Recyclable Consumption and Its Implications for Sustainable Development in the EU

Abstract

The transition to a circular economy is imperative in order to confer considerable benefits upon the environment, the economy, and society. The present study aimed to analyse the interdependence and causal relationships between recyclable material consumption as the dependent variable and other independent variables, including the raw material footprint, the trade in recyclable materials, greenhouse gas emissions, investments in the circular economy sectors, the real GDP per capita, renewable energy sources, the circular material use rate, and the population within the 27 EU Member States from 2013 to 2021. In order to achieve the objective, a two-stage economic model was constructed using a panel approach. The research findings indicate a direct and positive correlation between the consumption of recyclable materials and all the aforementioned independent variables, with the exception of greenhouse gas emissions. This study confirms that innovation and investment significantly reduce environmental degradation, and, moreover, the efficiency of investment remains unaffected. A further relationship that emerged from this study is that developed countries have higher resource consumption, which is consistent with the cause of increased consumption being the rapid growth of the middle class around the world. The main conclusion is that Europe cannot achieve sustainable development without a circular economy.

1. Introduction

As the global population continues to expand and all future growth is anticipated to occur in urban areas [1], it is imperative that governments take immediate action to ensure the sustainable development of urban areas, with a particular focus on smart cities [2,3].

Sustainable development involves meeting the needs of present generations without compromising the resources available to future generations. Therefore, it is believed that the circular economy is a suitable approach to achieve this objective. The European Commission has set a target of achieving climate neutrality by 2050. To achieve this goal, the commission believes that transitioning to circular systems in production and consumption is essential. In line with this objective, on 11 March 2020, the European Commission launched a new Circular Economy Action Plan titled “For a cleaner and more competitive Europe” [3]. Today, the circular economy is a relevant topic and is increasingly the focus of researchers, academics, and scientists.

The term “circular economy” is defined as a regenerative economic system that is based on the reduction, reuse, remanufacturing, and recycling of existing materials and products as much as possible in order to minimise waste and the environmental impact. Under a circular economy model, resources are kept in use for as long as possible, the value of products and materials is retained in the economy, and waste generation is minimised.

More specifically, the circular economy is understood as a critical priority in terms of our relationship with the environment and natural resources and represents a solution to the constraints of the take–make-use–dispose linear economy that reconciles ambitions for economic growth and environmental protection [4,5,6]. In this respect, the European Union is becoming a central pillar that desires more and more protection against climate change, an increase in the use of renewable energy sources in the European energy mix, and a reduction in carbon dioxide emissions. All these initiatives can be achieved through the transition to a circular economy, wherein strategic plans for the use of recyclable materials, increasing recycling rates, and reducing waste are promoted [7].

In the intricate domain of economic science, the notion of the circular economy is gaining traction as a compelling proposition. This concept advocates for a multidisciplinary approach to consumption and production processes, emphasising the utilisation of renewable energy sources and the promotion of environmental sustainability. The circular economy model emphasises the integration of an economic system that is designed to maximise the efficient and responsible use of resources.

The primary objective of this study was to investigate and examine the effects of short- and long-term interdependencies on the consumption of recyclable materials, which is also considered a relevant and important indicator in terms of the circular economy [4,6]. To this end, a two-stage economic model was constructed to demonstrate how the consumption of recyclable materials is influenced and affected by the dynamics of the following selected explanatory variables: the raw material footprint, the trade in recyclable materials, greenhouse gas emissions, investments in circular economy sectors, the real GDP per capita, renewable energy sources, the circular material use rate, and the population. The model was applied to the 27 EU Member States, with the period analysed being 2013–2021, to better understand the correlations between these variables and to formulate recommendations for economic policies to achieve sustainable development.

The distinguishing characteristic of this research is its multivariate treatment of issues posed by the new facet of the circular economy. It emphasises the demonstration, investigation, and evaluation of the effects of short- and long-term interdependent relationships on the dependent variable, namely, the consumption of recyclable materials. Consequently, the empirical findings from the developed model underscore the role of this variable and its place in the circularity levels within the European Union.

