Please use this identifier to cite or link to this item: http://ir.library.ui.edu.ng:8080/jspui/handle/123456789/1227
Title: EFFICIENCY OF MEDIAN PER CAPITA EXPENDITURE METHOD FOR MEASURING THE LEVELS OF POVERTY IN NIGERIA
Other Titles: A DISSERTATION IN THE DEPARTMENT OF STATISTICS SUBMITTED TO THE FACULTY OF SCIENCE IN PARTIAL FULFILMENT OF THE REQUIREMENT FOR THE DEGREE OF MASTER OF PHILOSOPHY OF THE UNIVERSITY OF IBADAN
Authors: NWOKE, F. C.
Keywords: Poverty line
Poverty depth
Median per capita expenditure
Household expenditure
Issue Date: 2014
Abstract: Measurement of levels of poverty is a powerful instrument for focusing the attention of policy makers on the living conditions of the poor. Previous studies on measurement of poverty in Nigeria have adopted two-third of mean per capita expenditure method for defining the poverty line. This has failed to capture the number of poor skewness in household expenditure data. Therefore, this study was designed to access the efficiency of Median Per Capita Expenditure (MPCE) method in measuring poverty in Nigeria. Expenditure data from the 2006 general household survey from National Bureau of Statistics (NBS), Nigeria were collected. A standard distribution software was used to fit four distributions (log normal, log logistic, gamma and frechet) to per capita expenditure using Maximum Likelihood Estimation (MLE) method. The distributions were ranked using Kolmogorov Smirnov goodness-of-fit statistic. Any distribution with p value < 0.05 was considered good. The MPCE and the corresponding value for two-third mean per capita expenditure were obtained for comparison. Foster-Greer-Thorbecke poverty index was used to determine the proportion of poverty for the states. Differences in the lowest poor and highest poor states were investigated using the Z test. Using two-third mean and MPCE, a bootstrap simulation was performed on the expenditure data to obtain precision estimates for poverty headcount index. This was carried out for the purpose of comparing the two-third mean per capita and the MPCE methods as well as the relative gain in efficiency. Log logistic distribution with parameter estimates (α=1.0452, β=3169.2) performed best, (p < 0.000014805) in fitting per capita expenditure data. The MPCE was N2,550.00 while the Two-third Mean Per Capita Expenditure (TMPCE) was N3,613.75 per month respectively. The proportion of the poor was 50.1% using MPCE but rose to 61.6% when TMPCE was used. Across states, the proportion of the poor was highest in Katsina state (52.7%) and lowest in Anambra state (42.9%) using MPCE method. However, poverty was highest in Kwara state (73.9%) and lowest in Anambra state (48.3%) using TMPCE method. The difference between the proportions were highly significant (P < 0.003) for MPCE and not significant (P < 0.246) for two-third mean per capita expenditure. The precision estimates were 2.50 x 10-05 and 9.78x10-05 for median and two-third mean per capita methods respectively and the relative gain in efficiency of the median method over two-third of mean method was 25.5%. Median rather than two-third of mean per capita expenditure was a more efficient method for measuring poverty. Therefore, for improved precision of poverty measurements in Nigeria, median per capita expenditure would be a better approach.
URI: http://ir.library.ui.edu.ng:8080/jspui/handle/123456789/1227
Appears in Collections:Academic Publications in Statistics

Files in This Item:
There are no files associated with this item.


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.