Introduction

Poverty is a multidimensional, dynamic, complex, gendered and location specific phenomenon (World Bank 2000). Due to its multidimensional nature, different disciplines adopt different approaches. A geographer would approach the subject from the spatial perspective, location–allocation and distribution of infrastructures, access to welfare services and information. Identifying spatial patterns of poverty status provide new insights into the socio-spatial dynamics of poverty, poverty severity. Poverty rates and the dynamics differ considerably in different parts of the country and the number of poor Nigerians living below the poverty line has grown measurably (NBS 2010). Although, poverty is a rural phenomenon but the share of the poor living in urban areas is rising with urbanization, both the quantitative and qualitative measurements attest to the growing incidence and depth of poverty in the country (Okunmadewa et al. 2005).

This poverty situation however presents a paradox, considering the vast human and physical resources that the country is endowed with. Alayande and Alayande (2004) states that Nigeria suffers from high levels of poverty in spite of her enormous wealth of human and material resources. The National Bureau of Statistics (2012) report indicates that about 112.52 million Nigerians live below the poverty line. This represents about 68.7% of the Nigerian population. Poverty has been massive, pervasive, and has engulfed a large proportion of the Nigerian society. The scourge of poverty in Nigeria is an incontrovertible fact, which results in hunger, ignorance, malnutrition, disease, unemployment, poor access to credit facilities, low life expectancy as well as a general level of human hopelessness Abiola and Olaopa (2008). Despite government’s effort to reduce poverty incidence through poverty alleviation programmes, strategies and the quest to be one of the 20 largest economies by the year 2020, Nigeria continues to be one of the poorest countries in the world.

Previous efforts at analyzing poverty used uni-dimensional measures and did not investigate issues of socio-spatial dynamics of urban poverty, poverty severity at the household level using a single cross-section in a multi-dimensional approach. Researches on multi-dimensional poverty have largely focused on the socio-economic determinants of poverty in Nigeria. For instance (Akerele and Adewuyi 2011; Edoumiekumo et al. 2013; Oyekale and Oyekale 2013; Ifelunini et al. 2013; Alkire and Housseini 2014) provides an overview of multidimensional poverty—levels and trends—in sub-Saharan Africa (SSA), using the most recent estimations and analyses of the global multidimensional poverty index (MPI). Apata et al. (2010) and Adeoti (2014) examined the trends and determinants of poverty in rural Nigeria. Yakubu et al. (2014) focused on multidimensional poverty analysis and informal sector in Nigeria. Deinne and Ajayi (2017) examined the spatial dynamics of urban poverty in Nigeria. While these studies have focused on patterns and determinants of poverty, the severity of multi-dimensional poverty at the household level measured multi-dimensionally has not been given much attention.

This paper examines the socio-spatial factors influencing the variation in the severity of multi-dimensional poverty at the household level. Specifically, this paper examines the socio-spatial factors (location, accessibility, age at marriage, household structure, health status, employment status, ownership of assets, livelihood loss, level of education and remittances) influencing the severity of multi-dimensional poverty and determines the severity of multi-dimensional poverty among households in Delta state, Nigeria. The only hypothesis tested is that poverty severity experienced vary significantly among households in the study area.

Apart from this introduction, the rest of the paper is in five sections. Section two provides the theoretical and conceptual clarifications. The study area and methodology are provided in sections three and four while section five focuses on the various dimensions of poverty. The conclusion is provided in section six.

Theoretical and conceptual clarifications

The theoretical basis for this paper is derived from the sustainable livelihood framework in the absence of a grand theory to explain socio-spatial dynamics of multidimensional poverty. Conceptual clarification of approaches to poverty is also undertaken in this section.

