COMPARING THE FORECAST ABILITY OF CONSTRUCTION MATERIALS SECTOR INDEXES IN THE STOCK EXCHANGE OF THAILAND BASED ON STATISTICAL MODELS

Abstract

The objective of this study is to compare for predicting value of the construction material index (CONMAT) by using three forecasting models: ordinary least square (OLS), vector error correction model (VECM) and Box-Jenkins concept. Eight independent variables are selected in the models namely: construction material price index (MPI), real effective exchange rate (REER), minimum loan rate (MLR), cement price index (CEMENT), energy index (ENERG), consumer price index (CPI), private investment index (PII) and land price index (LAND). Price of all indexes is on a monthly and data set covers the period from January 2010 to December 2017, totally 96 months. The results indicate that Box-Jenkins model suitable for predicting the construction material index (CONMAT) by using root mean square error (RMSE) criteria.  For impulse response analysis test, minimum loan rate (MLR) is the most impact to the CONMAT when comparing between those independent variables.