Quantifying the Impact of Coal Transition on GDP Growth through System Dynamics: The Case of the Region of Western Macedonia, Greece

Abstract

The transition from coal to more sustainable energy sources represents a critical shift for economies reliant on coal production. To investigate the intricate processes involved in such a transition, the use of powerful analytical tools is essential. This study assesses the impact of the delignification process on GDP growth over a 20-year horizon (2015–2035) in the Region of Western Macedonia, Greece, using the Vensim PLE Plus 9.0.1 software, a robust tool for system dynamics modeling. By developing a dynamic model that captures the key variables and feedback loops associated with coal transition, this research examines economic, social, and investment variables, emphasizing their causal relationships. The study integrates societal, economic, and educational impacts on production transition, addressing issues such as unemployment, financial support, and investments in human resources and R&D. Additionally, it considers the influence of climate change on GDP. The model highlights population dynamics, economic development, and education as critical factors. Scenarios explore the impact of increased funding on education, research, and financial aid efficiency, providing insights into enhancing GDP in decarbonizing regions. The study reveals that increased investment in education and human capital leads to slight improvements in local GDP, though the effects are not immediate. Enhanced efficiency in government and European spending significantly boosts local GDP by creating strong value chains and local economies of scale. It is found that the increase in financial support to the regions in transition is of the utmost importance and has a multiplicative nature, something that should encourage the European Union to increase its financial support tools. The model’s simulations align closely with historical GDP data, validating its accuracy. The contributions of the present work offer valuable insights to policymakers and stakeholders engaged in the transition processes.

1. Introduction

The combustion of fossil fuels is the primary contributor to environmental pollution, particularly air pollution [1], causing various respiratory problems; it is also one of the primary causes of mortality [2]. Considering the challenges of climate change and aiming at ensuring a safer environment for European citizens, the EU has taken a number of drastic actions, such as several energy initiatives aimed at stabilizing Europe’s energy sector, including the Energy Efficiency Directive, and incorporating aspects of the plans to transition to a green economy [3]. Within the framework of the European Green Deal [4], many regions across Europe have been undergoing an energy transition stage to meet the environmental targets set by the EU and change the energy mix. The regions most affected are those relying on coal and coal products for electricity generation. The European Union is committed to supporting regions affected by and undergoing energy transition, recognizing that the specific regions will suffer substantial economic and social pressures, such as income loss, job displacement, and marginalization. This commitment was underscored in January 2020, when the EC’s pledge to support the transition in fossil-fuel-dependent regions gained further importance with the establishment of a Just Transition Mechanism [5]. In this context, a comprehensive and multifactorial transition impact analysis carried out in a fair and efficient manner is critical.

Dynamic and integrated assessment models are considered valuable supporting tools contributing to decision-making processes and provide a secure method to explore alternative scenarios and policies [6]. Among the first dynamic models were Urban Dynamics and the World Dynamics designed by Jay W. Forrester [7]. Based on the analysis of dynamic systems, the specific analysis involves examining cause-and-effect relationships among variables, aiming at a quantitative system analysis, information feedback, and the dynamic relationship between function and performance [8,9,10]. Xu et al. [11] state that simulation methods are considered ideal as they enable understanding and exploring the dynamics and long-term impacts of organizational, social, and economic systems with long-term and cyclical characteristics. Overall, system dynamics theory is widely recognized as an approach to estimating and analyzing the internal dynamic structure and feedback mechanisms of a system [12].

The present research employs the Vensim PLE Plus 9.0.1 software, developed by Ventana Systems, which primarily supports continuous simulation (system dynamics) with some distinct modeling capabilities, the most significant of which involve connections to data and design as well as the monitoring of dynamic models. This specific tool is used to enable the creation of interdependent relationships between variables, mainly focusing on local gross domestic product growth in the Region of Western Macedonia. Vensim is ideal for capturing the dynamic behavior of complex systems involving feedback loops, time delays, and non-linear relationships. It allows for the modeling of intricate interactions between various components of a system over time. Additionally, it is ideally suited for simulating and analyzing the impact of different policy scenarios and interventions over time. It helps in the understanding of the long-term effects and unintended consequences of policy decisions. Vensim, has been used in research addressing environmental, societal, and economic issues, such as urban sustainability performance [13], sustainability of the economy–nitrogen resource–environment system [14], recycling and the collection of waste material [15], waste management policies [16], decision-making and the socio-economic benefits of infrastructure projects [17,18,19], policy implications [20], innovation performance studies [21], socio-economic development [22], the creation of greener economies [23], renewable energy sources [24], energy networks [25], and air pollution [26]. In addition, Vensim has recently been proposed by the United Nations as a tool to assess the socioeconomic impact of climate information services used in connection with disaster risk reduction initiatives in Africa [27].