Sustainable livelihood framework (SLF)

The sustainable livelihood framework which comprises people’s assets (natural, physical, human, financial, social and information) and the complementary building blocks for their livelihoods in Fig. 1 provides the theoretical basis for this paper. Livelihood assets provide both a proxy for absolute poverty and a broader poverty measure (Erenstein 2009). Chambers and Conway (1992: 1) noted that “a livelihood comprises the capabilities, assets and activities required for a means of living”. The approach serves as an instrument for the investigation of poor peoples’ livelihoods while visualizing the main factors of influence which can be understood by qualitative and quantitative analysis at the local or communal level, regardless of where (i.e. which sector or geographical space,…) these factors occur.

Fig. 1
figure 1

Source: Adapted from Ashley and Carney (1999) and DFID (1999)

Sustainable livelihoods framework (SLF). NR natural resources based livelihood strategies (farming, fishing, mining etc.); Non-NR non-natural resources based livelihood strategies (business, civil servants etc.).

Full size image

The five building blocks of sustainable livelihood

Sustainable livelihood is multi-dimensional, comprehensive and people-centred (Chang and Tipple 2009). The livelihood assets were conceptualised under the five building blocks of: physical asset/natural asset (shelter, energy, accessibility, sanitation and safe water, health care); social asset (membership of co-operative society, family support, relationship, friendship); financial asset (access to credit, loans, regular flow of money, savings and remittances); and human asset (knowledge, good health, skills, leadership). A major critique of sustainable livelihood is that there are too many components to address. Therefore, it has made it impossible to go in depth with any of these livelihoods thereby making the framework too broad and superficial (Clarke and Carney 2008).

The theory implies that access to the livelihood assets is mediated and influenced by transforming structures at different levels of government, private sector and civil society, processes such as laws and policies. Loan schemes and poverty alleviation strategies are also perceived to be contributing factors to household poverty status and poverty severity (see Fig. 1). Inadequate access to the assets and interest free loan schemes as a results of lack of information, lack of geographic access due to the swampy terrain and environmental issues, conflicts, illiteracy and ignorance lead to poor health, high incidence of poverty and multiple deprivations experienced. In urban areas, the burden of increased demand and high expenditure on services has effects on access to education, nutrition, housing, health care and poor sanitations.

Conceptual clarification of approaches to poverty

Monetary and sustainable livelihood approaches are very central to dynamic and multi-dimensional poverty studies.

Monetary approaches to poverty

Monetary approaches are mostly used with poverty lines and can either be income-based or consumption-based. Currently, there are two estimates of income-poverty in wide use for assessing poverty especially in poor countries. The use of US $1.25 a day by the World Bank is meant to encapsulate what income is required to purchase the absolute basics; the US $2.00 a day poverty line is estimated by factoring in national poverty line data especially from lower–middle-income countries, thereby to some extent taking a less severe view on what are the socially and culturally acceptable minima. Given that exchange rates do not accurately reflect the differential purchasing power across countries, different purchasing power estimates are commonly made and factored into the raw income figures to produce income level based on purchasing power parity (PPP).

Sustainable livelihoods approaches to poverty

Sustainable livelihoods became the focus of Department for International Developments (DFID) poverty alleviation policy (Solesbury 2003a, b). Sustainable livelihood approach is participatory and multi-dimensional in nature. According to Geiser et al. (2011), DFID explicitly aimed at “a refocus on assistance to the poor”. The pro-active, self-help image of the poor, the sustainable livelihoods thinking fits very well with the study of multi-dimensional poverty. The livelihood approach focused on how people organised their lives, on opportunities and on agency. The livelihood approach adopted in Delta State was also strongly motivated by the need to develop more effective poverty alleviation policies such as the introduction of multidimensional poverty reduction policies and strategies to help improve peoples’ livelihood and reduce poverty. More effectiveness was expected to come from bottom-up and participatory methods, putting emphasis on poor people’s lives and daily needs rather than from the top-down interventionist methods practiced so widely. In this respect, the livelihood approach was much indebted to the work and inspiration of Sen (1981); Chambers et al. (1989); Chambers and Conway (1992). It includes complex, diverse and dynamic strategies developed by households to meet their needs Gaillard et al. (2009).

