Placide Aime Kwizera
Abstract: The aim of this paper is to introduce GARCH-modelling and its application on Rwanda food inflation data spanning from January, 2004 to December, 2018 (180-observations). On the basis of estimation results of various GARCH models and diagnostic check has shown that the AR-GARCH with Gaussian distributed innovations is most appropriate specification for modeling food inflation volatility in Rwanda. The study finds no evidence of asymmetry in the response of food inflation volatility to negative and positive shocks. Simulation on estimated AR-GARCH with Gaussian distributed innovations and a 6-years (72months) forecast from January 2019 to December 2024 were also made. Hence, the AR-GARCH with Gaussian distribution of innovations could be a widely useful tool for modelling the food inflation volatility in Rwanda.
Keywords: Inflation, volatility, GARCH, Leverage effect