The purpose of the present research is to investigate how and to what extent external factors can influence the development of local GDP in regions undergoing transition and fossil fuel phase-out. Local GDP is vital to understanding economic trends, making informed policy decisions, guiding investment, and assessing the standard of living and economic health of nations and their specific regions. We have chosen GDP as the central variable in our dynamic system for a number of reasons. Despite its imperfections, GDP is widely recognized as the most comprehensive economic indicator for exploring a region’s economy, growth, and development. It represents the total value of goods and services produced over a specific period, providing a reflection of economic performance and productivity. More importantly, GDP uniquely allows models to capture an economy’s aggregate effects and interdependencies, making it a valuable proxy for understanding various economic dynamics, such as consumer behavior, investment trends, government spending, and net exports. In system dynamics models involving feedback loops and time delays, GDP provides a holistic measure that can reflect the outcomes of different economic policies or external shocks. Furthermore, GDP is the best indicator for comparative benchmarking. It is a standardized and widely recognized metric for comparing countries and periods. This standardization greatly facilitates benchmarking and comparative analysis, which is crucial for understanding relative economic performance and competitiveness. Researchers and policymakers can leverage GDP data to compare financial conditions and policy outcomes across different contexts, thereby making the results of their models more relevant and actionable. Finally, GDP as an indicator presents huge data availability and reliability. GDP data are widely available, regularly updated, and collected using consistent methodologies across countries and regions by reputable institutions like the World Bank, the International Monetary Fund (IMF), and national statistical agencies. This ensures that models can be calibrated and validated using accurate and up-to-date information, thereby enhancing the credibility and robustness of the model outcomes. In conclusion, the comprehensive nature of GDP, its utility for comparative analysis, and the availability of reliable data collectively enhance the models’ robustness, relevance, and credibility, making GDP a critical variable in examining economic dynamics and policy impacts. Each nation’s statistical agency calculates the gross value added (GVA) at the regional level annually. GVA stands for the net value of production, estimated based on the retraction of the intermediate consumption valued at market prices from basic prices. GVA is calculated before the consumption of fixed capital. After GVA is determined and considering the regional population data, it is possible to calculate local GDP and GDP per capita.

One basic and two alternative scenarios are applied to assess alternative policy actions. There is no similar approach, to the best of our knowledge, to provide a substantial representation of the transition to a new production system by means of a dynamic approach.

The present work makes a significant contribution by providing a robust methodology for estimating GDP in regions undergoing a complete transition in energy production. This study stands out by developing a comprehensive system dynamics model that rigorously explores the interactions among various factors influencing regional GDP. The focus on investments, the impact of renewable energy sources (RESs), technological innovation, GDP growth and decline, and climate effects offers a holistic perspective on how these elements interact within the context of decarbonizing regions.

This integrative approach not only provides nuanced insights into the economic strategies and policy implications necessary for such transitions but also addresses a critical gap in the existing literature. By considering the complex interdependencies and feedback loops among these variables, the study enhances our understanding of regional economic dynamics. This model is the first of its kind in Greece to examine the social, environmental, and economic aspects of regional GDP in such a comprehensive manner, potentially serving as a catalyst for a series of future studies in the field of dynamic systems.

The decision to investigate this issue is driven by the growing need to understand how regional economies can adapt and thrive amid European and global decarbonization efforts. With climate change and sustainability at the forefront of policy discussions worldwide, this study’s examination of how investments, technological advancements, and the adoption of renewable energy can drive economic growth while mitigating negative climate impacts is both timely and essential.

Ultimately, the study aims to provide actionable insights for policymakers and stakeholders, focusing on enhancing financial resilience and sustainability in transitioning regions. The complexity of the interactions among the studied variables necessitates the dynamic modeling approach uniquely provided by this work, offering a valuable tool for guiding economic development in the face of climate and energy challenges.

