A Comprehensive Analysis of Renewable Energy Based on Integrating Economic Cybernetics and the Autoregressive Distributed Lag Model—The Case of Romania

Abstract

Renewable energy represents a crucial resource in the efforts to combat climate change and reduce dependence on fossil fuels. In the past few decades, Romania has lessened its environmental footprint and played an important role in actions against climate change. In this research, the energy sector in Romania is analyzed from a holistic perspective as a complex adaptive system by using econometric tools. The purpose of the research is to analyze the Romanian energy sector as a cybernetic system and to study the long-run and the short-run causal impact of greenhouse gas emissions (GHG) and renewable energy (RE) on real GDP per capita. The causality among GHG, foreign direct investment (FDI), RE, and real GDP is checked by means of the autoregressive distributed lag model (ARDL). The time series are extracted from Eurostat and OECD databases and cover the period 2000-2021. The results reveal that (i) the variables are cointegrated according to the ARDL bounds test; (ii) in the long run, GHG negatively impacts GDP, RE positively impacts GDP; (iii) in the short run, GHG and RE positively impact GDP; (iv) the speed of adjustment is around 32%. The study holds significance both for scholars and the policy makers from the governmental environment agencies that should decide how to effectively reduce GHG emissions, promote renewable energy adoption, and design policies to facilitate the transition to a low-carbon economy.

1. Introduction

Renewable energy in Romania has gained more and more importance in recent years, following efforts to diversify the energy source and reduce dependence on fossil fuels.

From an economic perspective, Romania has significant potential in terms of renewable resources, such as solar energy, wind energy, biomass, and hydropower [1,2,3]. Also, the country has extensive surface areas with favorable exposure to the sun and wind, rivers, and lakes, with potential for hydropower plants and biomass resources from agriculture and industry [4]. Additionally, the utilization of renewable energy can lead to prospects for economic growth and employment creation in Romania [5]. The renewable energy industry encompasses various activities, such as designing, constructing, and maintaining the infrastructure and the related equipment. This might encourage steady economic growth and diversification of the economy [5,6].

Using an economic cybernetics methodology and an econometric analysis of renewable energy consumption, the current study aims to give a thorough overview of the Romanian renewable energy industry. It is a novel topic that our research addresses by assessing the state of the art in the field. This study, by the holistic approach from the angle of economic cybernetics and econometric models on the Romanian renewable energy sector in Romania, fills in the gap in the existing literature.

Renewable energy in the five EU member states in Southeast Europe—Bulgaria, Croatia, Greece, Romania, and Slovenia—has experienced remarkable growth in recent years and is poised to reach unprecedented levels, as indicated by the regional report [7] “Renewable Energy in Southeast Europe–a regional report by SeeNext”. According to the report [7], the sector witnessed a 30% increase in operating revenues in 2021, amounting to EUR 12.9 billion, surpassing the previous year’s figures.

A more in-depth comprehension of the interactions and dynamics of this field can be obtained by looking at renewable energy in Romania by automatically incorporating the renewable energy industry into a complex adaptive system from the viewpoint of economic cybernetics.

To increase the effectiveness and sustainability of this industry, economic cybernetics can be used in Romania to monitor and manage energy systems, model the economy, optimize its performance, and integrate smart technology.

Economic cybernetics can assist in locating the most effective options for the production, storing, and distribution of renewable energy in Romania through economic modeling and optimization.

Through optimization algorithms, the best locations for wind or solar farms can be determined, considering variables such as wind intensity, solar radiation, and associated costs. Monitoring and controlling energy systems through advanced technologies and sensors can enable the collection of data on energy production, consumption, and the state of energy infrastructure. This information can be used to optimize the operation of systems, identify potential faults, and adapt production according to demand. In a complex and dynamic context, this guarantees the efficient and consistent operation of the renewable energy sector.

One of the fundamental characteristics that we need to identify in the renewable energy sector in order to start analyzing it as a complex adaptive system is to examine the involved agents and the feedback mechanisms of interaction among them. Regarding the integration of this sector in Romania, the energy market, regulatory policy makers, and the environment interact with each other and form feedback mechanisms for regulation and self-regulation at the level of the renewable energy sector.

Political, scientific, and commercial advancements can have unforeseen and unanticipated effects on the renewable energy industry; as a result, this system’s adaptability and resilience are essential for overcoming challenges and seizing opportunities. So, understanding and enhancing the performance of this sector can be aided by the integration of renewable energy into a complex adaptive system in Romania and the economic cybernetics approach to the current study problem. Through mathematical models, econometrics, optimization algorithms, advanced monitoring, and systems analysis, we can contribute to increasing energy efficiency and sustainability in a constantly changing economic environment.

The connection between the economic cybernetics approach and ARDL model lies in their application in analyzing and modeling economic systems. The economic cybernetics approach focuses on understanding the feedback mechanisms and dynamic interactions within an economic system. ARDL, on the other hand, provides a methodological framework to estimate the long-term and the short-term relationships between time series. The economic cybernetics approach and ARDL model are connected through their shared objective of studying the dynamics of economic systems and their feedback mechanisms. The ARDL approach serves as a statistical tool to estimate the long-term and the short-term relationships, which can be integrated into the broader framework of economic cybernetics to optimize economic systems.

The research hypotheses are described by the way significant relationships are constructed between key economic variables and the development of renewable energy sources in Romania, illustrated through the ARDL econometric model. They are also shaped by the integration of economic cybernetics concepts into the analysis of renewable energy, which can provide a holistic and efficient perspective on the evolution of the energy sector in Romania.

The research presents the following structure: Section 2 presents the current state of knowledge in the field. The most relevant research in the sphere of econometric models applied to the renewable energy sector has been selected, along with specific studies analyzing this sector from the holistic perspective of economic cybernetics. Considering the novelty of this article in approaching a holistic perspective from the viewpoint of cybernetic systems, the current literature can be complemented with such a perspective. In Section 3, the cybernetics approaches from a conceptual perspective are described, which will be used in Section 4, as well as the methodology of the ARDL model used in Section 5. Section 4 and Section 5 highlight the practical case study of the research. The research concludes with the limitations, future research, and conclusions section.

2. The Stage of Knowledge in the Field

The EU is in a transition to low-carbon economy, which means lower amounts of GHG emissions generated from fossil fuel combustion. Climate change is mainly caused by GHG emissions. These emissions, generated from human activities such as burning fossil fuels, deforestation, and industrial processes, trap heat in the Earth’s atmosphere, leading to high temperatures and climate changes. The main GHGs include CO₂, methane, nitrous oxide, and water vapor. Low-carbon energy comes from various sources such as wind, hydro, geothermal, solar and nuclear power, biomass, and biofuels. Fossil fuels are the source of 71% of GHG emissions [8]. GHG emissions can be reduced by withdrawing gradually the use of fossil fuels and using instead renewable energy resources. The main types of fossil fuels are gas (natural gas), oil (petroleum), and coal. These fuels are nonrenewable resources, used for electricity generation, transportation, and industrial processes. Their combustion releases CO₂ and other GHGs, contributing to climate change and environmental pollution. Decreasing the dependence on fossil fuels and mitigating their detrimental impact on the environment have become a priority,

Androniceanu and Sabie [9] make an extensive overview of the renewable energy literature and their contribution to sustainable development. It is stated that from 2019 until 2021, the share of low-carbon energy sources increased from 15.7% to 22% as primary energy. There are several directions in the literature on renewable energy. One direction deals with renewable energy policies such as financial incentives, loans, and renewable portfolio standards (RPS) [10]. Tax incentives are the main financial incentives applied by governments to stimulate the development of renewable energy. According to Ogunlana and Goryunova [11], the US is the first country to apply the tax incentive mechanism. According to the same study, some European countries use property-tax deductions to promote investments in low-carbon technologies. He et al. [12] analyze the efficiency of bank loans on renewable energy investment in China and conclude that the effect is negative by reducing loan issuance. Bento et al. [13] find that for the US, RPS can stimulate resources or energy saving, but not both. This is because RPS can be accomplished either by stimulating renewable energy or by reducing conventional sources of energy production. Bara et al. [14] study the long-run and short-run causality between the electricity price, energy consumption, EU inflation, and gas quantity during 2019–2022 by means of the ARDL technique.