From a methodological perspective, the discussion focused on methodologies and techniques that are specific to the analysis of panel data. These methodologies and techniques included the following: the common-effects model, the fixed-effects model, the random-effects model, the cointegration regression model, and the autoregressive model, which was implemented in the VAR/VEC form.

In accordance with the applied model, seven research hypotheses were proposed to test the relationships between the consumption of recyclable materials and the other exogenous variables. This study is a relatively recent addition to this field of research, with the objective of shedding light on the way that the consumption of recyclable materials is perceived by the general population and policymakers in the European Union (EU). The study aimed to ascertain the extent to which this perception can influence their behaviour and decision-making processes, adopting a holistic approach.

The consumption of recyclable materials is the dependent and central variable in the present analysis, and for this reason, it is considered a relevant measure of the circular economy, with the capacity for exponential growth and development in the coming years [8,9]. It must also be borne in mind that the EU’s priorities are focused on climate change and the environment, where the elements of the circular economy are coming to the fore as one of the central pillars of sustainability.

For these reasons, the model under consideration also incorporates additional variables that are specific to the circular economy, including the material footprint, the rate of use of recyclable materials, and the dynamics and evolution of the trade in recyclable materials. Similar to the line of research developed in [10,11], the GDP per capita, renewable energy sources, population, and greenhouse gas emissions per capita were chosen as the independent variables to determine the short-term and long-term economic, environmental, and social impacts of the consumption of recyclable materials in European countries between 2013 and 2021. Different from some of the studies analysed, the model developed uses the variable consumption of recyclable materials as the dependent variable. This is employed to illustrate how it quantifies the part of circularity that is proposed by the multidisciplinary approach of the circular economy in European countries.

According to statistical data provided in [12,13], the European Union is making considerable efforts to increase the competitiveness and development of sectors specific to the circular economy, wherein the attraction of investments and investment projects is a determining factor for the employment and upgrading rates.

Moreover, it is anticipated that the present study will provide valuable insights on how the consumption of recyclable materials constitutes a transmission channel of the circular economy by revealing those causal relationships and interactions that propagate differently across European countries [6,8,10,14]. The empirical evidence obtained from the application of the model provides and generates a starting point for new interests and concerns in the study of the comprehensive phenomenon of the circular economy. It may also be suitable and appropriate in the formulation of new future directions of environmental policy and the promotion of climate neutrality by European institutions.

The remainder of the paper is structured as follows: Section 2 discusses the main findings from the literature review; Section 3 describes the model, data, variables, econometric methods, methodological approach, and research design; Section 4 presents the results and findings according to the analysis; Section 5 highlights the implications and discusses the results; and Section 6 outlines the potential directions for future research and concludes the research paper.

2. Literature Review

The circular economy is a highly relevant topic at the global level, generating several interactions and interdependencies from an economic point of view. The scientific literature to understand, define, and raise awareness of this complex phenomenon is now expanding.

Therefore, there has been a great deal of research and discussion on the benefits of circularity for society, the economy, and the environment, as well as on consumer awareness, and on the development of innovative circular business models and government policies to support them. Similarly, this is taken note of in Ref. [15], which provides a conceptual analysis of the term circular economy. The main findings highlight a better understanding of the definition and specific objectives of the circular economy, with the authors concluding that the term represents a combination of elements related to sustainability, recycling, recyclable consumption, the material footprint, and sustainable reduction.

2.1. Investments in the Circular Economy

To achieve growth, a company must make significant investments, and for sustainable development, these investments must be in circular economy technologies. However, there is concern that such investments may not be effective. Ref. [16] argues that there is a direct link between sustainability and efficiency. Also, Ref. [17] suggests a positive relationship between sustainable investments and firm efficiency, particularly in developing countries. However, Ref. [18] has suggested a negative correlation between firms focused on their corporate sustainability performance or social responsibility (including the environment) and return-on-investment reduction. The authors of [19] conducted a bibliometric and systematic literature review investigating the relationship between the sustainable behaviours of SMEs and their financial performances. Most of the studies reviewed confirmed a positive correlation between the two.