2. Materials and Methods

The present research employs the method of system dynamics based on the Vensim software, aiming to address a deeper analysis of gross domestic product (GDP). The approach followed involves analyzing a series of correlated variables of economic, social, research, and investment aspects in regions undergoing a transitional stage of changing their production model and adopting decarbonization processes. It also highlights the causal relationship among the correlated variables. As such, the methodology employed in formulating the specific model integrates the fundamental pillars of societal, economic, and educational impacts on production transition and conducts a thorough analysis of GDP dimensions, implementing a system dynamics approach to address real-world problems [28].

A decline in local GDP is a common problem encountered by regions undergoing a fossil fuel phase-out process and it was the most significant factor in choosing it as the target variable, highlighting the importance of estimating its future trends. Potential decline entails business closures, unemployment, economic downturns, and various socio-economic challenges. Therefore, estimates of GDP are of great importance, especially in challenging situations such as shifts in the energy sector. In the present work, a GDP estimation is formulated as a function of variables or real-world quantities that are expected to have a strong influence on it.

To formulate the estimation model employed in this study, we have identified key factors influencing GDP, drawing on insights from previous research. The most significant factor is unemployment, particularly resulting from business closures, which has been shown to be a major determinant of local GDP [29,30,31]. A decline in local GDP is often observed in regions with high unemployment rates. However, accurately assessing unemployment requires consideration of various underlying factors, including population dynamics influenced by mortality, birth rates, immigration, and emigration. Financial support provided to regions in transition is another critical determinant; greater financial assistance can alleviate economic pressure, leading to quicker recovery, reduced unemployment, and improved GDP outcomes. In addition to unemployment, exports, imports, and investments play vital roles in local economic development, which in turn affects unemployment and GDP [31,32]. In this study, we aim to account for the impact of unemployment indirectly by considering the variables influenced by it and their aggregate effect on GDP. Beyond investments, technology emerges as a crucial driver of regional development, particularly in areas undergoing a transition in their production models. This includes investments in human capital and research and development [33]. Furthermore, the supply of and demand for energy, influenced by the adoption of renewable energy sources (RESs) as part of decarbonization efforts, are expected to impact GDP [34,35]. Finally, climate change is incorporated into our model as a significant factor affecting GDP [36,37]. Given its role in shaping socio-economic and environmental conditions, it is essential to include climate-related variables in our analysis. Additionally, GDP growth and GDP loss variables are included to capture the effects of unknown factors and uncertainties, thereby providing a more comprehensive understanding of the influences on GDP in this context.

Having therefore identified the main factors that potentially affect GDP, we consider the following estimate and calculate its evolution using the mathematical formula:

$$GDP\left(t\right)={\int }_{t_0}^t(I\left(t\right)+R\left(t\right)+TI\left(t\right)+GG(t)-GL(t)-CI\left(t\right))dt$$

(1)

where each variable corresponds to the following quantities:

• I(t) = INVESTMENTS
• R(t) = RES IMPACT ON GDP
• TI(t) = TECHNOLOGY AND INNOVATION
• GG(t) = GDP GROWTH
• GL(t) = GDP LOSS
• CL(t) = CLIMATE IMPACT ON GDP

The Euler integration of the sum of the selected variables is numerically performed in the Vensim software. Vesim has been used in different cases in the past to simulate system dynamics in various scientific fields [38,39]. The actual GDP estimation accuracy and applicability of the selected variable combination is one the main contributions of the present work. Additional details and analysis on the background behind the selection and interaction of the quantities are given in the following subsections.

2.1. Theoretical Analysis

The approach combines a series of variables, creating a dynamic system and examining how all the variables affect local GDP. A part of the system involves studying local population growth, especially in areas undergoing radical industrial and economic changes, which result in depopulation [40], caused by compulsory labor migration, early retirements, and general migration. A decline in the size of the local population is directly conducive to economic stagnation, wealth loss, and other macro-economic effects [41]. Thus, the model encompasses population growth and a set of underlying variables, including working population and unemployment rates, as well as immigration and emigration. In addition, as demonstrated in the relevant research, population growth and economic growth are positively related [42,43].

