beegle et al., 2016; Munoz Boudet et al., 2018). The variety of household-headship designations in existing studies has led to mixed results regarding poverty feminization and dynamic patterns. Our reading of some selected studies in the past two decades suggests that while FHHs are not observed to be poorer than non-FHHs in many cases, FHHs can be poorer or have lower consumption depending on the specific type and country context (Appendix A, Table A.1). Advancing an FHH typology can thus be critical for clearing the apparent inconsistencies and for re-classifying households with what may be considered “de facto female heads” (based on demographic or socioeconomic characteristics), as opposed to “de 6 jure female heads” (based on official status or self-reported information). This also has important implications for poverty reduction efforts targeted at vulnerable population groups. Figure 1 presents our proposed typology of FHHs, which consists of several layers. For the first layer, existing studies can be broadly grouped under two categories: “de jure FHH” and “de facto FHH” (second row). For the second layer, we consider three approaches under these two groups: the self-reported approach (under “de jure FHH”), and the demographic approach and the socioeconomic approach (under “de facto FHH”) (third row). For the third layer, we consider four main types of FHHs under these three approaches: i) Type 1: self-reported FHHs, ii) Type 2: FHHs defined using the majority share of females among adults in the household (i.e., majority-female- adult FHHs), iii) Type 3: potential FHHs, and iv) Type 4: FHHs defined as those where the most- educated adult member is female and no working-age employed male is present (i.e., most- educated-female-adult FHHs) (fourth row). Furthermore, under these four main types, we also consider five alternative sub-types, which include de jure and married FHHs (under self-reported FHHs), FHHs defined using the majority share of employed females in the household (under majority-female-adult FHHs), and asset and core FHHs (under potential FHHs) (fifth row). 4 Finally, all these types of FHHs should be considered separately with or without any children (last row), since the presence of children plays a crucial role in FHHs’ poverty as discussed below. The typology is motivated by both our review of the literature and our empirical analysis for each type of FHHs in the Arab region. Figure 1 briefly refers to some illustrative studies that employ these approaches and we elaborate below on this new typology. “De jure” FHHs: Self-reported approach 4 While we propose these three approaches and types for classification purposes, they are not mutually exclusive and existing studies have combined one or more in defining female headship. 7 A natural departure point to analyzing FHHs is to adopt the self‐reported identification of the head by the survey respondent (our first type, self-reported FHHs), which falls under the de jure FHHs group. Marital status is a key characteristic in this respect. A large share of FHHs are formed as the result of a major marital shock such as divorce or widowhood. If, prior to the shock, the husband was the primary income-earner, the newly-formed FHH may be more likely to fall into poverty (Brown and van de Walle, 2021). FHHs formed through widowhood, especially at a young age with children present, can exhibit both more poverty and higher persistence of poverty (Appleton, 1996; Dreze and Srinivasan, 1997; Horrell and Krishnan, 2007; van de Walle, 2013; Munoz Boudet et al., 2018; Brown and van de Walle, 2021) than FHHs formed largely “by choice”, through divorce or migration of the male spouse (Quisumbing, et al., 2001; Klasen et al., 2015; Beegle et al., 2016; Bradshaw et al., 2017). Females who never marry or who seek divorce might have chosen this status because they have strong prospects for supporting their newly-formed households on their own, such as higher personal incomes or enabling family-support systems. Ignoring such considerations could mask differences between self-reported FHHs that are financially secure and those that are economically vulnerable (Kabeer, 1997; van de Walle, 2013; Milazzo and van de Walle, 2017). Consequently, it may also be useful to consider alternative types of households based on their marital status— never married, married, divorced or separated, and widowed. In our samples, most self-reported MHHs are married, and this group is the largest in the sample. By contrast, from 69 to 77 percent of self-reported FHHs in all six countries are widowed, except for Mauritania and the West Bank and Gaza, where 35 and 53 percent, respectively, of self- reported FHHs are widowed households when considering all years together (Appendix B, Tables B.1-B.7). The second-largest group of self-reported FHHs have married heads, but this share 8 typically remains about 20 percent or below in all countries and years, except for Mauritania where it rose rapidly to almost 40 percent after 2008. 5 “De facto” FHHs: Demographic and socioeconomic approaches The term “head” is a loaded term carrying strong connotations about household decision- making power that has traditionally been given to the oldest-male member (regardless of his breadwinner status). This is certainly an issue in the Arab region, where the traditional patriarchal system may preclude designating the female as “head” in the presence of a disabled adult male or a son (regardless of age), even if the woman is the main income-earner. A de facto FHH can thus be defined as one where the male head is temporarily or regularly absent, or (if co-resident) is not the main breadwinner (Buvinic and Youssef, 1978; Buvinic et al., 1983; Klasen et al., 2015). De facto headship accounts for the demographic composition of the household, as well as the socioeconomic circumstances determining the respective members’ relative contributions to household resources (Rosenhouse, 1989; Handa, 1994, 1996; Rogers, 1995; Varley, 1996; Buvinic and Gupta, 1997; Fuwa, 2000; Posel, 2001; Budlender, 2008; Grown and Valodia, 2010; Chant, 2010; Rogan, 2013; Klasen, 2015). A de facto FHH may be more vulnerable to poverty for several reasons. In many societies, the absence of male connections to local economic and social institutions can be debilitating. De facto 5 One complication in classifying married self-reported FHHs arises where one spouse works overseas and sends home remittances, which is common in the region. If the overseas spouse is the husband, the stay-behind spouse might or might not designate herself as the household head in his absence. This can underestimate poverty among “true” FHHs, where the female head does not rely on others for support, since some of the self-declared female heads or main income-earners are in fact temporary designees while the main income-contributing spouses are overseas. In the surveys for all years, remittances are major sources of income for self-reported FHHs, consisting for example of 68% of the consumption per capita in Egypt, and 37% of the consumption per capita in Jordan (Appendix B, Tables B.1 and B.4, all years columns). However, the surveys lump together domestic and overseas remittances and do not allow any further breakdown or provide information on the amount of overseas remittances. The surveys do not identify the relations between the remitters and the households, which complicates matters as such remittances might be alimony or in-kind support. 9 FHHs residing with the female heads’ fathers or older sons may still be better off than FHHs who do not have support of working-age males; for example, agricultural production may become especially harder due to fewer working-age household members working on the farm (Rogan, 2013; Brown and van de Walle, 2021). These FHHs may also have less access to productive assets such as livestock or extension services. Moreover, women in Arab labor markets have far fewer job opportunities compared to men. Their labor-force participation rates are currently among the lowest in the world; their unemployment rates are also four times the world average (UNDP, 2022). When they do work, they tend to face wage and occupational discrimination (Elhamidi and Said, 2008) or are overrepresented in the informal sector with low pay and no social insurance. Residing in a majority-female household, or in one where the majority of earners are female, affects the propensity of being poor (Rogan, 2013; Munoz Boudet et al., 2018). Ideally, an objective criterion would be used to assign headship to the family member whose income or decision-making contributes most to maintaining the family. For instance, Gammage (1998) found that using the maintenance criteria to define FHHs in El Salvador and Costa Rica results in a markedly higher percentage of such female-maintained households (FMHs) compared to the de jure FHH group, and higher poverty incidence. Unfortunately, household surveys in the Arab region do not provide information about individual incomes or total earnings, only aggregates for the household. In the absence of such information, based on our review of the literature (Table A.1 in Appendix A), we propose several alternative definitions of de facto FHHs using demographic and socioeconomic criteria to provide a more multifaceted understanding of FHHs. Regarding household composition, we consider households where the proportion of females among (working- age) adults exceeds 0.5 (second type, majority-female-adult FHHs). We define potential FHHs as 10 those households where there are no working-age males present (third type, potential FHHs). The final type of FHHs combines the demographic and socioeconomic criteria and consists of households with no employed males, whose most-educated adult member is female (fourth type, most-educated-female-adult FHHs). Notably, these main types of FHHs can also include subcategories. For example, under the second-type majority-female-adult FHHs, we can consider those where the proportion of employed females exceeds that of employed males (majority-employed-female-adult FHHs). Similarly, under the third-type potential FHH, we can consider a subcategory called core FHH that encompasses only the potential FHHs with employed females, and another subcategory called asset FHH encompassing only households with females who have ownership rights over the dwellings they reside in. Key confounding factors: Presence of children To account for additional household circumstances interacting with household poverty status, we should distinguish FHHs with and without children. Access to childcare affects women’s labor force participation (LFP) in many countries around the world, rich and poor alike (Akgunduz and Plantega, 2018; Clark et al., 2019). In European countries, childless women (with or without a partner) and single mothers have higher personal earnings than women whose family trajectories combined parenthood and partnership (Muller et al., 2020). In Egypt, childcare similarly presents a considerable barrier to women’s employment (Caria et al., 2022). Yet, only a handful of previous studies have examined how poverty differs with and without children, but mostly for self-reported FHHs (Medeiros and Costa, 2008; Liu et al., 2017; AlAzzawi, 2018). Exceptions include Munoz Boudet et al. (2018) and Munoz Boudet et al. (2021), who look at household gender composition 11 with and without children. Importantly, the common finding in these few studies is that FHHs with children are generally poorer than FHHs without children. Furthermore, a related economics literature on equivalence scales suggests that scale adjustments (for different numbers of adults and children) could have substantial effects on poverty and profiles of the poor for various countries at different income levels (Lanjouw and Ravallion, 1995; Newhouse et al., 2017; Abanokova et al., 2022). This is especially relevant for FHHs; for example, FHHs tend to have lower numbers of household members, but higher dependency ratios (World Bank, 2011; Klasen et al., 2015). In our sample (Appendix B, Table B.7), across all years and countries, the average size of FHHs is 5.7, while that of MHHs is 7.4. The average age of female heads is much higher than that of male heads (56 for FHHs vs. 48 for MHHs). Female heads are also mostly widowed (70.7%, compared to 1.3% for male heads). For widowed FHHs, their offspring are typically already-grown, independent adults who might be contributing to household expenses from their own earnings. The current welfare of these female heads is likely a function of their accumulated earnings, or more likely those of their deceased or living spouses and adult children, and thus are not strictly comparable to (male or female) heads with young children who rely on current labor market earnings to support themselves and their families. This distinction is especially pertinent to dynamic analysis. If the full sample of female or male heads were treated as a single group, this would unduly bias the results in favor of the elderly, widowed female heads without children, and against the much younger working male heads with children. Consequently, it is important to examine poverty incidence and dynamics among FHHs with or without children. In summary, our new typology consists of four main types of FHHs and their associated five sub-types (variants). We investigate poverty trends and dynamics of these four types of FHHs for 12 households with and without children, further differentiating between those with different numbers of children. 2.2. Empirical framework We provide both static and dynamic analyses of FHH (headcount) poverty in the Arab region. For static analysis, we examine the differences in poverty between different types of FHHs and non-FHHs. Specifically, we estimate the following linear probability model ′ ℎ = ℎ ℎ + ℎ ℎ ∗ ℎℎ + ℎ ℎ + + + ℎ (1) where ℎ is a binary variable representing the poverty status (i.e., equals 1 if poor and 0 otherwise) for household i, i= 1,..., n in survey round j, j= 1 or 2, country c, c= 1,…, 6, for FHH type h (ℎ ). ℎℎ is the number of young children age 0-14 in the household (who are generally not old enough to enter the labor force). ℎ is a vector of control variables, including household employment and demographic characteristics and residence area (i.e., urban/ rural residence). and are respectively the country and survey round (year) fixed effects that control for unobserved macro factors that can affect the whole country or outcomes in specific years, and ℎ is the error term. In Equation (1), ℎ and ℎ are the coefficients of interest. Compared to non-FHHs, ℎ presents the association between poverty and different types of FHHs without any children, ℎ + ℎ presents this association for FHHs with exactly one child, and so on. For easier interpretation, we can also fix the number of children at the mean (���������������� ℎℎ ) and consider the association between poverty and different types of FHHs with the average number of children as ℎ + ℎ ���������������� ℎℎ . 13 It is useful to estimate and compare two different versions of Equation (1), one without the control variables ℎ and one with these control variables. If the estimates for ℎ considerably change (or weakens) if the control variables are added, this indicates that the specified FHH type’s exposure to poverty is sensitive to these control variables. Put differently, this presents a test whether the specified FHH type can capture a relationship with poverty that is not explained by the control variables (i.e., how good the definition of the specified FHH type is). The findings based on our review of the literature suggest that FHHs’ exposure to poverty (ℎ and to some extent ℎ ) are likely sensitive to household composition and employment characteristics. For the dynamic analysis, let ℎ represent type-h FHHs’ household consumption (or income) per capita, and ℎ be the poverty line in period j for country c. We are interested in knowing the unconditional measures of upward poverty mobility such as (1ℎ < 1 2 > 2 ) (2) which represents the percentage of type-h FHHs that are poor in the first survey round (year) but nonpoor in the second survey round, or the conditional upward mobility measures such as (2ℎ > 2 | 1ℎ < 1 ) (3) which represents the percentage of poor households in the first round that escape poverty in the second round. If true panel data were available, we could straightforwardly estimate the quantities in (2) and (3); but in the absence of such data, we can use synthetic panels to study mobility. We employ recent advances with synthetic panel data methods (Dang and Lanjouw, 2023) to construct synthetic panel data and provide more insights into the dynamics of poverty for FHHs over time. 6 6 Recent validations and applications of the synthetic-panel methods by various researchers for different country contexts in Africa, Latin America, the Middle East, and Europe have been encouraging regarding accurate projections 14 Different from traditional pseudo-panel methods that require multiple rounds of cross-sectional surveys to study poverty mobility at the cohort level, the method that we apply works with as few as two survey rounds and provides poverty estimates at the more disaggregated household level. This method essentially exploits the time-invariant variables across the cross-sectional surveys to link different cohorts, in combination with additional cohort-based assumptions about the error terms, to construct the synthetic panels. Further discussion of this method and detailed estimates are provided in Appendix C. 3. Data and Descriptive Statistics 3.1. Data We analyze 20 survey rounds from six countries: Egypt, Iraq, Jordan, Mauritania, the West Bank and Gaza, and Tunisia. For Egypt, we use the Household Income, Expenditure and Consumption Surveys (HIECs) for 2012-2013, 2015, 2017-2018, and 2019-2020; for Iraq, the Household Socio-Economic Survey (IHSESs) for 2007 and 2012; for Jordan the Household Expenditure and Income Surveys (HIESs) for 2010-2011 and 2013-2014; for Mauritania, the Permanent Survey of Living Conditions of Households (EPCVs) for 2004, 2008, 2014, and 2019; for the West Bank and Gaza, the Expenditure and Consumption Survey (PECSs) for 2007, 2009, 2011, and 2016-2017; and for Tunisia, the National Survey on Household Budget, Consumption and Standard of Living (NSHBCs) for 2005, 2010, 2015, and 2021. These surveys provide rich information on household expenditures and various household and individual characteristics for the different household types. of economic status (Ferreira et al., 2012; Beegle et al., 2016; UNDP, 2016; OECD, 2018; Salvuci and Tarp, 2021; Ghomi, 2022). 15 Several of these surveys were harmonized by the Economic Research Forum (Egypt’s 2012- 2013, 2015, 2017-2018 HIECSs; Iraq’s 2007 and 2012 IHSESs; Jordan’s 2010-2011, and 2013- 2014 HIESs; the West Bank and Gaza’s 2009 and 2011 PECSs; and Tunisia’s 2005 and 2010 NSHBCs. The most recent surveys for Egypt (2019-2020), the West Bank and Gaza (2016-2017), Tunisia (2015 and 2021), and the Mauritanian EPCVs were obtained from national statistical agencies CAPMAS, PCBS, INS and ONS, respectively. We implemented careful harmonization of these surveys with the previous survey years and translated the variables from Arabic or French to English. We present the poverty lines for the six countries in Tables A.2-A.7 (Appendix A), compiling them from official sources and World Bank publications. We used region-specific poverty lines within each country to account for spatial differences in consumption (expenditure) patterns and price levels. 7 Since our focus is on poverty analysis, we used consumption values and national poverty lines in local currency units and in survey-year prices to sidestep conversion issues (e.g., with the PPP or market exchange rates) and adjustment for inflation. 3.2. Living-standards indicator Expenditure is widely regarded as a better indicator of permanent income when households, particularly in poorer countries, exercise consumption smoothing and use savings to augment unstable incomes due to seasonal or informal employment or unexpected shocks (Deaton, 1997; Deaton and Zaidi, 2002; Mancini and Vecchi, 2022). We use household consumption expenditures 7 We were able to do this for all countries in our sample except for Jordan. According to DOS reports, Jordan’s Department of Statistics (DOS) did not publish region-specific poverty lines and used a single poverty line for all of Jordan in 2010 and 2013. Jordan’s DOS does not publish region-specific Consumer Price Indices so we were unable to take spatial price differences into consideration. 