2.2. Trade in Recyclable Materials

Some studies suggest a positive correlation between trade and sustainable development. However, the same studies indicate a negative correlation between trade and environmental sustainability [20,21,22]. Therefore, our aim was to test whether the trade in recyclable materials has a positive correlation with both sustainable economic growth and the environment using the model.

2.3. Renewable Energy

The consumption of conventional energy, particularly fossil fuels, has a negative impact on environmental efficiency [21,23]. So, the effect of renewable energy on environmental sustainability is being investigated, as demonstrated in the study [24]. Ref. [25] provides the basis for building an economic model that considers a set of relevant circular economy indicators (i.e., the recycling rate, human capital, resource productivity, and green energy use) applied at the level of the 27 countries of the European Union from 2008 to 2017. The specific results demonstrate that the environmental circular economy factors were significant indicators of economic growth in all 27 EU countries. The degree of environmental innovation and the use of renewable energy play greater roles in terms of the economic growth impact rate compared to the impact of the GDP per capita and increasing human capital in renewable energy.

2.4. Greenhouse Gas Emissions

There are studies that suggest that urbanisation is one of the main factors contributing to CO₂ emissions [26,27]. For sustainable urban development, it is important to consider the impact of greenhouse gases on the circular economy. Empirical evidence in Refs. [28,29] shows that in areas relevant to the circular economy, such as waste management, energy consumption, and material use, population is a key factor. For cities and regions, the circular economy is an opportunity to rethink production and consumption models, services, and infrastructure.

2.5. Population

The inclusion of population as an independent variable in the model proposed in this study is justified by the literature. For example, this variable is found in the study carried out by the authors of [30], who studied the increase in recycling rates at the European level and how it is influenced by behavioural, economic, institutional, and social factors. The analysis, carried out for the period 2010–2018 for 20 EU countries, highlights the benefits of implementing green solutions. Countries with higher levels of renewable energy and lower CO₂ emissions also tend to have higher rates of e-waste recycling. The same study also showed that educating the population to recycle and participating in recycling campaigns are active and behavioural factors. Another recent study [31] shows the direct relationship between the consumption of recyclables, recycling rates, and population in a panel analysis of 25 European countries over the period 2010–2019.

2.6. GDP per Capita

The studies elaborated in [32,33,34] also aimed to examine the impact of the circular economy indicators on the level of economic growth for 25 European countries over the period 2010–2018. After implementing the VEC panel model, these authors found that an important factor in the trade-off towards circular practices that also generate economic growth is the recycling rate. In addition, the same authors consider that the circular economy is still at an early and preliminary stage, but that financial support in the form of investment projects from the European Union is essential and necessary to stimulate economic sectors to reach/achieve the ambitious environmental goal of zero waste generation.

Another interesting insight into the circular economy at the level of EU countries comes from the perspective of environmental factors and indicators. In this regard, the authors of [35,36] present an extended version of the Mankiw–Romer–Weil model of economic growth for the study of the circular economy for the 27 EU countries in the period 2007–2016. The results of the model suggest that the levels of productivity, environmental employment, environmental innovation, and the recycling rate are relevant factors of the circular economy and have a statistically significant impact on economic growth. Furthermore, these results are in line with the European Union’s target of increasing resource productivity by 30% by 2030.

Based on artificial intelligence techniques (i.e., the Data Envelopment Analysis Method and Factor Analysis), it could be shown that the implementation of circular economy measures in the case of European Union countries is closely related to the level of the GDP per capita [37]. In addition, the results statistically confirmed that countries with higher GDPs per capita perform better in terms of the circular economy objectives. It was also concluded that indicators specific to the circular economy have a positive impact not only on economic growth and development, but also, more importantly, on the population, society, and the natural environment.

Ref. [38] examines the manifestation and implementation of the circular economy objectives for the Chinese state. The primary conclusions of the study indicated that the circular economy policy was selected as a fundamental component of the national sustainable development policy framework. The authors of [39] conducted an analysis of the implementation of the circular economy in EU countries in the period 2008–2016. The findings suggest that the application of the circular economy concept can ensure economic growth and GDP growth while reducing the use of natural resources and better protecting the environment.

