Menu
Menu
Menu
Menu
Background: The global poverty profile shows that Africa and Asia bear the highest burden of multidimensional child poverty. Child survival and development therefore depend on socioeconomic and environmental factors that surround a child.The aim of this paper is to measure multidimensional child poverty and underpin what drives it among children aged 5 to 18 years in a resource poor region of Burkina Faso. Methods: Using primary data collected from a cross sectional study of 722 households in the Mouhoun region of Burkina Faso, the Alkire-Foster methodology was applied to estimate and decompose child poverty among children aged 5-18 years. Seven broad dimensions guided by the child poverty literature, data availability and the country’s SDGs were used. A binary logistic regression model was applied to identify drivers of multidimensional child poverty in the region. Results: The highest prevalence of deprivations were recorded in water and sanitation (91%), information and leisure (89%) followed by education (83%). Interestingly, at k = 3 (the sum of weighted indicators that a child must be deprived to be considered multidimensionally poor), about 97% of children are deprived in at least three of the seven dimensions. At k = 4 to k = 6, between 88.7 and 30.9% of children were equally classified as suffering from multidimensional poverty. The odds of multidimensional poverty were reduced in children who belonged to households with a formally educated mother (OR = 0.49) or stable sources of income (OR = 0.31, OR = 0.33). The results equally revealed that being an adolescent (OR = 0.67), residing in the urban area of Boromo (OR = 0.13) and rural area of Safané (OR = 0.61) reduced the odds of child poverty. On the other hand, child poverty was highest among children from the rural area of Yé (OR = 2.74), polygamous households (OR = 1.47, OR = 5.57 and OR = 1.96), households with an adult head suffering from a longstanding illness (OR = 1.61), households with debts (OR = 1.01) and households with above five number of children/woman (OR = 1.49). Conclusion: Child poverty is best determined by using a multidimensional approach that involves an interplay of indicators and dimensions, bearing in mind its causation.
Located in the northwest of Burkina Faso, the Boucle du Mouhoun region (12°30′N; 3°30′W) whose headquarter is Dédougou, occupies about 12.6% of the national territory (about 34,333 km2). It encompasses approximately 10% of the total population of the country with a density of 53 inhabitants per square kilometre [24]. The country is bounded by the Republic of Mali, Ghana, Niger, Togo and Cote d’Ivoire. Boucle du Mouhoun region was selected for several reasons including (i) the high prevalence of poverty (one of the poorest region in the country); (ii) higher proportion of young people under the ages 15 years (49.9%) and 25 years (68.1%); (iii) the negative migration balance: the region is a ‘hot spot’ for intra and inter rural/urban migration in the country with far reaching implications for national development. The study used primary data from a UNICEF-Save the Children sponsored project on child poverty profiling and vulnerability in Burkina Faso. Five communities, including two urban communities (Dédougou and Boromo), and three rural communities (Safané, Kona and Yé) were purposively selected. These communities are located in three provinces (Balé, Mouhoun and Nayala) of the Boucle du Mouhoun region (Fig. 1). The five communities were selected on the basis of high poverty incidence as recommended by the office of UNICEF Ouagadougou. A total of 20 enumerated areas (EAs) were randomly selected from all 5 communities. Given that 60% of the population of the region were from rural areas, 12 out of 20 EAs were randomly selected from Safané (6), Kona (2) and Yé (4), while 8 EAs were selected from the urban areas of Dédougou (6) and Boromo (2). Overview of the Study Area. The map was created by the authors showing the Boucle du Mouhoun region and its three provinces, Nayala, Mouhoun and Bale. The survey was conducted in the two urban areas of Boromo and Dédougou, and three rural areas of Kona, Safané and Yé The required primary sampling units for the EAs were numbered, and households were then randomly selected. The sample size calculation was based on the number of children including teenagers in the region, making a total of approximately 81,818 children. Applying the Taro Yamane formula [25] with a 5% margin of error, the minimum sample required for the study was calculated as follows: Where n is the sample size to be estimated, N is the population size and e is the error margin (e = 0.05).Based on this specification, we obtained a minimum sample size of about 794 children aged 0–18 years. We avoided the error in non-response by adjusting the sample size by 20%. This resulted to a sample size of 952 respondents, approximated to 1000 respondents aged 0–18 years. However, the inclusion criteria for this study involved children aged 5 to 18 years, which reduced the sample size to 722 children. The survey used a structured interviewer administered questionnaire divided into three parts; a section for household characteristics, children’s characteristics and mother’s characteristics. The household heads were directed to the sections specified for household characteristics. The mothers responded to their specific sections and the children’s section if the child was below 10 years of age. Adolescents responded to the children’s section with occasional interventions from the mothers when needed. The questions in the study tool were adopted from developing countries National Living Standard Survey Measurements (NLSS), OPHI modules, ‘Bristol Approach’ by UNICEF, Multiple Indicator Cluster Surveys (MICs) by Alkire and Foster including other national surveys in Burkina Faso. The questions were equally adapted to suit the content of the study. Data quality was ensured by doing a pilot study to test the survey instruments and identify potential errors for corrections. There were in total 11 trained graduates as enumerators and 3 field supervisors. The field supervisors in the beginning conducted 4 interviews per day with the enumerators to monitor their progress and check for data inconsistencies. Data entry was simultaneously done alongside data collection in case errors were identified. Digitalizing the