Another direction in the literature concerns the correlation between renewable energy consumption and economic growth. Study [15] explores a panel of 28 EU countries finding a unidirectional causality from GDP to RE. This implies that the amount of renewable energy consumed by developed countries is higher. Study [16] finds that energy consumption and GDP are not correlated in Turkey.

According to Matei [17], the study on OECD countries spanning 1990 to 2014 emphasizes a negative relation between economic growth and renewable energy consumption. This could be motivated by inadequate investments or the existence of underdeveloped energy industries. Armeanu et al. [18] explore the causality between renewable energy sources and economic growth for EU countries during 2003–2014. Among the sources of renewable energy, biomass has the most important influence on economic growth. Shpak et al. [19] discover that the macroeconomic determinants of CO₂ emissions, imports, exports, inflation rate, and unemployment are better explained for EU countries than for USA. Krzywda et al. [20] discuss the decarbonization process of Poland, which has a tradition of using coal as a main energy resource. Androniceanu et al. [21] study how to increase energy efficiency with the Kaizen approach.

Another direction in the literature is focused on the relationship between air pollutants and economic growth. This connection has been a subject of extensive debate since the 1970s. Grossman and Krueger [22] introduced the environmental Kuznets curve (EKC) as a graphical representation of this relationship, depicting an inverted U-shaped curve describing the dependence between GDP and environmental pollution. The dynamics between air pollutants and economic growth are intricate. Some research suggests a positive correlation, implying that economic growth increases with an increase in CO₂ emissions [23,24]. Other studies propose a decoupling, leading to a negative correlation [25], indicating that economic growth can occur without an increase in CO₂ emissions. The complexity of this relation stems from numerous factors, such as variation in industrial structure, energy efficiency, policy measures, technological advancements. The study by Kim et al. (2020) [26] examines the impact of various factors on Korean GHG emissions during 1981–2014 by means of an ARDL model. The factors under investigation include FDI, GDP, renewable and nuclear energy, and urbanization.

This study continues the recent research on renewable energy in Romania [14,27,28]. Chirita et al. [27] explore the renewable energy in Romania in the framework of the circular economy as a cybernetic system. Also, they discover a long-run negative correlation between GDP and renewable energy consumption in Romania [27,28] from 2000 to 2021.

Regarding cybernetic perspectives, Xu et al. [29] state that climate change, caused by the increase in anthropogenic carbon dioxide (CO₂) emissions from large-scale human activities, has led to significant environmental and humanitarian crises in recent years. Some of these extreme weather events are so severe that they threaten development in many regions around the world. If CO₂ emissions are allowed to continue increasing without the imposition of appropriate restrictions, climate change and related natural disasters will persist in impacting the environment, posing a serious threat not only to human existence but to all life on Earth. The authors argue that regional development cybernetics is an innovative approach that utilizes the principles of cybernetics to analyze and address issues related to the low-carbon economy. This approach focuses on using systemic models to develop a problem framework and modeling system, aiming to understand and integrate the complex relationships among various components of the low-carbon economy. The primary objective of this approach is to drive economic growth, improve social progress, and maintain ecological balance through the promotion of renewable energies and the reduction of CO₂ emissions [29]. Thus, the use of cybernetics at the regional level provides an efficient method to overcome challenges related to climate change and promote sustainable development with reduced carbon emissions.

Transitions to renewable energy represent a global challenge driven by climate change. These transitions are complex and nonlinear, requiring bold leadership and innovative policies to address environmental risks and navigate sociopolitical and technological environments. In this context, Person and Bardsley [30] integrate theories of complex adaptive systems and risk society to explore regional sociotechnical transitions. South Australia serves as a relevant example, leading energy transitions in the country. Using the complex adaptive cycle, its evolution towards a multidimensional modernity is highlighted, going beyond initial simplistic approaches. Sociotechnical transitions are influenced by responses to risks. While there is no standard approach to navigate these transitions, it is proposed that risk society theory can provide an important dimension for understanding fundamental changes in sociotechnical systems.

In his research, Naude [31] presents complex adaptive systems (CAS) and their principles as an alternative to traditional management models to enhance sustainable development (SD). The author proposes management and leadership suggestions to implement CAS principles for improving SD. The CAS approach allows for adaptability and emergence, aligning with the ever-changing nature of SD challenges in the organization’s internal and external environments. Translating this approach may involve cognitive, structural, and political changes in thinking and understanding how to address SD.

3. Materials and Methods

3.1. Models and Approaches of Economic Cybernetics

Economic cybernetics is an approach that utilizes the principles of cybernetics to understand and solve complex economic issues. It focuses on using systemic models to develop problem structures and modeling systems to scientifically describe the relationships between different components of the economy and integrate their functions. In the context of renewable energy, economic cybernetics can be employed to evaluate and enhance the efficiency of using renewable sources and address challenges related to transitioning towards clean energies.

Cybernetics of economic systems is a field that applies principles, methods, and specific models of cybernetics to study, analyze, and manage complex economic systems in dynamic and complex environments [32]. According to this field, to understand these complex systems, the emphasis is placed on understanding the interactions between them and the mechanisms of feedback loops that form within economic systems [33]. The goal of a holistic analysis from the perspective of cybernetics in economics is to develop models and approaches that aid in optimizing and regulating economic processes, improving decision making, and achieving desired economic outcomes. By using the tools and concepts of cybernetics, economic cybernetics aims to address issues related to economic stability, efficiency, and adaptability [32,33,34]. It provides a framework for analyzing and predicting the behavior of economic systems and for developing strategies to enhance their performance [33,34].

One important category of systems, which is the subject of the analysis and diagnostics of economic systems, is that of cybernetic systems capable of self-regulation through internal factors encapsulated in a block called the decision subsystem [35,36,37]. Understanding and diagnosing an economic system, based on system analysis methodologies, involves studying the system itself, its inputs, outputs, and the concrete ways in which inputs are transformed into outputs (system functionality). The inputs and outputs of a system, analyzed as causal relationships between systems and/or subsystems, constitute the structural-functional connections of the analyzed system’s elements. Studying these connections, an essential element of analysis and diagnosis, holds great importance in identifying the internal and external behavior of the economic system. The systemic approach is possible by resorting to an isomorphic representation of the real system, which takes the form of a conceptual, graphical, mathematical model, or in a modern and current sense an agent-based model [35,36,37].

Figure 1 illustrates the way in which a system is analyzed from a cybernetic perspective, in which the different feedback mechanisms are identified that the control operator processes, either for the purpose of stabilizing the system, or to multiply a certain effect.

The current object of study of economic cybernetics is the application of principles, methods, and concepts of cybernetics to analyze, model, and manage complex economic systems in dynamic and uncertain environments in which they operate [32,33,37]. All these are described by complex adaptive systems (CAS) and the properties of these types of systems.

Complex systems represent a new scientific frontier that has emerged in recent decades with technological advancements and the study of new parametric domains in natural systems. These systems present difficulty in predicting their future behavior due to interactions between their components, which conceal specific individual features. They challenge reductionism, which applies only to complex systems [38]. Although it is not yet clear whether complex systems follow strict laws, there are methods that allow for the treatment of some dynamic properties. These methods focus on the self-organization of systems during the transition from complicated to complex, utilizing new paradigms of representing trajectories associated with natural system invariants. These methods are qualitative in nature and prioritize robustness to compensate for uncertain variations in system parameters and functions. Control plays a significant role in complex systems, particularly in industrial applications, where transforming a complicated system into a complex one can enhance overall performance. Achieving this requires greater intelligence delegation to the system and establishing a well-defined control law. The method involves the use of the concept of an equivalence class, where the system is constrained to stay within a specific framework due to the action of the control law constructed from the mathematical representation of the system’s dynamics [38].

Many eminent scientists have been drawn to the study of complex adaptive systems (CAS) over the past 30 years, including notable Nobel laureates like Murray Gell-Man, Ilya Prigoine, and Thomas Shelling, among others. When seen as a complex adaptive system, the economy is composed of many different cybernetic systems, from small businesses to the entire global economy [33,39,40]. Each of these cybernetic systems exhibits the well-known properties of complex adaptive systems (interdependence and connectivity, coevolution, dissipative structure, far-from-equilibrium behavior, historicity, fundamental feedback processes, trajectory dependence, self-organization, emergence, and adaptation). However, each of these systems also possesses specific properties that make them diverse and determine their behavior and functioning toward achieving their own objectives and purposes.