2.7. The Material Footprint

The material footprint (MF) is also referred to as the (RMC) and has increased exponentially since the year 2000. This has led to the prioritisation of reducing the MF as a key aspect of sustainable development in megacities [40]. Urban circular economy policies should be at least partly aimed at reducing the material footprint (MF) in cities [41]. The authors of References [41,42] demonstrated that infrastructure development, urbanisation, and economic growth have a significant impact on the MF. However, green innovation has been shown to have a substantial effect on reducing the MF, and this has been identified as the most prominent factor in environmental degradation, particularly in the context of economic development and infrastructure expansion.

The material footprint constitutes a pivotal component in the analysis of the circular economy model, as evidenced by the extant research [32,33]. The incorporation of this explanatory variable into the proposed model is substantiated by the findings of these studies, which underscore the imperative to comprehend the potential interrelationships between the circular economy and the material footprint, particularly in the context of attaining a pristine and sustainable environment. In this regard, the extant literature has employed panel regression methodologies at the level of the 23 EU Member States, with the principal empirical findings indicating the necessity for the EU to foster the raw material markets, thereby leading to a substantial decrease in the material footprint.

Research has demonstrated a direct correlation between the circular economy and the material footprint. Ref. [43] demonstrated that measures specific to the circular economy have reduced the material footprint in certain cities by more than 25%. Furthermore, this study posits that the footprint exerts an influence on the nation’s consumption (and, consequently, GDP).

The findings of the studies [44,45] demonstrate that the promotion of the circular use of materials and the undertaking of circular economy activities have been shown to contribute to a reduction in the raw material footprint.

2.8. The Circular Material Use Rate

The significance of another indicator used, the circular material use rate (CMR), is evidenced by its role as a metrics tool utilised by the EU to assess the progression of the circular economy [46]. The EU’s 2020 document articulates its objective to augment its circular material use rate, with a stated ambition to achieve a twofold increase within the ensuing decade.

The circular material use rate was 11.2% in the given year (2020), and it was anticipated that this would reach 22.4% by the year 2030. However, it should be noted that in 2010, when the measurement was taken for the first time, the rate was 10.7%, and over a decade, it increased by only 0.5%. It is evident that, subsequently, the rate of increase was negligible (a rise to 11.8% was only achieved in 2023). Therefore, it may be concluded that the EU has not only failed to achieve the aforementioned target, but has also, de facto, abandoned the monitoring of this specific indicator [47,48,49].

Despite the fact that it is no longer subject to monitoring by the EU, the objective was to ascertain whether there is a positive and significant correlation between the two variables in order to assign this indicator the importance it merits and enable its monitoring once more.

The study formulated by the authors of [49] aimed to examine the relationship between the macro-level circularity rate and various macroeconomic variables across a sample of 28 European countries, employing panel data. The findings suggest a robustly positive relationship between the real GDP and the circularity rate over the long run, while higher environmental taxes are associated with an increase in the circularity rate.

The study [50] indicates that it is logical that if the materials used are no longer wasted and their recycling rate increases, the demand for newly raw materials will decrease, and that there is a positive relationship between the quantities of recycling and the circular material use rate. Furthermore, the findings of other studies [51] indicate that the macroeconomic determinants of the CMR are domestic material consumption (DMC) and recycling. The results of these studies suggest the presence of a direct link between these indicators, with the link being stronger for DMC and weaker for recycling.

Another study [52] claims that the circular material use rate is high in developed EU countries (Netherlands, Belgium, France, etc.) and suggests that there may be a link between the circular material use rate and raw material consumption. The study concludes that there is a strong relationship between the rate of circular material use and the final energy consumption in households. The study [34] found that the circular material use rate (CMUR) has a strong negative correlation with GHG emissions. They also found that processing recycled materials requires less energy than processing raw materials, resulting in lower GHG emissions.