data minimized error risks during data processing, i.e. entering the correct codes for the responses. This was done using data capture mask designed with Census and Survey Processing System (CSPro) software package version 5.0. Data was cleaned and analysed using SPSS (IBM SPSS Statistics for Windows, Version 20.0) and STATA 13. Measuring multidimensional child poverty and deprivation requires the identification of relevant dimensions in relation to public ideals. In this study, seven broad dimensions were identified for the multidimensional frame work as shown in Table 1. The selected dimensions were specifically chosen to capture progress in the country’s MDGs. These include nutrition, health, education, water and sanitation, housing, information and material deprivation (per capital income). Each of the dimensions were measured using well-defined indicators drawn from the literature on child poverty [26]. Note that each indicator was assigned equal weights assuming that each counts equally in a child’s wellbeing and development in the society as suggested by the Convention on the Rights of the Child [9]. Dimensions specific deprivation cutoffs and weights for children aged 5–18 years Family income is used as a dimension because a stable income provides family security and influences child development and growth [27, 28]. In using multiple dimensions to define child poverty, it is important to include the dimension, income that offers command over non-market goods [29]. Children from low-income households in Burkina Faso run the risk of engaging in child labour activities like mining, hawking to supplement the family budget. By so doing, they are often exposed to the risk of unwanted pregnancy, juvenile delinquent behaviours and poor school attendance increasing the tendency to drop out of school. In this analysis, children are considered deprived in income if they come from households that fall within the last two quintiles of household per capital income distribution. A second dimension is housing. A child’s dwelling can affect his psychosocial well-being as well as expose him to certain health risks [30]. Burkina Faso is a country with very harsh weather conditions especially during the rainy seasons where houses are often flooded with water and debris, increasing the risks of infectious disease spread. Individuals may lose poorly built homes to strong winds and flooding, putting the family at risk of migrating and squatting from one home to another. Additionally, electricity is an indicator included in this dimension not only because it offers some form of family satisfaction but rather a booster to a child’s performance at school. A child is considered deprived in housing if he or she lives in a house without electricity, or the house is not made of formal roofing or wall construction materials, or sleeping in an overcrowded house (i.e., 4 or more persons per room). Person’s per room is a measure of the indicator overcrowding and has been a subject of debate for over a decade [31]. Some scholars refer to it as an objective variable that must take into consideration the age difference of occupants in the room, the space and size of the room. What others see as overcrowding may not necessarily be overcrowding in another context. This study uses the UN definition and other previous studies on child poverty to define overcrowding. That is, 4 or more persons living in a tiny room thereby increasing the risk of infectious disease spread and violence [32, 33]. The dimension water and sanitation include provision of clean drinking water and availability of improved toilet facilities. These are the most basic and cost-effective ways of improving health in impoverished communities. Children are deprived in water and sanitation if they use unprotected well/rainwater or river/stream/ lake /pond as main water source and have no toilet facilities or share toilet, use unimproved pit latrines or practice open defecation. We used the Composite Index of Anthropometry Failure (CIAF) to assess the nutritional status of children by forming a composite nutritional index, Under-nutrition [34]. For older children (5–9 years), the recommended nutritional assessment is BMI-for-Age (BAZ), an indicator for Wasting or Thinness and Height-for-Age (HAZ), an indicator for Stunting [35]. The World Health Organization’s Anthoscore software was downloaded and used to construct the indicators for Wasting and Stunting. A child is defined as suffering from Wasting or Stunting if he falls − 2 standard deviations (SD) below the referenced population mean. A child was suffering from Undernutrition if he was either wasted or stunted or suffering from both (Table 2). Composite index for anthropometry failure (CIAF) To reduce maternal and child mortality rates in the country, a free health initiative for pregnant women and children under-5 years of age was implemented in the country in 2015/2016. Prior to this, the Integrated Community Case Management (iCCM) of childhood illnesses has been a strategy implemented at community level to provide healthcare services in hard to reach areas. This intervention aimed at improving access to healthcare services and thus improve child survival. However, it mainly focuses on children under 5 years while little is known about health care access for older children. It is therefore interesting to determine the extent of healthcare access among older children who rely on out of pocket payment for medical expenses. A child is considered deprived in health if he or she did not get healthcare when last needed or if the child is from a household with an incidence of child mortality. With regards to information, children need the media to improve on their intellectual capacity as well as shape certain behavioural norms. It is essential for children to live in households with access to phones especially for school emergencies. A child is thus, classified as deprived in information if he or she lives in a household without radio or television or from a household the lacks access to a mobile phone. The importance of child education cannot be overemphasized as it improves an individual’s social status and standard of living later in life. It is not enough to enrol a child in school but equally important to monitor school frequency and school dropout among this vulnerable age group. A child is considered deprived in education, if