Complex adaptive systems are found all around us, and the sciences of complexity assert that the vast majority of real-world systems are complex. Due to the abundance and diversity of these systems, defining them becomes quite challenging, and perhaps even more difficult, attempting to establish the general principles and properties that give them specificity within the broader category of complex systems. For example, Dooley [33,41,42] states that three principles form the basis of defining a complex adaptive system: emergence of order and control, irreversibility of history, and uncertainty of the future.

The creation of order and control in the Romanian renewable energy sector, for instance, is a result of the intricate interactions between various renewable energy sources, the energy infrastructure, and the energy policy. As different components of the sector interact, models and structures can be developed to enable more efficient control and organization of the renewable energy system. Moreover, decisions made in the past, such as investments in specific types of energy sources, infrastructure, or technologies, can have long-term effects and influence the future development of the sector, highlighting the principle of historical irreversibility. In a complex adaptive system, the future is challenging to predict due to complex interactions and multiple influences on the system. The third characteristic proposed by Dooley can be highlighted by the fact that energy prices are subject to uncertain and continuous changes in the surrounding environment, impacting the renewable energy sector. Illustrating the system’s intricacy and adaptability as well as the implications for management and decision making in the search for effective and sustainable energy solutions, the ideas advanced by Dooley can thus be applied to the Romanian renewable energy industry.

Another researcher, Mitleton-Kelly [43], considers that a CAS is described by ten generic characteristics: connectivity, interdependence, coevolution, historicity, trajectory dependence, far-from-equilibrium behavior, exploration of possibility space, feedback, self-organization, and emergence.

Another valuable tool provided by the field of economic cybernetics is the viable system model (VSM) [44,45]. It was initially used in the cybernetic analysis and synthesis of enterprises because it allows for a qualitative evaluation of strategies used to deal with the inherent complexity of cybernetic systems. As a result, it enables the development of useful tools for operational diagnosis and optimal decision-making processes [45]. VSM was developed by the pioneer of managerial cybernetics, Stafford Beer [46], who built upon the cybernetic concepts of Norbert Wiener, McCulloch, and Ross Ashby. Stafford Beer’s studies led him to envision a connection between the functioning of the human organism and the functioning of an organization. He viewed the human organism as composed of three interdependent parts: organs and muscles, nervous systems, and the external environment (body, brain, environment). Later, he identified five interactive systems within the human body (S1. totality of muscles and organs; S2. sympathetic nervous system; S3. the basic brain; S4. the middle brain, diencephalon; S5. the functions of the higher brain, cortex), which he analogically adapted to five systems applicable to an organization:

➢

System 1: Primary Activities;

➢

System 2: Conflict Resolution;

➢

System 3: Internal Regulation, Optimization, Synergy;

➢

System 4: Adaptation, Action Planning, Strategy;

➢

System 5: Policies, Final Decisions, Identity.

Today, by applying the viable system model (VSM), complex systems can identify and address the challenges of complexity and uncertainty, ensuring their adaptability and long-term success. Although the model was initially used for the analysis and design of enterprises, it can be applied in other contexts as well, such as strategic decision making and operational diagnosis within organizations or complex systems [47].

To complete a comprehensive diagnosis of an economic system from a cybernetic perspective, it is essential to conduct an analysis of both the internal environment (such as organizational structure, human resources, internal processes, etc.) and the external environment in which the system operates. The PESTELE analysis [35,48,49], which entails assessing and comprehending the political, economic, social, technical, environmental, legal, and ethical elements that can affect the economic system, can be used to analyze the external environment.

PESTELE analysis [35,48,49] is an important strategic tool that helps identify and understand the risks and opportunities in the external environment of an organization or economic system. This analysis assists managers in anticipating changes and trends in the economic environment and making informed decisions to address challenges and capitalize on existing opportunities.

Political factors can influence the business environment through government policies, legislation, regulations, and political stability. Economic factors relate to aspects such as economic growth, inflation, interest rates, exchange rates, and the overall economic health of the country or region in which the economic system operates.

Social factors include demographic changes, consumer preferences and cultural values, behavioral trends, and lifestyles. Technological factors cover technological advancements, innovations, and their impact on the business and industry.

Environmental factors pertain to aspects related to sustainability and environmental protection, while legal factors focus on laws and regulations that can affect the activities of the economic system. Additionally, ethical analysis can assess the ethical behavior and practices of the economic system, as well as its social and environmental impact.

By combining the PESTELE analysis with the viable system model (VSM), a comprehensive and coherent diagnosis of the economic system can be achieved, enabling the identification of issues and opportunities and the formulation of appropriate strategies to enhance the performance and resilience of the system in response to changes in the external environment.

For a more detailed understanding of the external environment in which an economic system operates, the analysis of Porter’s five forces [35,37] can also be conducted. Porter’s five forces is an analytical framework used to comprehend the competitiveness of an economic sector and its influence on the performance of other actors within that sector. These forces include the threat of new entrants, bargaining power of buyers, bargaining power of suppliers, threat of substitute products or services, and rivalry among existing competitors. Together, Porter’s five forces, the viable system model, and PESTELE analysis can provide a more comprehensive and holistic perspective on the business environment and the economic system, enabling companies and organizations to identify opportunities and threats and develop appropriate strategies to enhance their competitiveness and adapt to changes in the external environment [49].

3.2. ARDL (Autoregressive Distributed Lag) Econometric Model

The ARDL econometric model is a statistical method used to analyze long-term relationships between economic variables. It enables researchers to assess the long-term impact of certain factors on other variables and identify causal links between them.

In economic theory, the long-run causality among the variables can be expressed by constant mean and variance over time. The long-run causality is modeled by the cointegration technique. Cointegration assumes that in the long term, the time series evolution will be similar, even if there are periods where they suffer changes, caused by shocks. The time series nonstationarity may lead to spurious regression. Cointegration is a tool when dealing with nonstationary time series.

Over time, several cointegration techniques have been employed, including those introduced by Granger [50], Engle and Granger [51], Johansen and Juselius [52], ARDL developed by Pesaran and Shin [53], and Pesaran et al. [54]. The ARDL model is applied for variables stationary at I(0) or I(1) or a combination of I(0) and I(1).

The advantages of the ARDL model over other econometric techniques are, according to Adebayo et al. [55], that it comprises both long-run and short-run coefficients, it is successfully applied for small size datasets, and it eliminates the issue of autocorrelation.

The generalized ARDL model has the following equation:

$$B_t=a_0+{\sum }_{i=1}^p{b}_i{B}_{t-i}+{\sum }_{i=0}^q{c}_i{A}_{t-i}+{\epsilon }_t$$

(1)

where $B_t$ is the dependent variable expressed as a function of its lagged values and the lagged values of the independent variables comprised in vector $A_t$. p and q are often determined by the Akaike information criterion (AIC), which is more commonly used for small sample sizes of fewer than 60 observations [56]. b and c are constants, called short-run coefficients. p and q are optimal lag numbers for the dependent and the independent variables. ${\epsilon }_t$ is the error term, the errors being serially uncorrelated or independent.

Steps of the implemented methodology:

i.

Checking the stationarity:

Stationarity is an important feature of time series, meaning that mean and variance are constant over time. The common practice for checking data stationarity involves applying ADF test [57] or PP test [58].

ii.

Cointegration:

A significant contribution to the ARDL model is the bounds test, which checks out whether the mixture of I(0) and I(1) variables have a long-run relationship, i.e., they are cointegrated. The ARDL bounds test model has the form (2):

$$\mathsf{\Delta }{B}_t={\lambda }_0+{\sum }_{i=1}^p{\lambda }_i\mathsf{\Delta }{B}_{t-i}+{\sum }_{i=1}^q{\mu }_i\mathsf{\Delta }{A}_{t-i}+{\phi }_1{B}_{t-1}+{\phi }_2{A}_{t-1}+{\zeta }_t$$

(2)

where ${\sum }_{i=1}^p{\lambda }_i\mathsf{\Delta }{B}_{t-i}+{\sum }_{i=1}^q{\mu }_i\mathsf{\Delta }{A}_{t-i}$ represent the short-run terms and ${\phi }_1{B}_{t-1}+{\phi }_2{A}_{t-1}$ represent the long-run terms.