Another fundamental aspect of the approach, economic development, is a primary force in generating wealth, fighting unemployment, and fostering socio-economic prosperity on a local and national scale. The research also includes the variable of the impact of renewable energy sources (RESs) on the local GDP. As investments (domestic or foreign) play a crucial role in economic growth [32,44,45], special emphasis should be placed on investments in RESs, as these are the predominantly investments made in regions undergoing decarbonization.

Finally, it is worth noting that the research also investigates the Education and Human Capital variable. The variable Education and Human Capital facilitates retraining and upskilling the local workforce and is often accomplished through collaboration with local universities and other educational institutions. Notably, a strong university fostering the education of new scientists, as well as a retrained workforce qualified with considerable industry- and energy-related work experience and culture become vital assets to attract new business activity [46].

In the extant literature, there are various scenarios contributing to the growth of coal-reliant regions, including biomass production in lignite mines [47], modeling the transition from coal to renewable energy sources [48], supporting the development of the energy sector [49], and exploring regional sustainable development using the Urbanization and Eco-environment Coupler [50]. However, there are also research efforts which explore coal-based energy and sustainable development from the perspective of the spatial field [51] rather than coal phase-out, which is more typical.

The research focuses on specific scenarios to explore the impact of increased funding of education and research, as well as the increase in financial aid from European and state resources and their efficiency. The selection of the variables was based on their ability to be more readily influenced and yield faster results within the model, compared to other, more complex variables. In addition, the outcomes of these scenarios can be used as political leverage to determine the allocation and size of funds. Based on scientific research, transition regions, bearing a substantial economic and social burden, can increase funding from European and national resources, and, thus, outweigh the total loss of local GDP. The urgency and importance of retraining and reskilling human resources cannot be overstated, especially in understanding the impact on local GDP when workers affected by decarbonization are retrained. From this perspective, the pivotal role of the university and its effectiveness in supporting the transition phase of the production model must be underscored. To reflect these critical aspects, the present research has incorporated the variables EDUCATION AND HUMAN CAPITAL and GOVERNMENT AND EUROPE SPENDING in the proposed scenarios.

2.2. Cause Tree of the Variable GDP

In the Vensim software, the “cause tree”, which is a valuable tool for visualizing and understanding the relationships between variables within a model, graphically illustrates how different variables are connected through cause-and-effect relationships, highlighting the dependencies and influence that one variable exerts over others within the system.

Figure 1 shows the relationships between the key variable, GDP, and the other variables used in the model. The cause tree is a simple tool, which provides a real and easy-to-understand picture of the model under consideration. The cause tree analysis methodology is widely used in statistics [52], machine learning, and engineering [53].

2.3. Casual Loop Diagram (CLD)

Casual loop diagrams are designed to provide a simpler and more understandable representation of the dynamic model, illustrating the positive or negative relationships between indicators and how they mutually influence each other. Therefore, the casual loop diagram is a conceptual tool, which facilitates the understanding of feedback loops and causal relationships in a system.

The analysis incorporates a casual loop diagram (CLD) to visually represent the cause-and-effect relationships between various variables affecting GDP (Figure 2), which is the major variable under examination. The arrows shown in the diagram describe the relationship between each factor within the system, highlighting their impact on system behavior. The plus symbol signifies a positive impact of one variable on another, whereas the minus indicates a negative one. The CLD illustrates both positive and negative impacts of variables on GDP, elucidating their influence on system behavior. Variables negatively affecting GDP include GDP LOSS, unemployment, the GDP decline index, and the impact of climate change. Positive impacts include investments, European and national funds, net exports, renewable energy footprint, and technology and innovation.

2.4. Stock and Flow Diagram

Although the casual loop diagram (CLD) provides a high-level understanding of the system structure and feedback loops, the stock and flow diagram can be used to add quantitative details about how variables change over time. Stock and flow diagrams are used to provide more detailed insights by offering a quantitative approach to system dynamics. Thus, each diagram serves a distinct purpose, complementing the others in the overall analysis.

The final structure of the setup is demonstrated using the stock and flow diagram in Figure 3. The flows, relationships, and fundamental structural elements of the system dynamics methodology were identified based on the Casual Loop Diagram (Figure 2) discussed earlier.

2.5. Data Collection

Data collection is based on various reliable sources, including Statista, Eurostat, Hellenic Statistical Authority, and the World Bank.