Study area

Delta State is named after the delta of the River Niger and is situated in the region known as the Niger-Delta. The state is ethnically diverse and lies approximately between longitude 5°00′ and 6°45′ East of the Greenwich meridian and latitude 5°00′ and 6°30′ North of the Equator. The state is bounded by Ondo State to the northwest, Edo State in the north, Anambra State and River State to the east, Bayelsa State to the south and on the south western flank is the Bight of Benin. The choice of the study area is based on the fact that despite being one of the main source of oil exports for the country, this region suffers from poor infrastructure, sanitation, access to healthcare, and educational opportunities, difficulty in satisfying household needs as well as inadequate investment in development. Environmental degradation from oil exploration and exploitation has contributed to and exacerbated a number of these problems. Among the states in Nigeria, Delta State has the highest level of unemployment with 20.8% in 2010 which increased to 27.2% in 2011, (NBS 2012), as unemployment increases wellbeing decreases more persons fall into poverty and increased deprivations (see Fig. 2).

Fig. 2
figure 2

Study area

Full size image

Research methodology

A research survey design was adopted, while multi-dimensional poverty was measured based on deprivation headcount index of the global multi-dimensional poverty (MPI) along education, health and living standard.

Sampling technique

A multi-stage sampling technique involving four steps was used to select households for the survey. At the first stage, Delta State was purposively selected, while the twenty-five (25) local government areas that make up the state were involved in this study. At the second stage, a town in each local government area was selected by systematic random sampling. At the third stage, 2521 households involved in the study were randomly selected while at the fourth stage, the questionnaire was administered to household heads in each of the selected households. Households in each local government area across the state were further re-classified into three senatorial districts to ease analysis and interpretations.

Measurement of poverty indicators

There are three main ways of measuring what happens to poverty over space and time. They include appropriate questions in a single cross-section survey, repeated cross-sectional surveys, and panel data. Given the attrition and paucity of longitudinal survey data needed for this kind of analysis, this paper adopted cross-section survey to generate data on the sampled households at a single point in time. The household survey was preceded by a pilot survey and training of field assistants to enhance the validity of survey instruments.

Multidimensional poverty index (MPI)

The multidimensional poverty index (MPI) was adopted. The rationale for the use of MPI thresholds are diverse: MPI assesses the nature and severity of poverty at the individual and household level, with poor people being those who are deprived and the extent or severity of their poverty being measured by the extent of their deprivations. MPI incorporates key services such as water and sanitation, electricity, education level, and housing which are not consistently captured in all income and consumption surveys. The MPI has ten indicators, two each for health domain and education domain, while six for livelihood asset domain (UNDP 2012).

Furthermore, the chosen indicators were affected by the available household survey data derived from urban households in the study area. Multidimensional poverty construct is measured through the aggregation of different deprivation variables experienced by the households. The index reflects deprivations and various patterns of poverty severity and show aspects in which they are deprived and help reveal the interconnections among deprivations.

Instead of the ten (10) indicators used in the global MPI, twenty-three (23) indicators were used in this paper, three proxy indicators of health deprivation were used under the health dimension (health status/health related losses experienced and nutrition measured as consumption expenditure of household heads), and two indicators of education deprivation under the education dimension measured as (primary education and no formal education), while eighteen (18) indicators were used within the standard of living dimension such as: unimproved cooking fuel, unimproved sanitation, unimproved energy source, lack of access to safe clean water, accessibility/good road, quality housing, poor sanitation, livelihood assets, unemployment, dependence on financial assistance and remittances for survival.

These deprivations are defined based on the Harmonised Nigeria Living Standard Survey (2010); Millenium Development Goals (2010); UNDP (2010) definitions. While an equal weight of 1 was assigned to all the variables used in the multiple deprivation index. The MPI was developed by Alkire and Santos (2010) and presented in the Human Development Report (2012).

Calculation of headcount ratio (H) and poverty severity (A)

The MPI is calculated as the product of two numbers: The percentage of people involved and the “average intensity of deprivation” which reflects the number of dimensions in which households are deprived. The proportion of the population that is multi-dimensionally poor determines the incidence of poverty or headcount ratio (H), while the proportion of indicators in which the poor are deprived determines the severity of their poverty (A).

