Use of Information Variables in Inflation Forecasting(Vol.9 No.1)
Inflation can be seen as a phenomenon caused by a mixture of various international and domestic economic factors including demand, supply, and cost conditions. Consequently, extensive use of information variables that reflect key economic trends and factors is critical for accurate inflation forecasting. Despite their importance, studies on information variables - especially on the concept and their practical application - are relatively rare.
This paper analyses the scope for enhancing prediction accuracy in inflation forecasting by employing out-of-sample predictability tests, when information taken by way of principal component analysis from 42 individual economic indicators of the real, external, price, and financial-market sectors is used appropriately. The empirical experiments show that the accuracy of the model incorporating information variables is substantially superior to other time-series models. These results support the notion that efforts to apply more sophisticated processes to a broad spectrum of information variables are required in order to increase the predictability of inflation forecasting.