Journal of Statistics Applications & Probability


Inflation is a major economic problem in emerging market economies and requires accurate models to avoid high volatility and long periods of inflation. This paper is aimed at evaluating a Functional Time Series (FTS) model as compared to other models in forecasting inflation in Uganda. The monthly Time Series (TS) data for the general Consumer Price Index (CPI) was used during the period of Jul-2005 to Jun-2020. Box-Jenkins’ Auto Regressive Integrated Moving Average (ARIMA) and Seasonal Autoregressive Integrated Moving Average (SARIMA) methodologies are explored to evaluate the FTS method of forecasting the general CPI where their accuracies are compared and validated using Mean Squared Error (MSE), Root Mean Square Error (RMSE), Akaike Information Criterion (AIC), and Bayesian Information Criterion (BIC) criteria. Existing inflation models in Uganda are outdated by structural changes in the economy igniting the need for a novel accurate model for forecasting inflation. FTS technique is overall considered accurate and particularly used to model high-frequency data such as Uganda general CPI data modeled as a functional observation after smoothing by kernel smoothing methods compared to traditional methods. Business operations and consumers normally base their decisions on modeled and forecasted inflation with their decisions affected by inflation uncertainties that hinder their motivations to invest and save in a given country as they try to avoid inflation-related risks. Findings therefore show FTS having great accuracies and recommended the method for forecasting Uganda inflation. This opens a new framework for extending the Box and Jenkin’s methodology to the functional setting.

Digital Object Identifier (DOI)