Table 1. Categorization of the variables.

Dimensions	Variables
Economic	GDP Growth, Exports, Imports, Government and European Spending, Investments, GDP Loss
Social	Births, Immigration, Deaths, Emigration, Population, Population Growth Rate, Labor Force Participation, Labor Force, Employed, Employment Rate, Unemployment, Unemployment Rate
Environment	Climate Change Impact on GDP
Technology and Education	Technology and Innovation, Educational and Human Capital, Educational and Human Capital Multiplier, R&D Investment, R&D Investment Multiplier, RES Investments, RES Multiplier, RES Impact on GDP

presents the categorization of the variables used and facilitates understanding the structure of the generated model. Estimations of indicators such as Export-to-GDP Ratio and Import Dependency Ratio, were employed to assess the impact of international trade on the economy.

More specifically, exports and imports of goods were obtained from Statista [54] unemployment rates from Eurostat [55], labor force-related information from the Hellenic Statistical Authority, and additional data from the World Bank [56].

The modeling process involves estimating a series of indicators, as defined in

Table 2. List of ratios.

Dimensions	Variables
Export-to-GDP Ratio (EGR)	EGR = (Total Exports/GDP) × 100
Import Dependency Ratio (IDR)	IDR = (Total Imports/GDP) × 100
Rate of Change of Unemployment (RCU)	RCU = [(FINAL/INITIAL) − 1] × 100
Labor Force Participation Rate (LFPR)	LFPR = LF/WAP × 100

. The Export-to-GDP ratio indicates the degree to which a country/region relies on international trade to drive its economy. A higher ratio generally suggests a greater reliance on exports in the economy, whereas a lower ratio implies that the economy is less dependent on foreign trade. A high Import Dependency Ratio suggests that a substantial part of a country’s economic activity relies on imported goods and services, which can have both positive and negative financial implications. As regards the Labor Force Participation Rate, LF represents Labor Force, defined as the sum of employed individuals and those actively seeking employment (unemployed), and WAP denotes Working Age Population, which includes individuals aged 16 or older.

2.6. Multipliers

The model also employed various multipliers, which are coefficients quantifying the relationship between an independent and a dependent variable and indicate the degree to which changes in independent variables affect dependent variables. In detail, RES Multiplier involves the economic impact of investments or changes in the renewable energy sector on a country’s GDP. Notably, calculating the RES Multiplier can be a complicated process, requiring the use of various methods and models by economists and policymakers. In addition, the Educational and Human Capital Multiplier, which represents the economic impact of investments in education and the development of human capital on a country’s GDP and overall economic growth, reflects how improving education and skills in the workforce can lead to increased productivity, innovation, and economic prosperity. Finally, the R&D Investment Multiplier measures the economic impact of investments in research and development activities on a country’s GDP and overall economic growth and reflects how investments in R&D can lead to technological advancements, innovation, increased productivity, and economic prosperity.

3. Results

The proposed application was designed with a time horizon of 20 years, from 2015 to 2035, with time step 1 year and initial values of the main variables (year 2015) as presented in

Table 3. Initial Values of Variables for the Region of Western Macedonia.

Variable	Initial Value
GDP	4,700,000,000 EUR
R&D INVESTMENTS	50,000,000 EUR
EDUCATION AND HUMAN CAPITAL	5,000,000 EUR
POPULATION	283,689 persons
UNEMPLOYMENT	24,845 persons
EMPLOYED	103,225 persons

, based on the data collected from ELSTAT and the World Bank for the Region of Western Macedonia. The application places particular emphasis on the impact of the GDP variable, which is focal in the present research. The GDP variable, which involves the wealth produced locally, is particularly underscored as a major indicator of local prosperity. When the wealth produced in a region increases, the multiplier effects on local economic development, investment growth, employment, and innovation also increase.

The applied parameters are categorized into two groups. The first group comprises parameters which remain constant, such as imports, exports, births, deaths, etc., whereas the second includes parameters which are subject to change, such as government and European spending, multipliers, unemployment, etc. The relevant simulations and scenarios enable examining the impact of possible policy actions and concern the case of the Region of Western Macedonia. The proposed scenarios (

Table 4. Scenarios and changing variables for the Region of Western Macedonia.