The technique for calculating the MPI poor is given by:

$${\text{MPI}}\, = \,{\text{H}} \times {\text{A}}$$
(1)

where H = q, q = number deprived or poor, n = population size

$$\begin{aligned} & {\text{Percentage}}\,{\text{of}}\,{\text{household}}\,{\text{deprived}}\,{\text{in}}\,{\text{each}}\,{\text{indicator}}\, \\ & \quad = \,\frac{{{\text{Number}}\,{\text{of}}\,{\text{poor}}\,{\text{or}}\,{\text{deprived}}}}{{{\text{Total}}\,{\text{Number}}\;{\text{of}}\,{\text{households}}\,{\text{surveyed}}}} \times 100 \\ \end{aligned}$$
(2)

The proportion of indicators in which the poor are deprived (A) is computed thus:

$${\text{A}}\, = \,\frac{{{\text{Number}}\,{\text{of}}\,{\text{indicators/dimensions}}\,{\text{in}}\,{\text{which}}\,{\text{the}}\,{\text{poor}}\,{\text{are}}\,{\text{deprived}}}}{{{\text{Sum}}\,{\text{of}}\,{\text{all}}\,{\text{indicators/dimensions}}}}$$
(3)

Percentage of the proportion of indicators in which the poor are deprived is:

$${\text{A}}\left( \% \right)\, = \,\frac{{{\text{proportion}}\,{\text{of}}\,{\text{indicators/dimensions}}\,{\text{in}}\,{\text{which}}\,{\text{the}}\,{\text{poor}}\,{\text{are}}\,{\text{deprived}}}}{{{\text{Sum}}\,{\text{of}}\,{\text{all}}\,{\text{indicators/dimensions}}}} \times 100$$
(4)

One implication of the approach is that households are described by counting the number of deprivations suffered (headcount index). A general equal weight of one (wj = 1j) was assigned to all variables within each dimension as measures of deprivation based on the rationale that an equal weight gives equal importance to each dimension. The justification for applying equal weight of (1) one is that each indicator is weighted equally along each dimension irrespective of its nature and the number of indicators used to describe each dimension introduces objectivity and prevents one dimension from being too important than other dimensions and eases analysis and interpretation (see Table 1).

Table 1 Thresholds and weights of variables
Full size table

In addition, poverty cut-off of 50% was adopted which represents k value—the minimum proportion of weighted domains in which a person has to be deprived to be identified as multi-dimensionally poor. The 50% deprivation cut-off indicates multi-dimensionally poor households are deprived in more than half of the indicators. Hence, the higher the deprivation cut-off the more severe the poverty situation of deprived households (Alkire and Foster 2010).

Discussion of results and findings

Deprivations in Education Dimension

The extent of human poverty in each household is represented by current and future levels of education deprivations. Two measures, illiteracy or heads of household without formal education and household heads with just primary school education are indicators of education deprivation included in the construction of education dimension of MPI poverty. The most deprived households in the education dimension are concentrated in Burutu local government with a headcount index of (28) representing 0.6% of multi-dimensionally poor households followed by households in Ughelli south with a headcount index of (21) representing 0.4% of multi-dimensionally poor households, Ethiope east (16) 0.3%, Ughelli north (15) 0.3% Warri south (10) 0.2% and Ethiope west (10) 0.2% respectively.

On the other hand, the least deprived and less multi-dimensionally poor households in the education dimension are found in the town of Asaba and Ibusa in Oshimili north and Oshimili south with a head count index of (1) in each town representing 0.02% of multi-dimensionally poor households. Accessibility to basic amenities like education facilities in Asaba and Ibusa communities influenced the low level of deprivation experienced in the educational domain compared to communities with limited access to basic amenities which results to high level of deprivation experienced in the educational domain (see Fig. 3).

