Spatial Concentration Of Poverty And Its Determinants In Nigeria

ABSTRACT

Poverty reduction programmes in Nigeria have not had significant intended effects.

This can be attributed to the non-consideration of the heterogeneous nature of poverty and

spatial contiguity of geographical units in their designs. There is scarce information on spatial

decomposition and spillover of poverty across the Senatorial Districts (SD) in Nigeria.

Therefore, the spatial concentration of poverty and its determinants were investigated.

The study employed secondary data from Nigeria Living Standard Survey (NLSS)

and Core Welfare Indicators Questionnaire (CWIQ) survey conducted by National Bureau of

Statistics (NBS), Independent National Electoral Commission (INEC) and Food and

Agriculture Organization (FAO). The NLSS and CWIQ were conducted in 2004 and 2006

respectively. The national sample sizes for NLSS and CWIQ were 22,200 and 77,400

household units respectively. Following the elimination of households with missing values,

samples considered for the study were 18,760 and 54,536 households for NLSS and CWIQ

respectively. The Poverty Rate (PR) per SD was obtained from household consumption

expenditure data sourced from NLSS. Data on Household Size (HS), Household Membership

of Association (HMA), Households’ Access to Health Facilities (AHF), People Employed in

Agriculture (PEA), Access to Credit Facilities (ACF) and Literate Adult (LA) were obtained

from CWIQ. Data on Number of Years Spent in the National Assembly by Senators

(NYSNAS) (1999 – 2004) and soil fertility classification of Nigeria were sourced from INEC

and FAO respectively. These variables and spatial dimension were hypothesized to influence

PR. Data were analysed using descriptive statistics, Foster Greer and Thorbeck model, spatial

regression, local indicator of spatial association and spatial probit at p = 0.05.

Mean annual household per capita consumption expenditure was

N28475.01 ± N11967.5. Percentage of PEA in the SD was 44.2 ± 18.4% while mean HS was

6.5 ± 1.5. Mean values of NYSNAS, ACF and AHF were 4.3 ± 0.5years, 10.5 ± 7.4% and

51.6 ± 18.2% respectively. Fifty-six percent of the SD had fertile soils. Average national PR

of the SD was 56.03 ± 24.1%. Fifty three of the SD had PR below the national average. The

Moran’s I value (3.4) indicated that spillover of poverty existed among SD. Ten percent

increase in PR in one SD resulted in 3.1% increase in PR in the neighbouring SD ( = 0.3).

Fifty-two percent of the SD with significant spatial association had low PR neighboured by

low PR SD, 41.03% of the SD with high PR were neighboured by high PR SD. The PR in

high-high SD was significantly reduced by HMA (-0.9), AHF (-0.3), ACF (-0.9), LA (-1.1),

fertile soil (-5.2) and NYSNAS (-6.6). Poverty rate was significantly increased by PEA (0.4)

UNIVERSITY OF IBADAN LIBRARY

iv

and HS (5.5). Mean PR in high-high and low-low SD was 82.6% and 31.8% respectively.

Household’s probability of being poor was higher in high-high SD (0.8) compared to low-low

(0.08). Poverty incidence in a senatorial district influenced the neighbouring senatorial

district. Reduction in poverty incidence would be achieved through households’ membership

of associations, improved access to health and credit facilities.