Variable	Basic Scenario	First Scenario	Second Scenario
EDUCATION AND HUMAN CAPITAL	EUR 5,000,000	EUR 15,000,000	-
GOVERNMENT AND EUROPE SPENDING	0.3 (index for resource efficiency)	-	0.33 (index for resource efficiency)

) assume that all model determinants remain stable, ceteris paribus, whereas only one specific variable is modified, thus enabling the investigation of the exact impact on the dependent variable, which is GDP.

The three possible scenarios proposed in the model are demonstrated in Figure 4, which provides a better visual representation and enables a comparative analysis between scenarios. The vertical axis represents local GDP growth, measured in billions of euros, whereas the horizontal axis the time horizon, i.e., the period of model development. Notably, Figure 4 demonstrates the basic scenario, the first scenario, in which all determinants are stable (ceteris paribus), except for the variable “EDUCATION AND HUMAN CAPITAL”, which exhibit resources increase, and, finally, the second scenario, in which the performance indicator of the variable “GOVERNMENT AND EUROPE SPENDING” is modified.

In detail, the basic scenario demonstrates a decline in the local GDP in the Region of Western Macedonia until 2020, followed by an upward trend until 2030, and next a downward trend. Analyzing the local GDP curve reveals three distinct stages. In the first stage, there is a decline resulting from the sudden and radical phase-out of lignite, resulting in deterioration of the local economy, business closures, and an increase in local real unemployment. In the second stage, from 2020 to 2030, the local GDP curve rises, as a result of the activation of European transition funds (the Just Transition Fund, etc.), when investments and employment in the specific region increase, mostly in energy-related projects, which involve wind turbines and photovoltaics. In the third stage, the local GDP curve exhibits a downward trend as European transition funds are exhausted, and there are no enterprises comparable to the Public Power Corporation, with a large workforce and a high value chain. Investments in RESs create numerous short-term job opportunities during construction; however, the majority of them are lost upon completion of the projects. Following the specific consideration, in the first scenario, the research investigates the impact on the local GDP growth resulting from an increase in capital for the variable “EDUCATION AND HUMAN CAPITAL”. In effect, the research aims at exploring the effect on local GDP when expenditure for the local population’s education and current employees’ retraining and reskilling in declining sectors are increased. Retraining and reskilling programs are mainly offered by the University of Western Macedonia and other regional educational institutions.

The initial amount in the proposed model scenario for the variable “EDUCATION AND HUMAN CAPITAL” was EUR 5,000,000 and increased to EUR 15,000,000, as shown in Table 2. As demonstrated, the GDP curve in the first scenario follows a similar trend to the curve of the basic scenario, whereas there is a variation during the first years, which implies that increased resources do not have a direct impact; subsequently, there is a slight improvement of local GDP.

The second scenario is a more extreme case, as performance indicators for the variables “GOVERNMENT AND EUROPE SPENDING” were modified. Remarkably, in the second scenario there is a change in examining resource efficiency, which implies exploring whether European and national funds can be used more effectively to create investments with strong value chains and local economies of scale. The results of the specific enhanced efficiency are illustrated as Scenario 2 (Figure 4), in which, despite having a similar overall curve behavior, higher levels of local wealth are achieved. In addition, the results of the different scenarios do not have a significant impact during the first years (i.e., the first 5 years) of the model implementation, and they reach a peak in about 2030, when the impact of various policies becomes more pronounced.

To test the accuracy and feasibility of the model, we compared the simulated results and historical data of our primary variable, local GDP, for the Region of Western Macedonia (Figure 5). The statistical analysis carried out here employs the following statistical measures: Mean of Observations, Mean of Estimations, Standard Deviation of Observations, Standard Deviation of Estimations, Correlation Coefficient, Mean Absolute Error, Root Mean Squared Error, and Bias (

Table 5. Statistical measures.

