Modelling monthly non food component of the colombo consumer price index (ccpi) using vector auto regressive (var) approach

Abstract

This study attempts to model the non food component of monthly Colombo Consumer Price Index (CCPI) in Sri Lanka using multivariate generalization of the univariate ARIMA model known as vector auto regressive (VAR) modeling approach. The data used are monthly series of Colombo Consume Price Index from year 2008 to 2015 and corresponding monthly series of data related to non food items. The structure of model is a linear function of past lags of itself and past lag of the other variables. All series were stationary for the corresponding first difference of log series and confirmed that the existence of long run dynamic relationship among all variables. The significant variables identified are clothing and footwear, housing water electricity gas and fuel, health, education, furnishing, communication, transport, recreation and culture and miscellaneous goods and services. These non food sub categories in the CCPI can be forecast using the developed model. The results would be useful when analyzing the key indicators in the economic sphere. Furthermore, the results of this study emphasize the need to put in place a stable macroeconomic policy environment to maintain price stability, since low inflation would enhance economic growth.