Based on Pesaran et al.’s findings [54], if the calculated F-value is significant, we will proceed to construct the error correction model (ECM) and analyze both the long-run and short-run dynamics.

F-calculated is compared to the critical values from [54]. The ECM equation has the form (3):

$$\mathsf{\Delta }{B}_t={\lambda }_0+{\sum }_{i=1}^{p-1}{\lambda }_i\mathsf{\Delta }{B}_{t-i}+{\sum }_{i=1}^{q-1}{\mu }_i\mathsf{\Delta }{A}_{t-i}+\phi EC{T}_{t-1}+{\xi }_t$$

(3)

In Equation (2), the lagged variables $A_{t-i}$, $B_{t-i}$ are replaced in Equation (3) with error correction term (ECT). The coefficient of ECT should be negative, larger than −2, and statistically significant for the convergence of the model to equilibrium. If $\phi$ fulfills these conditions, then the variables are cointegrated. The ECT coefficient represents the speed of adjustment to equilibrium.

According to Asteriou and Hall [59], ECM has some advantages. It measures the correction from disequilibrium; in its expression appear first differences that remove the variables’ trend. The ECM corresponds to the parsimonious ARDL model. According to Asteriou and Hall [59], the most important property of ECM is that the error term is a stationary variable; in this way, stationarity prevents long-term errors from having higher values.

iii.

Diagnostic and stability tests:

The Breusch–Godfrey serial correlation LM test, the Breusch–Pagan Godfrey heteroskedasticity test, and the Jarque–Bera normality test are performed to check the robustness of the ARDL model. The Ramsey reset test estimates the structural stability of the model. CUSUM and CUSUMSQ tests are graphical representations that confirm the stability of the model at 5% level of significance. If the blue line lies within the dotted red line, it means that the model is stable.

iv.

Variance decomposition and impulse response functions:

Variance decomposition and impulse response functions (IRFs) are applied as an indispensable tool in an ARDL model in the economic growth–energy consumption relation [60]. IRFs reveal the dynamic behavior of the variables and show the effect of a shock of the independent variables to the dependent one. To obtain a stable model, IRFs must converge to 0 in time horizon. IFRs are used to predict the effects of shocks and in policy control. Forecast error variance decomposition shows the variation in a variable explained by different impulses.

The ARDL approach can be a useful tool in analyzing the renewable energy sector, as it presents a range of advantages, such as treating cointegration, data structure, flexibility, and both short- and long-term analysis [61,62]. Treating cointegration among time series is important in the analysis of the renewable energy sector, where long-term relationships between variables can be significant.

In cases of limited or evolving data, as is common in the renewable energy sector, the ARDL approach can operate effectively with small-to-moderate dataset sizes. Furthermore, ARDL can be employed to analyze both short-term and long-term relationships, enabling the investigation of the influence of renewable energy in both the immediate and distant future.

However, like any statistical method, the ARDL approach may also come with certain limitations or disadvantages [61,62]. ARDL relies on certain assumptions, such as the presence of white noise error terms, the normality of errors, and the absence of autocorrelation. These assumptions need to be verified before interpreting the results. On the other hand, determining the optimal order of lag can be challenging, and an incorrect choice could impact the analysis outcomes. Furthermore, ARDL may require a significant number of observations to ensure robust results. Limited data can influence the precision of the analysis.

4. Cybernetics Analysis and Diagnosis of the Renewable Energy Sector in Romania

4.1. Energy Sector Overview in Romania

Romania is endowed with abundant energy resources, which positions it with the potential for energy self-sufficiency for many years to come. Compared to other European countries, it possesses considerable reserves of fossil fuels, including domestic sources of natural gas, crude oil, and coal (mainly lignite) [63].

Currently, Romania’s national energy sector is confronted with diverse challenges on both global and local levels: ensuring electricity supply security, intensifying competition, and mitigating environmental impact by reducing greenhouse gas emissions. Romania aims to bridge the economic gap with more developed EU countries. Aligned with EU directives (Renewable energy directive [64], Energy efficiency directive [65], Electricity market design [66], Directive (EU) 2018/2001 [67]), one of Romania’s principal challenges is to deliver competitive and environmentally friendly electricity, given climate change, the escalating global demand for electricity, and uncertainties surrounding conventional power sources. Globally, investments in green electricity are on the rise, and Romania is following this trend. In the immediate future, substantial financial efforts are needed to foster renewable energy development, driven by the ongoing green certificate scheme promoting renewable energy production. In the long term, research and development investments render renewable energy a viable energy supply solution. By 2030, Romania intends to reach a new goal of 30.7% renewable energy in gross final energy consumption [2].

In recent years, Romanian investments in renewable energy have moved from hazy goals to sound financial foundations.

The principal energy sources used in Romania to produce electricity in 2021 showed the following average specific values at the national level, according to the report of the National Energy Regulatory Authority (NERA) in Romania [4]: CO₂ emissions were 217.24 g/kWh, and radioactive waste resulting from electricity production registered a value of 0.003 g/kWh. These values allow electricity suppliers to elaborate labels to inform end customers whether the electricity supplied in 2021 had an impact on the environment below or above the national average.

The Figure 2 shows that renewable energy sources accounted for 45.47% of the total electricity produced in Romania in the year 2021. The largest contribution comes from the renewable source of hydropower, followed by wind and solar energy sources. The highest CO₂ emissions are attributed to coal processing, followed by oil shale processing.

Throughout 2021, there has been an increasing interest among investors to develop new energy capacities, either based on renewable energy sources or conventional sources utilizing modern technologies with a lower environmental impact. This trend is evidenced by NERA issuing eight establishment permits for energy projects, amounting to approximately 80 MW in total, compared to only four permits totaling 13.59 MW issued in 2020.

The Green Certificates system established by Law No. 220/2008 provides coverage for electricity produced from renewable resources such as hydro, wind, solar, biomass, and biogas [68,69]. It covers electricity delivered to the grid or directly to consumers from new or modernized power plants, including the commissioning period, and own consumption locations connected to the power plant’s buses. From the perspective of the promotion system for electricity produced in renewable energy power plants with an installed electric power of up to 100 kW, this applies to prosumers who do not benefit from the Green Certificate promotion system. Prosumers with renewable energy power plants of up to 100 kW installed capacity can sell the electricity produced and delivered to the electricity suppliers’ grids with whom they, as final consumers, have concluded or will conclude electricity supply contracts, at the price stipulated by Law no. 220/2008 [68,69,70]. According to Figure 3, electricity from renewable energy sources (E-RES) is being used more and more frequently in Romania’s energy industry. Romania’s extensive natural resources, including hydroelectric, solar, wind, biomass, and others, offer tremendous potential for the development and use of renewable energy sources [71].

Presently, Romania’s energy sector is undergoing steady growth, with renewable energy sources becoming more significant components of the nation’s energy mix. Romania possesses considerable potential for harnessing and incorporating renewable energy, owing to its abundant natural resources. In the domain of energy policy, Romania has adopted various measures and programs to foster the advancement of renewable energy. Notably, the Green Certificates scheme and other support mechanisms have been used to attract investors and stimulate the progress of renewable energy initiatives.

However, the energy sector in Romania also faces challenges, such as the modernization of energy infrastructure, diversification of the energy mix, and improvement of energy efficiency. Currently, the government and relevant authorities are working to promote a sustainable transition to a low-carbon economy and encourage investments in renewable energy projects in line with the EU’s objectives to combat climate change and transition towards a green economy [72,73,74].

4.2. The Holistic Approach from the Perspective of Economic Cybernetics on the Renewable Energy Sector

Modern cybernetics focuses on studying complex adaptive systems [32,33,34]. Therefore, the sector will be further analyzed from the standpoint of such a system by defining and detailing the features of the complex adaptive system present in the renewable energy business.