2.9. Econometric Model

In consideration of the importance of the model used, there is growing interest in economic modelling for the estimation and prediction of indicators that revolve around the concept of the circular economy. Along this line, most studies focus on examining the impact of the circular economy at the level of European Union countries, like Ref. [14], which statistically analyses the concept of the circular economy and the main indicators of the circular economy in the 28 EU countries. The analysis was carried out over the period 2010–2018, and its main objective was to provide an aggregate index measuring the level of development in the implementation of strategic economic practices and policies at the European level. The main findings show that the European Union is increasingly concerned about the transition to a circular economy, and that most European countries have adopted strategic action packages to ensure easy access to a circular economy.

Recent studies [53,54] have established different quantitative and qualitative models to measure the impact of the circular economy, especially in European countries. The main findings were that the transition to a circular economy has a positive impact on the rate of economic growth; that the increase in the recycling rates in European countries is due to changes in the attitudes and behaviour of European citizens, who are becoming increasingly responsible and aware of environmental protection; that investment in circular sectors is a crucial factor for sustainability and for achieving environmental goals.

This is considered in the research developed by the authors of [55], who developed a model for the analysis of the circular economy. The variables used were the polluting and recyclable inputs in the case of China. The main results indicated that improving the environmental quality, as measured by the reduction in pollution, can only be achieved by increasing the rate of self-renewal or recycling. There are some studies that show that the emergence of circular economy activities is a key driver of sustainability and has been the subject of increasing attention in recent years. Similarly, in 2015, Refs. [56,57] showed that the recycling of mixed paper and mixed plastic and organic waste and the use of the recycled materials in industrial production can potentially reduce CO₂ emissions by about 69 kt and the amount of incineration ash sent to landfill by 8 kt, using a case study of Kawasaki in Japan.

The autoregressive model in modified VEC form and the FMOLS cointegration regression method were used to reveal the causal and interdependent relationships in the short and long term on the consumption of recyclable materials in the 27 European countries analysed. From this perspective, it is considered that the results obtained are suitable and consistent enough to extend the literature.

The study’s innovation (in addressing a research gap) lies in the construction of a two-stage model, which facilitates the analysis and determination of the dynamic relationships between the consumption of recyclable materials and the remaining integrated explanatory variables. The econometric tools and methods employed in this study were deemed suitable for the present context.

2.10. Hypothesis

As a result, to explore and investigate the potential impacts and effects of the consumption of recyclable materials on the economic, social, political, and environmental impacts in the European countries during the period under study, the following research hypotheses were developed to validate or invalidate them in relation to the results of the model:

Hypothesis 1 (H₁):

The consumption of recyclable materials, the raw material footprint, and the circular material use rate are expected to be positively correlated;

Hypothesis 2 (H₂):

The consumption of recyclable materials is positively influenced by the trade in recyclable materials;

Hypothesis 3 (H₃):

The consumption of recyclable materials is positively influenced by the renewable energy sources;

Hypothesis 4 (H₄):

The consumption of recyclable materials is negatively influenced by greenhouse emissions;

Hypothesis 5 (H₅):

The consumption of recyclable materials will be positively affected by investments in circular economy sectors;

Hypothesis 6 (H₆):

The consumption of recyclable materials is positively influenced by the level of GDP per capita;

Hypothesis 7 (H₇):

The consumption of recyclable materials is positively influenced by the populations in the EU countries.

3. Research Methodology

3.1. Data and Proposed Model

The model developed in the present research presents theoretical foundations found in other previous research that focuses on the introduction of sound economic models for a better awareness of the circular economy. In this context, the primary dependent variable is the consumption of recyclable materials, which is considered a proxy for a circular and sustainable economy, while the raw material footprint, trade in recyclable materials, greenhouse gas emissions, investments in circular economy sectors, real GDP per capita, renewable energy sources, circular material use rate, and population are the explanatory variables that can have positive or negative effects and influences on the dynamics and evolution of the consumption of recyclable materials.