Name	Formula	Value	Unit
Mean of Observations	${\displaystyle \frac{{\sum }_{i=1}^n{X}_o}{n}}$	4.067	billion EUR
Mean of Estimations	${\displaystyle \frac{{\sum }_{i=1}^n{X}_p}{n}}$	4.336	billion EUR
Standard Deviation of Observations	$\sqrt{{\displaystyle \frac{{\sum }_{i=1}^n{(X_o-{\overline{X}}_o)}^2}{n-1}}}$	0.478	billion EUR
Standard Deviation of Estimations	$\sqrt{{\displaystyle \frac{{\sum }_{i=1}^n{(X_p-{\overline{X}}_p)}^2}{n-1}}}$	0.317	billion EUR
Mean Absolute Error	${\displaystyle \frac{{\sum }_{i=1}^n|X_p-X_o|}{n}}$	0.622	billion EUR
Root Mean Squared Error	$\sqrt{{\displaystyle \frac{{\sum }_{i=1}^n{(X_p-X_o)}^2}{n}}}$	0.689	billion EUR
Mean Bias	${\displaystyle \frac{{\sum }_{i=1}^n(X_p-X_o)}{n}}$	0.268	billion EUR
Correlation Coefficient	${\displaystyle \frac{{\sum }_{i=1}^n \left[\right(X_p-{\overline{X}}_p)·(X_o-{\overline{X}}_o)]}{\sqrt{ {{\sum }_{i=1}^n (X_p-{\overline{X}}_p)}^2 · {{\sum }_{i=1}^n (X_o-{\overline{X}}_o)}^2 }}}$	−0.466	-

where: X_p represents the predicted values, X_o represents the observed values, n is the number of observed values.

). The results suggest that model validity is satisfactory. The model overestimated GDP, starting from 2020, during the COVID-19 pandemic. However, both the simulated and actual GDP values exhibit the same trend and rate of change.

4. Discussion

4.1. Findings Analysis

This study, employing a system dynamics model, investigates the factors that influence regional GDP during the delignitization phase. By integrating societal, economic, and educational impacts on the production transition, the study addresses issues such as unemployment, financial support, and investments in human resources and R&D. It also takes into account the influence of climate change on GDP. The results, as depicted in the developed scenarios, underscore the importance of the study’s findings. They highlight the impact of increased funding on education, research, and financial aid efficiency, offering insights into enhancing GDP in regions undergoing decarbonization. The study reveals that a more significant investment in education and human capital leads to modest improvements in local GDP, though these effects take time. Enhanced government and European spending efficiency substantially boost local GDP by creating robust value chains and regional economies of scale. The findings indicate that increasing financial support to transitioning regions is paramount and has a multiplicative effect.

Although this study offers insightful information about how GDP is affected by some specific variables, it must be noted that it has several limitations. Despite the fact that the implementation of dynamic and integrated assessment models (IAMs) offer numerous advantages as an optimal approach to simulate and validate social, economic, and environmental conditions, they also possess inherent weaknesses, which are difficult to overcome, such as the formulation and acceptance of assumptions, often considered questionable. More specifically, IAMs often rely on simplifying complex environmental and socioeconomic processes due to the inherent difficulty of accurately modeling these systems, which can lead to underestimation or misrepresentation of critical dynamics and feedback loops. In addition, IAM outputs are highly dependent on the assumptions and scenarios used as inputs, such as those concerning population growth, economic development, technological progress, or policy decisions. Uncertainties in these inputs can propagate through the models and significantly affect outcomes, sometimes deriving speculative results or becoming overly dependent on initial assumptions. IAM outputs can also have a significant impact on policy decisions; however, the interpretation of these models can be contingent on political or ideological biases, which results in selective use or misrepresentation of model outputs. Most researched phenomena are multi-factorial, making it impossible to incorporate all variables into a model, pertinent constraints of data and information, etc.

To illustrate, with regard to renewable energy sources, most models do not take into account the potential impacts of future investments in power generation on the energy production system, to achieve the transition to renewable energy sources [57]. The specific models tend to rely on assumptions and variables as it is difficult to incorporate technological advancements, innovations, socio-economic shifts, and availability of raw materials [58]. In addition, most models are constrained to specific information, such as transportation or local energy systems modeling [59,60].

In the framework of the coal phase-out in Western Macedonia, there is no similar tool to provide a substantial representation of the transition to a new production system by means of a dynamic approach. The specific model could function as a tool for various transition observatories, such as the METAVASI S.A. (https://metavasi-dam.gr/en/), the upcoming Lignite Phase-out Observatory, etc., as well as for relevant ministries and policymakers.

However, there are other factors which can positively or negatively affect local GDP, such as the number of business closures or start-ups, global factors such as the COVID-19 pandemic [61,62,63] with an impact on global economy, the Ukraine-Russia war afflicting the energy and grain markets, etc. As such, the model can be further elaborated, by allowing additional variables in any case and region.