Understanding the interconnections and dependencies among the many components of the renewable energy sector from a cybernetic and complex adaptive system perspective is essential. This method sees the renewable energy industry as a complex adaptive system, where various components interact and change in response to the environment. In this approach, communication and information flow are key factors. The cybernetics system efficiently gathers, processes, and coordinates data on renewable energy supply and demand, available resources, and technology. For the best planning and administration of the production and distribution of renewable energy, this information is crucial.

Another important aspect is the feedback and interaction between the system components. In a complex adaptive system, there are feedback loops that allow the system to respond to changes in the environment and adapt accordingly [32,33]. For example, the demand for renewable energy can influence investment in new production capacity, and the level of production can affect prices and energy policy. These interactions and feedback are essential to maintaining the renewable energy sector’s efficient operation and achieving sustainability goals. Cybernetics system analysis of the renewable energy business can potentially reveal new features and characteristics of the system as a whole. For example, the interconnection of renewable energy networks can create synergies and additional benefits, such as balancing supply and demand, reducing dependence on traditional energy sources and creating a more resilient system. A complex adaptive system, according to Mitleton-Kelly [43], has ten general characteristics, as mentioned before. The most significant characteristics of complex adaptive systems that are highlighted in Romania’s renewable energy sector are examined in the analysis that follows.

Connectivity: Connectivity in the context of renewable energy refers to the interconnectivity and interdependence of parts and entities. In order to facilitate the efficient flow of information, resources, and energy, a network or integrated system must be built. Connectivity improves the renewable energy system’s effectiveness, affordability, dependability, and stability, which aids in a sustainable and effective transition to a new energy source.

Interdependence: The strong ties and mutual interactions between various renewable energy sources and components define the field of renewable energy. This means that the operation and performance of one element or system can directly or indirectly influence others within the global energy system. For instance, an integrated renewable energy network can demonstrate interdependence, with solar and wind energy complementing each other based on climate variations and resource availability. Interdependence can also exist between renewable energy sources and storage equipment, as energy storage relies on the production and availability of solar or wind energy. Additionally, interdependence extends to policy and regulation. Decisions made in one country regarding renewable energy promotion can impact other countries’ energy markets and infrastructure through energy interconnection and exchange.

Coevolution: The development of various components and actors engaged in the generation, distribution, and consumption of renewable energy is referred to as this property in the sector. This idea highlights the dynamic interactions and interdependencies among technology, regulations, markets, society, and the environment that have shaped the development of the renewable energy system. Technological developments that are driven by market demand and, in turn, are influenced by technology are clear examples of coevolution.

Far-from-equilibrium: Describes the dynamic state of the renewable energy industry, where interactions between components are constantly being changed and altered rather than existing in a stable equilibrium. This system experiences continuous influences and disturbances, such as weather-related fluctuations in solar or wind energy production, changes in energy demand, technological innovations, and shifts in energy policies. Operating far from equilibrium allows the renewable energy system to achieve increased performance and efficiency, quickly adapting to changing conditions and optimizing the use of renewable sources in a sustainable and cost-effective manner. This characteristic, a key aspect of complex adaptive systems, enables efficient and flexible operation through multiple interactions and feedback loops between its components.

Feedback: Feedback mechanisms are interactions that affect how a system performs and evolves in the field of renewable energy. Feedback mechanisms have the role of stabilizing or accelerating an effect, such as increasing renewable energy sources, injecting financial incentives into the sector, and analyzing performance data to identify and correct discrepancies.

Emergence: In the renewable energy sector, emergence refers to the appearance of new and unexpected properties resulting from complex interactions. These emerging properties can lead to significant changes or innovations in renewable energy production, distribution, and use. For instance, the emergence of a decentralized prosumer network, where users generate and share renewable energy through distributed generation technologies, can create a flexible and sustainable energy infrastructure. This sector’s emergence can also foster technological innovations, such as energy storage, promoting efficient renewable energy consumption, and reducing dependence on traditional sources.

This study may help in locating and defining newly emerging properties in the field of renewable energy. This refers to the emergence of unexpected characteristics or behaviors resulting from complex interactions and interdependencies within the system. Finding these emergent properties can open up possibilities for innovation and optimization in the design and implementation of policies for renewable energy.

This research’s conclusions can act as a sound basis for decision making when creating and putting into practice renewable energy policy. Highlighting complex interactions and identifying new emergent properties can guide informed decision making, supporting sustainable and efficient development initiatives within the sector. The application of economic cybernetics in analyzing the renewable energy sector in Romania can lead to a better understanding of the complexity and interdependencies within the system. This can shed light on challenges and provide robust foundations for developing sustainable and effective policies and initiatives in the field.

The so-called father of managerial cybernetics, Stafford Beer, developed in 1960 the viable system model (VSM) [45,46,75], which has since been further expanded to define a comprehensive view on the systems in Romania’s renewable energy sector. This model is a complex cybernetic approach that can be used to analyze and manage complex systems, like the Romanian renewable energy industry. It enables an understanding of the interactions and interdependencies between the system’s constituent parts and the external environment.

The following systems within the renewable energy sector have been identified in order to develop this model:

• System of Renewable Energy Production: This covers clean energy sources like biomass, windmills, solar panels, etc. Its function is to produce clean, renewable energy to run the electrical grid.
• Renewable Energy Distribution and Transport System: This system deals with transporting the produced renewable energy to consumption points. It includes the infrastructure of the distribution and transmission grid, including smart grids for energy flow management.
• Renewable Energy Consumption and Utilization System: Represents the end-users who utilize renewable energy for various purposes, such as homes, businesses, institutions, etc. Includes energy-efficient technologies and demand management to optimize energy utilization.
• Renewable Energy Storage System: Involves technologies to store surplus renewable energy for use during periods of low production. Storage technologies may include batteries, hydrogen systems, thermal storage, etc.
• Renewable Energy Policies and Regulations System: Represents the governing, financial, and policy frameworks that regulate the production and consumption of renewable energy. includes renewable energy promotion-related policies such as feed-in tariffs, mandatory renewable energy quotas, subsidies, and other incentives.
• Research and Development System for Renewable Energy: Focuses on the development of innovative renewable energy methods and technology. Involves academic institutions, research centers, and companies contributing to advancing the field.

The relationships and interactions between multiple systems can be examined using the Stafford Beer-developed viable system model, providing a thorough understanding of how the Romanian renewable energy industry functions and grows. The effectiveness, viability, and development potential of the nation’s entire renewable energy system may be assessed using this model.

According to Figure 4, the following advantages can be attained by using the VSM to analyze the Romanian renewable energy sector:

• Comprehending complex relationships: We can recognize and understand the intricate relationships between the various components and stakeholders involved in the generation, distribution, and use of renewable energy thanks to the viable system model. This detailed understanding helps us identify opportunities for optimization and improvement within the system.
• Adaptability and resilience: By using VSM principles, the renewable energy sector in Romania is made more flexible and resistant to environmental changes, such as variations in renewable energy production and changes in energy policies. For the continued stability and effectiveness of the system, this is essential.
• Coordination and synergy: Within the renewable energy industry, VSM can improve coordination and synergy between diverse parts and systems, resulting in more effective operations and the best use of resources.
• Continuous monitoring and learning: The viable system model incorporates a system of continuous monitoring and learning, enabling us to assess system performance and adapt strategies based on obtained results. This aspect is crucial for addressing challenges and consistently identifying opportunities for improvement.
• Integration of policies and strategies: VSM facilitates the integration of policies and strategies across the entire renewable energy system. This contributes to the coherence and efficiency of energy policies and supports the achievement of sustainable development objectives.

Figure 4 describes the five systems developed by Stafford Beer. System 3* highlights a specific level of complexity and adaptability that underscores the importance of this level of adaptability in maintaining the viability and efficient functioning of the system in the face of changes and challenges in the environment.

To achieve an objective diagnostic framework and capture all factors that can influence the renewable energy sector in Romania, a further analysis of the following factors was conducted: political, economic, social, technological, environmental, legal, and ethical. This analysis method, as described in the previous section, is known as PESTELE analysis [35]. The ethical factor is a newly introduced element that complements this analysis, considering the present socioeconomic context. We can identify the PESTELE elements that have an impact on the growth, implementation, and regulation of the renewable energy industry in Romania, and their effective management can hasten the switch to a more efficient, sustainable, and clean energy source. In order to recognize and comprehend the crucial elements that may affect the growth and performance of this industry, a PESTELE study must be conducted in Romania’s renewable energy sector. A complete picture of the environment in which the renewable energy business operates can be acquired by analyzing the political, economic, social, technological, ecological, legal, and ethical issues. This offers crucial data for the creation of relevant strategies and action plans to support the sector’s sustainable development. Making strategic decisions and creating policies in the Romanian renewable energy sector are possible with the help of a PESTELE study. This analysis enables the anticipation and management of risks and opportunities, as well as the identification of gaps and challenges facing the sector.