The period considered was 2013–2021, the data were annual, and the model was applied to the 27 EU Member States. We used econometric methods specific to panel data, i.e., regression models that focus on the interdependencies between variables (i.e., the common-effects model, fixed-effects model, and random-effects model), regression methods that highlight the long-term association between variables (i.e., panel FMOLS regression), the implementation of the multivariate autoregressive model in a modified form (i.e., VEC model), and the performance of econometric analyses specific to the autoregressive model applied (i.e., impulse response functions, variance decomposition, and Granger causality analysis).

The effects of short- and long-term interdependencies on the consumption of recyclable materials can be captured using an empirical economic model. In this context, the common-effects, fixed-effects, and random-effects regression models were applied to determine the short-term impact on the consumption of recyclable materials, while the FMOLS regression method, VEC model implementation, impulse response analysis, variance decomposition, and Granger causality test were applied to highlight the series of long-term relationships and interdependencies between the variables.

Table 1. Definitions of variables, units of measurement, and data sources.

Variable	Acronym	Definition	Measurement Unit	Data Source	Form of Variable
Consumption of recyclable materials	RECYCL	The indicator measures the annual quantity of recyclable materials that is consumed and subsequently utilised, collected, processed, and returned to the economy as raw materials or products.	Tonnes	Eurostat 1	Dependent variable
Raw material footprint	RMF	The indicator shows the amount of extraction required to produce the products demanded by end users in the geographical reference area, regardless of where in the world the material was extracted from the environment.	Tonnes per capita	Eurostat 2	Independent variable
Trade with recyclable materials	TRADE	This indicator measures the quantities of recyclable waste, scrap, and other secondary raw materials (by-products) transported between EU Member States (intra-EU) and across EU borders (extra-EU).	Tonnes	Eurostat 3	Independent variable
Greenhouse gas emissions	GHGE	This indicator covers greenhouse gas emissions from all production activities, including the production of goods and services.	Kilograms per capita	Eurostat 4	Independent variable
Investments in circular economy sectors	INV	This indicator includes “Gross investment in tangible goods” and “Value added at factor costs” in the following three sectors: the recycling sector, repair and reuse sector, and rental and leasing sector.	EUR million	Eurostat 5	Independent variable
Real GDP per capita	GDP_CAP	This indicator is calculated as the ratio of the real GDP to the average population each year.	EUR per capita	Eurostat 6	Independent variable
Renewable energy sources	RENEW	This indicator measures the share of renewable energy consumption in the gross final energy consumption.	Percentage	Eurostat 7	Independent variable
Circular material use rate	CMR	This indicator measures the proportion of total material used that is recycled and returned to the economy, thereby avoiding the extraction of primary raw materials.	Percentage	Eurostat 8	Independent variable
Population	PPL	This indicator refers to population data for each country in the European Union.	Number of people	Eurostat 9	Independent variable

Note: ¹ Published online on Eurostat and OECD. Retrieved from https://ec.europa.eu/eurostat/databrowser/product/page/ENV_AC_MFA; https://ec.europa.eu/eurostat/databrowser/view/nrg_cb_rw$defaultview/default/table (accessed on 10 January 2025). ² Published online at Eurostat. Retrieved from https://ec.europa.eu/eurostat/databrowser/view/ENV_AC_RME/default/table?lang=en (accessed on 10 January 2025). ³ Published online at Eurostat. Retrieved from https://ec.europa.eu/eurostat/databrowser/view/CEI_SRM020/default/table (accessed on 10 January 2025). ⁴ Published online at Eurostat. Retrieved from https://ec.europa.eu/eurostat/databrowser/view/cei_gsr011/default/table (accessed on 10 January 2025). ⁵ Published online at Eurostat. Retrieved from https://ec.europa.eu/eurostat/databrowser/view/cei_cie012$defaultview/default/table (accessed on 10 January 2025). ⁶ Published online at Eurostat. Retrieved from https://ec.europa.eu/eurostat/databrowser/view/sdg_08_10/default/table (accessed on 10 January 2025). ⁷ Published online at Eurostat. Retrieved from https://ec.europa.eu/eurostat/databrowser/view/sdg_07_40/default/table (accessed on 10 January 2025). ⁸ Published online at Eurostat. Retrieved from https://ec.europa.eu/eurostat/databrowser/view/cei_srm030/default/table (accessed on 10 January 2025). ⁹ Published online at Eurostat. Retrieved from https://ec.europa.eu/eurostat/databrowser/view/demo_pjan$defaultview/default/table (accessed on 10 January 2025). Source: Authors’ work.

shows the variables included in the applied model and their descriptions and definitions, units of measurement, and data sources. The variables used have an annual frequency for the period 2013–2021 and are available for each of the 27 EU Member States.

