Econometric techniques for estimating construction demand in Australia

An understanding of future trend in the demand for construction could influence investment strategies for a variety of parties, including construction developers, suppliers, property investors and financial institutions. Although there are clear needs for achieving sustainable development in the long-run, the Australian construction industry still lacks an established strategy to estimate demand in both national and regional construction markets. The overall aim of this research is therefore, to develop advanced econometric estimation models to estimate construction demand for both national and regional construction markets. Although construction demand estimation has been studied in several countries using various techniques, research has hardly considered the impact of external factors, such as the recent global financial crisis, in the estimation models, and previous used estimating methods are not applicable due to the volatility and fluctuations of the construction market during the economic downturn. In this study, global economic events as intervention dummies, together with identified determinants of construction demand, are considered in a vector error correction model to accurately estimate the movement of construction demand in the national construction market. The numerical forecasting results generated by the VEC model with dummy variables are compared with those from a conventional VEC model and other commonly used temporal estimating techniques, namely multiple-regression and Box-Jenkins. The estimation results indicate that the VEC model with dummy variables is more effective and reliable for estimating construction demand than the conventional VEC, BJ and MR modelling techniques. Previous studies in the field of quantitative forecasting for construction demand employed mainly temporal estimation models, such as ARIMA, MR and VEC models, and most of them have been carried out at the national level. However, in Australia, construction markets cannot be simply considered as a national aggregate, but are better represented as a series of interconnected regional and local markets. Using Australian state-level data, two regional estimating techniques - panel vector error correction (P-VEC) model and panel vector error correction model with dummy variables - were developed, in view of regional disparities and global financial crisis impact among sub-national construction markets in the regional demand estimation. Spatial linkages among regional construction markets in Australian were found in the study. Two spatial panel estimating models - spatial panel vector error correction (SP-VEC) model and SP-VEC model with dummy variables - were developed for estimating regional construction demand in view of linkages among regional construction markets. The estimation results of two spatial panel models were further compared with two proposed P-VEC models in order to select the most appropriate and accurate forecasting model for the regional construction markets. The comparison results suggest that spatial panel vector error correction model with dummy variables outperforms other three proposed regional forecasting models (i.e. P-VEC, P-VEC model with dummy and SP-VEC model).