Forecasting Consumer Price Index Inflation in India: Vector Error Correction Mechanism Vs. Dynamic Factor Model Approach for Non-Stationary Time Series
Publication dateOct, 2020
DetailsNIPFP Working Paper No. 323
AuthorsRudrani Bhattacharya, and Mrigankshi Kapoor
Short to medium term forecasting of inflation rate is important for economic decision making by economic agents and timely implementation of monetary policy. In this study, we develop two alternative forecasting models for Year-on-Year (YOY) inflation in Consumer Price Index (CPI) in India using a large number of macroeconomic indicators. The YOY CPI inflation and its predictive indicators are found to be nonstationary and cointegrated. To address this issue, we employ Vector Error Correction Model (VECM) and Dynamic Factor Model (DFM) modified for non-stationary time series to forecast CPI inflation. We find that in terms of Root Mean Square Error (RMSE), the VECM model performs marginally better than the DFM model. However, both models are found to have the same predictive accuracy using Diebold-Mariano test.
JEL Classification: C32, C53.
Keywords: CPI Inflation, India, Forecasting, Vector Error Correction Model, Dynamic Factor Model.