16 per capita as the welfare measure underlying poverty analysis. 8 This includes all monetary expenditures on consumer goods and non-monetary consumption, such as imputed rents, use-value of durables, own production and in-kind transfers (i.e., gifts) received by households. Food consumption includes food that the household has purchased, grown and received from other sources. Non-food consumption is the sum of expenditure on all non-food items, including expenditure on fuel, clothing, schooling, health and miscellaneous items, and in-kind transfers. It can be useful to ensure comparability of household expenditures across different contexts to account for potential differences in households’ age and size compositions, as well as economies of scale in consumption. Studies have examined individual-level, rather than household-level, consumption to better disaggregate expenditures by gender (Dunbar et al., 2013; De Vreyer and Lambert, 2021). Unfortunately, the available surveys provide data on household consumption aggregates rather than individual-level consumption; therefore this approach cannot be applied to the available data. Another approach is to calculate the Adult Equivalent Expenditure (AEE) (or income) for each household, which gives smaller weight to children than adults and takes economies of scale into consideration. For example, Deaton and Paxson (1998) suggest using a parametric form of an equivalence scale, where a child is assumed to require a fraction of what an adult needs, and where the elasticity of needs with respect to adjusted household size is a constant . This gives rise to the following formula ∗ = �( � (4) + ) 8 This is also the most common approach employed in recent studies of poverty in countries in the Middle East (Marotta et al., 2011; CAPMAS, 2013). 17 where ∗ is the AEE for household i in survey j, which is an adjusted version of household expenditure conditional on the number of adults and children (we suppress the country and FHH type indexes for less cluttered notation). The smaller is, the smaller the relative weight of children; the higher is , the smaller the degree of economies of scale assumed. 9 We construct several different AEE levels for each household based on this method, using different values for the weight of children () and degree of economies of scale ( ) and show the results in Figure D.1, Appendix D. The relationship between household size and poverty dynamics reveals varying scenarios for FHHs, with FHHs generally having a higher likelihood of escaping poverty than non- FHHs when assessing consumption on a per capita basis. 10 3.3. Descriptive statistics Table 1 presents some key sample statistics on the prevalence of FHHs defined according to our proposed typology (Section 2.1) across the six Arab countries and different (survey) years. The four main types of FHHs are shown in bold while the alternative sub-types are shown in regular font. The shares of self-reported FHHs remain relatively stable over time in most countries, except for Mauritania. In recent years, this share hovers from around 10 percent (Iraq, the West Bank and Gaza) to 13 percent (Jordan) and 18 percent (Egypt, Tunisia). 11 Mauritania has the largest share of self-reported FHHs, which has almost doubled from 18.9 percent in 2004 to 36.6 percent in 2019. The shares of majority-female-adult FHHs are significantly higher in all countries, 9 When = 1 and = 1, we have per capita expenditure, which assumes no economies of scale and an equal weight for children and adults in the household. 10 These results are consistent with Abanokova et al.’s (2022) finding regarding the sensitivity of income dynamics to scale parameters, showing persistent upward mobility when income is evaluated on a per capita basis for the Russian Federation. 11 These shares are lower than the corresponding figure of 26 percent for African households observed in Milanzo and van de Walle (2017). 18 ranging from 21 percent (Egypt) to 44 percent (Mauritania) in the most recent years. Potential FHHs are as prevalent as those identified by self-reporting in all the countries except Iraq, where they are half as prevalent. Finally, most-educated-female-adult FHHs have relatively low prevalence rates, ranging from around 6 percent in Iraq to 25 percent in Mauritania. There is a weak-to-medium correlation among the four FHH types (i.e., ranging from 0.27 to 0.51; Appendix A, Table A.8), suggesting that each of the proposed FHH types captures different information about female headship. Compared with the main four types, the alternative subtypes all provide lower-to-almost- negligible prevalence of FHHs. For example, under the self-reported FHH type, while de jure FHHs account for between 8 percent and 17 percent of households for all countries and years, the corresponding figures for married FHHs are between 1 percent and 4 percent for all the countries, except for Mauritania in 2008 and later years. Under the majority-female-adults FHH type, the sub-type majority-employed-female-adult FHHs, however, yield a much smaller group of FHHs (ranging from around one-half to two-thirds as small). This is expected given the very low female LFP rates in the region, especially in such countries as Iraq and Jordan where they are among the lowest globally. Figure 2 illustrates the trends in poverty headcount ratios by country for the four main types of FHHs against those of the whole population for each country. This figure shows that different FHH types display clear differences regarding poverty levels and trends. Specifically, while potential FHHs (purple line) show faster poverty decreases in Iraq, Jordan, the West Bank and Gaza, and Tunisia, most-educated-female-adult FHHs (pink line) show slightly opposite trends from those of the whole population for Iraq. This contrasts with self-reported FHHs (green lines) 19 and potential FHHs, which predominantly have less poverty than the whole population for almost all the country-year observations. 12 For each country, Figure 3 and Figure 4 present the poverty differences between FHHs and non-FHHs for the four main FHH types respectively by year and by the number of children (age 0-14). Figure 3 indicates that the self-reported and potential FHH types typically have lower poverty ratios than non-FHH households across most years and countries. However, majority- female-adult FHHs tend to be poorer than the respective non-FHHs in most years and countries, except in Egypt 2017–2020. Most-educated-female-adult households have systematically more poverty prevalence than the corresponding non-FHHs in Iraq, Jordan, the West Bank and Gaza and Tunisia, but less poverty prevalence in Egypt and Mauritania. 13 Figure 4 shows that the presence of children is positively associated with poverty prevalence among FHHs for most of the countries, except for Egypt and Mauritania. 4. Estimation Results 4.1. Cross-sectional poverty Table 2 provides the estimation results for the associations between four main FHH types and poverty (ℎ in Equation (1)), without and with the household employment, demographic 12 Pooling data for all years and countries, we further show the FHH–non-FHH poverty differences for all FHH types and by the number of children in households in Figures A.2 and A.3 (Appendix A). These figures indicate that the self-reported and potential FHH types tend to have less poverty than non-FHH households across most years and countries, but the relationship between the number of children and poverty varies across countries. 13 Table A.9 (Appendix A) provides cross-sectional poverty rates for different household types over time in six countries. Panels B and C additionally report these poverty rates for rural and urban subgroups, and Panels D and E report the poverty rates for households with children under 14 and without children under 14. Poverty rates are typically higher in rural area than in urban areas, except for the West Bank and Gaza, and higher for households with younger children. Given the consistently high poverty rates among FHHs defined by the share of women among adults, we also assess the poverty rates among self-reported MHHs according to the number of female adults present in Figure A.1 (Appendix A). Poverty almost always increases with the number of females in all six countries, validating the central finding from Figure 3. 20 characteristics and residence-area variables shown respectively in the first four columns and the second four columns (Appendix A, Table A.10 offers the full results). Several interesting results stand out. First, the estimated �ℎ is negative and strongly statistically significant for three FHH types: self-reported, potential, and most-educated-female-adult FHHs, suggesting that these three FHH types are associated with less poverty prevalence. Majority-female-adult FHHs, in contrast, are associated with more poverty. This is generally consistent with our earlier discussion for Figure 3, indicating that these types of FHHs can serve as useful definitions. �ℎ increases for self-reported and most-educated-female- Second, the absolute magnitude of adult FHHs but decreases for majority-female-adult and potential FHHs when the control variables are added. The t-tests for these changes are statistically significant. This suggests that, consistent with our earlier discussions of the literature (Sections 2.1 and 2.2), FHHs’ exposure to poverty is also affected by the control variables, including household employment, demographic characteristics, and residence area variables. Indeed, prior research for various countries suggests that FHHs are not systematically poorer or more vulnerable (Fuwa, 2000; Klasen et al., 2015; Munoz Boudet et al., 2018; Brown and Van de Walle, 2021). Liu et al. (2017) find that in eight of 14 Latin American countries, FHHs more likely live in poorer conditions, but these gaps either disappear or reverse when controlling for other household and demographic characteristics. Table 2 shows that self-reported FHHs are about 1 percent (without control variables) to 4 percent (with control variables) less likely to be poor than non-FHHs if there are no children in the household. The corresponding changes are about 3 percent (without control variables) to 2 percent (with control variables) for potential FHHs, and 1 percent (with control variables) for most- educated-female-adult FHHs. However, majority-female-adult FHHs are 5 percent (without 21 control variables) to 3 percent (with control variables) more likely to be poor than non-FHHs if there are no children. � ) is Finally, the estimated interaction term between FHH types and the number of children ( positive for three of the four FHH types (self-reported, potential, and most-educated-female-adult �ℎ are FHHs), but negative for majority-female-adult FHHs. While the absolute magnitudes of small, around 1 percent (i.e., one more child is associated with a 1-percent change in the probability of the household being poor), it is strongly statistically significant. Furthermore, when we fix the number of children at the mean of the estimation sample (i.e., 1.81 children), self-reported FHHs become 1 percent more likely to be poor (without control variables) and 3 percent less likely to be poor (with control variables). The corresponding probabilities, without and with control variables, become 2.5-2.8 percent more likely to be poor for most-educated-female-adult FHHs and 4-0.7 percent more likely to be poor for majority-female-adult FHHs. However, potential FHHs are 0.2 percent (without control variables) less likely to be poor and are 0.3 percent (with control variables) more likely to be poor. In addition, having more children (or larger household sizes) is associated with greater poverty risks (Appendix A, Table A.10). This result concurs with the finding by Munoz Boudet et al. (2018) and Munoz Boudet et al. (2021) that adult couple households with children are the largest and overrepresented group among poor households. This provides supportive evidence for our proposed typology that considers children when defining FHH types. The five remaining FHH subtypes offer qualitatively similar results, showing that most FHH types are associated with less poverty, except for majority-employed-female-adult FHHs where the opposite result holds (Appendix A, Table A.11). This table also shows the interaction terms between FHH types and the number of children, which are mostly statistically significant. The results using the alternative logit model are qualitatively similar, albeit somewhat weaker for the 22 most-educated-female-adult FHHs (Appendix A, Tables A.12 and A.13). 14 We further consider the overlaps of three main FHH types (self-reported, potential, and most-educated-female-adult FHHs) and all four main FHH types and show the estimation results in Appendix A, Table A.14, which remain qualitatively similar. 4.2. FHH poverty dynamics based on synthetic panels We turn next to discussing the results on poverty dynamics based on synthetic panels. For each country, Figure 5 reports the conditional upward mobility rates in the second survey year for the four main FHH types, considered separately with and without any children (Equation (3)). Figure 5 shows considerable (conditional) upward mobility at the national average level (dashed line) for some countries. In particular, the upward mobility rate is 45 percent in Iraq during 2007-2012 (i.e., 45 percent of the initial poor in the first year escape poverty in the second year), 54 percent in Jordan during 2010-2013, and 41 percent in Mauritania during 2014-2019. Still, a significant degree of immobility exists in Egypt, the West Bank and Gaza, and Tunisia, where most of the population remained poor in both years and only about one-third (or less) of the poor escaped poverty in the most recent year: 29 percent for Egypt during 2017–2020, 31 percent in the West Bank and Gaza during 2011–2017, and 21 percent in Tunisia during 2015–2021. 15 Unsurprisingly, non-FHHs have upward mobility rates that are almost the same as the national averages, given their large shares in the population (Table 1). But interestingly, FHHs without 14 The estimated marginal effects for the interaction terms with children are qualitatively similar (using the Stata command “ginteff” (Radean, 2023)). 15 The survey period lengths generally differ for the six countries so the estimated mobility rates are not exactly comparable across countries or to those in other studies. For a rough reference, Dang and Ianchovichina (2018) obtain a regional upward mobility rate around 52 percent in the early 2000s and 2010s. However, if we assume a similar rate of change for mobility across the years for all countries, we can obtain the average mobility per year for each country (Appendix A, Table A.15). 23 children are most likely to experience upward mobility. Out of 24 FHH types across six countries, the probabilities of FHHs without any children escaping poverty are higher than the national averages in 22 cases. The exceptions are self-reported and potential FHHs in Jordan during 2010- 2013, which have similar upward mobility rates as the national average. However, FHHs with children have upward mobility rates that are clearly higher than the national averages in five cases (self-reported, majority-female-adult, and most-educated-female-adult FHHs in Egypt during 2017-2020, and self-reported FHHs in Iraq during 2007-2012 and in Mauritania during 2014-2019) and clearly lower than the national averages in six cases (self-reported, majority-female-adult, and most-educated-female-adult FHHs in Jordan 2010-2013 and Tunisia 2015-2021). FHHs with children have similar upward mobility as the national averages for the remaining cases. Overall, across all countries and four main FHH types, the upward mobility rates for FHHs without children, FHHs with children, and non-FHHs are respectively 42 percent, 37 percent, and 36 percent. Figure 6 plots the conditional downward mobility (i.e., falling into poverty in the second year when being initial non-poor in the first year). The results are consistent with those shown in Figure 5, with FHHs without children experiencing the least downward mobility, followed by FHHs with children and non-FHHs. Overall, the downward mobility rates for FHHs without children, FHHs with children, and non-FHHs are respectively 14 percent, 17 percent, and 19 percent. As an alternative to Figure 5, we plot the results of locally weighted regressions of upward mobility on the number of children (Appendix A, Figure A.3). This figure also shows that the number of children is negatively associated with upward mobility for most countries and FHH types. The results for other sub-types of FHHs are qualitatively similar, with FHHs without children having the most upward mobility (Appendix A, Figure A.4). The results for preceding 24 years are, however, somewhat mixed. FHHs without children had the strongest upward mobility for Egypt and Tunisia, but had similar upward mobility as FHHs with children for Mauritania and the West Bank and Gaza (Appendix A, Figures A.5-A.8). Finally, we plot the results for upward and downward mobility for FHH types, with and without children considered together, in Figures A.9 and A.10 (Appendix A). These figures show that FHHs have higher upward mobility and lower downward mobility than MHHs across all FHH types and countries, except for Jordan. 16 5. Conclusions and Policy Implications The dramatic events of the Arab Spring and the following decade of structural reforms and sectoral developments inter alia have brought to the fore the importance of better understanding gender inequalities. We offer new analysis on the feminization of poverty as related to FHHs, using 20 survey rounds spanning the past two decades for six countries across the Arab region— namely Egypt, Iraq, Jordan, Mauritania, the West Bank and Gaza, and Tunisia—an understudied set of countries. We propose and evaluate a new typology of FHHs consisting of four main types (and several sub-types) of FHHs with a new focus on the presence of children, which offers policy- relevant insights regarding the trends and dynamics of poverty feminization. We assemble and harmonize the available cross-sectional data and construct synthetic panels to address the lack of actual panels. We find that different FHH types display clear differences regarding poverty levels and trends. In particular, self-reported FHHs, potential FHHs, and most-educated-female-adult FHHs are less likely to be poor than non-FHHs for the six countries, while the opposite holds for majority-female- 16 While Milazzo and Van de Walle (2017) find self-reported FHHs to be generally poorer, they also find these households to contribute more to the overall decline in poverty in Africa. 