he or she is a school drop-out, or was not enrolled in a school, or was enrolled in school late (age 7 years and above) or does not go to school daily. In this section, we calculate the multidimensional poverty index (MPI). The first question asked is, who are the poor? Bourguignon and Chakravarty [36] identified the poor as those deprived in any of the dimensions being explored. While this is a useful place to start, it does not look across dimensions to label individuals as poor. Alkire and Foster (AF) [23] use a more practical approach in measuring multidimensional poverty, which takes into consideration the number of dimensions an individual is deprived. The two methods of identifying the poor include the union and the intersection approach. In the union approach, an individual is considered poor if deprived in at least one dimension. This is theoretically intuitive but practically improbable because almost everyone will be considered poor if studying a large population. Thus, it represents a bias of inclusion [37]. The latter on the other hand considers a person as poor if deprived in all dimensions. Again, this method fails to identify persons who are deprived in certain dimensions and not in the other. For instance, a healthy child may not be considered poor if he did not go to school or lives in a low-income household. There is therefore the tendency to underestimate poverty. These two approaches are balanced in the AF’s dual cut-off approach, which builds on Sen’s two basic principle namely; identifying the poor and constructing an index to determine the extent of poverty [38]. As the name implies, two cut-offs are established to define multidimensional poverty. First is the deprivation cut-offs, that determines if a person is deprived in any of the dimensions and then the poverty cut-off which determines how extensively deprived a person should be to be considered poor [39]. The AF’s methodology goes through a series of steps namely, defining the indicators used, setting the level of deprivation cut-offs for each indicator, assigning equal or differential weights to the indicators and summing each up to one, ascertaining if an individual is deprived or not, creating a weighted sum of deprivations for everyone and lastly determining the poverty cut-off that identifies an individual as multidimensionally poor. This phase is also known as the identification phase. The next phase, aggregation phase, calculates the following; the head count ratio (HO) which identifies the proportion of individuals who are multidimensionally poor, the intensity of multidimensional poverty (A) defined as the average share of weighted indicators in which poor children are deprived in, the adjusted head ratio (MO) calculated as product of the head count ratio and intensity of multidimensional poverty (HxA). To compute the multidimensional measures, the paper uses the cut-off value k, which by definition, is the sum of weighted indicators that a child must be deprived in order to be considered multidimensionally poor [40]. It is equally seen as a policy variable describing the range of deprivations each poor child must have to be classified as being deprived. Following Alkire and Santos definition, a child is considered multidimensionally poor if the weighted indicator (k) of which he or she is deprived is greater than or equal to 33.3% [41]. In this paper, we differentiate three broad categories of poverty based on similar precepts. That is; The Non-Poor Children (k = 1), Children who are Vulnerable to Poverty (k = 2), and Multidimensionally Poor Children (k ≥ 3) [42]. We identify factors associated with Multidimensional Poverty using binary logistic regression models at 5% level of significance. The dependent variables for the binary models are poverty/deprivation used in computing the headcount (HO) for each of the poverty construct (k = 3, k = 4, k = 5 and k = 6). Four models were used to obtain a comprehensive picture of drivers of child poverty. A deprived child has a value of ‘1’ while a child who is not deprived has a value of ‘0’. Some of the predictors explored in this analysis include the age of household head, adult health, child age, area of residence, household size, education status of household head, marital status and household debts status among others. It is hypothesized from the literature that these household characteristics were associated with child poverty. The explanatory variables were measured thus: Adult health and mother’s health status were defined as those diagnosed with longstanding illness in the past 12 months. The conditions assessed were Diabetes, Asthma, Low Back Pain, Hypertension, Angina, Depression, Arthritis, Chronic Obstructive Pulmonary Disease, Cancer and Others to specify. Those in the ‘others category’ with HIV and other long-term ill-health from unknown cause were included in the yes group. Adults and mothers with longstanding illness were coded ‘1’ meaning ‘Present’ and those without as ‘0’ meaning ‘Absent.’ An indebted household was defined as households where the household head or other members of the household were in debt. This measure was included because previous studies show that households with unmanageable debts have higher chances of compromising a child’s general well-being [43, 44]. Given that over 80% of the indigenes in the region were engaged in agriculture, income sources were categorized into three categories. Those who had never worked or having any source of income what so ever (no income), income derived from non-farm activities in both private and public sectors including transfers (non-farm incomes), and incomes from agricultural activities (farm incomes). Adult education status was grouped into no formal education, formal education (primary, secondary or tertiary school attendance) and informal education (koranic or adult education). A household size of below eight members was considered normal in the African setting where nuclear families often live with extended family members. In Burkina, the average rural household size is normally 8 persons [45]. Despite global decline in fertility rates, SSA still experience a slow decline in fertility rates. The average woman in SSA desire to have 4 to 5 children [46]. It is on this precept that number of children per woman was coded ‘0’ if the number of children/woman was between 1 and 5 and ‘1’ if 6 and above.
N/A
