A TIME SERIES MODEL OF HUMAN IMMUNODEFICIENCY VIRUS(HIV) AND TUBERCULOSIS (TB) COINFECTIONS AMONG FARMERS IN BENUE STATE

ABSTRACT

This dissertation applied time series model to study the impact of HIV and Tuberculosis with implications on  productivity of Benue state farmers using cointegrated autoregressive model and granger causality test. The study utilize monthly time series data on the serologically confirmed and reported cases of HIV/AIDS and tuberculosis (TB) in Benue state of Nigeria covering the fiscal year from January, 2007 to Decembers, 2016 and employed dickey-fuller Generalized least square unit root test, KPSS stationarity test, Johansson cointegration, unrestricted vector autoregressive model and granger causality based on Toda-Yamamoto procedure as methods of analysis. Results showed that HIV and TB as study variables are integrated of order one, I(1). HIV is found to have positive and significant impact of TB which indicates that increases in the number of HIV patients in Benue State leads to increases in the number of TB patients. The study found no long-run stable relationship existing between HIV and TB meaning that the study variables are not cointegrated. In the estimated result of the unrestricted VAR model, HIV and TB appeared to have high predictive information on one another in the short-run. The impulse responses of the VAR model showed that positive shocks to HIV persistently raises TB in the long run and HIV is found to have predictive information on TB while TB has no predictive information on HIV in the long-run. The granger causality test base on the Toda-Yamamoto procedure result revealed that HIV Granger causes TB whereas TB does not granger cause HIV implying that HIV has been the driving force behind the persistence increase of TB infection among Benue state farmers. The study provides some policy recommendations.