From gender issues to gender statistics on household-level income/consumption poverty: illustrative examples
Sources of data
What types of female-headed households have a higher risk of poverty? Do they fare better or worse than similar households headed by males? Are lone mothers with children more likely to be poor than lone fathers with children? Are old women living alone more likely to be poor than old men living alone?
Equivalized household income/consumption disaggregated by detailed types of household and sex of the head of household; poverty threshold.
Household surveys designed to measure poverty, such as living standards surveys, household income and expenditure surveys, household budget surveys, etc.
170. Certain types of households headed by women are more vulnerable to poverty than those headed by men (United Nations, 2010). This is the case, for example, of households of lone mothers with children compared to those of lone fathers with children. Likewise, households of women living alone tend to be poorer than those of men living alone. The higher risk of poverty for these types of households headed by women is linked to the gender gap in access to economic resources. For example, women participate less than men in the labour market and when they do participate, their earnings are usually lower than men's (United Nations, 2010). Older women are covered less than older men by the pension system and when they are covered, the pension that they receive is much lower than the pension that men receive (United Nations, 2009b). Lower access to economic resources increases women's economic dependency on men. When men with higher earnings are not around anymore because of divorce, migration or death, women as lone mothers and older women living alone have a higher risk of poverty. Households of lone mothers and single women may also be more vulnerable to economic or environmental shocks, owing to the obstacles faced by women in accessing the economic resources necessary for recovery and survival (United Nations, 2009b).
171. However, an overall higher vulnerability to poverty for female-headed households compared to male-headed households cannot be generalized, mainly because female- and male-headed households cover a broad range of situations in terms of demographic and economic composition (Lampietti and Stalker, 2000; United Nations, 2010). Examples of female-headed households include one-person households, such as older women with a small or no income and young women who are economically independent, households of lone mothers with children who may or may not receive financial support from the father, households where the male partner is temporarily absent and contributes remittances, households of visiting unions where the male partner is polygamous and does not provide economic support and households of couples with or without children where the woman rather than the man is reported as the household head. Because of their different demographic and economic composition, these types of female-headed households tend to have a different risk of poverty. Similarly, different types of male-headed households have a different risk of poverty depending on their demographic and economic composition. Furthermore, as the share of the detailed types of female- and male- headed household in total households varies across countries and within countries, the overall gap between the poverty of female-headed households and that of male-headed households also varies, to the extent that, in many developing countries, the overall poverty incidence of female-headed households is lower than that of male-headed households (Lampietti and Stalker, 2000; United Nations, 2010).
172. Household-level consumption or income adjusted for differences in the age and sex composition of households and the poverty line chosen to separate poor from non-poor are the basis for measuring poverty. On the basis of this information, several measures of poverty can be constructed. Those frequently used are headcount poverty index, poverty gap (or poverty depth) index and severity of poverty index.3
173. Poverty measures should be calculated separately for various types of female- and male-headed households. These types of households can be identified on the basis of the demographic composition of the household or a combination of the demographic and economic characteristics of the household members. For instance, households of one female or male adult with or without children, such as lone mothers with children, lone fathers with children, female one-person households, male one-person households, older female one-person households, older male one-person households, can be identified on the basis of information on sex and age for all household members and family relations among them.
174. Depending on the context, additional types of female- and male-headed households can be distinguished on the basis of the individual demographic, social and economic characteristics of all household members. For example, based on marital status, a distinction can be made between female heads of households who do not have a partner, female heads who are in visiting unions and whose partners are usually living in another household and female heads whose formal partners are temporarily absent. Similarly, a distinction can be made between monogamous and polygamous male-headed households.
175. Taking into account the age and economic characteristics of all members of the household, households may be classified further by the level of dependency burden. Statistics on the age, education and economic characteristics of the household head can be used for more detailed analysis. Additional breakdowns commonly used in poverty analysis, such as urban/rural areas or geographical areas, are also useful.
Sources of data
176. Household surveys, such as household income and expenditure surveys, household budget surveys, household integrated surveys and living standards/living conditions surveys, are the main source of income or consumption data. However, these surveys are not equally adequate for poverty measurement. In general, shorter recall periods tend to be sufficient for the estimation of averages of income or expenditure, whereas longer recall periods are often needed for adequate poverty estimates. Living standards surveys are one of the types of surveys used to collect data on consumption that are based on longer recall periods. They are therefore more adequate for poverty analysis. Living standards surveys are also used to collect more comprehensive information on the individual characteristics of all household members, thereby allowing for an analysis of household-level poverty by detailed types of female- and male-headed households as well as a broader analysis of non-consumption individual-level indicators of poverty.
Conceptual and measurement issues
177. A lack of explicit criteria in identifying the household head impacts the adequacy of poverty statistics for female- and male-headed households. The traditional notion of head of household assumes that one person has primary authority over and responsibility for household affairs and is, in the majority of cases, its chief economic support. However, where spouses are considered equal in household authority and responsibility and may share economic support, the concept of head of household is no longer considered valid. In countries where this is the case, either the concept should no longer be used or provisions for joint headship should be made. Even in the many countries where the traditional concept of head of household is still relevant, it is important to recognize that the procedures followed in applying the concept may distort the true picture, particularly with regard to female heads of households. The most common assumption that can skew the facts is that no woman can be the head of a household that also contains an adult male. This fact is often neglected, resulting in a biased interpretation of the association between gender and poverty.
178. The use of different criteria in defining the household headship leads to the identification of different sets of households with different poverty rates (Fuwa, 2000). The criteria used to identify a household head should be clearly specified in the survey design so that they are the same for all households surveyed. Depending on the criteria selected, the identification of the household head can be done at the time of the interview or at the analysis stage. For example, at the time of the interview, the head of the household may be defined as the person considered by the household members as the main economic provider with most authority and decision-making power on economic resources, while at the analysis stage, the head of the household may be defined as the person with the highest income or the person with a regular stable income. Analysis of poverty differences between female- and male-headed households based on self-reporting should be avoided when the respondents interviewed were not given any criteria for identifying the household head.
179. The current practice of disaggregation of household-level poverty data by sex of the household members gives only a poor measure of the gender gap in poverty. Poverty is traditionally measured on the basis of income or expenditure at the household level, whereas the number of poor people (women or men) is calculated as the number of people living in households below a poverty line. The inequality within the household in satisfying individual basic needs is not taken into account, mainly because it is difficult to measure how household income is spent or consumed on an individual basis and how expenditures are distributed to each household member. The results of disaggregation of household-level poverty data by sex of the household members will not reflect possible gender inequality within households, only the distribution of population by sex in poor households. If, in the same household, the women consume or spend less than what they need to function properly physically and socially (and are therefore considered poor), while the men consume or spend what they need or more (and are therefore considered non-poor), they will still be considered to have the same poverty status, either poor or non-poor, depending on the average consumption estimated at the household level. This approach may lead to the undercounting of women in poverty, because additional poor women might be found in some non-poor households.
180. In addition, the gender gap measured on the basis of simple poverty counts by sex is heavily influenced by country-specific living arrangements and ageing factors (United Nations, 2010). Poverty rates for women may appear higher than those for men, especially in countries with a significant proportion of households with an overrepresentation of adult women, such as households of lone mothers with young children and female one-person households, in particular one-person households of older women. Such countries are more likely to be located in the more developed regions, where, by other standards of well-being, such as education and health, women enjoy an improved status and less gender discrimination.
3 For more information on measures of poverty, their definition and method of calculation, see Haughton and Khandker (2009).