Nexus between construction sector and economic indicators for Turkey and European Union evidenced by panel data analysis

Abstract

PurposeThe research aims to study the regression, cointegration and causality between the construction sector (CS) and the Gross Domestic Product (GDP), considering other variables in the study such as interest rate, taxation, industry sector, investment and Foreign Direct Investment (FDI), which are analyzed through unique panel models. The study was conducted in Turkey and the ten other countries of the European Union (EU) from 1988 to 2019.Design/methodology/approachRegression, cointegration and causality methods were used to investigate the different types of relationships between variables in the models. Data were obtained from official databases and the study contains four main stages, which are explained in detail in the methodology section.FindingsThe study used the analysis methods of regression, cointegration and causality tests and found that the CS and GDP have long-run estimates and the relationship between the two for different countries is negative in a two-way direction. Results are detailed in the analysis section.Research limitations/implicationsNo data were available for the variables before 1988 for most countries, which led to a limited number of observations and issues in statistical analysis methods.Originality/valuePreviously, only input and output tables were used in the analysis. The impact of interest rate, taxation, investment and FDI has not been analyzed. Key variables are very relevant for Turkey, which suffers from chronical inflation and taxation regimes. These show variability with the EU countries for comparative analysis and have not been explored to date, remaining as a major gap for the construction industry. No attempts were made to use regression, cointegration and causality methods with variables. These analysis methods enable an understanding of the differences in variance (heteroscedasticity) and the presence of cross-sectional dependence (CSD), both critical for the reliability of the comparison of data sets and analysis.