Univariate Time Series Analysis of Consumer Price Index on Food and Non-alcoholic Beverages

Abstract

The paper examined univariate time series forecast of consumer price index on the consumption of food and nonalcoholic beverages in Nigeria. It filled the knowledge gap by explicitly modeling and forecasting consumer price index in Nigeria using the univariate ARIMA model. The work was restricted to Nigerian Consumer Price Index. It was also restricted to food consumption (FC) data and food & nonalcoholic beverages consumption (FNBC) data from 1995-2021. This paper analyses were carried out using gretl 2019c, Minitab 16 and Micro software Excel (2010). The monthly and yearly means plots were done, so as to have a better understanding of the series behaviours. The series plots points to the fact that there is possibility that the time series are integrated of order 1 for food consumption series with no seasonality, while integrated of order 1 for food & nonalcoholic beverages consumption series with seasonality of order 12. Stationarity after second difference of the first differencing was obtained for both series. A suitable ARIMA Model was obtained for both series and was used for models forecast computation. Hence, the computed result suggested that ARIMA(0,1,1) and ARIMA(0,1,1)(0,0,0)12 model were the best model for estimating and forecasting the two time series, using model selection criteria and accuracy measures. The plots of the forecasts generated for the FC and FNBC shows that the two variables are dependent and also shows that any gradual increase in the food consumption tends to pave way for increase in the food & nonalcoholic beverages consumption or a drastic drop in the food consumption will also drop the food & nonalcoholic beverages consumption in the same manner. It seems reasonable to conclude that there is significant relationship between the food consumption and food & nonalcoholic beverages consumption series. It was recommended that more detailed work should be carried out in the area of co-integration analysis of the two variables to enhance a better understanding and prediction distribution in Nigeria.