Inflation Forecasting in Ghana - Artificial Neural Network Model Approach

Loading...
Thumbnail Image
Date
2013-03-12
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
Price stability is the primary objective of monetary policy in Ghana. To achieve this mandate, the Bank of Ghana has adopted inflation targeting as its monetary policy framework. Indeed, inflation targeting requires good forecasting ability for the monetary authorities. Two approaches i.e. econometric and artificial neural network (ANN) models have been used by some central banks and researchers to forecast inflation. However, the Bank of Ghana and researchers in Ghana have used only econometric models to forecast inflation. This thesis uses both the econometric and ANN methods to predict inflation in Ghana. The econometric models (AR and VAR) and the ANN models (NAR and NARX) were applied to the monthly year-on-year inflation data from Jan. 1991 to Dec. 2011. The models were estimated using the data from Jan. 1991 to Dec. 2010 so as to forecast for the period Jan. 2011 to Dec. 2011. It was found that the forecast errors of the ANN models were lower than those of the econometric models; thus, the ANN predicts inflation better than the econometric models. The policy implication is for the Bank of Ghana and researchers in Ghana to use the ANN model in addition to the econometric models to forecast macroeconomic variables such as the inflation.
Description
A thesis presented to the Department of Economics, K.N.U.S.T, in partial fulfillment of the requirements for the award of Master of Philosophy (Economics) Degree, March-2013
Keywords
Citation