25 adult FHHs. Yet, more children are associated with more poverty for the former three types of FHHs and less poverty for the last type of FHHs. We also find considerable (conditional) upward mobility, ranging between 21 and 54 percent of the initially poor in a country. But country heterogeneity exists, with Iraq, Jordan, and Mauritania having relatively more upward mobility, while Egypt, the West Bank and Gaza, and Tunisia have relatively less upward mobility. While most types of FHHs more likely experience upward mobility out of poverty (or less likely fall into poverty), FHHs without children have the strongest upward mobility (or the least downward mobility), followed by FHHs with children, and non-FHHs. Our proposed typology aligns with recent calls to go beyond identifying headship based on the gender of the head alone. For example, Beegle and van de Walle (2019) argue that since many women live in MHHs, especially in Sub-Saharan Africa, if resources are unequally shared among household members, simply comparing FHHs and MHHs based on heads’ gender can provide biased results. Summarizing opinions from experts on gender issues and survey design on the topic, Buvinic and van de Walle (2019) similarly call for other definitions based on other household characteristics including demographic and gender composition. Other concerns were also raised about practical survey challenges with headship (e.g., when the male head temporarily lives away from the household). These discussions do not just serve academic purposes but have practical policy implications. Governments in the region strive to identify various vulnerable FHHs for effective social- protection interventions aimed at targeting vulnerable groups and reducing gender inequalities. For example, Egypt’s largest poverty-targeting cash-transfer program, Takaful, uses proxy means testing to target households and the criteria include a much lower threshold for FHHs. In 2017, the 26 poverty threshold used to determine eligibility was raised considerably for MHHs while it was kept constant for FHHs (ESCWA, 2021), resulting in the share of beneficiaries who were FHHs almost doubling from 48% to 92%. In Jordan, the National Aid Fund targets several categories of “vulnerable” FHHs such as widows with children, those without “support”, and divorced female heads, not just poor FHHs, while its poverty-reduction program directly targets poor FHHs (NAF, 2020; ESCWA, 2021; World Bank, 2022). In Lebanon, and Tunisia, FHHs, especially widows, are also prioritized (Nasri, 2020; ESCWA, 2021). Against this background, our findings offer highly relevant policy inputs and run against the conventional wisdom that FHHs are typically poorer than non-FHHs, which appears to be the implicit assumption underlying many targeting programs in the region and elsewhere. In contrast, we find majority-female-adult households or households with more children more vulnerable to (remaining in) poverty. While these results suggest that female headship definition using gender composition can offer an alternative approach—and potentially help identify a more vulnerable group—for poverty targeting, they also highlight the need for a more nuanced understanding of how female headship can be defined, especially in the presence of children. Furthermore, we also need to better understand the extent to which the different types of FHHs’ exposure to poverty can change, depending on various other factors such as whether we examine households’ static or dynamic poverty status, whether other household demographic and employment characteristics are considered, and last but not least, the country-specific contexts. 27 References Abanokova, K., Dang, H.A.H. and Lokshin, M., (2022). “Do adjustments for equivalence scales affect poverty dynamics? Evidence from the Russian federation during 1994–2017”. Review of Income and Wealth, 68, pp.S167-S192. AbdelLatif, Lobna M., Mohamed Ramadan, and Sarah A. Elbakry. (2019). “How gender biased are female-headed household transfers in Egypt?” Middle East Development Journal, Vol. 11(2), pp.165-180. Akgunduz, Y. E., & Plantenga, J. (2018). Childcare prices and maternal employment: A meta‐ analysis. Journal of Economic Surveys, 32(1), 118-133. AlAzzawi, Shireen. (2018). “Do Endowments Matter? Exploring the Gender Dimensions of Poverty in Egypt.” Review of Income and Wealth, Vol. 64(S1), pp. S189-S224. AlAzzawi, Shireen and Vladimir Hlasny. (2022). “Youth Labor Market Vulnerabilities: Evidence from Egypt, Jordan and Tunisia.” International Journal of Manpower, Vol. 43(7), pp. 1670- 1699. Alon, Titan, Sena Coskun, Matthias Doepke, David Koll, and Michèle Tertilt. (2022). “From mancession to shecession: Women’s employment in regular and pandemic recessions.” NBER Macroeconomics Annual, Vol. 36(1), pp. 83-151. Amara, Mohamed, and Hatem Jemmali. (2018). “Household and Contextual Indicators of Poverty in Tunisia: A Multilevel Analysis.” Social Indicators Research, Vol. 137, pp. 113-138. Barrett, C. B. (2005). Rural poverty dynamics: development policy implications. Agricultural Economics, 32, 45-60. Beegle, K., & van de Walle, D. (2019). What can female headship tell us about women’s well- being? Probably not much. https://blogs.worldbank.org/impactevaluations/what-can-female- headship-tell-us-about-womens-well-being-probably-not-much Beegle, Kathleen, Luc Christiaensen, Andrew Dabalen, and Isis Gaddis. (2016). Poverty in a Rising Africa. Washington, DC: The World Bank. Bibi, Sami and Rim Chatti. (2010). “Gender Poverty in Tunisia: Is there a Feminization Issue?” Middle East Development Journal, Vol. 2(2), pp. 283-307. Brown, Caitlin, and van de Walle, Dominique. (2021). “Headship and Poverty in Africa.” World Bank Economic Review, Vol. 35(4), pp. 1038-1056. Buvinic, Mayra and Geeta Rao Gupta. (1997). “Female-Headed Households and Female- Maintained Families: Are They Worth Targeting to Reduce Poverty in Developing Countries?” Economic Development and Cultural Change, Vol. 45(2), pp. 259-280. 28 Buvinic, M., & van de Walle, D. (2019). The Debate about Headship in Poverty and Gender Studies. https://www.cgdev.org/blog/debate-about-headship-poverty-and-gender-studies Caria, S., Crepon, B., ElBehairy, H., Fadlalmawla, N., Krafft, C., Nagy, A., ... & El Assiouty, S. (2022). Child Care Subsidies, Employment Services and Women's Labor Market Outcomes in Egypt: First Midline Results. World Bank. Chant, Sylvia. (2003). “Female Household Headship and the Feminization of Poverty: Facts, Fictions and Forward Strategies”, LSE Gender Institute, New Working Paper Series, Issue 9 (London: London School of Economics). CAPMAS. (2014). “Poverty Indicators according to the Household Income, Expenditure and Consumption Survey 2012/2013”. http://capmas.gov.eg/pepo/a.pdf. Clark, S., Kabiru, C. W., Laszlo, S., & Muthuri, S. (2019). The impact of childcare on poor urban women’s economic empowerment in Africa. Demography, 56(4), 1247-1272. Dang, Hai‐Anh H., and Elena Ianchovichina. (2018). “Welfare dynamics with synthetic panels: the case of the Arab world in transition.” Review of Income and Wealth, Vol. 64, pp. S114- S144. Dang, Hai-Anh and Peter Lanjouw. (2023). “Measuring poverty dynamics with synthetic panels based on cross-sections.” Oxford Bulletin of Economics and Statistics, 85(3): 599-622. Dang, Hai-Anh and Cuong Viet Nguyen. (2021). “Gender Inequality during the COVID-19 Pandemic: Income, Expenditure, Savings, and Job Loss.” World Development, Vol. 140, 105296. Datt, G., D. Jolliffe, and M. Sharma. (2001). “A profile of poverty in Egypt.” African Development Review, Vol. 13(2), pp. 202-237. Deaton, A., & Paxson, C. (1998). Economies of scale, household size, and the demand for food. Journal of Political Economy, 106(5), 897-930. DeGraff, Deborah S., and Richard E. Bilsborrow. (1993). “Female-headed households and family welfare in rural Ecuador.” Journal of Population Economics, Vol. 6, pp. 317-336. De Vreyer, P., & Lambert, S. (2021). Inequality, poverty, and the intra-household allocation of consumption in Senegal. World Bank Economic Review, 35(2), 414-435. Deere, C. D., & León, M. (2003). The gender asset gap: Land in Latin America. World Development, 31(6), 925-947. Deere, C. D., Alvarado, G. E., & Twyman, J. (2012). Gender inequality in asset ownership in Latin America: Female owners vs household heads. Development and Change, 43(2), 505-530. 29 Dreze, J., & Srinivasan, P. V. (1997). Widowhood and poverty in rural India: Some inferences from household survey data. Journal of Development Economics, 54(2), 217-234. Dunbar, G. R., Lewbel, A., & Pendakur, K. (2013). Children's resources in collective households: identification, estimation, and an application to child poverty in Malawi. American Economic Review, 103(1), 438-471. El-Laithy, Heba. (2001). “The Gender Dimensions of Poverty in Egypt.” ERF Working Paper 127. ESCWA, 2021 “Targeted social protection in Arab countries before and during the Covid-19 crisis”. Economic and Social Commission for Western Asia. Ferreira, Francisco H. G., Julian Messina, Jamele Rigolini, Luis-Felipe López-Calva, and Renos Vakis. (2012). Economic Mobility and the Rise of the Latin American Middle Class. Washington DC: World Bank. Fuwa, Nobuhiko. (2000). “The poverty and heterogeneity among female-headed households revisited: the case of Panama.” World Development, Vol. 28(8), pp. 1515-1542. Gammage, Sarah. (1998). “The Gender Dimension of Household Poverty: Is Headship Still a Useful Concept?” International Center for Research on Women (Washington DC) Ghomi, M. (2022). Who is afraid of sanctions? The macroeconomic and distributional effects of the sanctions against Iran. Economics & Politics, 34(3), 395-428. Grant, Monica J., and Jere R. Behrman. (2010). “Gender Gaps in Educational Attainment in Less Developed Countries.” Population and Development Review, Vol. 36(1), pp. 71-89. Horrell, Sara, and Pramila Krishnan. (2007). “Poverty and productivity in female-headed households in Zimbabwe.” Journal of Development Studies, Vol. 43(8), pp. 1351-1380. IMF. (2011). Islamic Republic of Mauritania: Poverty Reduction Strategy Paper, IMF Country Report No. 11/252, August, www.imf.org/external/pubs/ft/scr/2011/cr11252.pdf. Klasen, Stephan, Tobias Lechtenfeld, and Felix Povel. (2015). “A feminization of vulnerability? Female headship, poverty, and vulnerability in Thailand and Vietnam.” World Development, Vol. 71, pp. 36-53. Lanjouw, P., & Ravallion, M. (1995). Poverty and household size. Economic Journal, 105(433), 1415-1434. Liu, Chia, Albert Esteve, and Rocío Treviño. (2017). “Female-headed households and living conditions in Latin America.” World Development, Vol. 90, pp. 311-328. 30 Mancini, Giulia, and Vecchi, Giovanni. (2022). On the Construction of a Consumption Aggregate for Inequality and Poverty Analysis. Washington, D.C.: World Bank Group. Marotta, D., Yemtsov, R., El-Laithy, H., Abou-Ali, H., & Al-Shawarby, S. (2011). Was growth in Egypt between 2005 and 2008 pro-poor? From static to dynamic poverty profile. World Bank Policy Research Working Paper 5589. Medeiros, Marcelo, and Joana Costa. (2008). “Is there a feminization of poverty in Latin America?” World Development, Vol. 36(1), pp. 115-127. Milazzo, Annamaria, and Dominique van de Walle. (2017). “Women left behind? Poverty and headship in Africa.” Demography, Vol 54(3), pp. 1119-1145. Muller, J. S., Hiekel, N., & Liefbroer, A. C. (2020). The long-term costs of family trajectories: Women’s later-life employment and earnings across Europe. Demography, 57(3), 1007-1034. Munoz Boudet, A. M., Buitrago, P., Leroy De La Briere, B., Newhouse, D. L., Rubiano Matulevich, E. C., Scott, K., & Suarez-Becerra, P. (2018). Gender Differences in Poverty and Household Composition through the Life-Cycle: A Global Perspective. Policy Research Working Papers 8360. Washington, D.C: The World Bank. Munoz Boudet, A. M., Bhatt, A., Azcona, G., Yoo, J., & Beegle, K. (2021). A global view of poverty, gender, and household composition. Policy Research Working Papers 9553. Washington, D.C: The World Bank. Nasri, Khaled (2020). Social Safety Nets in Tunisia: Do Benefits Reach the Poor and Vulnerable households at the Regional Level? GLO Discussion Paper, No. 440, Global Labor Organization (GLO), Essen NAF 2020. “Assessing Gender Equity in National Aid Fund Programs”. Jordan National Aid Fund https://naf.gov.jo/EN/List/Statistics_and_Data Nassar, Heba. (1997). “Feminization of Poverty.” The Egyptian Human Development Report; Research Papers Series, UNDP. Newhouse, D., Becerra, P. S., & Evans, M. (2017). New global estimates of child poverty and their sensitivity to alternative equivalence scales. Economics Letters, 157, 125-128. OECD. (2018). A Broken Social Elevator? How to Promote Social Mobility. Paris: OECD Publishing. https://doi.org/10.1787/9789264301085-en Palestinian Central Bureau of Statistics (PCBS) and World Bank. (2018). “Measuring poverty in West Bank and Gaza: Methodology review using PECS 2016.” Technical report. 31 Quisumbing, Agnes R., Lawrence Haddad, and Christine Peña. (2001). “Are women overrepresented among the poor? An analysis of poverty in 10 developing countries.” Journal of Development Economics, Vol. 66(1), pp. 225-269. Radean, M. (2023). ginteff: A generalized command for computing interaction effects. Stata Journal, 23(2), 301-335. Ravallion, M. (2016). The economics of poverty: History, measurement, and policy. Oxford University Press. Rogan, Michael. (2013). “Poverty and headship in post-apartheid South Africa, 1997–2006.” Social Indicators Research, Vol. 113, pp. 491-511. Rogers, Beatrice Lorge. (1995). “Alternative definitions of female headship in the Dominican Republic.” World Development, Vol. 23(12), pp. 2033-203 Salvucci, Vincenzo, and Finn Tarp. (2021). “Poverty and vulnerability in Mozambique: An analysis of dynamics and correlates in light of the Covid‐19 crisis using synthetic panels.” Review of Development Economics, Vol. 25(4), pp. 1895-1918. United Nations Development Programme (UNDP). (2016). Multidimensional Progress: Well- being beyond Income. New York: United Nations Development Programme. Van de Walle, Dominique. (2013). “Lasting Welfare Effects of Widowhood in Mali.” World Development, Vol. 51, pp. 1-19. World Bank. (2011). World development report 2012: Gender equality and development. World Bank: Washington, DC. ---. (2022). “Jordan Emergency Cash Transfer COVID19 Response Project (P173974) Project Information Document.” 32 Table 1. Share of Female-Headed Households in Six Arab Countries (percentages) Egypt, Arab Rep. Iraq Jordan Mauritania West Bank and Gaza Tunisia 2012 2015 2017 2020 2007 2012 2010 2013 2004 2008 2014 2019 2007 2009 2011 2017 2005 2010 2015 2021 17.80 17.86 18.43 17.59 11.30 9.63 13.88 13.24 18.92 31.30 30.19 36.58 9.12 10.00 11.07 10.05 17.01 14.85 16.24 18.52 Type 1. Self-reported FHH 1,346 2,104 2,265 1,967 1,906 2,531 402 669 1,860 4,273 3,033 3,654 108 388 505 397 2,128 1,704 4,088 3,153 13.98 14.96 15.92 14.03 10.43 8.22 11.29 11.05 16.89 18.90 17.23 16.98 8.26 8.79 9.93 8.04 13.73 11.95 14.12 16.62 Official FHH 1,058 1,741 1,956 1,599 1,721 2,152 335 592 1,674 2,688 1,839 1,834 97 341 447 308 1,671 1,347 3,462 2,788 0.48 0.37 0.48 0.59 0.46 0.20 0.89 0.89 0.44 0.45 0.50 0.60 0.59 1.07 1.09 0 1.01 0.56 1.29 1.51 Never married 35 43 57 67 78 101 17 43 44 67 48 72 7 45 56 0 113 65 282 261 0.95 1.33 1.55 1.80 0.78 0.62 0.51 0.66 6.15 6.98 6.22 6.64 0.99 0.65 1.14 0.82 1.28 1.18 1.53 1.90 Divorced/ separated 74 153 194 210 120 175 19 33 602 974 660 690 11 25 43 29 156 131 360 291 12.56 13.26 13.89 11.64 9.19 7.40 9.90 9.50 10.29 11.47 10.52 9.74 6.68 7.07 7.70 7.22 11.44 10.21 11.29 13.21 Widow only 949 1,545 1,705 1,322 1,523 1,876 299 516 1,028 1,647 1,131 1,072 79 271 348 279 1,402 1,151 2,820 2,236 3.80 2.89 2.52 3.53 0.87 1.41 2.58 2.19 1.92 12.20 12.96 18.45 0.87 1.22 1.14 2.01 3.28 2.90 2.12 1.91 Married FHH 287 363 309 365 185 379 67 77 176 1,558 1,194 1,714 11 47 58 89 457 357 624 365 22.96 22.29 22.46 20.87 24.02 28.92 25.80 22.80 31.50 40.40 39.98 43.96 21.52 19.53 20.97 24.28 29.20 28.98 26.17 27.21 Type 2. Share of female adults>0.5 1,712 2,659 2,797 2,352 4,547 6,525 728 1,105 3,075 5,606 3,897 4,393 265 751 919 955 3,710 3,363 6,907 4,740 Share of employed females> share of 5.42 6.22 6.12 6.02 4.81 3.81 6.02 7.56 11.40 12.90 14.30 14.68 7.65 6.72 6.54 4.59 10.94 9.01 10.17 11.04 employed males 406 733 757 660 916 1,070 178 291 1,056 1,888 1,422 1,454 88 269 300 176 1,313 987 2,382 1,854 16.68 16.49 19.37 18.86 5.73 3.59 13.25 12.78 17.01 27.09 25.02 27.88 11.09 10.35 10.55 9.88 18.30 17.54 18.13 26.75 Type 3. Potential FHH 1,244 1,935 2,344 2,087 1,007 1,621 359 576 1,643 3,694 2,345 2,824 127 398 482 397 2,280 1,974 4,609 4,618 3.06 3.17 3.49 3.75 1.65 0.86 2.52 3.52 7.07 8.07 9.68 9.12 3.04 2.34 2.84 1.82 5.07 4.11 4.27 5.67 Core FHH 229 364 419 407 283 388 77 129 635 1,160 939 912 34 97 130 74 620 457 1,022 947 11.77 11.70 14.05 13.10 4.00 2.61 12.02 10.90 14.85 24.30 22.63 24.67 9.24 8.76 8.52 N.A. 15.56 15.55 15.22 N.A. Asset FHH 881 1,459 1,765 1,489 783 1,258 321 510 1,417 3,287 2,095 2,523 106 338 402 N.A. 1,967 1,769 4,037 N.A. Type 4. Most educated adult member 12.92 13.88 14.30 13.98 8.54 5.87 18.49 18.48 8.76 16.47 13.83 24.67 16.87 15.25 14.82 11.52 15.42 17.75 17.84 20.11 is female & no employed males 951 1,610 1,749 1,554 1,485 1,990 499 932 969 2,384 1,378 2,490 201 595 643 419 1,936 2,024 4,479 3,520 Note: The main definitions of female-headed households are in bold, and the variant definitions are in regular font. The numbers in bold refer to the percent of the cross-sectional sample for each period. The numbers in italics refer to the sample size of each group. Type 1 self-reported FHHs are obtained from self-reporting information in the survey questionnaires. Type 2 majority-female-adult FHHs are defined as households where the proportion of females among (working age) adults exceeds 0.5. Type 3 potential FHHs are those households where there are no working-age males present. Type 4 most-educated-female-adult FHHs consist of households with no employed males, whose most educated adult member is female. Under Type 1 FHHs, the different sub-types are defined as in the survey questionnaires. Under Type 3 FHHs, sub-type 3 core FHH encompasses only the potential FHHs with employed females, and sub-type 3 asset FHH encompasses only households with females who have ownership rights over the dwellings they reside in. 33 Table 2. Probabilities of Being Poor Specification 1 Specification 2 FHH Type 2 FHH Type 4 FHH Type 4 FHH Type 1 Majority-female- FHH Type 3 Most-educated- FHH Type 1 FHH Type 2 FHH Type 3 Most-educated- Self-reported adult Potential female-adult Self-reported Majority-female-adult Potential female-adult -0.007*** -0.044*** Self-reported FHH (0.00) (0.00) 0.008*** 0.006*** Self-reported FHH # Number of children (0.00) (0.00) 0.052*** 0.032*** Share of female adults>0.5 (0.00) (0.00) Share of female adults>0.5# Number of -0.007*** -0.014*** children (0.00) (0.00) -0.027*** -0.017*** Potential FHH (0.00) (0.00) 0.014*** 0.011*** Potential FHH# Number of children (0.00) (0.00) 0.002 -0.006** Educated females (0.00) (0.00) 0.013*** 0.013*** Educated females# Number of children (0.00) (0.00) Household head`s characteristics N N N N Y Y Y Y Household characteristics N N N N Y Y Y Y Country FE Y Y Y Y Y Y Y Y Survey year FE Y Y Y Y Y Y Y Y N 214931 214931 214931 214931 211069 211069 211069 211069 Note: ***, **, and * denote statistical significance at the 0.01, 0.05, and 0.10 levels. Robust standard errors are in parentheses. The full regression results using the linear probability model are provided in Appendix A, Table A.10. 34 Figure 1. Diagram of FHH Types Note: Solid line and dashed line respectively indicate direct and indirect relationship. Some studies are shown for illustrative purposes and do not represent an exhaustive list. 35 Figure 2. Cross-sectional Headcount Poverty Rate (percentage), by Household Type, Regional Poverty Lines Egypt Iraq 15 20 25 30 35 Poverty rate (%) Poverty rate (%) 30 25 20 15 2012 2015 2017 2020 2007 2012 Jordan Mauritania 5 10 15 20 25 30 20 30 40 50 60 Poverty rate (%) Poverty rate (%) 2010 2013 2004 2008 2014 2019 Palestine Tunisia 20 40 60 80 100 10 15 20 25 30 Poverty rate (%) Poverty rate (%) 2007 2009 2011 2017 2005 2010 2015 2021 Whole population Self-reported FHH Share of females>0.5 Potential FHH Educated females Note: Population sampling weights are applied. 