Comparative analysis of profitability of real estate, industrial construction and infrastructure firms: evidence from India

This study aims to identify the most profitable segment of construction firms amongst real estate, industrial construction and infrastructure. This paper also examines the determinants of profitability of real estate, industrial construction and infrastructure firms.

The data of 67 firms (20 real estate, 21 industrial construction and 26 infrastructure) is collected for a 15-year period (2003–2017). Two models are created using total return on assets (ROA) and return on invested capital (ROIC) as dependent variables.. Leverage, liquidity, age, growth, size and efficiency of the firm are identified as firm-specific independent variables. Two economic variables, i.e. growth in GDP and inflation, are also used as independent variables. Initially, the models are tested for stationarity, multicollinearity and heteroscedasticity, and finally, the coefficients are estimated using Arellano–Bond dynamic panel data estimation to account for heteroscedasticity and endogeneity.

The results suggest that industrial construction is the most profitable segment of construction, followed by real estate and infrastructure. Their profitability is positively driven by liquidity, efficiency and leverage. The real estate firms are somewhat less profitable compared to industrial construction firms, and their profitability is positively driven by liquidity. The infrastructure firms have low ROA and ROIC.

The real estate, infrastructure and industrial construction drastically differ from each other. The challenges involved in real estate, infrastructure and industrial construction are altogether different. Therefore, authors present a comparative analysis of the profitability of real estate, infrastructure and industrial construction segments of the construction and compare their determinants of profitability. The results provided in the study are robust and reliable because of the use of a superior econometric model, i.e. Arellano–Bond dynamic panel data estimation with robust estimates, which accounts for heteroscedasticity and endogeneity in the model.