Estimating ecological footprints of OECD countries with panel data analysis and neural networks

The rapid increase in population, industrialization, urbanization, and changing consumption patterns has placed significant pressure on the ecological balance of our planet. This situation has led to the uncontrolled and unsustainable demand for natural resources. As ecological degradation becomes increasingly evident, the concept of the "Ecological Footprint" has emerged in line with the goals of sustainable development. This study analyzes the effects of renewable energy consumption, globalization, and economic growth on the ecological footprint. Using data from the period 1995–2021, estimations were conducted through panel data analysis and artificial neural network methods. Panel data analysis was employed to examine the impact of independent variables on the dependent variable in detail, and these variables were then used as inputs for the artificial neural network model. Consequently, a high-accuracy model was developed for predicting the ecological footprint. The findings indicate that the independent variables included in the model have significant and varying effects on the ecological footprint. In particular, Gross Domestic Product (GDP) and the KOF Globalization Index have been identified as key factors that positively influence the ecological footprint. On the other hand, renewable energy consumption has been determined as a crucial variable in reducing the ecological footprint. Based on the results, it has been concluded that increasing investments in renewable energy is essential for controlling the rise in the ecological footprint. During the modeling process, a feedforward artificial neural network approach was adopted, and the Levenberg-Marquardt algorithm was utilized for model training. The artificial neural network model demonstrated high reliability in forecasting ecological footprints for each OECD country. The limited number of studies in the literature that simultaneously employ panel data analysis and artificial neural network methods for predicting ecological footprints served as a primary motivation for this research. Therefore, this dissertation aims to contribute significantly to the literature both methodologically and in terms of content.