The panel was balanced and fixed, with data for each variable, country, and time under study. It is important to note that to ensure the robustness and appropriateness of the results, all variables incorporated into the empirical model were expressed in logarithmic form.

In the present article, the foundations are established for the construction of a two-stage model that is centred on measuring and estimating the consumption of recyclable materials at the level of the 27 European countries in the period 2013–2021.

The initial stage of the model (see Equation (2)) undertakes a static assessment of the impact of specific independent variables (e.g., the material footprint, renewable energy resources, the trade in recyclables, greenhouse gas emissions, etc.) on the dependent variable (i.e., consumption of recyclables). The subsequent stage of the model (see Equation (3)) undertakes a dynamic assessment of the impact of these independent variables with a lag (lag 1) on the dependent variable. The following explanations about the methodological framework can be found in Appendix D. The general form of the economic model is shown in Equation (1) below, and

Table 2. Descriptive statistics of the variables.

	RECYCL	RMF	TRADE	GHGE	INV	GDP_CAP	RENEW	CMR	PPL
Mean	10,311.34	18.1767	1,470,049.00	7694.312	3474.531	26,280.04	21.4802	8.7761	16,481,980.00
Median	430.2355	16.0560	501,411.00	6836.263	800.00	20,780.00	18.0010	7.00	8,772,865.00
Std.Dev.	42,246.81	7.8072	1,861,019.00	2937.259	6033.555	17,108.94	11.7135	6.292	21,768,937.00
Maximum	233,320.00	52.5260	8,062,841.00	16,698.00	34,489.00	86,690.00	62.5730	29.00	83,166,711.00
Minimum	0.2512	7.3500	1118.00	3786.840	33.00	5390.00	3.4940	1.30	422,509.00
Skewness	4.84	1.73	1.52	1.006	2.84	1.52	0.95	1.16	1.79
Kurtosis	24.69	6.87	4.51	3.24	11.57	5.64	3.70	3.89	5.02
Jarque–Bera (prob.)	5714.64 (0.00)	273.56 (0.00)	116.99 (0.00)	41.62 (0.00)	1071.62 (0.00)	164.87 (0.00)	41.93 (0.00)	62.70 (0.00)	172.29 (0.00)
Observations	243	243	243	243	243	243	243	243	243

Source: Authors’ work.

illustrates the descriptive statistics of the variables included in this study.

$${\mathrm{RECYCL}}_{\mathrm{it}}=\mathrm{f}\left({\mathrm{RMF}}_{\mathrm{it}},{\mathrm{TRADE}}_{\mathrm{it}},{\mathrm{GHGE}}_{\mathrm{it}},{\mathrm{INV}}_{\mathrm{it}},{\mathrm{GDP}\text{\_}\mathrm{CAP}}_{\mathrm{it}},{\mathrm{RENEW}}_{\mathrm{it}},{\mathrm{CMR}}_{\mathrm{it}},{\mathrm{PPL}}_{\mathrm{it}}\right)$$

(1)

where $\mathrm{i}$ indicates the country (i = 1 to 27), and $\mathrm{t}$ indicates the period (t = 2013 to 2021).