Fig. 3
figure 3

Deprivation in education

Full size image

Deprivations in Health Dimension

Three proxy indicators were used for the computation of health deprivation experienced by household members in the health domain are: The first two indicators looks at nutrition of household members measured as consumption expenditure of households at (a) $1.25 and (b) $2.00 respectively, while the third indicator was measured as health related losses experienced by household heads in the study area. The most deprived and multi-dimensionally poor households in the health dimension are concentrated in delta south senatorial district which is made up of Warri South with a headcount index of (157) representing 2.1% of multi-dimensionally poor households, followed by Burutu (134) 1.8%, Ewu (119) 1.6%, Ozoro (117) 1.5%, Effurun (115) 1.5%, Oleh (106) 1.4%, Ashaka (106) 1.4%, Owa (98) 1.2%, Obiaruku (94) 1.2%, while the least deprived households in the health domain are found in the towns of Asaba and Ibusa in delta north senatorial district with a headcount index of (20) representing 0.3% of multi-dimensionally poor households and (28) representing 0.4% of multi-dimensionally poor households respectively (see Fig. 4).

Fig. 4
figure 4

Deprivation in health

Full size image

Deprivations in the standard of living dimension

The standard of living thresholds based on MDG’s and HNLSS (2010) was used to identify household deprivations measured as multi-dimensionally poor, the deprivation scores for each household are summed to obtain the household deprivation in the standard of living domain. Household deprivations in the standard of living domain are measured as: households with no electricity, households using wood or kerosene, oil as cooking fuel, households with no safe drinking water and unimproved energy sources, no access to adequate sanitation such as unhygenic waste disposal methods, using ‘dirty’ cooking fuel, unimproved energy source (wood or charcoal), having a home with a dirty floor and non-ownership of house and non-ownership of essential household assets are added to the list of variables used to assess households’ poverty severity in the living standard domain.

Owa in Ika north east had the most deprived households in the living standard domain with a headcount index of (871) and a severity index of 2.1% of multi-dimensionally poor households, followed by Warri (685) 1.6%, Ughelli (619) 1.5%, Burutu (616) 1.4%, Ozoro (596) 1.4%, Ewu (565) 1.3%, and Oleh (505) representing 1.2% of multi-dimensionally poor households, while the least deprived and least multi-dimensionally poor households in the living standard dimension are found in the towns of Asaba and Ibusa in Oshimili north and Oshimili south with a headcount index of (99) and (144) and severity index of 0.2% and 0.3% of multi-dimensionally poor households respectively.

The variations observed among households deprivation in the living standard dimension is a reflection of administrative function and commercial function performed by Asaba as the seat of government and availability of socio-economic infrastructure which provides opportunities for the poor and improves livelihood, and the likelihood of transition from poverty and of course, the proximity of Ibusa to Asaba influences the relatively low level of deprivations and multi-dimensional poverty experienced in the standard of living domain (see Fig. 5).

Fig. 5
figure 5

Deprivation in living standard

Full size image

Disaggregated poverty severity

Table 2 disaggregates the distribution of the severity of poverty at the senatorial district level. Multi-dimensional poverty measured as multiple deprivations simultaneously experienced along education, health and living standard domains. Although, household poverty in the study area is essentially chronic and multi-dimensional in nature with most of the households being deprived in more than k-value of 50% of the multi-dimensional indicators adopted in this paper.

Table 2 Disaggregate of poverty severity at the senatorial district level
Full size table

The most deprived households are found in delta south senatorial district with a severity index of (0.1729) representing (17.29%) of sampled households, followed by delta central senatorial district with a severity index of (0.1418) representing (14.18%) of sampled households, while delta north senatorial district had a severity index of (0.1463) representing (14.63%) of sampled households. The average severity index of 0.01625 (1.62%), 0.01773 (1.78%) and 0.02162 (2.16%) was recorded for delta north, delta central and delta south respectively, while an average severity index of 0.018445 representing (1.84%) was computed for the state as a whole. This result reveals that households in delta south senatorial district are the most severely deprived along education, health and living standard domains adopted.