36 Figure 3. FHH–non-FHH Differences in Headcount Poverty Rate (percentage points) Panel A: Self-reported FHH Panel B: Share of females>0.5 2012 *** 2012 Palestine Jordan Egypt Palestine Jordan Egypt 2015 *** 2015 2017 *** *** 2017 2020 *** *** 2020 *** 2007 2007 Mauritania Iraq Mauritania Iraq 2012 *** 2012 *** 2010 *** 2010 2013 *** 2013 2004 2004 2008 *** *** 2008 2014 ** *** 2014 2019 *** *** 2019 *** 2007 2007 2009 2009 2011 *** 2011 2017 ** *** 2017 ** Tunisia Tunisia 2005 *** 2005 2010 ** *** 2010 2015 *** ** 2015 2021 *** *** 2021 *** -5 0 5 10 -15 -10 -5 0 5 Panel C: Potential FHH Panel D: Educated females 2012 *** *** 2012 Palestine Jordan Egypt Palestine Jordan Egypt 2015 *** *** 2015 2017 *** *** 2017 2020 *** *** 2020 2007 * 2007 Mauritania Iraq Mauritania Iraq 2012 *** 2012 2010 ** *** 2010 2013 *** *** 2013 2004 *** 2004 2008 * *** 2008 2014 *** *** 2014 2019 *** *** 2019 2007 *** *** 2007 2009 *** *** 2009 2011 *** ** 2011 2017 *** 2017 Tunisia Tunisia 2005 ** 2005 2010 *** 2010 2015 *** 2015 2021 *** *** 2021 -40 -30 -20 -10 0 -20 -10 0 10 Note: Headcount poverty rates are estimated using per capita household expenditures. Stars indicate significantly higher headcount poverty ratio between FHHs and non-FHHs in each category. ***, **, and * denote statistical significance at the 0.01, 0.05, and 0.10 levels. Negative difference means FHHs are less likely to be poor. 37 Figure 4. Differences in Headcount Poverty Rate for FHHs vs. non-FHHs by Number of Children under 14 (percentage points) Panel A: Self-reported FHH Panel B: Share of females>0.5 Egypt Egypt Iraq Iraq Jordan Jordan Mauritania Mauritania Palestine Palestine Tunisia Tunisia -10 0 10 20 30 40 -10 0 10 20 30 Panel C: Potential FHH Panel D: Educated females Egypt Egypt Iraq Iraq Jordan Jordan Mauritania Mauritania Palestine Palestine Tunisia Tunisia -20 0 20 40 -20 0 20 40 0 1 2 3 4 or more Note: Authors’ calculation based on pooled cross section. Headcount poverty rates are estimated using per capita household expenditures. The number of children are shown for 0, 1, 2, 3, and 4 or more children. The headcount poverty rates are shown on the x-axis, with positive numbers indicating more poverty for FHHs. The error bars are the 95% CIs. 38 Figure 5. Probabilities of Female-Headed Households Escaping Poverty in Second Year Conditional on Being Poor in First Year (percentage) Egypt 2017-2020 Iraq 2007-2012 40 40 45 50 55 60 Percentage (%) Percentage (%) 35 30 25 Self-reported Potential Self-reported Potential Females>0.5 Educated fem. Females>0.5 Educated fem. Jordan 2010-2013 Mauritania 2014-2019 50 60 Percentage (%) Percentage (%) 55 45 50 40 45 35 Self-reported Potential Self-reported Potential Females>0.5 Educated fem. Females>0.5 Educated fem. Palestine 2011-2017 Tunisia 2015-2021 30 Percentage (%) Percentage (%) 25 30 35 40 45 25 20 15 Self-reported Potential Self-reported Potential Females>0.5 Educated fem. Females>0.5 Educated fem. FHH wo/children FHH w/children Non-FHH Note: Estimates are obtained with synthetic panel data and weighted with population weights where the second survey round is used as the base year. The figure shows the percentage of the population that moves out of poverty in the second year. FHH characteristics are measured in second period. Dashed red lines represent the national average for each period. Household heads' ages are restricted to 25-55 for the first survey round and adjusted accordingly for the second survey round. Standard errors are estimated with 1,000 bootstraps. 39 Figure 6. Probabilities of Female-Headed Households Falling in Poverty in Second Year Conditional on Being Non-poor in First Year (percentage) Egypt 2017-2020 Iraq 2007-2012 10 12 14 8 10 12 14 16 Percentage (%) Percentage (%) 8 6 Self-reported Potential Self-reported Potential Females>0.5 Educated fem. Females>0.5 Educated fem. Jordan 2010-2013 Mauritania 2014-2019 10 12 14 16 8 10 12 14 Percentage (%) Percentage (%) 8 6 Self-reported Potential Self-reported Potential Females>0.5 Educated fem. Females>0.5 Educated fem. Palestine 2011-2017 Tunisia 2015-2021 30 35 40 45 50 55 5 6 7 8 9 10 Percentage (%) Percentage (%) Self-reported Potential Self-reported Potential Females>0.5 Educated fem. Females>0.5 Educated fem. FHH wo/children FHH w/children Non-FHH Note: Estimates are obtained with synthetic panel data and weighted with population weights where the second survey round is used as the base year. The figure shows the percentage of the population that enters poverty in the second year. FHH characteristics are measured in second period. Dashed red lines represent the national average for each period. Household heads' ages are restricted to 25-55 for the first survey round and adjusted accordingly for the second survey round. Standard errors are estimated with 1,000 bootstraps. 40 Appendix A: Additional Tables and Figures Table A.1. Overview of the key studies FHH Reference Studies Country Conclusions definition group Self- Self- FHHs have lower income, land ownership, and average level of education than reported reported MHHs and are less likely to be employed (with fewer hours if employed). Children FHH MHH of FHHs are significantly less likely to be enrolled in school than children of MHHs. Self- Self- reported Widowed or divorced FHHs have higher income and amount of land owned than reported MHH, other MHHs but lower children's school enrollment and are less likely to be in school DeGraff and widowed 1 Ecuador self-reported than are children of MHHs and children of other FHHs. Bilsborrow (1993) FHH FHH Self- Self- Married FHHs have lower income and amount of land owned than MHHs and reported reported enrollment rate similar to MHH, but children of married FHHs are significantly less MHH, other married likely to be enrolled in school than are children of MHHs but more likely than self-reported FHH widowed FHHs. FHH Self- Thirty-eight studies - FHHs are poorer than MHHs when poverty is measured by 65 studies reported Self- (total/per capita/per equivalent) household income and consumption expenditures, Buvinic and Gupta on 2 FHH, de reported access to services, and ownership of land and assets. Fifteen studies - certain types (1997) developing facto, de MHH of FHHs are more vulnerable to poverty than others. Eight studies – poverty in countries jure FHH FHHs is not higher than in MHH. Self- Self- FHHs are similar to MHHs when poverty is measured by per capita expenditure. reported reported The difference between FHHs versus non-FHHs does not change if using different FHH MHH poverty indicators. Self- Widows/divorced/separated FHHs have significantly higher headcount poverty in reported de Non-FHH indigenous areas when poverty is measured by per capita expenditure. FHHs have jure, de lower education than non-FHHs facto FHH Self- reported FHHs with unmarried partners have higher headcount poverty ratios in urban and 3 Fuwa (2000) Panama married indigenous areas when poverty is measured by per capita expenditure. The result is FHH, FHH Non-FHH robust to applying an equivalence scale using alternative poverty measures and with poverty lines. unmarried partners Potential FHHs are not poorer than non-FHHs when poverty is measured by per capita Non-FHH FHH expenditure. FHHs are similar to non-FHHs when poverty is measured by per capita expenditure. “Working” Non-FHH FHHs have higher education endowments than non-FHHs, except in indigenous FHHs areas. 41 FHHs are less poor than non-FHHs when poverty is measured by per capita core FHH Non-FHH expenditure. FHHs and individual females contribute disproportionately to overall poverty in 25- Self- 50% of the dataset when headcount poverty is measured by (total/per capita/per 10 Self- Quisumbing et al. reported equivalent) household income and consumption expenditures and are insensitive to 4 developing reported (2001) FHH, the poverty line. FHHs and individual females are similar to MHHs or males when countries MHH, males females using stochastic dominance criteria, but they are constantly worse off in Ghana and Bangladesh. Self- reported The income per capita/ adult equivalent is lower in widowed FHHs than in the widowed Self- MHHs. Horrell and Krishnan 5 Zimbabwe FHH reported (2007) Self- MHH The income per capita/ adult equivalent is higher in the de facto FHHs than in the reported de- MHHs. facto FHH Poverty is higher among FHHs, but there is no clear evidence of a recent and widespread feminization of poverty in Latin America. Differences in poverty FHH, among FHHs and MHHs increased in Argentina and Mexico, showing specific MHH, males females types of feminization of poverty. The results are robust to different values of 8 Latin Medeiros and Costa poverty lines, the use of equivalence scales, and the distribution of household 6 American (2007) income. countries Self- The insignificant increase in poverty indices when comparing FHHs without reported Couple HH children to couple-headed HH without children in Bolivia. The rise in poverty FHH w/o w/o children indices is significant at 5% when comparing FHHs with children to MHHs with children children in Costa Rica. Self- HHs where reported Latin The gender of the household head is a poor substitute for a gendered analysis of women have FHHs who 7 Deere et al. (2012) American asset ownership within and among households since an analysis based on headship ownership have countries tends to underestimate women’s ownership of assets. rights ownership rights Self- Self- reported reported widowed married Widowed FHHs have significantly lower consumption per capita than married FHH FHH FHHs, while MHHs do not have any significant differences in per capita Self- Self- consumption between widowed MHHs and married MHHs. reported reported 8 Van de Walle (2013) Mali widowed married MHH MHH Self- Self- Widowed FHH living in rural areas have lower per capita consumption than all reported reported other households living in rural areas. The gap between widowed FHHs and other widowed widowed HHs is lower for HHs residing in urban areas. FHH rural FHH urban 42 Self- Self- reported Per capita consumption of widowed FHHs is around 12% lower than that of all rural reported non- households. The results are robust to using an equivalence scale in measuring widowed widowed consumption. FHH rural FHH rural Self- Self- reported Per capita consumption of widowed FHHs is around 6% lower than that of all other reported non- urban households. The results are robust to using an equivalence scale in measuring widowed widowed consumption. FHH urban FHH urban Self- reported de jure FHH, Self- Poverty rates are higher in FHHs than in MHHs, irrespective of how headship is de facto reported defined. FHH, co- MHH resident South 9 Rogan (2013) FHH Africa Self- reported de co-resident jure FHH, Co-resident FHHs are less poor than other types of FHHs. FHH de facto FHH core FHH non-FHH FHH has the lowest risk of poverty Self- Self- No significant differences between FHHs and MHHs were found regarding reported reported consumption, the probability of shock exposure, or vulnerability to poverty in FHH MHH Thailand or Viet Nam. Self- Self- De jure FHHs have lower consumption than MHHs in Viet Nam. There are no reported de- reported significant differences between de-jure FHHs and MHHs regarding the probability jure FHH MHH of shock exposure or vulnerability to poverty in Thailand or Viet Nam. Thailand, Self- Self- De facto FHHs have higher consumption than MHHs in Thailand. There are no 10 Klasen et al. (2015) Viet Nam reported de- reported significant differences between de-facto FHHs and MHHs regarding the probability facto FHH MHH of shock exposure or vulnerability to poverty in Thailand or Viet Nam. Self- FHHs with an absent spouse have higher consumption levels than MHHs in reported Self- Thailand. Single FHH has a lower consumption level than Viet Nam. There are no single, reported significant differences between FHHs and MHHs regarding the probability of any widowed MHH shock exposure in Thailand or Viet Nam. Single FHHs are less vulnerable to FHH poverty in Thailand but more vulnerable to poverty in Viet Nam. Self- In eight of the 14 countries, FHHs are more likely to live in poor conditions. Self- reported However, MHHs are in more impoverished conditions than FHHs when married 14 Latin reported married w/o status, urban or rural setting, ownership, and the presence of children are controlled 11 Liu et al. (2017) American married spouse, in the regression. Generally, married FHHs with the spouse present are better off countries FHH with single, than any other category. The worst living conditions are associated with single, spouse separated, separated, divorced, or widowed FHHs. 43 widowed FHH Self- Self- While the share of FHHs in the population is growing during 1990-2012, poverty reported reported has been falling faster among FHHs. FHHs contributed more to the overall decline 20 FHH MHH in poverty despite their smaller overall population share. countries in Milazzo and van de Self- 12 Sub- Walle (2017) reported Self- Saharan The poverty trends of the various types of FHHs followed different paths across FHH w/o a reported Africa countries and periods, with no one type consistently outperforming the others. resident MHH adult male Self- Self- reported reported FHHs have a higher predicted poverty rate than MHHs in urban areas. The factors urban FHH urban MHH contributing to the poverty differential between FHH and MHH households are with with education, employment status, occupation, sector, and region of residence. Egypt, children children 13 Alazzawi (2018) Arab Rep. Self- Self- FHHs have a higher predicted poverty rate than MHHs in rural areas. Education, reported reported employment status, occupation, number of rooms per capita, and region of rural FHH rural MHH residence are factors that contribute to the poverty differential between FHHs and with with MHHs children children Adult couple households with children, children, and other adults (extended couple/singl family) are the most frequent among poor households. Poor and non‐poor women e females concentrate in the adult couple household with children. One adult female 71 w/o children household with children is more prevalent among the poor in Latin America, the Munoz Boudet et al. 14 developing other HH Caribbean, and Sub‐Saharan Africa. (2018) countries Male/female Poor women live in households with children and with children and earner earner with dependents, where the earner is a single male or a head couple. Single female- and w/o earner households comprise the largest percentage of poor households in Latin children America, the Caribbean, and Sub‐Saharan Africa. Self- FHHs have lower poverty rates than MHHs when using per capita welfare reported measures. FHHs are significantly worse than MHH when poverty is measured FHH Self- using consumption adjusted for economies of scale. Brown and Van de 43 African 15 Self- reported Walle (2021) countries reported MHH MHHs are poorer than married FHHs married FHH 44 Table A.2. Poverty Line, Arab Republic of Egypt by Region, in LCU, Per Capita Annual Consumption in Survey Year Prices Food poverty line Poverty line Region 2012/2013 2015 2017/2018 2019/2020 2012/2013 2015 2017/2018 2019/2020 Urban governorates 2748 4318 6065.3 7071 4320 6141 9280.1 11285 Urban lower Egypt 2484 3835 5667.6 6304 3840 5631 8536.9 9755 Rural lower Egypt 2568 3854 5901.7 6570 3852 5675 8673 10108 Urban upper Egypt 2568 3968 5752.1 6553 3972 5823 8728.5 10225 Rural upper Egypt 2496 3760 5896.5 6484 3756 5694 8865.6 10068 Urban frontier 2736 3990 5924.3 6696 3996 6247 8568.7 10409 Rural frontier 2688 3979 6304.7 7074 3984 5788 8979.3 10788 Total 2568 3921 5889.6 6604 3924 6141 8827 10279 Source: Compiled from various CAPMAS Poverty assessment updates. 45 Table A.3. Poverty Lines, Jordan, in LCU, Per Capita Annual Consumption in Survey Year Prices Food poverty line Poverty line 2010 2013 2010 2013 Jordan 336 383 814 929 Source: Jordan Department of Statistics: DOS https://jorinfo.dos.gov.jo/Databank/pxweb/en/Poverty/Poverty__Poverty-Indicators/ 46 Table A.4. Poverty Lines, Iraq, in LCU, Per Capita Annual Consumption in Survey Year Prices 2007 2012 Kurdistan 1212 1709 Baghdad 987 1391 Rest of Iraq 865 1220 Total 1073 1266 Source: World Bank “Poverty Estimates and Trends in Iraq” https://microdata.worldbank.org/index.php/catalog/2334/download/34771 47 Table A.5. Poverty Lines, the West Bank and Gaza, in LCU, Per Capita Annual Consumption in Survey Year Prices Food poverty line Poverty line 2009 2011 2017 2009 2011 2017 Gaza 567 570 567 712 714 710 West Bank 609 632 710 765 792 889 Total 603 620 668 757 776 836 Source: Compiled from various PCBS poverty reports. Spatial deflator provided by PCBS was used to calculate regional poverty lines for Gaza and the West Bank relative to the national poverty line available from PCBS publications. 48 Table A.6. Poverty Lines, Tunisia, in LCU, Per Capita Annual Consumption in Survey Year Prices Food poverty line Poverty line 2005 2010 2015 2021 2005 2010 2015 2021 Cities (metropolitan) 615 757 1085 1346.526 902 1038 1878 2682.997 Small & medium towns (urban) 596 733 1050.154 -- 818 941 1702.871 -- Noncommunal (rural) 466 571 951.668 1529.233 581 669 1500.530 2223.527 Source: World Bank (2016). “Tunisia Poverty Assessment 2015”. Table A1.3. 49 Table A.7. Poverty Lines, Mauritania, in LCU, Per Capita Annual Consumption in Survey Year Prices Extreme poverty line Poverty line 2004 2008 2004 2008 2014 2019 Total 70400 96000 94650 129000 169445 191000 Source: IMF (2011) Table 1.1. 50 Table A.8. Correlation between Main Types of Female-Headed Households FHH Type 1 FHH Type 2 FHH Type 3 FHH Type 4 Self-reported Majority-female-adult Potential Most-educated-female-adult FHH Type 1 1.000 0.265*** 0.415*** 0.298*** (0.000) (0.000) (0.000) FHH Type 2 1.000 0.319*** 0.304*** (0.000) (0.000) FHH Type 3 1.000 0.510*** (0.000) FHH Type 4 1.000 Note: ***, **, and * denote statistical significance at the 0.01, 0.05, and 0.10 levels. 51 Table A.9. Headcount Poverty Rate of Female and Non-female Headed Households by Headship Definition (percent) Egypt, Arab Rep. Mauritania (2004-2019) West Bank and Gaza (2012-2020) Iraq (2007-2012) Jordan (2010-2013) (2007-2017) Tunisia (2005-2021) FHH non-FHH FHH non-FHH FHH non-FHH FHH non-FHH FHH non-FHH FHH non-FHH Panel A: All Self-Reported FHH 23.32 32.01* 20.22 24.79* 14.46 14.94 29.52 38.03* 50.63 59.70* 16.04 17.34* Reported de jure FHH 21.51 31.92* 19.24 24.80* 16.08 14.80 29.66 36.80* 48.17 59.73* 13.46 17.57* Reported married FHH 29.15 31.01 25.96 24.36 7.76 15.03* 28.35 36.73* 62.71 59.10 26.55* 16.96 Potential FHH 17.41 32.41* 22.28 24.43 8.42 15.35* 31.43 36.80* 35.30 60.18* 13.46 17.69* Core FHH 19.53 31.25* 15.41 24.43* 4.16 15.11* 30.47 36.17* 34.38 59.47* 13.41 17.32* Asset FHH 16.33 32.05* 21.76 24.42 8.38 15.30* 32.88 36.37* 39.06 59.69* 15.78 17.28* Share of female adults>0.5 29.62 31.32* 25.94* 23.64 17.15* 14.19 37.36* 34.61 62.04* 58.33 19.10* 16.42 Share of employed 22.3 31.44* 22.11 24.48* 7.48 15.37* 33.11 36.18* 54.89 59.38* 16.68 17.24 females>employed males Most educated member is female 20.43 32.11* 29.12* 24.17 21.87 13.61 27.55 37.22* 64.52* 58.46 18.15* 17.01 adult & no employed males Panel B: Rural Self-Reported FHH 28.16 37.9* 36.29 35.34 13.20 17.70 32.42 44.98* 45.81 56.02* 23.95 25.50 Reported de jure FHH 25.89 37.69* 34.15 35.45 13.55 17.61 33.94 43.01* 42.56 56.07* 18.79 25.85* Reported married FHH 32.87 36.92* 43.82* 35.28 11.18 17.39 30.16 43.30* 61.12 55.38 33.52* 24.98 Potential FHH 22.05 38.24* 38.52 35.33 15.50 17.41 33.14 44.18* 37.66 56.25* 22.28 25.77* Core FHH 25.67 37.00* 28.11 35.42 1.93 17.50* 34.21 42.46* 37.72 55.73* 22.96 25.41 Asset FHH 20.52 38.07* 38.85 35.34 12.51 17.54 33.88 43.83* 40.33 55.89* 24.28 25.42 Share of female adults>0.5 36.19 36.9 38.08* 33.89 21.37* 16.00 42.93* 40.94 60.23* 54.05 27.51* 24.34 Share of employed 27.46 37.17* 34.11 35.46 13.03 17.51 39.58 42.13* 59.83 55.11 25.30 25.32 females>employed males Most educated member is female 24.25 37.90* 51.45* 34.76 23.55* 16.19 33.68 43.27* 60.50 54.97 28.20* 24.82 adult & no employed males Panel C: Urban Self-Reported FHH 16.96 23.08* 15.19 18.52* 14.70 14.36 26.68 30.24* 50.17 58.89* 12.29 13.27* Reported de jure FHH 16.96 22.97* 15.08 18.48* 16.61 14.21 25.75 29.90* 47.92 58.92* 11.59 13.33* Reported married FHH 16.81 22.39* 15.97 18.20 7.29 14.53* 26.42 29.62* 61.23 58.33 17.25* 13.09 Potential FHH 11.5 23.56* 14.49 18.25* 7.33 14.91* 29.05 29.28 32.54 59.48* 8.69 13.75* Core FHH 13.47 22.58* 9.54 18.21* 4.43 14.60* 25.24 29.51* 31.97 58.70* 8.68 13.32* Asset FHH 9.11 23.18* 12.56 18.25* 7.72 14.82* 31.32* 28.90 36.09 58.94* 9.64 13.37* Share of female adults>0.5 21.27 22.61 17.92 18.28 16.24* 13.81 31.12* 27.94 60.55* 57.77 14.40* 12.68 Share of employed 17.58 22.65* 11.31 18.40* 6.65 14.91* 26.20 29.73* 52.13 58.69* 12.50 13.23 females>employed males Most educated member is female 16.40 23.10* 18.57 18.15 21.52 13.06 20.58 30.74* 63.05* 57.76 13.38 13.11 adult & no employed males Panel D: Have children under 14 Self-Reported FHH 35.2 39.33* 22.15 25.99* 29.00* 18.18 31.91 40.75* 67.33 64.40 28.70* 22.48 Reported de jure FHH 37.62 39.05 21.27 26.01* 34.42* 18.11 33.39 39.21* 65.85 64.47 25.84* 22.81 Reported married FHH 31.51 39.24* 26.85 25.66 13.71 18.82 29.33 39.71* 72.35 64.45 33.15* 22.64 Potential FHH 32.83 39.34* 27.84 25.64 25.95* 18.55 33.94 39.50* 59.40 64.60 29.98* 22.55 Core FHH 30.60 39.13* 23.68 25.68 21.68 18.73 32.23 38.93* 48.35 64.63* 25.61 22.89 Asset FHH 32.24 39.24* 28.13 25.65 27.28* 18.55 35.51 39.04* 64.80 64.51 33.36* 22.60 Share of female adults>0.5 40.1 38.72 27.32 24.90 26.82* 16.80 39.83* 37.41 70.08* 63.15 27.90* 21.48 Share of employed 35.69 39.1* 25.57 25.68 17.43 18.79 35.55 38.89* 67.74 64.38 26.60* 22.68 females>employed males Most educated member is female 32.09 39.48 * 33.62 * 25.37 34.47 * 16.46 30.61 39.74 * 77.00 * 63.23 29.63 * 22.22 adult & no employed males Panel E: No children under 14 Self-Reported FHH 10.53 10.52 6.79 6.15 3.88 4.13 12.21 13.63 30.83 31.39 8.71 9.28 52 Reported de jure FHH 10.39 10.55 6.26 6.26 4.54 3.98 12.76 13.34 30.16 31.50 8.69 9.28 Reported married FHH 12.7 10.49 13.20* 6.17 0.00 4.20 9.25 13.42* 36.90 31.19 8.97 9.18 Potential FHH 6.24 11.58* 5.05 6.36 0.81 4.77* 11.70 13.58* 23.57 32.92* 7.74 9.57* Core FHH 7.59 10.65* 1.33 6.39* 0.00 4.33* 11.32 13.31 23.75 31.66* 7.38 9.29* Asset FHH 6.14 11.29* 4.76 6.36 0.68 4.72* 12.19 13.42 25.22 32.03* 8.97 9.21 Share of female adults>0.5 12.16* 9.89 8.69* 4.83 2.74 4.86* 15.30* 11.80 33.93* 29.72 11.40* 7.91 Share of employed 8.79 10.7* 3.09 6.57* 0.47 4.64* 10.29 13.61* 31.32 31.29 10.27* 8.99 females>employed males Most educated member is female 8.74 10.90 * 9.36 * 5.84 3.02 4.41 10.04 14.05 * 35.36 * 30.23 11.40 8.51 adult & no employed males Note: The data are pooled across all available years for each country. Headcount poverty rates are estimated using per capita household expenditures. Stars indicate statistically significant difference in headcount poverty between FHHs and non-FHHs in each category at the 5% or lower level. Population sampling weights are applied. 