The Impact of COVID-19 on Nigeria Consumer Price Index (CPI)

Abstract

This study models the Nigeria Consumer Price Index (CPI) and
forecasts its accuracy with COVID-19 impact on the data. The CPI
data is homoscedastic and having heteroscedastic in nature and as
such it is modelled using models such as Autoregressive (AR)
Autoregressive Conditional Heteroscedastic (ARCH), Generalized
Autoregressive Conditional Heteroscedastic (GARCH), Threshold
GARCH, ExponentialGARCH and Power ARCH. This study compares
these models using available data for Nigeria and found the AR as the
model of best fit according to the minimum information criteria. The
dynamic forecast evaluation reveals that GARCH-N has the minimum
forecast residuals by the Root Mean Square Error (RMSE), Mean
Absolute Error (MAE) while AR has the minimum Mean Absolute
Percentage Error (MAPE) among the models. On average AR has the
minimum forecast residuals. Therefore, employing AR model for
modelling and forecasting Nigeria CPI which results in low inflation
rate when compared with the results of other models considered. This
can improve the economy of the nation.