Deep understanding of the context and external influences can contribute to creating a favorable environment and promoting the sustainable and efficient development of renewable energy in Romania. Policy analysis can reveal government guidelines and energy policies, as well as relevant regulations and legislation. Economic factors can highlight investment and financing opportunities as well as the costs of renewable technologies. Social aspects can provide information on public awareness and acceptance, the impact on communities, and job creation. Technological factors can highlight technological innovations and advances in the field, as well as possibilities for integration into existing energy systems. The ecological analysis focuses on the impact on the environment and climate change, highlighting the advantages and benefits of using renewable energy. Legal aspects may include public procurement regulations and compliance with European rules. Finally, ethical factors are relevant to ensuring social responsibility and sustainability within the renewable energy sector. PESTELE analysis of the renewable energy sector in Romania has been conducted in

Table 1. PESTELE analysis of the renewable energy sector in Romania. Conceptualization of the authors.

Political	Economic	Social	Technological	Environmental	Legal	Ethic
Energy policy: Through the goals and plans set at the national and European levels, the Romanian government encourages the growth of renewable energy.	Investments and financing: Through European grants, government programs, and other sources of funding, there are finance and investment opportunities in the renewable energy sector.	Public awareness and acceptance: Society is becoming more aware of the effects of climate change and the necessity to convert to renewable energy sources.	Technological advances: Technology advancements like more efficient solar panels and bigger wind turbines boost the efficiency and economics of renewable energy sources.	Environmental protection and climate change: Using renewable energy lessens the impact on the environment and greenhouse gas emissions.	Public procurement regulations: Public procurement can greatly support the promotion of the use of renewable energy through specialized rules and selection criteria.	Social responsibility and sustainability: In line with social responsibility and sustainable development, the renewable energy industry works to lessen negative environmental effects while enhancing quality of life.
Regulations and legislation: There is a specific regulatory framework, which includes support mechanisms and green credentialing programs, for the promotion of renewable energy.	Costs of renewable technologies: Renewable energy is more cost-effective than conventional energy sources because of advancements in solar and wind technology.	Jobs and regional development: The renewable energy industry may help with both the economic growth of the regions where projects are carried out and the employment creation process.	Intelligent energy management systems: The most efficient use of renewable energy and its integration into existing electrical networks are made possible by modern monitoring and control technology.	Available renewable resources: Keeping up with the analysis of Romania’s potential for using renewable energy sources as solar, wind, hydroelectric, biomass, etc. examining the evolution and capabilities of these sources.	Compliance with European regulations: Regarding renewable energy and greenhouse gas emissions, Romania must adhere to European goals and rules.

Authors’ design.

.

The PESTELE analysis outlines the macroeconomic landscape in order to understand the context in which the industry is and the direction it is heading, the growth potential and possible risks to which the industry is exposed.

Being a strategic industry, the market is highly regulated, and any change has a strong impact on the business. At the same time, the industry is stable, as well as profitable in times of uncertainty, and technological advances continue to improve the efficiency and affordability of solar farm equipment. The high capital requirement and the challenging authorizations and licenses to obtain, however, create significant entry hurdles in this sector.

From a strategic analysis perspective, this can be illustrated by applying Porter’s five forces model, which analyzes the bargaining power of buyers and suppliers, the threat of new entrants to the market, the threat of substitute products or services, and the rivalry among existing competitors [35,49].

Thus, we have developed the competitive landscape analysis by addressing Porter’s five forces. As for the competitors, they are energy groups, renewable specialists, financial investors, or established entrepreneurs in the energy market, which are relatively constant and without much segmentation.

New players face high barriers to entry, which is why synergies, vertical and horizontal integration capabilities, and achieving economies of scale are crucial considerations for consolidation or entry decisions.

The threat of substitute products may arise during periods of cyclical production, such as during the day or year. For example, in summer periods when photovoltaic energy production is significantly increased, the availability of production capacities in certain areas may exceed the existing demand, resulting in energy losses for parks without storage capacities.

The buyer’s bargaining power is relatively low since energy is a utility constantly needed in day-to-day activities. However, price elasticity is high, depending on the customer’s level of knowledge and its size. In the context of renewable energy, end consumers such as companies and institutions may have specific requirements and preferences regarding energy sources and tariffs for renewable energy.

Additionally, suppliers have increased bargaining power due to the rapid pace of investments in this sector. The desire for energy independence of net energy-importing countries translates into increased demands for photovoltaic pack equipment. In the renewable energy sector in Romania, supplier bargaining power may be influenced by the availability and price of raw materials, such as solar panels or wind turbines.

5. Analysis of the Renewable Energy Sector in Romania Using the Econometric ARDL Model

In order to complement the previous analyses and achieve an integrated perspective on how macroeconomic factors and policies influence the renewable energy sector in Romania, we utilized the econometric ARDL model. This model offers the capability to investigate both short-run and long-run associations among economic variables, providing insights into the dynamics within the renewable energy sector.

The pressing challenges of climate change and global warning pose significant threats to the global economy, largely attributed to the rapid growth of economic activities leading to increased GHG emissions.

The study intends to explore the long-run and the short-run causality between GHG FDI, GDP, and RE for Romania during 2000–2021. By choosing this time span, 2000–2021, we intended to capture the recent transformations indicated by the variables during this interval. This research holds particular relevance to Romania’s ongoing sustainable reform efforts, aligning with the National Recovery and Resilience Plan (NRRP) established by the European Commission. The NRRP aims to stimulate economic growth while promoting a transition towards green and sustainable practices.

Table 2. The definitions of variables.

Variables	Abbreviation Symbol of the Variable	Unit of Measurement	Source
Real GDP per capita	GDP	Chained linked volumes (2000), euro per capita	Eurostat
Greenhouse gas emissions intensity of energy consumption	GHG	Index: 2000=100	Eurostat
Renewable energy	RE	(%) of primary energy supply	OECD
Foreign direct investments	FDI	Net inflows (%) of GDP	World Bank

Authors’ collection.

contains the definition of variables and their sources, Eurostat, World Bank, and OECD. GDP depends on GHG, FDI, and RE, as shown by Equation (4).

GHG is measured as the ratio between energy GHG emissions and the national energy consumption. Renewable sources of energy consist in hydro, geothermal, wind, solar, and biofuels.

The selection of variables used in the ARDL econometric model was based on both theoretical logic and the relevance and availability of data at the Romanian level. Real GDP per capita was chosen as it reflects the country’s level of economic development and can influence energy demand and the adoption of renewable energy sources. Considering the significant emphasis on sustainability and ecological impact by most international institutions, greenhouse gas emissions intensity of energy consumption was chosen to reflect the degree of sustainability and ecological impact of the energy sector. Furthermore, the renewable energy variable, being central to the objective of sustainable development, was selected because it can provide information about the progress and impact of the renewable energy sector. Last but not least, the foreign direct investments variable was chosen as it reflects the level of interest and involvement of foreign investments in the development and implementation of renewable energy technologies and infrastructure in Romania.

Table 3. Statistical summary of variables.

Summary Statistics	GHG	GDP	RE	FDI
Mean	95.055	6922.5	5373.713	3.681934
Median	96.55	6835	5447.178	2.883482
Maximum	101.5	9610	6192.833	9.020062
Minimum	83.8	4350	3749.226	1.230493
Std. Dev.	5.443536	1557.741	732.7453	2.329391
Skewness	−0.705601	0.098094	−0.726374	1.148407
Kurtosis	2.230542	2.156415	2.515779	3.104521
Jarque–Bera	2.152962	0.625105	1.954121	4.405231
Probability	0.340793	0.731577	0.376416	0.110514

Authors’ calculation.

contains the descriptive statistics of the original data. GHG, GDP, and RE have platykurtic distributions, while FDI has a mesokurtic distribution. The positively skewed variables GDP and FDI have a mean greater than their median, while the negatively skewed variables GHG and RE have a mean less than their median. The probability of the Jarque–Bera test shows that all variables have a normal distribution.