53 Table A.10. Probabilities of Being Poor, Linear Probability Model (Main FHH Types) Specification 1 Specification 2 FHH Type 1 FHH Type 2 FHH Type 3 FHH Type 4 FHH Type 1 FHH Type 2 FHH Type 3 FHH Type 4 Self-reported Majority-female-adult Potential Most-educated-female-adult Self-reported Majority-female-adult Potential Most-educated-female-adult -0.007*** -0.044*** Self-reported FHH (0.00) (0.00) Self-reported FHH # Number of children age 0.008*** 0.006*** 0-14 (0.00) (0.00) 0.052*** 0.032*** Share of female adults>0.5 (0.00) (0.00) Share of female adults>0.5# Number of -0.007*** -0.014*** children age 0-14 (0.00) (0.00) -0.027*** -0.017*** Potential FHH (0.00) (0.00) 0.014*** 0.011*** Potential FHH# Number of children age 0-14 (0.00) (0.00) 0.002 -0.006** Educated females (0.00) (0.00) Educated females# Number of children age 0- 0.013*** 0.013*** 14 (0.00) (0.00) Household head`s characteristics -0.001*** -0.001*** -0.001*** -0.001*** Head`s age (0.00) (0.00) (0.00) (0.00) -0.066*** -0.064*** -0.064*** -0.065*** Highest education level is primary (0.00) (0.00) (0.00) (0.00) -0.123*** -0.121*** -0.121*** -0.121*** Highest education level is secondary (0.00) (0.00) (0.00) (0.00) -0.178*** -0.176*** -0.176*** -0.176*** Highest education level is tertiary (0.00) (0.00) (0.00) (0.00) -0.020*** 0.007*** 0.005** 0.006*** Head is married (0.00) (0.00) (0.00) (0.00) -0.037*** -0.032*** -0.031*** -0.027*** Head is employed (0.00) (0.00) (0.00) (0.00) Household characteristics 0.022*** 0.024*** 0.023*** 0.023*** Household size (0.00) (0.00) (0.00) (0.00) 0.077*** 0.081*** 0.076*** 0.077*** 0.050*** 0.054*** 0.049*** 0.049*** Number of children age 0-14 (0.00) (0.00) (0.00) (0.00) (0.00) (0.00) (0.00) (0.00) -0.008*** -0.008*** -0.008*** -0.009*** Share of household members age 15-24 (0.00) (0.00) (0.00) (0.00) -0.022*** -0.019*** -0.017*** -0.020*** Share of household members age 60 and older (0.00) (0.00) (0.00) (0.00) -0.084*** -0.084*** -0.084*** -0.084*** Urban (0.00) (0.00) (0.00) (0.00) -0.149*** -0.152*** -0.149*** -0.148*** -0.181*** -0.182*** -0.179*** -0.180*** Iraq (0.00) (0.00) (0.00) (0.00) (0.00) (0.00) (0.00) (0.00) -0.144*** -0.147*** -0.144*** -0.146*** -0.092*** -0.094*** -0.090*** -0.092*** Jordan (0.01) (0.01) (0.01) (0.01) (0.01) (0.01) (0.01) (0.01) -0.155*** -0.160*** -0.156*** -0.158*** -0.207*** -0.208*** -0.213*** -0.212*** Mauritania (0.01) (0.01) (0.01) (0.01) (0.01) (0.01) (0.01) (0.01) 0.181*** 0.178*** 0.181*** 0.180*** 0.185*** 0.183*** 0.186*** 0.185*** West Bank and Gaza (0.01) (0.01) (0.01) (0.01) (0.01) (0.01) (0.01) (0.01) -0.060*** -0.063*** -0.060*** -0.061*** -0.054*** -0.055*** -0.054*** -0.054*** Tunisia (0.00) (0.00) (0.00) (0.00) (0.00) (0.00) (0.00) (0.00) _cons 0.127*** 0.112*** 0.133*** 0.126*** 0.312*** 0.261*** 0.270*** 0.267*** (0.00) (0.00) (0.00) (0.00) (0.01) (0.01) (0.01) (0.01) r2_a 0.16 0.17 0.16 0.16 0.21 0.21 0.21 0.21 N 214931 214931 214931 214931 211069 211069 211069 211069 Note: ***, **, and * denote statistical significance at the 0.01, 0.05, and 0.10 levels. Robust standard errors are in parentheses. All regressions control for survey rounds fixed effects. The reference groups for head`s education without formal education. The reference group for share of household members is share of members aged between 25 and 59 years. The reference country is the Arab Republic of Egypt. 54 Table A.11. Probabilities of Being Poor for Other FHH Types, Linear Probability Model Specification 1 Specification 2 (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) -0.003 -0.027*** De-jure FHH (0.00) (0.00) 0.013*** 0.009*** De-jure FHH # Number of children age 0-14 (0.00) (0.00) -0.031*** -0.058*** Married FHH (0.01) (0.01) 0.005** 0.006** Married FHH # Number of children age 0-14 (0.00) (0.00) 0.009** 0.008** Employed FHH (0.00) (0.00) -0.001 -0.004* Employed FHH # Number of children age 0-14 (0.00) (0.00) -0.019*** -0.018*** Asset FHH (0.00) (0.00) 0.014*** 0.011*** Asset FHH # Number of children age 0-14 (0.00) (0.00) -0.023*** -0.010** Core FHH (0.00) (0.00) 0.006** 0.007*** Core FHH # Number of children age 0-14 (0.00) (0.00) Household head`s characteristics -0.001*** -0.001*** -0.001*** -0.001*** -0.001*** Head`s age (0.00) (0.00) (0.00) (0.00) (0.00) -0.065*** -0.065*** -0.064*** -0.064*** -0.064*** Highest education level is primary (0.00) (0.00) (0.00) (0.00) (0.00) -0.122*** -0.122*** -0.121*** -0.121*** -0.121*** Highest education level is secondary (0.00) (0.00) (0.00) (0.00) (0.00) -0.177*** -0.177*** -0.176*** -0.177*** -0.176*** Highest education level is tertiary (0.00) (0.00) (0.00) (0.00) (0.00) -0.008** 0.011*** 0.006*** 0.006** 0.006*** Head is married (0.00) (0.00) (0.00) (0.00) (0.00) -0.033*** -0.038*** -0.032*** -0.031*** -0.032*** Head is employed (0.00) (0.00) (0.00) (0.00) (0.00) Household characteristics 0.023*** 0.023*** 0.023*** 0.023*** 0.023*** Household size (0.00) (0.00) (0.00) (0.00) (0.00) 0.077*** 0.078*** 0.078*** 0.077*** 0.078*** 0.050*** 0.050*** 0.050*** 0.049*** 0.050*** Number of children age 0-14 (0.00) (0.00) (0.00) (0.00) (0.00) (0.00) (0.00) (0.00) (0.00) (0.00) -0.009*** -0.008*** -0.009*** -0.009*** -0.009*** Share of household members age 15-24 (0.00) (0.00) (0.00) (0.00) (0.00) -0.021*** -0.021*** -0.021*** -0.018*** -0.021*** Share of household members age 60 and older (0.00) (0.00) (0.00) (0.00) (0.00) -0.084*** -0.084*** -0.084*** -0.084*** -0.084*** Urban (0.00) (0.00) (0.00) (0.00) (0.00) -0.149*** -0.150*** -0.149*** -0.149*** -0.149*** -0.181*** -0.182*** -0.181*** -0.180*** -0.181*** Iraq (0.00) (0.00) (0.00) (0.00) (0.00) (0.00) (0.00) (0.00) (0.00) (0.00) -0.145*** -0.145*** -0.145*** -0.144*** -0.145*** -0.091*** -0.093*** -0.093*** -0.090*** -0.092*** Jordan (0.01) (0.01) (0.01) (0.01) (0.01) (0.01) (0.01) (0.01) (0.01) (0.01) 0.122*** 0.123*** 0.123*** 0.125*** 0.124*** 0.058*** 0.059*** 0.059*** 0.061*** 0.059*** Mauritania (0.01) (0.01) (0.01) (0.01) (0.01) (0.01) (0.01) (0.01) (0.01) (0.01) 0.181*** 0.180*** 0.180*** 0.180*** 0.180*** 0.185*** 0.184*** 0.184*** 0.184*** 0.184*** West Bank and Gaza (0.01) (0.01) (0.01) (0.01) (0.01) (0.01) (0.01) (0.01) (0.01) (0.01) -0.060*** -0.061*** -0.061*** -0.060*** -0.060*** -0.054*** -0.054*** -0.054*** -0.054*** -0.054*** Tunisia (0.00) (0.00) (0.00) (0.00) (0.00) (0.00) (0.00) (0.00) (0.00) (0.00) _cons 0.126*** 0.126*** 0.125*** 0.126*** 0.127*** 0.289*** 0.284*** 0.272*** 0.268*** 0.273*** (0.00) (0.00) (0.00) (0.00) (0.00) (0.01) (0.01) (0.01) (0.01) (0.01) Adjuster R2 0.16 0.16 0.16 0.16 0.16 0.21 0.21 0.21 0.21 0.21 Number of observations 214931 214931 214931 214931 214931 211069 211069 211069 211069 211069 Note: ***, **, and * denote statistical significance at the 0.01, 0.05, and 0.10 levels. Robust standard errors are in parentheses. All regressions control for survey rounds fixed effects. The reference groups for head`s education without formal education. The reference group for share of household members is share of members aged between 25 and 59 years. The reference country is the Arab Republic of Egypt. 55 Table A.12. Probabilities of Being Poor, Logit Model Specification 1 Specification 2 FHH Type 1 FHH Type 2 FHH Type 3 FHH Type 4 FHH Type 1 FHH Type 2 FHH Type 3 FHH Type 4 Self-reported Majority-female-adult Potential Most-educated-female-adult Self-reported Majority-female-adult Potential Most-educated-female-adult -0.110*** -0.290*** Self-reported FHH (0.02) (0.03) Self-reported FHH # Number of children age 0.057*** 0.036*** 0-14 (0.01) (0.01) 0.415*** 0.323*** Share of female adults>0.5 (0.02) (0.02) Share of female adults>0.5# Number of -0.072*** -0.116*** children age 0-14 (0.01) (0.01) -0.339*** -0.165*** Potential FHH (0.02) (0.03) 0.124*** 0.075*** Potential FHH# Number of children age 0-14 (0.01) (0.01) -0.037 -0.040 Educated females (0.02) (0.03) Educated females# Number of children age 0- 0.086*** 0.074*** 14 (0.01) (0.01) Household head`s characteristics -0.009*** -0.009*** -0.007*** -0.008*** Head`s age (0.00) (0.00) (0.00) (0.00) -0.451*** -0.443*** -0.443*** -0.448*** Highest education level is primary (0.02) (0.02) (0.02) (0.02) -0.911*** -0.903*** -0.900*** -0.903*** Highest education level is secondary (0.02) (0.02) (0.02) (0.02) -1.572*** -1.568*** -1.563*** -1.562*** Highest education level is tertiary (0.03) (0.03) (0.03) (0.03) -0.097*** 0.083*** 0.059*** 0.068*** Head is married (0.03) (0.02) (0.02) (0.02) -0.256*** -0.213*** -0.212*** -0.178*** Head is employed (0.02) (0.02) (0.02) (0.02) Household characteristics 0.160*** 0.167*** 0.160*** 0.165*** Household size (0.00) (0.00) (0.00) (0.00) 0.468*** 0.508*** 0.456*** 0.470*** 0.301*** 0.338*** 0.298*** 0.293*** Number of children age 0-14 (0.00) (0.00) (0.00) (0.00) (0.01) (0.01) (0.01) (0.01) -0.092*** -0.095*** -0.095*** -0.100*** Share of household members age 15-24 (0.01) (0.01) (0.01) (0.01) -0.325*** -0.316*** -0.286*** -0.315*** Share of household members age 60 and older (0.02) (0.02) (0.02) (0.02) -0.614*** -0.610*** -0.611*** -0.611*** Urban (0.01) (0.01) (0.01) (0.01) -1.072*** -1.106*** -1.065*** -1.070*** -1.441*** -1.457*** -1.422*** -1.429*** Iraq (0.03) (0.03) (0.03) (0.03) (0.03) (0.03) (0.03) (0.03) -1.184*** -1.203*** -1.180*** -1.204*** -0.886*** -0.903*** -0.863*** -0.880*** Jordan (0.07) (0.07) (0.07) (0.07) (0.08) (0.08) (0.08) (0.08) -1.117*** -1.152*** -1.125*** -1.137*** -1.622*** -1.628*** -1.665*** -1.664*** Mauritania (0.05) (0.05) (0.05) (0.05) (0.05) (0.05) (0.05) (0.05) 0.796*** 0.784*** 0.797*** 0.795*** 0.887*** 0.880*** 0.896*** 0.892*** West Bank and Gaza (0.04) (0.04) (0.04) (0.04) (0.04) (0.04) (0.04) (0.04) -0.468*** -0.481*** -0.466*** -0.473*** -0.438*** -0.445*** -0.433*** -0.434*** Tunisia (0.03) (0.03) (0.03) (0.03) (0.03) (0.03) (0.03) (0.03) _cons -1.917*** -2.062*** -1.861*** -1.932*** -0.606*** -0.975*** -0.877*** -0.915*** (0.04) (0.04) (0.04) (0.04) (0.06) (0.06) (0.06) (0.06) N 214931 214931 214931 214931 211069 211069 211069 211069 Note: ***, **, and * denote statistical significance at the 0.01, 0.05, and 0.10 levels. Robust standard errors are in parentheses. All regressions control for survey rounds fixed effects. The reference groups for head`s education without formal education. The reference group for share of household members is share of members aged between 25 and 59 years. The reference country is the Arab Republic of Egypt. 56 Table A.13. Probabilities of Being Poor for Other FHH Types, Logit Model Specification 1 Specification 2 (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) -0.098*** -0.098** De-jure FHH (0.02) (0.05) 0.094*** 0.059*** De-jure FHH # Number of children age 0-14 (0.01) (0.01) -0.182*** -0.392*** Married FHH (0.05) (0.06) 0.024 0.029* Married FHH # Number of children age 0-14 (0.02) (0.02) 0.036 -0.135*** Employed FHH (0.03) (0.05) -0.002 0.058*** Employed FHH # Number of children age 0-14 (0.01) (0.02) -0.235*** -0.103*** Asset FHH (0.03) (0.03) 0.106*** 0.056*** Asset FHH # Number of children age 0-14 (0.01) (0.01) -0.259*** 0.028 Core FHH (0.04) (0.03) 0.060*** -0.016 Core FHH # Number of children age 0-14 (0.02) (0.01) Household head`s characteristics -0.008*** -0.009*** -0.008*** -0.008*** -0.008*** Head`s age (0.00) (0.00) (0.00) (0.00) (0.00) -0.443*** -0.449*** -0.442*** -0.442*** -0.442*** Highest education level is primary (0.02) (0.02) (0.02) (0.02) (0.02) -0.902*** -0.909*** -0.900*** -0.900*** -0.901*** Highest education level is secondary (0.02) (0.02) (0.02) (0.02) (0.02) -1.563*** -1.570*** -1.562*** -1.562*** -1.564*** Highest education level is tertiary (0.03) (0.03) (0.03) (0.03) (0.03) 0.061 0.104*** 0.065*** 0.065*** 0.066*** Head is married (0.04) (0.02) (0.02) (0.02) (0.02) -0.216*** -0.267*** -0.216*** -0.210*** -0.216*** Head is employed (0.02) (0.02) (0.02) (0.02) (0.02) Household characteristics 0.162*** 0.161*** 0.163*** 0.162*** 0.163*** Household size (0.00) (0.00) (0.00) (0.00) (0.00) 0.470*** 0.477*** 0.478*** 0.465*** 0.474*** 0.298*** 0.303*** 0.300*** 0.298*** 0.303*** Number of children age 0-14 (0.00) (0.00) (0.00) (0.00) (0.00) (0.01) (0.01) (0.01) (0.01) (0.01) -0.096*** -0.092*** -0.100*** -0.099*** -0.100*** Share of household members age 15-24 (0.01) (0.01) (0.01) (0.01) (0.01) -0.316*** -0.326*** -0.323*** -0.304*** -0.324*** Share of household members age 60 and older (0.02) (0.02) (0.02) (0.02) (0.02) -0.613*** -0.616*** -0.611*** -0.610*** -0.612*** Urban (0.01) (0.01) (0.01) (0.01) (0.01) -1.077*** -1.081*** -1.079*** -1.071*** -1.076*** -1.434*** -1.446*** -1.436*** -1.432*** -1.441*** Iraq (0.03) (0.03) (0.03) (0.03) (0.03) (0.03) (0.03) (0.03) (0.03) (0.03) -1.188*** -1.194*** -1.191*** -1.183*** -1.189*** -0.876*** -0.895*** -0.881*** -0.872*** -0.885*** Jordan (0.07) (0.07) (0.07) (0.07) (0.07) (0.08) (0.08) (0.08) (0.08) (0.08) -1.115*** -1.080*** -1.097*** -1.098*** -1.091*** -1.649*** -1.597*** -1.647*** -1.649*** -1.636*** Mauritania (0.05) (0.05) (0.05) (0.05) (0.05) (0.05) (0.05) (0.05) (0.05) (0.05) 0.795*** 0.794*** 0.794*** 0.786*** 0.795*** 0.889*** 0.883*** 0.889*** 0.887*** 0.887*** West Bank and Gaza (0.04) (0.04) (0.04) (0.04) (0.04) (0.04) (0.04) (0.04) (0.04) (0.04) -0.468*** -0.469*** -0.469*** -0.465*** -0.467*** -0.434*** -0.436*** -0.434*** -0.433*** -0.435*** Tunisia (0.03) (0.03) (0.03) (0.03) (0.03) (0.03) (0.03) (0.03) (0.03) (0.03) _cons -1.923*** -1.935*** -1.943*** -1.926*** -1.927*** -0.849*** -0.781*** -0.866*** -0.888*** -0.871*** (0.04) (0.04) (0.04) (0.04) (0.04) (0.07) (0.06) (0.06) (0.06) (0.06) N 214931 214931 214931 214931 214931 211069 211069 211069 211069 211069 Note: ***, **, and * denote statistical significance at the 0.01, 0.05, and 0.10 levels. Robust standard errors are in parentheses. All regressions control for survey rounds fixed effects. The reference groups for head`s education without formal education. The reference group for share of household members is share of members aged between 25 and 59 years. The reference country is the Arab Republic of Egypt. 57 Table A.14. Probabilities of Being Poor for Combination of Main Female-Headed Household Types, Linear Probability Model Specification 1 Specification 2 (1) (2) (3) (4) -0.037*** -0.033*** Overlap of self-reported FHHs, potential FHHs, and most-educated-female-adult FHHs (0.00) (0.00) 0.005** 0.007*** Overlap of self-reported FHHs, potential FHHs, and most-educated-female-adult FHHs # Number of children age 0-14 (0.00) (0.00) -0.037*** -0.033*** Overlap of self-reported FHHs, potential FHHs, majority of females and most-educated-female-adult FHHs (0.00) (0.00) 0.005** 0.007*** Overlap of self-reported FHHs, potential FHHs, majority of females and most-educated-female-adult FHHs # Number of children age 0-14 (0.00) (0.00) Household head`s characteristics -0.001*** -0.001*** Head`s age (0.00) (0.00) -0.064*** -0.064*** Highest education level is primary (0.00) (0.00) -0.121*** -0.121*** Highest education level is secondary (0.00) (0.00) -0.176*** -0.176*** Highest education level is tertiary (0.00) (0.00) 0.001 0.001 Head is married (0.00) (0.00) -0.033*** -0.033*** Head is employed (0.00) (0.00) Household characteristics 0.023*** 0.023*** Household size (0.00) (0.00) 0.078*** 0.078*** 0.050*** 0.050*** Number of children age 0-14 (0.00) (0.00) (0.00) (0.00) -0.009*** -0.009*** Share of household members age 15-24 (0.00) (0.00) -0.021*** -0.021*** Share of household members age 60 and older (0.00) (0.00) -0.084*** -0.084*** Urban (0.00) (0.00) -0.150*** -0.150*** -0.181*** -0.181*** Iraq (0.00) (0.00) (0.00) (0.00) -0.145*** -0.145*** -0.093*** -0.093*** Jordan (0.01) (0.01) (0.01) (0.01) -0.150*** -0.150*** -0.209*** -0.209*** Mauritania (0.01) (0.01) (0.01) (0.01) 0.180*** 0.180*** 0.184*** 0.184*** West Bank and Gaza (0.01) (0.01) (0.01) (0.01) -0.061*** -0.061*** -0.054*** -0.054*** Tunisia (0.00) (0.00) (0.00) (0.00) _cons 0.129*** 0.129*** 0.283*** 0.283*** (0.00) (0.00) (0.01) (0.01) Adjuster R2 0.16 0.16 0.21 0.21 Number of observations 214931 214931 211069 211069 Note: ***, **, and * denote statistical significance at the 0.01, 0.05, and 0.10 levels. Robust standard errors are in parentheses. 58 Table A.15. Average change for mobility per year between the survey rounds Upward mobility (%) Downward mobility (%) Jordan (2010-2013) 17.8 3.5 Egypt, Arab Rep. (2017-2020) 9.7 5.3 Iraq (2007-2012) 9.0 2.7 Mauritania (2014-2019) 8.1 2.8 West Bank and Gaza (2011-2017) 5.1 8.1 Tunisia (2015-2021) 3.5 1.3 Note: Countries are ranked in a decreasing order of upward mobility. 59 Figure A.1. Headcount Poverty Rates in Self-Reported Male-Headed Households (%), by Number of Female Adults Egypt Iraq 2012 2007 2015 2017 2012 2020 0 50 100 0 10 20 30 40 Jordan Mauritania 2004 2007 2008 2014 2012 2019 0 10 20 30 40 0 20 40 60 80 Palestine Tunisia 2007 2005 2009 2010 2011 2015 2017 2021 0 20 40 60 80 100 0 20 40 60 0 1 2 3 4 >4 Note: Headcount poverty rates are estimated using per capita household expenditures. The numbers of female adults are shown for 0, 1, 2, 3, 4, and more than 4 adults. The years are shown on the y-axis and the poverty rates are shown on the x-axis. 60 Figure A.2. FHH–non-FHH Differences in Headcount Poverty Rate in MENA, Pooled Cross Sections (percentage points) Egypt Iraq *** 0 5 *** difference (pps) difference (pps) *** -5 0 ** -10 -5 *** *** *** *** *** -15 *** *** -10 *** *** Self-reported Married Core *** Females>0.5 Educated Self-reported Married Core Females>0.5 Educated De jure Potential Asset Employed De jure Potential Asset Employed Jordan Mauritania 5 10 5 *** *** difference (pps) difference (pps) *** 0 0 *** *** -5 -5 *** *** ** *** *** *** *** -10 *** *** -10 *** *** Self-reported Married Core Females>0.5 Educated Self-reported Married Core Females>0.5 Educated De jure Potential Asset Employed De jure Potential Asset Employed Palestine Tunisia *** 10 10 *** *** difference (pps) difference (pps) -30 -20 -10 0 5 ** *** *** *** *** 0 *** *** ** *** *** *** *** *** -5 Self-reported Married Core Females>0.5 Educated Self-reported Married Core Females>0.5 Educated De jure Potential Asset Employed De jure Potential Asset Employed Note: Authors’ calculation based on pooled cross sections. The four main types of FHHs are shown in darker color, the five sub-types of FHHs are shown in lighter color. The headcount poverty rate is applied to per capita household expenditures, of FHHs versus the rest of the households. Stars indicate statistically significantly higher headcount poverty ratio between FHHs and non-FHHs in each category. ***, **, and * denote statistical significance at the 0.01, 0.05, and 0.10 levels respectively. 61 Figure A.3. Correlation Between Probabilities of Female-Headed Households Escaping Poverty in Second Year and Number of Children (percentage) Egypt 2017-2020 Iraq 2007-2012 .28 .3 .32 .34 .36 .38 .4 .42 .44 .46 .48 .5 Percentage (%) Percentage (%) 0 1 2 3 4 0 1 2 3 4 Number of children Number of children Jordan 2010-2013 Mauritania 2014-2019 .55 .4 .42 .44 .46 .48 Percentage (%) Percentage (%) .5 .45 .4 0 1 2 3 4 0 1 2 3 4 Number of children Number of children Palestine 2011-2017 Tunisia 2015-2021 .45 .25 Percentage (%) Percentage (%) .4 .2 .35 .15 .3 .1 0 1 2 3 4 0 1 2 3 4 Number of children Number of children Self-reported FHH Females>0.5 Potential FHH Educated females Note: Estimates are obtained with synthetic panel data and weighted with population weights where the second survey round is used as the base year. Household heads' ages are restricted to 25-55 for the first survey round and adjusted accordingly for the second survey round. 