ASYMMETRIC EFFECT OF CRUDE OIL PRICE CHANGES ON INFLATION IN SELECTED EAST AFRICAN COUNTRIES: APPLICATION OF PANELNONLINEAR AUTOREGRESSIVE DISTRIBUTED LAG MODEL (PNARDL)

Abstract

Inflationary pressure that has been started increasing in 2000 in global economies were sharp increasing in East Africa developing countries. Consumer price index included as the measures of inflation. This study examines the effects of crude oil price changes on inflationin selected East Africa'scountries with special focus on asymmetric relationship of crude oil price changes and consumer price index where others are control variables. Panel nonlinear autoregressive distributed lag model framework of pooled mean group and robust standerd fixed effect driscrol-Kray model was employed to consider cross sectional dependence.The longrun and shortrun crude oil price changes are introduced into the oil price-inflation model through the positive and negative partial sum decompositions of oil price changes. The appropriate panel nonlinear autoregressive distributed lag model was selected and estimated with pooled mean group estimation techniques to capture effect of unequal magnitude oil price changes. Fixed effect model introduced robustness and identified that the model is robust and normal with employed data, whereas Ramsey RESET suggests model validity. The study finds that, the long run positive and negative oil price changes significantly affect inflation. The long run asymmetry is discovered by identifying the unequal or different coefficient values between positive and negative oil price with oil price positive exerting more effect on inflation than its counterpart oil price negative and revealed significant speed of adjustment which converges to equilibrium slowly in the long run. The result confirms the large increase and small increase behaviors of positive and negative change of crude oil price. The study recommends consideration on real crude oil price, exchange rates, real gross domestic product, to East Africa policy makers. This study will call upcoming research to model crude oil price asymmetries and inflation using granger causality withdeterminants such as assets, with improving the cross sectional nonlinear autoregressive distributed lag estimation methods for modeling inflation in East Africa