A Time Series Model of Poverty Incidence in Nigeria

Abstract This paper attempts to search for an optimal Autoregressive Integrated Moving Average model that best forecast absolute poverty incidence in Nigeria. The study uses absolute poverty data in Nigeria for 35 years from 1980-2014. The data was obtained as secondary data from Central Bank of Nigeria, Federal Office of Statistics, National Bureau of Statistics and International Monetary Fund World Economic Outlook. Time series plots, Ng & Perron modified unit root test and KPSS stationarity test were employed to check the graphical and statistical properties of the series. The results indicate that the series is integrated of order one, I(1) and the ACF and PACF plots of the stationary series suggest a mix ARMA (p,q) model for the series. ARIMA (p,d,q) model in line with Box-Jenkins procedure were then employed to model the poverty time series data. The result shows that ARIMA (4,1,4) was the best candidate to model poverty incidence in Nigeria. It was generally observed from the tests of residuals of the modeled equation that, the model was good, valid and adequate in describing absolute poverty situation in Nigeria. Accuracy measures such as Root Mean Square Error, Mean Absolute Error, Mean Absolute Percentage Error and Theil Inequality Coefficient were used to evaluate the forecast ability of the model and an out-of sample forecast mode was best for the model. The modeled ARIMA (4,1,4) was then used to forecast future poverty values in Nigeria. The forecast indicates a linear growth in poverty level in Nigeria.
[File Type:PDF]

Updated: June 25, 2023 — 2:40 pm