62 Figure A.4. Probability of Other FHH Types Escaping Poverty in Second Year Conditional on Being Poor in First Year (percentage) Egypt 2017-2020 Iraq 2007-2012 25 30 35 40 45 70 Percentage (%) Percentage (%) 60 50 40 De-jure Employed fem. Core De-jure Employed fem. Core Married Assets Married Assets Jordan 2010-2013 Mauritania 2014-2019 70 40 42 44 46 48 50 Percentage (%) Percentage (%) 60 50 40 De-jure Employed fem. Core De-jure Employed fem. Core Married Assets Married Assets Palestine 2011-2017 Tunisia 2015-2021 50 25 Percentage (%) Percentage (%) 40 20 30 15 20 10 De-jure Employed fem. De-jure Employed fem. Married Core Married Core FHH wo/children FHH w/children Non-FHH Note: Estimates are obtained with synthetic panel data and weighted with population weights where the second survey round is used as the base year. The figure shows the percentage of the population that moves out of poverty in the second year. FHH characteristics are measured in second period. Dashed red lines represent the national average for each period. Household heads' ages are restricted to 25-55 for the first survey round and adjusted accordingly for the second survey round. Standard errors are estimated with 1,000 bootstraps. 63 Figure A.5. Probabilities of Female-Headed Households Escaping Poverty in Second Year Conditional on Being Poor in First Year (percentage), Arab Republic of Egypt Egypt 2012-2015 Egypt 2015-2017 Egypt 2017-2020 40 40 40 30 30 30 Percentage (%) Percentage (%) Percentage (%) 20 20 20 10 10 10 .5 l .5 l .5 l . . . d d d ia ia ia m m m rte rte rte >0 >0 >0 nt nt nt fe fe fe po po po te te te es es es ed ed ed Po Po Po re re re al al al at at at lf- lf- lf- m m m uc uc uc Se Se Se Fe Fe Fe Ed Ed Ed FHH wo/children FHH w/children Non-FHH Note: Estimates are obtained with synthetic panel data and weighted with population weights where the second survey round is used as the base year. The figure shows the percentage of the population that moves out of poverty in the second year. FHH characteristics are measured in second period. Dashed red lines represent the national average for each period. Household heads' ages are restricted to 25-55 for the first survey round and adjusted accordingly for the second survey round. Standard errors are estimated with 1,000 bootstraps. 64 Figure A.6. Probabilities of Female-Headed Households Escaping Poverty in Second Year Conditional on Being Poor in First Year (percentage), Mauritania Mauritania 2004-2008 Mauritania 2008-2014 Mauritania 2014-2019 50 50 50 45 45 45 Percentage (%) Percentage (%) Percentage (%) 40 40 40 35 35 35 30 30 30 .5 l .5 l .5 l . . . d d d ia ia ia m m m rte rte rte >0 >0 >0 nt nt nt fe fe fe po po po te te te es es es ed ed ed Po Po Po re re re al al al at at at lf- lf- lf- m m m uc uc uc Se Se Se Fe Fe Fe Ed Ed Ed FHH wo/children FHH w/children Non-FHH Note: Estimates are obtained with synthetic panel data and weighted with population weights where the second survey round is used as the base year. The figure shows the percentage of the population that moves out of poverty in the second year. FHH characteristics are measured in second period. Dashed red lines represent the national average for each period. Household heads' ages are restricted to 25-55 for the first survey round and adjusted accordingly for the second survey round. Standard errors are estimated with 1,000 bootstraps. 65 Figure A.7. Probabilities of Female-Headed Households Escaping Poverty in Second Year Conditional on Being Poor in First Year (percentage), West Bank and Gaza Palestine 2007-2009 Palestine 2009-2011 Palestine 2011-2017 45 45 45 40 40 40 Percentage (%) Percentage (%) Percentage (%) 35 35 35 30 30 30 25 25 25 .5 l .5 l .5 l . . . d d d ia ia ia m m m rte rte rte >0 >0 >0 nt nt nt fe fe fe po po po te te te es es es ed ed ed Po Po Po re re re al al al at at at lf- lf- lf- m m m uc uc uc Se Se Se Fe Fe Fe Ed Ed Ed FHH wo/children FHH w/children Non-FHH Note: Estimates are obtained with synthetic panel data and weighted with population weights where the second survey round is used as the base year. The figure shows the percentage of the population that moves out of poverty in the second year. FHH characteristics are measured in second period. Dashed red lines represent the national average for each period. Household heads' ages are restricted to 25-55 for the first survey round and adjusted accordingly for the second survey round. Standard errors are estimated with 1,000 bootstraps. 66 Figure A.8. Probabilities of Female-Headed Households Escaping Poverty in Second Year Conditional on Being Poor in First Year (percentage), Tunisia Tunisia 2005-2010 Tunisia 2010-2015 Tunisia 2015-2021 60 60 60 50 50 50 Percentage (%) Percentage (%) Percentage (%) 40 40 40 30 30 30 20 20 20 10 10 10 .5 l .5 l .5 l . . . d d d ia ia ia m m m rte rte rte >0 >0 >0 nt nt nt fe fe fe po po po te te te es es es ed ed ed Po Po Po re re re al al al at at at lf- lf- lf- m m m uc uc uc Se Se Se Fe Fe Fe Ed Ed Ed FHH wo/children FHH w/children Non-FHH Note: Estimates are obtained with synthetic panel data and weighted with population weights where the second survey round is used as the base year. The figure shows the percentage of the population that moves out of poverty in the second year. FHH characteristics are measured in second period. Dashed red lines represent the national average for each period. Household heads' ages are restricted to 25-55 for the first survey round and adjusted accordingly for the second survey round. Standard errors are estimated with 1,000 bootstraps. 67 Figure A.9. Probabilities of Female-Headed Households Escaping Poverty in Second Year Conditional on Being Poor in First Year (percentage) Egypt 2017-2020 Iraq 2007-2012 34 50 Percentage (%) Percentage (%) 32 48 30 46 28 44 Self-reported Potential Self-reported Potential Females>0.5 Educated fem. Females>0.5 Educated fem. Jordan 2010-2013 Mauritania 2014-2019 38 40 42 44 46 46 48 50 52 54 56 Percentage (%) Percentage (%) Self-reported Potential Self-reported Potential Females>0.5 Educated fem. Females>0.5 Educated fem. Palestine 2011-2017 Tunisia 2015-2021 45 Percentage (%) Percentage (%) 19 20 21 22 23 40 35 30 Self-reported Potential Self-reported Potential Females>0.5 Educated fem. Females>0.5 Educated fem. FHH Non-FHH Note: Estimates are obtained with synthetic panel data and weighted with population weights where the second survey round is used as the base year. The figure shows the percentage of the population that moves out of poverty in the second year. FHH characteristics are measured in second period. Dashed red lines represent the national average for each period. Household heads' ages are restricted to 25-55 for the first survey round and adjusted accordingly for the second survey round. Standard errors are estimated with 1,000 bootstraps. 68 Figure A.10. Probabilities of Female-Headed Households Falling in Poverty in Second Year Conditional on Being Non-poor in First Year (percentage) Egypt 2017-2020 Iraq 2007-2012 12 13 14 15 16 17 10 11 12 13 14 Percentage (%) Percentage (%) Self-reported Potential Self-reported Potential Females>0.5 Educated fem. Females>0.5 Educated fem. Jordan 2010-2013 Mauritania 2014-2019 9 10 11 12 13 11 12 13 14 15 16 Percentage (%) Percentage (%) Self-reported Potential Self-reported Potential Females>0.5 Educated fem. Females>0.5 Educated fem. Palestine 2011-2017 Tunisia 2015-2021 50 8.5 Percentage (%) Percentage (%) 45 8 7.5 40 7 35 Self-reported Potential Self-reported Potential Females>0.5 Educated fem. Females>0.5 Educated fem. FHH Non-FHH Note: Estimates are obtained with synthetic panel data and weighted with population weights where the second survey round is used as the base year. The figure shows the percentage of the population that enters poverty in the second year. FHH characteristics are measured in second period. Dashed red lines represent the national average for each period. Household heads' ages are restricted to 25-55 for the first survey round and adjusted accordingly for the second survey round. Standard errors are estimated with 1,000 bootstraps. 69 Appendix B: Additional Descriptive Statistics for Self-Reported FHHs and MHHs Table B.1. Descriptive Statistics, Arab Republic of Egypt 2012-2020 Egypt, Arab Rep. 2012 2017 2017 2020 Pooled 2012-2020 MHH FHH MHH FHH MHH FHH MHH FHH MHH FHH 47.17 54.11 48.71 56.05 49.99 57.19 47.01 55.49 48.37 55.91 Heads age (13.44) (15.13) (13.13) (14.43) (12.87) (14.01) (13.79) (16.04) (13.34) (14.89) 0.40 0.69 0.36 0.63 0.34 0.63 0.31 0.58 0.35 0.63 Head does not complete primary school (0.49) (0.46) (0.48) (0.48) (0.47) (0.48) (0.46) (0.49) (0.48) (0.48) 0.13 0.09 0.15 0.12 0.15 0.11 0.14 0.11 0.15 0.11 Head's highest education level is primary (0.33) (0.29) (0.36) (0.32) (0.36) (0.31) (0.35) (0.31) (0.35) (0.31) 0.28 0.14 0.29 0.16 0.30 0.15 0.33 0.19 0.30 0.16 Head's highest education level is secondary (0.45) (0.35) (0.45) (0.36) (0.46) (0.36) (0.47) (0.39) (0.46) (0.37) 0.19 0.08 0.20 0.09 0.20 0.10 0.22 0.12 0.20 0.10 Head's highest education level is tertiary (0.40) (0.27) (0.40) (0.29) (0.40) (0.30) (0.41) (0.33) (0.40) (0.30) 0.02 0.03 0.01 0.02 0.01 0.03 0.01 0.03 0.01 0.03 Head is never married (0.13) (0.16) (0.12) (0.14) (0.11) (0.16) (0.12) (0.18) (0.12) (0.16) 0.95 0.21 0.93 0.16 0.95 0.14 0.95 0.20 0.95 0.17 Head is mono married (0.22) (0.41) (0.25) (0.37) (0.22) (0.34) (0.21) (0.40) (0.23) (0.38) 0.01 0.00 0.02 0.00 0.00 0.00 0.00 0.00 0.01 0.00 Head is poly married (0.07) (0.00) (0.12) (0.00) (0.06) (0.00) (0.00) (0.00) (0.08) (0.00) 0.01 0.05 0.01 0.07 0.01 0.08 0.01 0.10 0.01 0.08 Head is divorced/separated (0.08) (0.22) (0.09) (0.26) (0.09) (0.28) (0.09) (0.30) (0.09) (0.27) 0.02 0.71 0.03 0.74 0.03 0.75 0.02 0.66 0.03 0.72 Head is widowed (0.15) (0.46) (0.17) (0.44) (0.16) (0.43) (0.15) (0.47) (0.16) (0.45) 0.86 0.19 0.83 0.20 0.81 0.19 0.85 0.21 0.84 0.20 Head is employed (0.34) (0.40) (0.37) (0.40) (0.39) (0.39) (0.36) (0.41) (0.37) (0.40) 0.01 0.00 0.01 0.00 0.01 0.01 0.01 0.02 0.01 0.01 Head is unemployed (0.09) (0.05) (0.07) (0.06) (0.08) (0.08) (0.08) (0.13) (0.08) (0.08) 0.00 0.20 0.00 0.48 0.00 0.51 0.00 0.49 0.00 0.44 Head is homemaker/housewife (0.00) (0.40) (0.00) (0.50) (0.00) (0.50) (0.00) (0.50) (0.00) (0.50) 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.01 0.00 0.00 Head is student (0.05) (0.00) (0.04) (0.04) (0.01) (0.04) (0.03) (0.07) (0.03) (0.05) 0.11 0.51 0.15 0.31 0.18 0.29 0.14 0.27 0.15 0.33 Head is pensioner/retired/disabled (0.31) (0.50) (0.36) (0.46) (0.38) (0.46) (0.35) (0.45) (0.36) (0.47) 0.02 0.09 0.01 0.00 0.00 0.00 0.00 0.00 0.01 0.02 Head is other activities (0.13) (0.29) (0.08) (0.06) (0.07) (0.02) (0.06) (0.03) (0.08) (0.13) 6,718.96 8,409.76 10,221.28 13,114.56 14,350.90 19,356.74 16,744.03 21,530.11 12,483.20 16,279.27 Per capita consumption (5,294) (6,273) (10,335) (12,830) (11,865) (13,604) (18,062) (16,090) (13,131) (14,000) 1,201.68 5,234.46 2,281.21 8,402.00 3,864.17 12,784.78 4,060.60 15,846.50 3,007.98 11,034.81 Per capita transfers (3,334) (5,406) (5,429) (9,957) (8,474) (11,940) (8,253) (15,900) (7,046) (12,376) 4.63 2.97 4.54 3.00 4.49 2.72 4.35 2.63 4.50 2.82 Household size (1.81) (1.85) (1.71) (1.92) (1.71) (1.73) (1.60) (1.67) (1.70) (1.80) 1.47 0.72 1.43 0.72 1.39 0.56 1.53 0.69 1.45 0.67 Number of children age 0-14 (1.35) (1.14) (1.39) (1.21) (1.40) (1.04) (1.38) (1.18) (1.38) (1.15) 0.20 0.31 0.21 0.32 0.25 0.36 0.21 0.36 0.22 0.34 Number of seniors (0.49) (0.47) (0.50) (0.47) (0.55) (0.49) (0.51) (0.49) (0.51) (0.48) 0.12 0.42 0.13 0.42 0.15 0.49 0.14 0.50 0.14 0.46 1-2 adults, no child (0.32) (0.49) (0.34) (0.49) (0.36) (0.50) (0.35) (0.50) (0.34) (0.50) 0.23 0.18 0.19 0.16 0.16 0.15 0.25 0.18 0.20 0.16 1-2 adults, 1-2 children (0.42) (0.38) (0.39) (0.37) (0.36) (0.35) (0.43) (0.38) (0.40) (0.37) 0.23 0.11 0.25 0.10 0.26 0.09 0.25 0.11 0.25 0.10 1-2 adult, 3 or more children (0.42) (0.31) (0.43) (0.30) (0.44) (0.29) (0.43) (0.31) (0.43) (0.30) 0.23 0.20 0.25 0.21 0.26 0.20 0.21 0.16 0.24 0.19 3 adults or more, 0-1 child (0.42) (0.40) (0.43) (0.41) (0.44) (0.40) (0.41) (0.36) (0.43) (0.40) 0.14 0.07 0.14 0.08 0.14 0.05 0.12 0.04 0.14 0.06 3 adults or more, 2-3 children (0.35) (0.25) (0.34) (0.28) (0.35) (0.23) (0.33) (0.21) (0.34) (0.24) 0.05 0.02 0.04 0.02 0.03 0.01 0.03 0.01 0.04 0.02 3 adults or more, 4 children or more (0.21) (0.14) (0.19) (0.15) (0.18) (0.12) (0.17) (0.11) (0.19) (0.13) 0.56 0.55 0.55 0.52 0.55 0.49 0.57 0.54 0.56 0.52 Rural area (0.50) (0.50) (0.50) (0.50) (0.50) (0.50) (0.50) (0.50) (0.50) (0.50) 0.44 0.45 0.45 0.48 0.45 0.51 0.43 0.46 0.44 0.48 Urban area (0.50) (0.50) (0.50) (0.50) (0.50) (0.50) (0.50) (0.50) (0.50) (0.50) Note: Household sampling weights are applied. Standard deviations are in parentheses. FHHs and MHHs are self-reported. 70 Table B.2. Descriptive Statistics, West Bank and Gaza, 2007-2017 West Bank and Gaza 2007 2009 2011 2017 Pooled 2007-2017 MHH FHH MHH FHH MHH FHH MHH FHH MHH FHH 44.45 57.75 43.99 57.96 45.06 58.27 45.37 59.97 44.77 58.61 Heads age (13.29) (16.31) (13.21) (15.03) (12.95) (13.61) (12.96) (13.16) (13.08) (14.16) 0.15 0.52 0.14 0.53 0.12 0.50 0.12 0.43 0.13 0.49 Head does not complete primary school (0.35) (0.50) (0.35) (0.50) (0.33) (0.50) (0.32) (0.50) (0.34) (0.50) 0.52 0.24 0.48 0.28 0.49 0.31 0.49 0.37 0.49 0.31 Head's highest education level is primary (0.50) (0.43) (0.50) (0.45) (0.50) (0.46) (0.50) (0.48) (0.50) (0.46) 0.16 0.12 0.17 0.11 0.18 0.10 0.16 0.08 0.17 0.10 Head's highest education level is secondary (0.36) (0.32) (0.38) (0.31) (0.38) (0.30) (0.37) (0.28) (0.38) (0.30) 0.17 0.11 0.21 0.08 0.21 0.09 0.23 0.11 0.21 0.10 Head's highest education level is tertiary (0.38) (0.32) (0.41) (0.28) (0.41) (0.29) (0.42) (0.31) (0.41) (0.30) 0.01 0.06 0.01 0.11 0.01 0.10 0.00 0.00 0.01 0.07 Head is never married (0.10) (0.25) (0.09) (0.31) (0.08) (0.30) (0.00) (0.00) (0.08) (0.26) 0.96 0.09 0.97 0.12 0.97 0.10 0.99 0.20 0.97 0.13 Head is mono married (0.19) (0.29) (0.18) (0.33) (0.17) (0.30) (0.10) (0.40) (0.16) (0.34) 0.02 0.00 0.01 0.00 0.01 0.00 0.00 0.00 0.01 0.00 Head is poly married (0.12) (0.00) (0.11) (0.00) (0.10) (0.00) (0.00) (0.00) (0.09) (0.00) 0.00 0.11 0.00 0.06 0.00 0.10 0.01 0.72 0.00 0.26 Head is divorced/separated (0.00) (0.31) (0.05) (0.25) (0.06) (0.30) (0.09) (0.45) (0.06) (0.44) 0.01 0.73 0.01 0.71 0.01 0.70 0.00 0.08 0.01 0.53 Head is widowed (0.11) (0.44) (0.10) (0.46) (0.10) (0.46) (0.05) (0.27) (0.09) (0.50) 0.76 0.21 0.76 0.23 0.78 0.24 0.77 0.23 Head is employed (0.43) (0.41) (0.43) (0.42) (0.42) (0.43) (0.42) (0.42) 0.12 0.02 0.08 0.01 0.08 0.02 0.08 0.02 Head is unemployed (0.32) (0.14) (0.27) (0.11) (0.27) (0.15) (0.28) (0.13) 0.00 0.56 0.00 0.47 0.14 0.08 0.06 0.29 Head is homemaker/housewife (0.05) (0.50) (0.05) (0.50) (0.34) (0.28) (0.24) (0.45) 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 Head is student (0.06) (0.00) (0.03) (0.06) (0.04) (0.04) (0.04) (0.04) 0.08 0.18 0.10 0.27 0.00 0.00 0.05 0.13 Head is pensioner/retired/disabled (0.27) (0.39) (0.30) (0.44) (0.00) (0.00) (0.22) (0.33) 0.04 0.03 0.06 0.02 0.01 0.66 0.03 0.33 Head is other activities (0.20) (0.18) (0.24) (0.13) (0.10) (0.48) (0.18) (0.47) 9,147.45 14,179.85 11,758.84 15,823.72 13,164.33 16,983.40 12,202.69 16,567.52 12,095.51 16,304.18 Per capita consumption (9,668) (13,757) (10,971) (12,594) (12,794) (13,025) (9,668) (13,757) (11,135) (12,622) 6.62 3.84 6.31 3.44 6.27 3.63 5.76 3.34 6.17 3.51 Household size (2.75) (2.86) (2.67) (2.34) (2.54) (2.86) (2.31) (2.51) (2.55) (2.62) 2.69 0.99 2.42 0.71 2.38 0.81 2.31 0.70 2.40 0.77 Number of children age 0-14 (1.95) (1.60) (1.93) (1.41) (1.85) (1.52) (1.86) (1.32) (1.89) (1.44) 0.22 0.40 0.19 0.42 0.19 0.42 0.17 0.41 0.19 0.42 Number of seniors (0.51) (0.50) (0.50) (0.51) (0.49) (0.53) (0.46) (0.50) (0.49) (0.51) 0.07 0.37 0.07 0.42 0.07 0.44 0.07 0.43 1-2 adults, no child (0.25) (0.49) (0.25) (0.49) (0.25) (0.50) (0.25) (0.49) 0.09 0.15 0.13 0.08 0.10 0.08 0.11 0.09 1-2 adults, 1-2 children (0.29) (0.36) (0.34) (0.28) (0.31) (0.27) (0.32) (0.28) 0.37 0.12 0.33 0.11 0.34 0.09 0.34 0.10 1-2 adult, 3 or more children (0.48) (0.32) (0.47) (0.32) (0.47) (0.29) (0.47) (0.31) 0.13 0.22 0.14 0.26 0.17 0.22 0.15 0.23 3 adults or more, 0-1 child (0.33) (0.42) (0.35) (0.44) (0.37) (0.42) (0.36) (0.42) 0.15 0.07 0.15 0.08 0.14 0.09 0.15 0.09 3 adults or more, 2-3 children (0.36) (0.25) (0.36) (0.28) (0.35) (0.29) (0.35) (0.28) 0.20 0.08 0.18 0.04 0.18 0.07 0.18 0.06 3 adults or more, 4 children or more (0.40) (0.27) (0.38) (0.20) (0.38) (0.26) (0.38) (0.24) 0.29 0.26 0.17 0.22 0.17 0.18 0.18 0.17 0.18 0.20 Rural area (0.45) (0.44) (0.37) (0.42) (0.38) (0.39) (0.38) (0.37) (0.39) (0.40) 0.57 0.60 0.74 0.68 0.74 0.72 0.73 0.72 0.72 0.70 Urban area (0.50) (0.49) (0.44) (0.47) (0.44) (0.45) (0.44) (0.45) (0.45) (0.46) Note: Household sampling weights are applied. Standard deviations are in parentheses. FHHs and MHHs are self-reported. 71 Table B.3. Descriptive Statistics, Tunisia 2005-2021 Tunisia 2005 2010 2015 2021 Pooled 2005-2021 MHH FHH MHH FHH MHH FHH MHH FHH MHH FHH 52.12 57.74 53.44 59.52 53.38 60.99 55.40 61.95 53.72 60.31 Heads age (14.07) (15.18) (13.74) (15.60) (13.65) (14.86) (13.53) (14.47) (13.78) (15.04) Head does not complete 0.75 0.92 0.77 0.93 0.19 0.57 0.13 0.48 0.42 0.68 primary school (0.43) (0.27) (0.42) (0.26) (0.39) (0.50) (0.34) (0.50) (0.49) (0.47) Head's highest education level 0.04 0.01 0.03 0.02 0.39 0.25 0.42 0.30 0.25 0.17 is primary (0.20) (0.12) (0.18) (0.12) (0.49) (0.43) (0.49) (0.46) (0.43) (0.38) Head's highest education level 0.12 0.05 0.11 0.04 0.29 0.13 0.32 0.16 0.22 0.10 is secondary (0.32) (0.21) (0.31) (0.20) (0.46) (0.34) (0.47) (0.36) (0.42) (0.31) Head's highest education level 0.09 0.02 0.09 0.02 0.13 0.05 0.13 0.06 0.11 0.04 is tertiary (0.28) (0.14) (0.29) (0.14) (0.33) (0.22) (0.33) (0.23) (0.31) (0.20) 0.02 0.06 0.01 0.04 0.03 0.08 0.03 0.08 0.02 0.07 Head is never married (0.13) (0.24) (0.11) (0.19) (0.17) (0.27) (0.17) (0.27) (0.15) (0.25) 0.96 0.19 0.97 0.20 0.95 0.13 0.95 0.10 0.95 0.15 Head is mono married (0.20) (0.39) (0.18) (0.40) (0.22) (0.34) (0.22) (0.30) (0.21) (0.35) 0.01 0.08 0.00 0.08 0.01 0.09 0.01 0.10 0.00 0.09 Head is divorced/separated (0.07) (0.26) (0.06) (0.27) (0.07) (0.29) (0.07) (0.30) (0.07) (0.29) 0.02 0.67 0.02 0.69 0.02 0.70 0.02 0.71 0.02 0.69 Head is widowed (0.14) (0.47) (0.14) (0.46) (0.13) (0.46) (0.13) (0.45) (0.13) (0.46) 0.74 0.24 0.71 0.17 0.72 0.21 Head is employed (0.44) (0.43) (0.46) (0.38) (0.45) (0.40) 0.02 0.01 0.02 0.01 0.02 0.01 Head is unemployed (0.15) (0.10) (0.15) (0.07) (0.15) (0.09) Head is 0.00 0.41 0.01 0.49 0.00 0.45 homemaker/housewife (0.05) (0.49) (0.07) (0.50) (0.06) (0.50) 0.00 0.01 0.00 0.00 0.00 0.00 Head is student (0.04) (0.09) (0.02) (0.02) (0.03) (0.06) Head is 0.22 0.32 0.26 0.33 0.24 0.32 pensioner/retired/disabled (0.42) (0.46) (0.44) (0.47) (0.43) (0.47) 0.00 0.02 0.00 0.01 0.00 0.01 Head is other activities (0.07) (0.15) (0.05) (0.08) (0.06) (0.12) 2,027.76 2,252.64 2,823.94 3,156.40 4,310.95 4,892.92 5,999.25 6,886.89 3,976.12 4,634.09 Per capita consumption (2,206.66) (2,473.81) (2,604.83) (2,640.10) (4,590.12) (4,002.85) (7,378.56) (6,515.51) (5,136.62) (4,910.49) 4.77 3.25 4.55 3.11 4.26 2.79 3.97 2.61 4.35 2.89 Household size (1.88) (1.96) (1.73) (1.75) (1.60) (1.60) (1.45) (1.39) (1.68) (1.67) 1.21 0.57 1.01 0.50 1.08 0.40 0.95 0.36 1.05 0.44 Number of children age 0-14 (1.26) (1.07) (1.19) (0.99) (1.23) (0.86) (1.18) (0.83) (1.21) (0.93) 0.38 0.43 0.38 0.46 0.35 0.47 0.37 0.45 Number of seniors (0.66) (0.52) (0.66) (0.54) (0.66) (0.53) (0.66) (0.53) 0.10 0.38 0.11 0.40 0.19 0.52 0.14 0.45 1-2 adults, no child (0.31) (0.49) (0.32) (0.49) (0.39) (0.50) (0.35) (0.50) 0.20 0.12 0.20 0.13 0.18 0.09 0.19 0.11 1-2 adults, 1-2 children (0.40) (0.32) (0.40) (0.33) (0.38) (0.29) (0.39) (0.31) 0.17 0.08 0.14 0.08 0.10 0.03 0.13 0.06 1-2 adult, 3 or more children (0.37) (0.28) (0.35) (0.27) (0.30) (0.17) (0.34) (0.23) 0.32 0.32 0.39 0.33 0.44 0.32 0.39 0.32 3 adults or more, 0-1 child (0.47) (0.47) (0.49) (0.47) (0.50) (0.47) (0.49) (0.47) 0.16 0.08 0.14 0.06 0.09 0.04 0.13 0.06 3 adults or more, 2-3 children (0.37) (0.27) (0.35) (0.24) (0.28) (0.19) (0.33) (0.23) 3 adults or more, 4 children or 0.04 0.02 0.02 0.01 0.01 0.00 0.02 0.01 more (0.20) (0.14) (0.15) (0.08) (0.07) (0.04) (0.14) (0.09) 0.32 0.33 0.32 0.34 0.30 0.26 0.31 0.27 0.31 0.29 Rural area (0.46) (0.47) (0.47) (0.47) (0.46) (0.44) (0.46) (0.44) (0.46) (0.45) 0.68 0.67 0.68 0.66 0.70 0.74 0.69 0.73 0.69 0.71 Urban area (0.46) (0.47) (0.47) (0.47) (0.46) (0.44) (0.46) (0.44) (0.46) (0.45) Note: Household sampling weights are applied. Standard deviations are in parentheses. FHHs and MHHs are self-reported. 