TEMPORAL MODELLING OF CURRENCY IN CIRCULATION IN GHANA

Abstract

The Currency in Circulation is the outstanding amount of notes and COIns circulated in the economy and are the most liquid monetary aggregate. In this study, data on monthly Currency in Circulation obtained from the Bank of Ghana database was modelled using both SARIMA model and Regression model with ARIMA errors. The results revealed that ARIMA (0, 1, 1)(0, 1, 1)12 model was the best SARIMA model for the Currency in Circulation. This model has the least AIC of -372.16, AICc of -371.97 and BIC of -363.53. Also, regression model with ARIMA (0, 0, 1) errors was identified as the best regression model with ARIMA errors. This model has an AIC of -417.39, AICc of -416.57 and BIC of - 396.60. Diagnostic checks of both models with the Ljung-Box test and ARCHLM test revealed that both models are free from higher-order serial correlation and conditional heteroscedasticity respectively. A comparative analysis of the forecasting accuracy of these models with the Diebold-Mariano test revealed that there is no significant difference in the forecasting performance of the two models. The two models were therefore proposed for predicting Currency in Circulation in Ghana. However, the Currency in Circulation is volatile and subject to several unobservable developments in the economy. Therefore continuous monitoring of the forecasting performance of these models, review of market conditions and necessary adjustments are required to make the use of these models more realistic.