Based on the dependence between GHG emissions, economic growth, and renewable energy, we will test the following hypotheses:

H₁. 

GHG is positively linked with GDP.

H₂. 

RE is positively linked with GDP.

These two hypotheses will be tested by means of the ARDL model. The most flexible econometric technique proved to be ARDL, since the energy–growth relation is characterized by shifts and shocks [57], when data exhibit nonstationarity and cointegration specific to economic and environmental variables.

We will follow the same sequence of steps described in the methodological Section 3. We work with logarithmized absolute-value time series presented in Table 2. The ARDL dependence of the variables is given by Equation (4):

$$GD{P}_t=f\left(GH{G}_t,R{E}_t,FD{I}_t\right)+{\epsilon }_t$$

(4)

where ${\epsilon }_t$ represents the error term.

i.

Stationarity

Data stationarity is checked by ADF unit root test (Dickey and Fuller, 1979) [57].

In

Table 4. ADF unit root test.

Variables	Level	First Difference	Order of Integration
	T-Statistics	T-Statistics
GDP	−1.30 (0.60)	−3.49 (0.02)	I(1)
GHG	0.26 (0.97)	−5.47 (0.00)	I(1)
RE	−1.29 (0.61)	−6.17 (0.02)	I(1)
FDI	−1.28 (0.29)	−4.61 (0.00)	I(1)

Authors’ calculation.

, one can see that all variables are integrated of order 1. All the p-values are less than 0.05. Thus, the time series in first difference are stationary, fulfilling the application of the ARDL model.

The bounds cointegration test is performed next to see whether there is a long-run causality. The selected model is ARDL (2,2,1,0).

ii.

Cointegration

In

Table 5. Results of cointegration bounds test.

Statistic Test	Value	K (Number of Regressors)
F-Statistic	12.49	3
Critical Value Bounds (Finite Sample N = 35)
Significance	I(0)	I(1)
10%	2.37	3.2
5%	2.71	3.67
2.5%	3.15	4.08
1%	3.65	4.66

Authors’ calculation.

are reported the cointegration bounds test results.

F-calculated = 12.49 is above the critical upper bound I(1); therefore, we have long-run causality from RE, FDI GHG to GDP.

Table 6. The long-run estimated coefficients (authors computation).

Variables	Coefficient	T-Statistics	Prob.
GHG	−2.14	3.2	0.005
RE	1.53	3.67	0.008
FDI	0.15	4.08	0.227
C	5.41	4.66	0.379

Authors’ calculation.

reports the long-run estimated coefficients.

The relationship between GHG and GDP is complex and can vary with respect to several factors. In general, this relation is positive, because economic activity, such as industrial production, transportation, and energy generation depend on fossil fuels combustion, which results in GHG emissions. This positive correlation is not always true. More efficient technologies and policies are implemented to mitigate climate changes. These actions can lead to decoupling economic growth from GHG, making the relation negative.

In our study, we find that in the long run, hypothesis $H_1$ is not validated. A 1% increase in GHG leads to a 2.14% decline in GDP. GHG including mainly CO₂ emissions is generated from fossil fuel combustion and cement manufacturing. The negative association is validated by the studies of Martinez-Zarzo and Bengochea-Morancho [76] and Azam et al. [77] for low-income countries. This can be explained by the progress Romania has made in decoupling GHG emissions from economic growth.

In the long run, RE positively influences GDP, confirming hypothesis H₂. A one-unit increase in RE causes a 0.95% rise in GDP. This means that RE contributes to economic growth in the long run, a relation also obtained in [78].

On the other hand, FDI does not have a statistically significant impact on GDP. Several factors could contribute to this insignificant relationship. If a considerable portion of FDI is directed towards nonproductive sectors, its effect on GDP growth may not be substantial. Additionally, a country’s ability to absorb FDI inflows is crucial; lacking the necessary infrastructure, skilled labor, or appropriate regulations can lead to an insignificant correlation, as noted by Simionescu et al. [79].

Furthermore, the unfavorable macroeconomic conditions caused by the COVID-19 pandemic, such as political instability, fluctuations in exchange rates, or inflation rates, may introduce uncertainties in the relationship between FDI and GDP.

To analyze the dynamics and short-run relationships between variables, an ARDL short-run model was constructed. In the context of energy and economic analysis, the ARDL model for the short term is often used to study the short-term effects of changes in independent variables (such as energy consumption, renewable energy utilization, or economic indicators) on a dependent variable (such as GDP or energy prices). The short-run ARDL coefficients are captured in

Table 7. ARDL short-run model.

Variables	Coefficient	T-Statistics	Prob.
D(GDP(-1))	−0.29	−2.49	0.03
D(GHG)	1.09	5.03	0.00
D(GHG(-1))	1.05	3.97	0.00
D(RE)	0.33	4.13	0.00
CointEq(-1)	−0.32	−9.50
R-squared	0.85
Adjusted R-squared	0.80

Authors’ calculation.

.

ECT equals −0.32, which is significant and belonging to the interval [−1, 0], providing evidence of long-run cointegration. The speed of adjustment towards the long-run equilibrium is 32%. The coefficient of determination of 85% indicates a good fit of the short-run model. At the same time, hypothesis $H_1$ is confirmed in the short term, by the positive coefficients of the first difference and the lag of the first difference of GHG. Hypothesis $H_2$ is also confirmed in the short term. Some factors that contribute to this short-term positive relation are industrial production, energy consumption, transportation and agricultural practices. From Table 7, one remarks a short-term positive relation between RE and GDP. This relation can be observed through various channels: investment in infrastructure, job creation, energy cost savings, and technological innovation. Investments in infrastructure such as solar panels and energy turbines produce economic activity.

iii.

Diagnosis and Stability Tests

The null hypotheses $H_0$ of the Breusch–Godfrey serial correlation LM test, the Breusch–Pagan–Godfrey heteroscedasticity test, and the Jarque–Bera normality test are shown in Table 6. Their p-values are greater than 0.05, which means accepting the null hypotheses. The model does not have autocorrelation and heteroscedasticity. The residuals are normally distributed. The Ramsey reset test (

Table 8. Diagnostic and stability tests (authors computation).

Test	${\mathit{H}}_0$	Statistics	p-Value	Decision
SC*	There is no serial correlation in the residual	0.75	0.50	$\mathrm{Accept}{H}_0$
HE**	There is no autoregressive conditional heteroscedasticity	0.78	0.62	$\mathrm{Accept}{H}_0$
NO**	Normal distribution	Jarque–Bera 1.20	0.54	$\mathrm{Accept}{H}_0$
RR**	Absence of model misspecification	0.40	0.69	$\mathrm{Accept}{H}_0$

SC*—serial correlation, HE**—heteroscedasticity, NO**—normal distribution, RR**—Ramsey reset. Authors’ calculation.

) shows that the model is correctly specified.

To detect structural changes or variations in the parameters of a time series model, CUSUM and CUSUM of squares tests were conducted. These tests are considered powerful tools in the literature for identifying structural changes in time series data and model stability. They can be applied in various fields, including economics, finance, and engineering, to detect changes in underlying patterns, trends, or relationships in the data. The results are presented in Figure 5 below.

Both tests indicate the moments when the variation in time series data or model parameters has undergone significant changes. Considering that no large fluctuations or abrupt increases have been observed, the model can be considered stable, and there are no significant changes in the model parameters. The small fluctuations can be explained by economic, political, and social events that have influenced Romania’s macroeconomic indicators, such as the COVID-19 pandemic, the inflation surge, and the Russia–Ukraine conflict. Thus, the statistical tests CUSUM and CUSUMSQ fall within the confidence interval marked by the blue lines, so we can state that the model is stable.

iv.

Variance decomposition and impulse response function

The ARDL technique was applied to show the impact of RE, FDI, and GHG on GDP, and at the same time, the cointegration inside the sample, not outside it [80]. We used VAR for variance decomposition and impulse response function. We applied variance decomposition for 20 periods, the approximate length of our time series (

Table A1. Variance decomposition (author’s computation).