The most severely poor households were found in delta south which is made up of Warri South with multi-dimensional poverty index of 0.0371 representing (3.71%) of sampled households. However, the least severely deprived households are found in delta north which is made up of Oshimili south with a severity index of 0.0034 representing (0.34%) of sampled households along education, health and living standard dimensions. Multiple deprivations in the living standard domain contributed most to the severity of poverty among the three senatorial districts followed by deprivations in the health domain. A significant proportion of the households are multiply deprived in the following indicators under the living standard domain: sanitation and toilet facilities, unsafe water sources, no assets, experienced livelihood losses (job, spouse, tool, equipment etc.), unimproved energy sources, and unemployment. Household poverty indices, like housing, health, drinking water, sanitation facilities and garbage collection system, are in deplorable conditions in most of these urban centres (see Table 2, for more details).

Socio-spatial dynamics of poverty severity

The hypothesis that the severity of household poverty varies significantly among households in the study area is tested using analysis of variance (ANOVA), while PCA reduced the observed variables into a smaller number of principal components that accounted for most of the variance in the observed variables. The principal component analysis reduced the observed variables into seven orthogonal variables which together accounted for 54.5% of the total variations. The extracted principal components 1, 2, 3, 4, 5, 6, 7, explained 10.451, 10.085, 9.309, 8.431, 6.793, 6.389 and 6.081% respectively. The first component can be described as a measure of sanitation, as it correlates strongly with unimproved sanitation facilities, waste disposal methods and unhealthy water source with component loadings of (0.772, 0.687 and 0.625). These loadings imply that increased unimproved sanitation facilities, poor water sources and waste disposal methods suggest a high deprivations and increased severity of poverty.

The second principal component is strongly correlated with unemployed adults above 18 years of age with a component loading of 0.737, poor nutrition 0.550 and location of households with a component loading of − 0.601. This implies that poverty severity increases with an increase in the number of unemployed adults and increased malnutrition, while poverty severity decreases with improved accessibility and good location with basic amenities. The second component could be described as a measure of wellbeing and livelihoods of households. The third principal component correlates with household composition measured as marital status and number of children with component loadings of 0.714 and 0.699.

The fourth principal component correlates with age of marriage with a loading of 0.527 which implies that poverty severity is significantly influenced by age at marriage, while employment of household heads decreases the severity of poverty with a component loadings of − 0.719. This is an indication that poverty severity decreases with increase in the employment of household heads. The fifth principal component correlates with the domestic remittances received by households with a loading of 0.611, dependency on remittances increases the severity of poverty.

The sixth principal component correlates with gender of households with a loading of 0.868, which implies that gender significantly influences poverty severity, while the seventh principal component correlates with the health status of household heads with a loading of 0.854 which implies that ill-health of household heads influences the severity of poverty (see Table 3).

Table 3 Summary of  the Principal component analysis
Full size table

Table 4 shows the result of the analysis of variance (ANOVA) based on further analysis of the extracted components in the principal component analysis (PCA). The result shows that the contribution of the high component loadings to the variations in the distribution of household poverty severity is statistically significant at (p ≤ 0.001) significance level with F value of (F = 527.305).

Table 4 Summary of ANOVA
Full size table

Conclusion

A significant proportion of severely deprived and multi-dimensionally poor households are concentrated in communities of delta south senatorial district and delta central senatorial districts based on the MPI multi-dimensional measures adopted. This study found that the geographic variation in urban poverty severity is due to variations in the deprivations of socio-spatial dynamics, such as availability of natural resource, geographic access to basic facilities, location, distance to water source and sanitation facilities. Proper understanding of the different poverty categories, multi-dimensional nature of poverty and identification of the chronic poor and multi-dimensionally poor to enable the inclusion of the poor in the formulation of pro-poor policies is vital to successful implementation of poverty reduction programmes in Nigeria. This study, apart from contributing to the growing literature on poverty studies in Nigeria, identified the geographic dimension of multi-dimensional poverty over space in a cross-sectional household survey.