Comparing the fixed and the time varying parameter model in forecasting the inflation rate of Malaysia

• Main Authors: Nur A’mirah Mohd Yaziz, Mohd Rosli Mohamad, Hasif Rafidee Hasbollah, Nurul Ashykin Abd Aziz
• Format: Conference or Workshop Item
• Language: English
• Published: 2014
• Online Access: http://discol.umk.edu.my/id/eprint/9118/1/Conference%202.pdf; http://discol.umk.edu.my/id/eprint/9118/

Inflation is one of the most substantial elements in the analysis of a country’s economy. It describes the general increment of the price of goods and services in a country. The interaction between inflation and economic growth are closely related. A high inflation rate would bring a negative impact on a country. Economists agreed that price stability is a prerequisite for a country’s rapid growth and important for a balanced development mechanism. Great economic decision can be made with a good forecasting values associated with low forecast errors. This research aimed to identify the best model in forecasting inflation using the Consumer Price Index (CPI) data as an indicator of inflation. Three models were considered based on univariate and multivariate time series. Models based on the fixed parameter and the time-varying parameter were also considered. The best model chosen is based on out-of-sample forecasting framework starting from January 2009 until December 2012, while the model was fitted for period January 1997 to December 2008. The starting point of comparison is Naive Model acting as benchmark while the other models, Autoregressive Distributed Lag (ADL), and Time-Varying Parameter (TVP) model were carried out. The results showed that, TVP Model outperformed all the other competed models. Thus, it is the best model to forecast inflation in Malaysia.