In this framework, the present research serves a dual purpose, to highlight the results of simulating a crucial socio-economic problem, such as lignite phase-out, and to explore it by employing the system dynamics methodology. In addition, future research should be carried out to create more enhanced dynamic models and enable efforts to establish a platform for communication between the private and public sector, promote socio-economic modeling, and contribute to developing cutting-edge tools both on a national and international scale. Tο generalize the results: future research should expand the study’s geographical scope to include various regions with varying technological and economic contexts. Additionally, incorporating more dynamic policy variables would enhance the model’s relevance and predictive accuracy. A deeper understanding of the strategic planning of regional transition may be gained by examining socio-technical approaches [64], new regional innovations systems (RISs), as a strategic tool for enhancing a region’s competitiveness [65], the long-term effects of massive investments in RESs [66,67], and how they affect the environment and the sustainability of regions.

This research significantly contributes to sustainability by examining the inter-connected factors influencing regional GDP within the context of decarbonization and sustainable development. The study underscores the importance of investing in human capital and technology, demonstrating that such investments lead to sustainable long-term economic growth and enhanced resource efficiency. By highlighting the need for efficient use of government and European funds, the research shows how optimized financial aid and public investments can strengthen value chains and economic resilience. It also addresses climate impacts by offering insights into building climate-resilient economies through targeted investments and adaptation strategies. Additionally, the research advocates for economic diversification away from fossil fuels, encouraging the development of green industries that drive sustainable economic growth. The comprehensive system dynamics model used in the study integrates these factors, providing a holistic framework that captures their complex interdependencies and offers scalable, transferable insights. This empowers policymakers to design and implement effective sustainable development strategies globally, ensuring that sustainability considerations are embedded in economic planning and promoting resilient, low-carbon economies that support both environmental and economic well-being.

4.2. Policy Implications

Based on the findings from the simulations, we can make the following recommendations for policymakers in the Region of Western Macedonia. On the investment side, it is crucial to implement tax incentives, subsidies, or grants to encourage private investment in critical sectors such as infrastructure, renewable energy, and technology. These measures, along with the development of favorable policies and a stable regulatory environment, are urgent to attract foreign investments that can bring in capital, technology, and expertise. The potential benefits of these incentives are significant, including the reduction of VAT to a rate of 2% or 5% for the first ten years of the investment, the electricity subsidy in the first years of operation, the simplification of bureaucratic procedures, and the increase in financing rates for the establishment of new businesses. In terms of the research of technology and innovation, increasing funding for research and development is vital to spur innovation across various sectors of the economy. We also propose exceptional support for startup businesses and innovation hubs, which act as accelerators of innovation and further economic development. Diversifying the economy is another critical strategy that encourages diversification to reduce reliance on a single industry or sector, creating a more resilient economy. Furthermore, supporting the unemployed with subsidized training programs in reskilling and upskilling and subsidizing businesses to hire the long-term unemployed are necessary actions that policymakers must take to support the local GDP and address the problems of unemployment and forced immigration. By addressing these factors through targeted policies, regions can enhance their economic resilience, promote sustainable growth, and improve overall well-being, underlining the situation’s urgency.

5. Conclusions

Regions currently undergoing or anticipated to undergo decarbonization processes will face significant financial and social changes. In this study, the Vensim software, a robust tool for system dynamics modeling, is employed to evaluate the extent to which external factors can influence the development of local GDP in the Region of Western Macedonia, Greece—a region in transition and phasing out fossil fuels. GDP is used here as it is the most comprehensive economic indicator for exploring economy, growth, and development.

The results indicate that increased investment in education and human capital leads to a modest improvement in local GDP, while enhanced efficiency in government and European spending significantly boosts local GDP. Additionally, increased financial support to regions in transition proves to be of paramount importance due to its multiplicative effects, underscoring the need for the European Union to augment its financial support mechanisms.

The methodology employed in this study offers a valuable tool for economic development in regions undergoing transition and fossil fuel phase-out. However, it is important to note that this study is based on specific assumptions and simplifications. Future applications can be refined by incorporating additional variables and development scenarios, making the approach adaptable to various contexts and regions.