72 Table B.4. Descriptive Statistics, Jordan 2010-2013 Jordan 2010 2013 Pooled 2010-2013 MHH FHH MHH FHH MHH FHH 47.62 58.63 47.49 58.96 47.55 58.80 Heads age (14.13) (13.68) (14.04) (13.24) (14.08) (13.45) 0.15 0.53 0.12 0.49 0.13 0.51 Head does not complete primary school (0.36) (0.50) (0.32) (0.50) (0.34) (0.50) 0.47 0.27 0.47 0.30 0.47 0.28 Head's highest education level is primary (0.50) (0.44) (0.50) (0.46) (0.50) (0.45) 0.15 0.08 0.14 0.09 0.15 0.09 Head's highest education level is secondary (0.36) (0.28) (0.34) (0.28) (0.35) (0.28) 0.22 0.12 0.28 0.12 0.25 0.12 Head's highest education level is tertiary (0.42) (0.33) (0.45) (0.33) (0.44) (0.33) 0.02 0.06 0.02 0.07 0.02 0.07 Head is never married (0.13) (0.25) (0.14) (0.25) (0.14) (0.25) 0.97 0.19 0.96 0.17 0.97 0.18 Head is mono married (0.17) (0.39) (0.20) (0.37) (0.18) (0.38) 0.00 0.00 0.00 0.00 0.00 0.00 Head is poly married (0.06) (0.00) (0.06) (0.00) (0.06) (0.00) 0.00 0.04 0.00 0.05 0.00 0.04 Head is divorced/separated (0.03) (0.19) (0.06) (0.22) (0.05) (0.20) 0.01 0.71 0.01 0.72 0.01 0.72 Head is widowed (0.08) (0.45) (0.11) (0.45) (0.10) (0.45) 0.66 0.04 0.67 0.05 0.66 0.05 Head is employed (0.47) (0.20) (0.47) (0.22) (0.47) (0.21) 0.04 0.01 0.07 0.04 0.06 0.02 Head is unemployed (0.20) (0.09) (0.26) (0.19) (0.24) (0.15) 0.00 0.79 0.00 0.75 0.00 0.77 Head is homemaker/housewife (0.02) (0.41) (0.00) (0.43) (0.01) (0.42) 0.00 0.00 0.00 0.00 0.00 0.00 Head is student (0.01) (0.00) (0.05) (0.04) (0.04) (0.03) 0.12 0.15 0.14 0.14 0.13 0.14 Head is pensioner/retired/disabled (0.33) (0.35) (0.35) (0.34) (0.34) (0.35) 0.17 0.02 0.12 0.02 0.14 0.02 Head is other activities (0.38) (0.13) (0.32) (0.15) (0.35) (0.14) 2,063.45 2,916.21 2,238.36 3,100.82 2,155.59 3,010.97 Per capita consumption (3,263.14) (2,417.35) (1,578.64) (2,442.96) (2,521.48) (2,431.02) 377.77 1,107.40 469.79 1,114.52 426.24 1,111.05 Per capita transfers (703.55) (1,489.48) (839.28) (1,364.42) (779.30) (1,425.88) 5.66 3.64 5.36 3.49 5.50 3.56 Household size (2.18) (2.28) (2.12) (2.51) (2.15) (2.40) 1.93 0.59 1.80 0.57 1.86 0.58 Number of children age 0-14 (1.70) (1.14) (1.66) (1.42) (1.68) (1.29) 0.26 0.38 0.23 0.41 0.24 0.40 Number of seniors (0.57) (0.49) (0.54) (0.51) (0.55) (0.50) 0.07 0.36 0.09 0.39 0.08 0.37 1-2 adults, no child (0.26) (0.48) (0.28) (0.49) (0.27) (0.48) 0.13 0.14 0.15 0.10 0.14 0.12 1-2 adults, 1-2 children (0.34) (0.35) (0.35) (0.30) (0.35) (0.32) 0.30 0.06 0.29 0.06 0.29 0.06 1-2 adult, 3 or more children (0.46) (0.25) (0.45) (0.23) (0.46) (0.24) 0.23 0.29 0.25 0.34 0.24 0.32 3 adults or more, 0-1 child (0.42) (0.46) (0.43) (0.47) (0.43) (0.47) 0.15 0.10 0.13 0.08 0.14 0.09 3 adults or more, 2-3 children (0.36) (0.30) (0.34) (0.27) (0.35) (0.29) 0.11 0.05 0.09 0.04 0.10 0.04 3 adults or more, 4 children or more (0.31) (0.22) (0.29) (0.18) (0.30) (0.20) 0.16 0.13 0.17 0.19 0.17 0.16 Rural area (0.37) (0.34) (0.37) (0.39) (0.37) (0.37) 0.84 0.87 0.83 0.81 0.83 0.84 Urban area (0.37) (0.34) (0.37) (0.39) (0.37) (0.37) Note: Household sampling weights are applied. Standard deviations are in parentheses. FHHs and MHHs are self-reported. 73 Table B.5. Descriptive Statistics, Iraq 2007-2013 Iraq 2007 2012 Pooled 2007-2013 MHH FHH MHH FHH MHH FHH 45.15 54.29 46.93 54.28 46.73 54.28 Heads age (13.71) (13.04) (13.06) (12.51) (13.14) (12.58) 0.31 0.75 0.34 0.72 0.34 0.72 Head does not complete primary school (0.46) (0.43) (0.47) (0.45) (0.47) (0.45) 0.41 0.17 0.41 0.20 0.41 0.20 Head's highest education level is primary (0.49) (0.38) (0.49) (0.40) (0.49) (0.40) 0.11 0.02 0.09 0.03 0.09 0.03 Head's highest education level is secondary (0.31) (0.14) (0.28) (0.17) (0.28) (0.17) 0.17 0.06 0.16 0.05 0.16 0.05 Head's highest education level is tertiary (0.38) (0.23) (0.36) (0.22) (0.37) (0.22) 0.02 0.04 0.01 0.02 0.01 0.02 Head is never married (0.13) (0.20) (0.09) (0.14) (0.10) (0.15) 0.95 0.08 0.95 0.15 0.95 0.14 Head is mono married (0.21) (0.27) (0.22) (0.35) (0.22) (0.34) 0.01 0.00 0.03 0.00 0.03 0.00 Head is poly married (0.11) (0.00) (0.17) (0.00) (0.16) (0.00) 0.00 0.07 0.00 0.06 0.00 0.06 Head is divorced/separated (0.04) (0.25) (0.04) (0.25) (0.04) (0.25) 0.01 0.81 0.01 0.77 0.01 0.77 Head is widowed (0.12) (0.39) (0.10) (0.42) (0.11) (0.42) 0.79 0.19 0.78 0.15 0.78 0.16 Head is employed (0.41) (0.40) (0.42) (0.36) (0.42) (0.36) 0.03 0.02 0.02 0.01 0.02 0.01 Head is unemployed (0.18) (0.13) (0.15) (0.08) (0.15) (0.09) 0.00 0.69 0.00 0.53 0.00 0.55 Head is homemaker/housewife (0.02) (0.46) (0.03) (0.50) (0.03) (0.50) 0.00 0.00 0.00 0.00 0.00 0.00 Head is student (0.04) (0.00) (0.04) (0.02) (0.04) (0.02) 0.14 0.08 0.18 0.30 0.18 0.27 Head is pensioner/retired/disabled (0.35) (0.27) (0.39) (0.46) (0.38) (0.45) 0.03 0.03 0.02 0.01 0.02 0.01 Head is other activities (0.17) (0.16) (0.12) (0.08) (0.13) (0.10) 1,878,839.75 2,011,498.91 2,855,613.71 3,072,404.17 2,746,443.13 2,933,690.42 Per capita consumption (1,662,430) (1,562,966) (2,612,256) (2,728,793) (2,542,574) (2,630,373) 166.71 409.56 249.20 647.95 239.98 616.79 Per capita transfers (415.53) (941.59) (902.11) (1,940.46) (861.86) (1,842.51) 7.00 5.76 8.49 7.64 8.33 7.39 Household size (3.43) (3.46) (4.21) (4.22) (4.16) (4.18) 2.69 1.64 3.46 2.75 3.37 2.61 Number of children age 0-14 (2.12) (2.00) (2.55) (2.68) (2.51) (2.63) 0.22 0.27 0.24 0.24 0.24 0.25 Number of seniors (0.51) (0.49) (0.52) (0.46) (0.52) (0.46) 0.04 0.14 0.01 0.04 0.01 0.05 1-2 adults, no child (0.20) (0.35) (0.11) (0.19) (0.12) (0.22) 0.14 0.08 0.06 0.05 0.07 0.06 1-2 adults, 1-2 children (0.34) (0.28) (0.25) (0.23) (0.26) (0.23) 0.29 0.13 0.27 0.15 0.27 0.15 1-2 adult, 3 or more children (0.45) (0.34) (0.44) (0.35) (0.44) (0.35) 0.15 0.30 0.11 0.21 0.11 0.22 3 adults or more, 0-1 child (0.35) (0.46) (0.31) (0.41) (0.32) (0.42) 0.16 0.18 0.18 0.26 0.18 0.25 3 adults or more, 2-3 children (0.37) (0.38) (0.39) (0.44) (0.39) (0.43) 0.22 0.16 0.36 0.29 0.35 0.27 3 adults or more, 4 children or more (0.42) (0.37) (0.48) (0.45) (0.48) (0.45) 0.27 0.20 0.33 0.22 0.32 0.22 Rural area (0.44) (0.40) (0.47) (0.42) (0.47) (0.41) 0.73 0.80 0.67 0.78 0.68 0.78 Urban area (0.44) (0.40) (0.47) (0.42) (0.47) (0.41) Note: Household sampling weights are applied. Standard deviations are in parentheses. FHHs and MHHs are self-reported. 74 Table B.6. Descriptive Statistics, Mauritania 2004-2019 Mauritania 2004 2008 2014 2019 Pooled 2004-2019 MHH FHH MHH FHH MHH FHH MHH FHH MHH FHH 48.10 53.29 48.22 47.33 49.11 48.05 49.63 46.83 48.83 48.01 Heads age (13.60) (14.09) (14.06) (15.24) (14.49) (15.80) (14.29) (15.25) (14.16) (15.40) 0.77 0.93 0.72 0.85 0.70 0.81 0.48 0.50 0.66 0.72 Head does not complete primary school (0.42) (0.25) (0.45) (0.35) (0.46) (0.39) (0.50) (0.50) (0.47) (0.45) 0.08 0.03 0.11 0.09 0.11 0.11 0.26 0.33 0.15 0.18 Head's highest education level is primary (0.28) (0.17) (0.31) (0.29) (0.32) (0.32) (0.44) (0.47) (0.35) (0.38) 0.11 0.03 0.12 0.05 0.13 0.07 0.18 0.16 0.14 0.09 Head's highest education level is secondary (0.31) (0.18) (0.33) (0.21) (0.34) (0.25) (0.38) (0.36) (0.34) (0.29) 0.04 0.01 0.05 0.01 0.06 0.01 0.08 0.01 0.06 0.01 Head's highest education level is tertiary (0.19) (0.08) (0.22) (0.08) (0.24) (0.08) (0.27) (0.10) (0.23) (0.09) 0.03 0.02 0.02 0.01 0.03 0.02 0.03 0.02 0.03 0.02 Head is never married (0.17) (0.15) (0.16) (0.12) (0.17) (0.13) (0.17) (0.13) (0.16) (0.13) 0.94 0.10 0.95 0.39 0.94 0.43 0.88 0.50 0.93 0.41 Head is mono married (0.24) (0.30) (0.22) (0.49) (0.23) (0.50) (0.32) (0.50) (0.26) (0.49) 0.00 0.00 0.00 0.00 0.00 0.00 0.06 0.03 0.02 0.01 Head is poly married (0.00) (0.00) (0.00) (0.00) (0.00) (0.00) (0.23) (0.17) (0.12) (0.11) 0.02 0.33 0.01 0.22 0.02 0.21 0.02 0.18 0.02 0.21 Head is divorced/separated (0.14) (0.47) (0.12) (0.42) (0.13) (0.40) (0.13) (0.39) (0.13) (0.41) 0.01 0.55 0.01 0.37 0.01 0.35 0.01 0.27 0.01 0.35 Head is widowed (0.11) (0.50) (0.11) (0.48) (0.11) (0.48) (0.12) (0.44) (0.11) (0.48) 0.86 0.54 0.78 0.40 0.87 0.40 0.81 0.37 0.83 0.41 Head is employed (0.35) (0.50) (0.41) (0.49) (0.33) (0.49) (0.39) (0.48) (0.37) (0.49) 0.02 0.02 0.04 0.02 0.01 0.02 0.01 0.01 0.02 0.02 Head is unemployed (0.15) (0.12) (0.19) (0.13) (0.08) (0.13) (0.11) (0.11) (0.14) (0.12) 0.12 0.45 0.18 0.59 0.12 0.58 0.18 0.61 0.15 0.58 Head is not searched and not work (0.32) (0.50) (0.38) (0.49) (0.32) (0.49) (0.38) (0.49) (0.36) (0.49) 184,984.25 147,570.28 240,551.50 219,258.68 341,658.92 350,431.69 373,240.01 403,696.14 294,662.55 315,850.25 Per capita consumption (2587509) (205663) (234972) (168764) (299510) (247036) (264805) (268529) (1,189,714) (252,048) 5.94 4.40 5.87 4.68 6.05 5.03 6.46 5.59 6.10 5.08 Household size (2.77) (2.44) (2.88) (2.47) (3.44) (2.61) (3.59) (2.98) (3.24) (2.73) 2.56 1.63 2.53 2.12 2.68 2.33 2.86 2.66 2.67 2.32 Number of children age 0-14 (2.01) (1.73) (2.03) (1.84) (2.29) (1.98) (2.41) (2.01) (2.21) (1.96) 0.19 0.27 0.21 0.20 0.25 0.23 0.27 0.23 0.23 0.23 Number of seniors (0.44) (0.46) (0.47) (0.41) (0.51) (0.44) (0.56) (0.46) (0.50) (0.44) 0.08 0.19 0.10 0.13 0.10 0.11 0.09 0.07 0.09 0.11 1-2 adults, no child (0.28) (0.39) (0.29) (0.33) (0.29) (0.31) (0.28) (0.25) (0.29) (0.31) 0.16 0.21 0.16 0.22 0.14 0.19 0.12 0.19 0.14 0.20 1-2 adults, 1-2 children (0.37) (0.40) (0.36) (0.41) (0.35) (0.40) (0.33) (0.39) (0.35) (0.40) 0.20 0.14 0.20 0.24 0.23 0.26 0.20 0.26 0.21 0.24 1-2 adult, 3 or more children (0.40) (0.35) (0.40) (0.43) (0.42) (0.44) (0.40) (0.44) (0.41) (0.43) 0.19 0.26 0.19 0.20 0.18 0.20 0.18 0.17 0.18 0.19 3 adults or more, 0-1 child (0.39) (0.44) (0.39) (0.40) (0.38) (0.40) (0.38) (0.37) (0.39) (0.40) 0.17 0.13 0.18 0.13 0.17 0.14 0.19 0.17 0.18 0.15 3 adults or more, 2-3 children (0.38) (0.34) (0.38) (0.34) (0.38) (0.34) (0.39) (0.37) (0.38) (0.35) 0.19 0.07 0.18 0.09 0.18 0.11 0.23 0.15 0.20 0.12 3 adults or more, 4 children or more (0.39) (0.25) (0.38) (0.28) (0.39) (0.31) (0.42) (0.36) (0.40) (0.32) 0.62 0.57 0.54 0.64 0.49 0.56 0.52 0.57 0.54 0.58 Rural area (0.49) (0.49) (0.50) (0.48) (0.50) (0.50) (0.50) (0.50) (0.50) (0.49) 0.38 0.43 0.46 0.36 0.51 0.44 0.48 0.43 0.46 0.42 Urban area (0.49) (0.49) (0.50) (0.48) (0.50) (0.50) (0.50) (0.50) (0.50) (0.49) Note: Household sampling weights are applied. Standard deviations are in parentheses. FHHs and MHHs are self-reported. 75 Table B.7. Descriptive Statistics, all countries - years MHH FHH 48.14 55.52 Heads age (13.61) (14.10) 0.36 0.70 Head does not complete primary school (0.48) (0.46) 0.38 0.19 Head's highest education level is primary (0.48) (0.40) 0.12 0.06 Head's highest education level is secondary (0.32) (0.23) 0.15 0.05 Head's highest education level is tertiary (0.36) (0.21) 0.01 0.04 Head is never married (0.11) (0.19) 0.95 0.17 Head is mono married (0.22) (0.37) 0.02 0.03 Head is poly married (0.15) (0.16) 0.01 0.25 Head is divorced/separated (0.07) (0.43) 0.01 0.71 Head is widowed (0.11) (0.45) 0.77 0.19 Head is employed (0.42) (0.39) 0.03 0.01 Head is unemployed (0.16) (0.10) 0.01 0.55 Head is homemaker/housewife (0.08) (0.50) 0.00 0.00 Head is student (0.04) (0.03) 0.18 0.27 Head is pensioner/retired/disabled (0.39) (0.45) 0.02 0.01 Head is other activities (0.15) (0.11) 2,034,242.98 1,731,419.27 Per capita consumption (2,496,523) (2,451,185) 204.12 471.50 Per capita transfers (805.10) (1,655.70) 7.39 5.75 Household size (4.05) (3.99) 2.85 1.89 Number of children (2.46) (2.38) 0.26 0.29 Number of seniors (0.54) (0.48) 0.04 0.16 1-2 adults, no child (0.19) (0.37) 0.10 0.09 1-2 adults, 1-2 children (0.29) (0.28) 0.25 0.13 1-2 adult, 3 or more children (0.43) (0.34) 0.16 0.25 3 adults or more, 0-1 child (0.37) (0.43) 0.17 0.19 3 adults or more, 2-3 children (0.38) (0.39) 0.28 0.19 3 adults or more, 4 children or more (0.45) (0.39) 0.32 0.27 Rural area (0.47) (0.45) 0.68 0.73 Urban area (0.47) (0.45) Note: Household sampling weights are applied. Standard deviations are in parentheses. FHHs and MHHs are self-reported. 76 Appendix C: Synthetic Panel Method This appendix offers a brief overview of the synthetic panel method based on Dang and Lanjouw (2023). Recent validations and applications of the synthetic panel methods by various researchers for different country contexts ranging from Africa to Latin America, the Middle East, and Europe have been encouraging in terms of accurate projections of economic status (Ferreira et al., 2012; Beegle et al., 2016; UNDP, 2016; OECD, 2018; Salvuci and Tarp, 2021; Ghomi, 2022). Let be a vector of household characteristics observed in survey round j (j= 1 or 2) that are also observed in the other survey round for household i, i= 1,…, N. 17 These household characteristics can include such time-invariant variables as ethnicity, religion, language, place of birth, parental education, and other time-varying household characteristics if retrospective questions about the round-1 values of such characteristics are asked in the second round survey. To reduce spurious changes due to changes in household composition over time, we usually restrict the estimation samples to household heads in a certain age range, say 25 to 55, in the first cross section and adjust this age range accordingly in the second cross section. This restriction also helps ensure certain variables such as heads’ education attainment remains relatively stable over time (assuming most heads are finished with their schooling). 18 This age range is usually used in traditional pseudo-panel analysis but can vary depending on the cultural and economic factors in each specific setting. Population weights are then employed to provide estimates that represent the whole population. Then let represent household consumption or income in survey round j, j= 1 or 2. The linear projection of household consumption (or income) on household characteristics for each survey round is given by ′ = + (C1) Let be the poverty line in period j. We are interested in knowing the unconditional measures of poverty mobility such as (1 < 1 2 > 2 ) (C2) which represents the percentage of households that are poor in the first survey round (year) but nonpoor in the second survey round, or the conditional measures such as (2 > 2 | 1 < 1 ) (C3) which represents the percentage of poor households in the first round that escape poverty in the second round. If true panel data are available, we can straightforwardly estimate the quantities in (C2) and (C3); but in the absence of such data, we can use synthetic panels to study mobility. To operationalize the framework, we make two standard assumptions. First, we assume that the underlying population being sampled in survey rounds 1 and 2 are identical such that their time- invariant characteristics remain the same over time. More specifically, coupled with equation (C1), this implies the conditional distribution of expenditure in a given period is identical whether it is conditional on the given household characteristics in period 1 or period 2 (i.e., 1 = 2 implies 1 |1 and 1 |2 have identical distributions) (Assumption 1). Second, we assume that i1 and i2 have a bivariate normal distribution with positive correlation coefficient ρ and standard deviations 1 and σ2 respectively (Assumption 2). Quantity (2) can be estimated by 17 We suppress the index for countries and FHH types to make notation less cluttered in this appendix. 18 While household heads may still increase their educational achievement in theory, this rarely happens in practice. 77 ′ 1 −1 ′ 2 −2 (1 < 1 2 > 2 ) = Φ2 � 2 ,− 2 , −� (C4) 1 2 where Φ2 (. ) stands for the bivariate normal cumulative distribution function (cdf), and 2 (. ) stands for the bivariate normal probability density function (pdf). Note that in Equation (1), the estimated parameters obtained from data in both survey rounds are applied to data from the second survey round (x2) (or the base year) for prediction, but we can use data from the first survey round as the base year as well. It is then straightforward to estimate quantity (C3) by dividing quantity ′ 1 −1 (C2) by Φ � 2 �, where Φ(. ) stands for the univariate normal cumulative distribution 1 function (cdf). In Equation (4), the parameters and are estimated from Equation (C1), and ρ can be estimated using an approximation of the correlation of the cohort-aggregated household consumption between the two surveys (1 2 ). In particular, given an approximation of 1 2 , where c indexes the cohorts constructed from the household survey data, the partial correlation coefficient ρ can be estimated by ′ ( ) 1 2 �(1 ) (2 )−1 2 = (C5) 1 2 An alternative way to estimate is to further assume that there is a cohort fixed effect in the error terms and aggregate all the time-invariant variables to the cohort level and use the following equation = ′ + (C6) where the error term includes a cohort fixed effect and the error . Note that the standard errors of estimates based on the synthetic panels can in fact be even smaller than that of the true (or design-based) rate if there is a good model fit (or the sample size in the target survey is significantly larger than that in the base survey; see Dang and Lanjouw, 2023, for discussion). Tables C.1-C.6 present the estimation results using Equation (C1) for all the countries and survey rounds. Tables C.7-C.12 present the descriptive statistics of the estimation sample. These tables show that while most of the time-invariant characteristics show similar distributions across survey rounds (and satisfy Assumption 1), some do not. For example, these include the shares of household heads achieving primary education or secondary education in the Arab Republic of Egypt during 2012-2015 (Table C.7). But the differences are practically very close to 0. Table C.13 presents the estimates for using Equations (C5) and (C6), where are estimated using Equation (C5) for all countries with cohorts being defined by age interacted with household heads’ education. We also provide alternative estimates for using Equation (C6). Using these estimates, Figures C.1 and C.2 offer qualitatively similar results to Figures 5 and 6. 78 Table C.1 First-stage regressions, Arab Republic of Egypt 2012-2015 2015-2017 2017-2020 2012 2015 2015 2017 2017 2020 Head`s age 0.006*** 0.012*** 0.010*** 0.011*** 0.009*** 0.012*** (0.00) (0.00) (0.00) (0.00) (0.00) (0.00) Head is female 0.141*** 0.141*** 0.136*** 0.234*** 0.226*** 0.198*** (0.02) (0.02) (0.02) (0.02) (0.02) (0.02) Highest education level is primary 0.142*** 0.126*** 0.123*** 0.097*** 0.099*** 0.091*** (0.02) (0.01) (0.01) (0.01) (0.01) (0.02) Highest education level is secondary 0.199*** 0.215*** 0.209*** 0.179*** 0.173*** 0.194*** (0.01) (0.01) (0.01) (0.01) (0.01) (0.01) Highest education level is tertiary 0.461*** 0.482*** 0.481*** 0.395*** 0.394*** 0.486*** (0.02) (0.01) (0.01) (0.01) (0.01) (0.02) Urban 0.241*** 0.210*** 0.215*** 0.126*** 0.126*** 0.162*** (0.01) (0.01) (0.01) (0.01) (0.01) (0.01) _cons 7.995*** 8.091*** 8.196*** 8.555*** 8.604*** 8.632*** (0.03) (0.03) (0.03) (0.03) (0.03) (0.03) adjusted R2 0.23 0.25 0.24 0.17 0.16 0.21 N 5102 8338 7836 8301 7799 7286 Note: ***, **, and * denote statistical significance at the 0.01, 0.05, and 0.10 levels. Standard errors are in parentheses. Regression is estimated using Weighted Ordinary Least Squares. Household heads’ ages are restricted to between 25 and 55 for the first survey round and adjusted accordingly for the second round. The reference groups are household with no primary education and living in rural areas. 79 Table C.2 First-stage regressions, Iraq 2007-2012 2007 2012 Head`s age -0.001** 0.005*** (0.00) (0.00) Head is female 0.012 0.114*** (0.02) (0.02) Highest education level is primary -0.015 0.069*** (0.01) (0.01) Highest education level is secondary 0.059*** 0.305*** (0.02) (0.02) Highest education level is tertiary 0.183*** 0.441*** (0.02) (0.01) Urban 0.326*** 0.317*** (0.01) (0.01) _cons 13.917*** 13.966*** (0.03) (0.03) adjusted R2 0.08 0.13 N 12895 18552 Note: ***, **, and * denote statistical significance at the 0.01, 0.05, and 0.10 levels. Standard errors are in parentheses. Regression is estimated using Weighted Ordinary Least Squares. Household heads’ ages are restricted to between 25 and 55 for the first survey round and adjusted accordingly for the second round. The reference groups are household with no primary education and living in rural areas. 