Variables	Period	D(GDP)	D(GHG)	D(RE)	D(FDI)
D(GDP)	1	100.00	0.00	0.00	0.00
	5	64.19	21.27	3.88	10.64
	10	61.44	21.80	6.31	10.43
	15	60.51	21.69	7.34	10.44
	20	60.19	21.64	7.70	10.45
D(GHG)	1	61.50	38.49	0.00	0.00
	5	64.04	26.04	4.94	5.86
	10	59.88	25.10	8.56	6.44
	15	57.77	24.74	0.90	6.58
	20	57.04	24.57	11.72	6.66
D(RE)	1	0.10	0.00	99.99	0.00
	5	34.27	4.78	51.29	9.64
	10	35.00	10.89	44.54	9.55
	15	33.98	11.47	45.14	9.39
	20	33.44	11.56	45.58	9.40
D(FDI)	1	12.17	36.64	0.04	51.14
	5	26.24	34.31	4.06	35.38
	10	22.62	35.45	13.65	28.26
	15	21.24	33.91	8.30	26.54
	20	20.70	33.23	20.00	26.05

from Appendix A). Variance decomposition measures the impact of a shock of one variable on the forecast error of another variable. In the first year, GDP is explained only by its own shocks (100%). GHG, RE, and FDI do not influence GDP in the short run. From periods 5 to 20, the forecast variance error of GDP explained by GDP itself decreases from 64% to 60%, so GDP exhibits a strong influence on itself from the short run to the long run. The influence of GHG increases over the 20 periods, while the influence of RE increases to 7.70%, which is still weak. The contribution of FDI to GDP increases up to 10.45% in 20 periods. When GHG is the dependent variable, the contribution of economic growth to GHG increases up to 57%, while the contribution of RE to GHG increases up to 11.7%. FDI contributes to GHG up to 6.66%. The RE forecast variance error decreases, being explained by itself around 45.48%, by GDP around 33.44% and by GHG around 11.56%, and by FDI around 9.40%. The most important contribution belongs to GHG (32.23%), followed equally by GDP and RE (20%) when FDI is the dependent variable.

Next, we study the reaction of a variable to another variable by means of the Cholesky impulse response function in Figure 6. The impulse response function measures the reaction of a variable as a result of a change in the standard deviation of another variable.

The response of GDP to GDP shocks reveals that GDP decreases abruptly in period 2, then becomes negative, having slight fluctuations around zero in the next periods. The response of one standard deviation innovation of GDP to GHG is in decline below zero in period 2, increasing above 0 in period 8, oscillating slowly, and finally disappearing in the time horizon. The response of GDP to RE is oscillating around 0. The response of GDP to FDI is above 0 during periods 2–4, then it tends to 0.

From Figure 6, one can see that IRFs have similar patterns, with some fluctuations at the beginning, followed by smaller fluctuations towards 0. The exception is made by the response of FDI to RE which presents several fluctuations. Some potential reasons for these fluctuations are the following. Regulations on renewable energy can cause FDI fluctuations when there are changes in subsidies or taxes from renewable energy projects. The investors’ interest is influenced by energy prices variations, market demand, inflation rate, or exchange rates. FDI in RE can vary due to regional factors: resource availability, political stability, and infrastructure.

These moves of IRFs are important in studying the dynamics of the system. A significant part of the variables dynamics is explained by their own shocks, then it becomes smaller in latter periods.

6. Limitations and Future Research

Our study has several limitations. Firstly, due to the relatively short time span of only 21 years, we could only include a limited number of variables in the analysis. Additionally, some variables were excluded from the model either due to multicollinearity issues or their statistical insignificance. Another limitation pertains to the consideration of CO₂ emissions from various economic sectors. Studying these emissions separately, rather than at an aggregated level, could provide more detailed insights.

Although we have identified six systems that impact the renewable energy sector through cybernetics analysis and diagnosis, the ARDL analysis encompasses only a restricted set of variables, thus failing to capture the model’s complexity. This integral aspect of this research remains one important limitation.

For future research, we recommend diversifying the econometric tools to model the nexus between economic growth and renewable energy consumption for Romania. For example, incorporating circular economy variables like the recycling rate of municipal waste and other pollutants could offer a more comprehensive understanding of the relationship.

Expanding the analysis to encompass a broader range of economic, social, and technological factors would also contribute to a more comprehensive assessment of the dynamics of the renewable energy sector.

7. Conclusions and Policy Recommendations

In regard to the comprehensive approach of the discussed subject in the study and the integration of economic cybernetics and the ARDL econometric model, we were able to grasp and assess how complex challenges related to this industry can be controlled by adopting an economic cybernetics approach to the analysis of Romania’s renewable energy sector. In this context, the Romanian energy market has been explored from the perspective of an adaptive complex system, allowing us to identify and highlight its characteristics. As a result, we conclude that this industry is a complex adaptive system, in which the numerous parts and pieces interact dynamically and change in response to the environment. Furthermore, the cybernetic viewpoint has enabled us to comprehend the key elements of this sector as well as how it can be considered as a comprehensive system. To understand its subcomponents, the viable system model (VSM) was applied as a managerial diagnosis framework, focusing on the renewable energy sector in Romania as a complex system. The analysis conclusions highlighted the advantages of this sector, which can be invested in to ensure sustainable development. These include adaptability and resilience, synergy, policy, and strategy integration, as well as coordination and learning. The VSM analysis was complemented by a holistic diagnosis of this sector to observe the influences of political, legal, environmental, social, technological, ethical, and economic regulations. The PESTELE analysis and Porter’s five forces were applied. The first analysis emphasized the ethical factor, which should be considered in this analysis and represents a novel element in the literature.

To evaluate the relationships between the supply of renewable energy and relevant economic, social, and technological aspects in the context of renewable energy, ARDL econometrics is a useful tool. In this work, the ARDL model was applied to analyze the relationship between Romania’s economic growth, RE, GHG, and FDI between 2000 and 2021. On the relationships between GHG, RE, and GDP, two hypotheses were examined. The results of the research proved that in the long term, $H_1$ is contradicted and $H_2$ is confirmed, while in the short run, both hypotheses are confirmed. The long-term negative correlation between GHG and economic growth proves that progress have been made in decoupling GHG emissions from economic growth in Romania. Continuous efforts are needed to sustain this decoupling trend and achieve emissions reductions. This asks for ongoing investment in clean energy infrastructure, and the adoption of innovative technologies. Regarding other directions and mechanisms that can stimulate the development of renewable energy both in Romania and in other countries, subsidies and financial incentives offered by the government, such as tax credits or deductions for investments in renewable technology, can be emphasized. Setting mandatory targets for the share of renewable energy in total production and consumption of energy, which can create increased demand for renewable energy sources, is another important direction. Another significant direction can be the promotion of education and public awareness regarding the importance and benefits of renewable energy, as well as the collaboration between the public and private sectors, which can create a platform for the development and implementation of renewable energy projects.

The Romanian government has implemented various support mechanisms to stimulate renewable energy investments, such as green certificate investments, feed-in tariffs, and competitive auctions. These measures intend to attract investments, including renewable energy capacity. Fossil fuel taxes are a common policy approach to encourage the usage of renewable energy sources. The relative cost competitiveness of renewable energy sources can be improved by raising the price of fossil fuels.

The novelty of this article and its significance in advancing the specialized literature lie in the holistic analysis from a cybernetic perspective and its integration with econometric models, such as the ARDL model.

Our study has some limitations. In the analysis, only a few variables are included. This arises on one hand from the fact that the time series span is too short, only 21 years, and on the other hand, from multicollinearity or from the statistical insignificance of other variables included in the model. Pesaran et al. (2001) [54] estimated F-test statistics for ARDL models for n > 30, even if some researchers may accept n > 20. Models with smaller n are less likely to adequately account for lagged variables, given the loss of degrees of freedom.

CO₂ emissions generated by different economic sectors could be studied separately, not at the aggregated level. Future research may consist in applying fiscal and monetary instruments to derive more details about the economic growth–renewable energy consumption nexus in Romania. Other variables such as recycling rate of municipal waste and other pollutants may be included in the model.