Inflation Forecasting Under Different Macroeconomic Conditions: A Case Study of Pakistan

Abstract

Inflation forecasting has been important task for monetary authorities, policy makers and government. Prediction about inflation confer us a precise image of how the economy is expected to accomplish in the future. It is essential job for researchers to examine which methods are suitable for inflation forecasting. We have used Naive model, ARIMA model, Philips curve model and Philips Curve (TAR) under different macroeconomic conditions with reference to real-time, revised and final data from 1974 to 2014 and predicted out-of-sample inflation forecast for 2015, afterward we roll-forward our regression from 1975 to 2015 to forecast inflation for 2016.We have analyzed naive model is superior to other modelsbecause RMSE and MAE of naive model are less than other models by using real-time, revised and final data for one year-ahead out-of-sample inflation forecasting. On the other hand for two years ahead out of sample inflation forecast, according to real-time data RMSE shows that naive model is most superior to other models whereas MAE shows that Philips curve Threshold auto regressive model is most superior to other models. According to revised and final data for two years ahead out of sample inflation forecasting both forecasting accuracy measures shows Naive model is most superior to other models.