80 Table C.3 First-stage regressions, Jordan 2010-2013 2010 2013 Head`s age -0.002 0.004*** (0.00) (0.00) Head is female 0.213*** 0.107*** (0.05) (0.03) Highest education level is primary 0.171*** 0.280*** (0.04) (0.03) Highest education level is secondary 0.320*** 0.449*** (0.05) (0.04) Highest education level is tertiary 0.666*** 0.729*** (0.05) (0.03) Urban 0.022 0.037* (0.03) (0.02) _cons 6.976*** 6.718*** (0.08) (0.06) adjusted R2 0.16 0.18 N 1873 3437 Note: ***, **, and * denote statistical significance at the 0.01, 0.05, and 0.10 levels. Standard errors are in parentheses. Regression is estimated using Weighted Ordinary Least Squares. Household heads’ ages are restricted to between 25 and 55 for the first survey round and adjusted accordingly for the second round. The reference groups are household with no primary education and living in rural areas. 81 Table C.4 First-stage regressions, Mauritania 2004-2008 2008-2014 2014-2019 2004 2008 2008 2014 2014 2019 Head`s age -0.007*** -0.003*** -0.006*** -0.004*** -0.007*** -0.004*** (0.00) (0.00) (0.00) (0.00) (0.00) (0.00) Head is female 0.022 0.075*** 0.065*** 0.128*** 0.103*** 0.130*** (0.02) (0.01) (0.01) (0.02) (0.02) (0.01) Highest education level is primary 0.187*** 0.143*** 0.146*** 0.073*** 0.079*** -0.014 (0.03) (0.02) (0.02) (0.02) (0.02) (0.02) Highest education level is secondary 0.384*** 0.383*** 0.382*** 0.223*** 0.245*** 0.145*** (0.03) (0.02) (0.02) (0.02) (0.02) (0.02) Highest education level is tertiary 0.708*** 0.609*** 0.621*** 0.382*** 0.408*** 0.382*** (0.04) (0.03) (0.03) (0.03) (0.03) (0.03) Urban 0.342*** 0.605*** 0.596*** 0.353*** 0.380*** 0.368*** (0.02) (0.01) (0.01) (0.01) (0.02) (0.01) _cons 11.532*** 11.710*** 11.819*** 12.279*** 12.356*** 12.421*** (0.05) (0.04) (0.04) (0.04) (0.04) (0.04) adjusted R2 0.18 0.32 0.32 0.16 0.18 0.18 N 6065 9269 9088 6672 6219 6425 Note: ***, **, and * denote statistical significance at the 0.01, 0.05, and 0.10 levels. Standard errors are in parentheses. Regression is estimated using Weighted Ordinary Least Squares. Household heads’ ages are restricted to between 25 and 55 for the first survey round and adjusted accordingly for the second round. The reference groups are household with no primary education and living in rural areas. 82 Table C.5 First-stage regressions, West Bank and Gaza 2007-2009 2009-2011 2011-2017 2007 2009 2009 2011 2011 2017 Head`s age 0.005 0.006*** 0.005*** 0.002 -0.000 0.012*** (0.00) (0.00) (0.00) (0.00) (0.00) (0.00) Head is female 0.226** 0.206*** 0.226*** 0.133*** 0.141*** 0.160*** (0.11) (0.05) (0.06) (0.05) (0.05) (0.06) Highest education level is primary 0.177** 0.220*** 0.214*** 0.218*** 0.210*** 0.260*** (0.07) (0.04) (0.04) (0.04) (0.04) (0.04) Highest education level is secondary 0.276*** 0.344*** 0.346*** 0.339*** 0.341*** 0.373*** (0.09) (0.04) (0.04) (0.04) (0.04) (0.05) Highest education level is tertiary 0.585*** 0.607*** 0.606*** 0.602*** 0.606*** 0.518*** (0.08) (0.04) (0.04) (0.04) (0.04) (0.04) Urban -0.034 0.014 0.012 -0.064** -0.070** -0.240*** (0.05) (0.03) (0.03) (0.03) (0.03) (0.03) Refugee -0.327*** -0.039 -0.034 -0.314*** -0.314*** -0.538*** (0.07) (0.05) (0.05) (0.04) (0.04) (0.05) _cons 8.288*** 8.388*** 8.457*** 8.718*** 8.828*** 8.381*** (0.15) (0.07) (0.07) (0.08) (0.08) (0.08) adjusted R2 0.08 0.09 0.09 0.09 0.09 0.11 N 962 2944 2938 3288 3229 2815 Note: ***, **, and * denote statistical significance at the 0.01, 0.05, and 0.10 levels. Standard errors are in parentheses. Regression is estimated using Weighted Ordinary Least Squares. Household heads’ ages are restricted to between 25 and 55 for the first survey round and adjusted accordingly for the second round. The reference groups are household with no primary education and living in rural areas. 83 Table C.6 First-stage regressions, Tunisia 2005-2010 2010-2015 2015-2021 2005 2010 2010 2015 2015 2021 Head`s age 0.004*** 0.008*** 0.005*** 0.008*** 0.006*** 0.014*** (0.00) (0.00) (0.00) (0.00) (0.00) (0.00) Head is female 0.143*** 0.053** 0.006 0.197*** 0.184*** 0.113*** (0.03) (0.02) (0.03) (0.02) (0.02) (0.02) Highest education level is primary 0.312*** 0.305*** 0.300*** 0.210*** 0.227*** 0.142*** (0.04) (0.04) (0.04) (0.01) (0.02) (0.02) Highest education level is secondary 0.386*** 0.293*** 0.266*** 0.443*** 0.460*** 0.354*** (0.02) (0.02) (0.02) (0.02) (0.02) (0.02) Highest education level is tertiary 0.945*** 0.764*** 0.754*** 0.904*** 0.911*** 0.704*** (0.03) (0.02) (0.03) (0.02) (0.02) (0.02) Urban 0.495*** 0.497*** 0.499*** 0.355*** 0.370*** 0.287*** (0.02) (0.01) (0.02) (0.01) (0.01) (0.01) _cons 6.547*** 6.747*** 6.881*** 7.006*** 7.084*** 7.183*** (0.05) (0.04) (0.05) (0.03) (0.03) (0.04) adjusted R2 0.30 0.29 0.29 0.29 0.29 0.21 N 6769 7507 6425 16456 13635 10520 Note: ***, **, and * denote statistical significance at the 0.01, 0.05, and 0.10 levels. Standard errors are in parentheses. Regression is estimated using Weighted Ordinary Least Squares. Household heads’ ages are restricted to between 25 and 55 for the first survey round and adjusted accordingly for the second round. The reference groups are household with no primary education and living in rural areas. 84 Table C.7. Descriptive statistics of estimation sample, Arab Republic of Egypt 2012-2015 2015-2017 2017-2020 Variables 2012 2015 diff 2015 2017 diff 2017 2020 diff 8.58 8.97 0.4*** 8.95 9.32 0.4*** 9.30 9.48 0.2*** Log of per capita consumption (0.01) (0.01) (0.0) (0.01) (0.01) (0.0) (0.01) (0.01) (0.0) 41.32 43.97 2.7*** 42.23 44.20 2.0*** 43.14 42.89 -0.3* Head`s age (0.12) (0.09) (0.1) (0.09) (0.09) (0.1) (0.09) (0.10) (0.1) 0.13 0.13 0.0 0.12 0.13 0.0 0.12 0.13 0.0 Head is female (0.00) (0.00) (0.0) (0.00) (0.00) (0.0) (0.00) (0.00) (0.0) 0.13 0.15 0.0*** 0.15 0.15 0.0 0.15 0.14 -0.0*** Head's highest education level is primary (0.00) (0.00) (0.0) (0.00) (0.00) (0.0) (0.00) (0.00) (0.0) 0.31 0.34 0.0** 0.35 0.36 0.0 0.37 0.37 0.0 Head's highest education level is secondary (0.01) (0.01) (0.0) (0.01) (0.01) (0.0) (0.01) (0.01) (0.0) 0.20 0.19 -0.0 0.19 0.18 -0.0 0.18 0.21 0.0*** Head's highest education level is tertiary (0.01) (0.00) (0.0) (0.00) (0.00) (0.0) (0.00) (0.00) (0.0) 0.42 0.42 -0.0 0.41 0.42 0.0 0.41 0.43 0.0** Urban area (0.01) (0.01) (0.0) (0.01) (0.01) (0.0) (0.01) (0.01) (0.0) Note: ***, **, and * denote statistical significance at the 0.01, 0.05, and 0.10 levels. Standard errors are in parentheses. The data are unweighted. Household heads' ages are restricted to between 25 and 55 for the first survey round and adjusted accordingly for the second survey round. The reference groups are household with no primary education and living in rural areas. 85 Table C.8 Descriptive statistics of estimation sample, Iraq 2007-2012 2007 2012 diff 14.22 14.67 0.1*** Log of per capita consumption (0.01) (0.01) (0.0) 40.46 44.28 1.7*** Head`s age (0.07) (0.06) (0.2) 0.09 0.09 0.0 Head is female (0.00) (0.00) (0.0) 0.41 0.40 0.0*** Head's highest education level is primary (0.00) (0.00) (0.0) 0.12 0.08 -0.0 Head's highest education level is secondary (0.00) (0.00) (0.0) 0.19 0.15 -0.0*** Head's highest education level is tertiary (0.00) (0.00) (0.0) 0.68 0.60 -0.1*** Urban area (0.00) (0.00) (0.0) Note: ***, **, and * denote statistical significance at the 0.01, 0.05, and 0.10 levels. Standard errors are in parentheses. The data are unweighted. Household heads' ages are restricted to between 25 and 55 for the first survey round and adjusted accordingly for the second survey round. The reference groups are household with no primary education and living in rural areas. 86 Table C.9 Descriptive statistics of estimation sample, Jordan 2010-2013 2010 2013 diff 7.29 7.35 0.1*** Log of per capita consumption (0.01) (0.01) (0.0) 41.00 42.72 1.7*** Head`s age (0.18) (0.14) (0.2) 0.09 0.09 0.0 Head is female (0.01) (0.00) (0.0) 0.51 0.55 0.0*** Head's highest education level is primary (0.01) (0.01) (0.0) 0.16 0.15 -0.0 Head's highest education level is secondary (0.01) (0.01) (0.0) 0.23 0.19 -0.0*** Head's highest education level is tertiary (0.01) (0.01) (0.0) 0.74 0.63 -0.1*** Urban area (0.01) (0.01) (0.0) Note: ***, **, and * denote statistical significance at the 0.01, 0.05, and 0.10 levels. Standard errors are in parentheses. The data are unweighted. Household heads' ages are restricted to between 25 and 55 for the first survey round and adjusted accordingly for the second survey round. The reference groups are household with no primary education and living in rural areas. 87 Table C.10 Descriptive statistics of estimation sample, Mauritania 2004-2008 2008-2014 2014-2019 2004 2008 diff 2008 2014 diff 2014 2019 diff 11.66 12.09 0.4*** 12.10 12.54 0.4*** 12.56 12.68 0.1*** Log of per capita consumption (0.01) (0.01) (0.0) (0.01) (0.01) (0.0) (0.01) (0.01) (0.0) 42.41 43.90 1.5*** 41.90 46.40 4.5*** 42.42 45.49 3.1*** Head`s age (0.10) (0.08) (0.1) (0.08) (0.10) (0.1) (0.10) (0.11) (0.1) 0.16 0.29 0.1*** 0.30 0.30 0.0 0.31 0.37 0.1*** Head is female (0.00) (0.00) (0.0) (0.00) (0.01) (0.0) (0.01) (0.01) (0.0) 0.10 0.11 0.0 0.12 0.12 -0.0 0.14 0.29 0.1*** Head's highest education level is primary (0.00) (0.00) (0.0) (0.00) (0.00) (0.0) (0.00) (0.01) (0.0) 0.13 0.13 -0.0 0.13 0.13 0.0 0.15 0.19 0.0*** Head's highest education level is secondary (0.00) (0.00) (0.0) (0.00) (0.00) (0.0) (0.00) (0.00) (0.0) 0.05 0.05 0.0 0.05 0.05 0.0** 0.05 0.06 0.0 Head's highest education level is tertiary (0.00) (0.00) (0.0) (0.00) (0.00) (0.0) (0.00) (0.00) (0.0) 0.51 0.48 -0.0*** 0.47 0.59 0.1*** 0.59 0.50 -0.1*** Urban area (0.01) (0.01) (0.0) (0.01) (0.01) (0.0) (0.01) (0.01) (0.0) Note: ***, **, and * denote statistical significance at the 0.01, 0.05, and 0.10 levels. Standard errors are in parentheses. The data are unweighted. Household heads' ages are restricted to between 25 and 55 for the first survey round and adjusted accordingly for the second survey round. The reference groups are household with no primary education and living in rural areas. 88 Table C.11. Descriptive statistics of estimation sample, West Bank and Gaza 2007-2009 2009-2011 2011-2017 2007 2009 difference 2009 2011 difference 2011 2017 difference 8.78 9.06 0.3*** 9.06 9.13 0.1*** 9.13 9.23 0.1*** Log of per capita consumption (0.02) (0.01) (0.0) (0.01) (0.01) (0.0) (0.01) (0.01) (0.0) 40.03 40.72 0.7** 39.80 41.65 1.9*** 40.71 44.94 4.2*** Head`s age (0.25) (0.15) (0.3) (0.15) (0.14) (0.2) (0.14) (0.16) (0.2) 0.06 0.06 0.0 0.06 0.07 0.0* 0.07 0.08 0.0 Head is female (0.01) (0.00) (0.0) (0.00) (0.00) (0.0) (0.00) (0.00) (0.0) 0.55 0.50 -0.0*** 0.50 0.51 0.0 0.51 0.51 -0.0 Head's highest education level is primary (0.02) (0.01) (0.0) (0.01) (0.01) (0.0) (0.01) (0.01) (0.0) 0.18 0.17 -0.0 0.17 0.18 0.0 0.18 0.16 -0.0** Head's highest education level is secondary (0.01) (0.01) (0.0) (0.01) (0.01) (0.0) (0.01) (0.01) (0.0) 0.17 0.22 0.0*** 0.21 0.22 0.0 0.22 0.23 0.0 Head's highest education level is tertiary (0.01) (0.01) (0.0) (0.01) (0.01) (0.0) (0.01) (0.01) (0.0) 0.54 0.70 0.2*** 0.70 0.53 -0.2*** 0.53 0.56 0.0*** Urban area (0.02) (0.01) (0.0) (0.01) (0.01) (0.0) (0.01) (0.01) (0.0) 0.18 0.12 -0.1*** 0.12 0.21 0.1*** 0.21 0.12 -0.1*** Refugee area (0.01) (0.01) (0.0) (0.01) (0.01) (0.0) (0.01) (0.01) (0.0) Note: ***, **, and * denote statistical significance at the 0.01, 0.05, and 0.10 levels. Standard errors are in parentheses. The data are unweighted. Household heads' ages are restricted to between 25 and 55 for the first survey round and adjusted accordingly for the second survey round. The reference groups are household with no primary education and living in rural areas. 89 Table C.12. Descriptive statistics of estimation sample, Tunisia 2005-2010 2010-2015 2015-2021 2005 2010 diff 2010 2015 diff 2015 2021 diff 7.23 7.58 0.4*** 7.55 7.98 0.4*** 7.96 8.36 0.4*** Log of per capita consumption (0.01) (0.01) (0.0) (0.01) (0.00) (0.0) (0.01) (0.01) (0.0) 43.55 46.72 3.2*** 44.24 47.23 3.0*** 44.32 48.40 4.1*** Head`s age (0.09) (0.09) (0.1) (0.09) (0.06) (0.1) (0.06) (0.08) (0.1) 0.12 0.11 -0.0* 0.11 0.11 0.0 0.10 0.12 0.0*** Head is female (0.00) (0.00) (0.0) (0.00) (0.00) (0.0) (0.00) (0.00) (0.0) 0.04 0.03 -0.0*** 0.03 0.46 0.4*** 0.44 0.44 0.0 Head's highest education level is primary (0.00) (0.00) (0.0) (0.00) (0.00) (0.0) (0.00) (0.00) (0.0) 0.14 0.11 -0.0*** 0.12 0.30 0.2*** 0.32 0.32 0.0 Head's highest education level is secondary (0.00) (0.00) (0.0) (0.00) (0.00) (0.0) (0.00) (0.00) (0.0) 0.08 0.09 0.0 0.09 0.11 0.0*** 0.12 0.12 0.0 Head's highest education level is tertiary (0.00) (0.00) (0.0) (0.00) (0.00) (0.0) (0.00) (0.00) (0.0) 0.65 0.66 0.0 0.65 0.62 -0.0*** 0.61 0.62 0.0** Urban area (0.01) (0.01) (0.0) (0.01) (0.00) (0.0) (0.00) (0.00) (0.0) Note: ***, **, and * denote statistical significance at the 0.01, 0.05, and 0.10 levels. Standard errors are in parentheses. The data are unweighted. Household heads' ages are restricted to between 25 and 55 for the first survey round and adjusted accordingly for the second survey round. The reference groups are household with no primary education and living in rural areas. 90 Table C.13. Estimated rho () from cross-sectional data Country Period Alternative 2012-2015 0.84 0.52 Egypt, Arab Rep. 2015-2017 0.89 0.46 2017-2020 0.79 0.61 2007-2009 0.54 0.56 West Bank and Gaza 2009-2011 0.62 0.66 2011-2017 0.34 0.59 2005-2010 0.57 0.67 Tunisia 2010-2015 0.73 0.65 2015-2021 0.89 0.61 2004-2008 0.77 0.57 Mauritania 2008-2014 0.63 0.56 2014-2019 0.70 0.61 Iraq 2007-2012 0.68 0.37 Jordan 2010-2013 0.63 0.63 Note: are estimated using Equation (C5) for all countries with cohorts being defined by age interacted with household heads’ education. Alternative ’s are estimated using Equation (C6). 91 Figure C.1. Probabilities of Female-Headed Households Escaping Poverty in Second Year Conditional on Being Poor in First Year (percentage) Egypt 2017-2020 Iraq 2007-2012 75 45 Percentage (%) Percentage (%) 70 40 65 35 60 30 Self-reported Potential Self-reported Potential Females>0.5 Educated fem. Females>0.5 Educated fem. Jordan 2010-2013 Mauritania 2014-2019 55 60 Percentage (%) Percentage (%) 55 50 50 45 45 40 Self-reported Potential Self-reported Potential Females>0.5 Educated fem. Females>0.5 Educated fem. Palestine 2011-2017 Tunisia 2015-2021 40 55 Percentage (%) Percentage (%) 35 50 30 45 25 20 40 Self-reported Potential Self-reported Potential Females>0.5 Educated fem. Females>0.5 Educated fem. FHH wo/children FHH w/children Non-FHH Note: Estimates are obtained with synthetic panel data and weighted with population weights where the second survey round is used as the base year. The figure shows the percentage of the population that moves out of poverty in the second year. FHH characteristics are measured in second period. Dashed red lines represent the national average for each period. Household heads' ages are restricted to 25-55 for the first survey round and adjusted accordingly for the second survey round. Standard errors are estimated with 1,000 bootstraps. ’s are estimated using Equation (C6). 92 Figure C.2. Probabilities of Female-Headed Households Falling in Poverty in Second Year Conditional on Being Non-poor in First Year (percentage) Egypt 2017-2020 Iraq 2007-2012 10 12 14 16 18 20 10 12 14 16 18 20 Percentage (%) Percentage (%) Self-reported Potential Self-reported Potential Females>0.5 Educated fem. Females>0.5 Educated fem. Jordan 2010-2013 Mauritania 2014-2019 18 14 Percentage (%) Percentage (%) 16 12 14 10 12 8 10 6 Self-reported Potential Self-reported Potential Females>0.5 Educated fem. Females>0.5 Educated fem. Palestine 2011-2017 Tunisia 2015-2021 25 30 35 40 45 16 Percentage (%) Percentage (%) 14 12 10 8 Self-reported Potential Self-reported Potential Females>0.5 Educated fem. Females>0.5 Educated fem. FHH wo/children FHH w/children Non-FHH Note: Estimates are obtained with synthetic panel data and weighted with population weights where the second survey round is used as the base year. The figure shows the percentage of the population that enters poverty in the second year. FHH characteristics are measured in second period. Dashed red lines represent the national average for each period. Household heads' ages are restricted to 25-55 for the first survey round and adjusted accordingly for the second survey round. Standard errors are estimated with 1,000 bootstraps. ’s are estimated using Equation (C6). 93 Appendix D: Further Analysis with Equivalence Scale Given our consistent observation of self-reported FHHs having an advantage in terms of greater mobility, the question arises as to whether this conclusion remains valid if we extend to selecting a measure of household members' welfare that goes beyond household expenditure per capita. Within the context of poverty dynamics, we show two scenarios for self-reported FHHs as an example: one in which FHHs have a greater probability to escape poverty compared to non-FHHs (denoted by the orange-shaded region in Figure D.1), and another where FHHs are less likely to escape poverty than non-FHHs (represented by the blue-shaded area in Figure D.1). Importantly, the selection of specific scale parameters can significantly alter the conclusions drawn regarding poverty dynamics among FHHs. In particular, when assessing consumption on a per capita basis (i.e., when β=1 and θ=1), self- reported FHHs consistently exhibit a higher probability of escaping poverty than non-FHHs and it holds true across all countries. Intriguingly, these findings align with those in Abanokova et al. (2022), which demonstrated a persistent upward mobility when income is evaluated on a per capita basis. The conclusions regarding poverty dynamics shift when adopting OECD-recommended (modified) equivalence scales, which assign a value of 0.3 to each child aged 0-13 (indicated by the green dashed line) and/or the "square root scale" set at 0.5 (represented by the red dashed line). Under the "square root scale," self-reported FHHs become less likely to escape poverty than non- FHHs in Jordan, the West Bank and Gaza, and Tunisia, regardless of the child parameter value. The use of a lower scale parameter than the “square root scale” alters the conclusion in the Arab Republic of Egypt, but the sensitivity to the child parameter is also observed. Significant sensitivity to the child parameter is found in the West Bank and Gaza. When the child parameter is set to 0.4 or lower, there is a shift in the scenario from FHHs experiencing upward mobility to FHHs facing downward mobility. However, varying the parameters of economies of scale and child parameters from 0 to 1 does not alter the conclusions regarding poverty dynamics for Mauritania, Iraq and Jordan. The absolute difference in the percentage of the population transitioning out of poverty between FHHs and non-FHHs is also influenced by the scale parameters. In the case of Mauritania, where self-reported FHHs are more likely to escape poverty than non-FHHs, fluctuations in scale parameters can result in significant changes in the percentage of self-reported FHHs escaping poverty. These variations can yield a discrepancy of up to 6.8 percentage points, depending on the scale parameters applied. The overarching finding is that the parameter dictating the economies of scale and the private– public nature of household consumption contributes non-trivially to the poverty ranking between FHHs and non-FHHs across most countries and FHH definitions, while the child parameter having a comparatively smaller impact compared to household size. These results mirror our earlier observation in Abanokova et al. (2022) regarding the sensitivity of income dynamics to scale parameters. 94 Figure D.1. Self-reported FHHs– non-FHHs Differences in Probabilities of Escaping Poverty in Second Year Conditional on Being Poor in First Year (percentage points), by Scale Parameters Note: Each figure shows 2-parameter equivalence scale that adjusts household consumption: ( + ) where – number of adults in the household, – number of children in the household, is “child parameter” that accounts for the needs of children aged 0-13 and is “size parameter” that measures the degree of economies of scale in household consumption. Both parameters are varying between 0 and 1. The blue zone indicates lower probabilities of escaping poverty among FHHs compared to non-FHHs. The orange zone indicates higher probabilities of escaping poverty among FHHs compared to non-FHHs. Each bar shows the difference in the percentage of the population that moves out of poverty among FHHs compared to non-FHHs in the second year (expressed in percentage points). We use OECD recommended (modified) equivalence scale that assigns a value of 0.3 to each child aged 0-13 (green dashed line) and “square root scale” that equals to 0.5 (red dashed line). The top right corner of the box (marked ×) illustrates the case when δ = 1 and = 1, which represents per capita expenditure (“Per Capita”). 95