To study and investigate the existence of interdependencies between the consumption of recyclable materials and the explanatory variables used in this context, we developed a two-stage model at the level of EU countries, covering the period from 2013 to 2021. The general estimation form for the two-stage model proposed in this study is shown in Equations (2) and (3):

$${\mathrm{R}\mathrm{E}\mathrm{C}\mathrm{Y}\mathrm{C}\mathrm{L}}_{\mathrm{i}\mathrm{t}}={\mathsf{\beta }}_0+{\mathsf{\beta }}_1{\mathrm{R}\mathrm{M}\mathrm{F}}_{\mathrm{i}\mathrm{t}}+{\mathsf{\beta }}_2{\mathrm{T}\mathrm{R}\mathrm{A}\mathrm{D}\mathrm{E}}_{\mathrm{i}\mathrm{t}}+{\mathsf{\beta }}_3{\mathrm{G}\mathrm{H}\mathrm{G}\mathrm{E}}_{\mathrm{i}\mathrm{t}}+{\mathsf{\beta }}_4{\mathrm{I}\mathrm{N}\mathrm{V}}_{\mathrm{i}\mathrm{t}}+{\mathsf{\beta }}_5{\mathrm{G}\mathrm{D}\mathrm{P}\_\mathrm{C}\mathrm{A}\mathrm{P}}_{\mathrm{i}\mathrm{t}}+{\mathsf{\beta }}_6{\mathrm{R}\mathrm{E}\mathrm{N}\mathrm{E}\mathrm{W}}_{\mathrm{i}\mathrm{t}}+{\mathsf{\beta }}_7{\mathrm{C}\mathrm{M}\mathrm{R}}_{\mathrm{i}\mathrm{t}}+{\mathsf{\beta }}_8{\mathrm{P}\mathrm{P}\mathrm{L}}_{\mathrm{i}\mathrm{t}}+{\mathsf{\epsilon }}_{\mathrm{i}\mathrm{t}}$$

(2)

where ${\mathsf{\beta }}_{0,}$is the intercept term; ${\mathsf{\beta }}_1\text{→}{\mathsf{\beta }}_8$ are the estimated coefficients; $\mathrm{i}$ indicates the country (i = 1 to 27); $\mathrm{t}$ is the period (t = 2013 to 2021); ${\mathsf{\epsilon }}_{\mathrm{it}}$ is the error term.

$$\begin{gathered} {\mathrm{RECYCL}}_{\mathrm{it}}={\mathsf{\beta }}_0+{\mathsf{\beta }}_1{\mathrm{RMF}}_{\mathrm{it}}+{\mathsf{\beta }}_2{\mathrm{TRADE}}_{\mathrm{it}}+{\mathsf{\beta }}_3{\mathrm{GHGE}}_{\mathrm{it}}+{\mathsf{\beta }}_4{\mathrm{INV}}_{\mathrm{it}}+{\mathsf{\beta }}_5{\mathrm{GDP}\text{\_}\mathrm{CAP}}_{\mathrm{it}}+{\mathsf{\beta }}_6{\mathrm{RENEW}}_{\mathrm{it}}+{\mathsf{\beta }}_7{\mathrm{CMR}}_{\mathrm{it}}+{\mathsf{\beta }}_8{\mathrm{PPL}}_{\mathrm{it}} \\ +{\mathsf{\beta }}_9{\mathrm{RECYCL}}_{\mathrm{it}-1}+{\mathsf{\beta }}_{10}{\mathrm{RMF}}_{\mathrm{it}-1}+{\mathsf{\beta }}_{11}{\mathrm{TRADE}}_{\mathrm{it}-1}+{\mathsf{\beta }}_{12}{\mathrm{GHGE}}_{\mathrm{it}-1}+{\mathsf{\beta }}_{13}{\mathrm{INV}}_{\mathrm{it}-1}+{\mathsf{\beta }}_{14}{\mathrm{GDP}\text{\_}\mathrm{CAP}}_{\mathrm{it}-1}+{\mathsf{\beta }}_{15}{\mathrm{RENEW}}_{\mathrm{it}-1}+ \\ {\mathsf{\beta }}_{16}{\mathrm{CMR}}_{\mathrm{it}-1}+{\mathsf{\beta }}_{17}{\mathrm{PPL}}_{\mathrm{it}-1}+{\mathsf{\epsilon }}_{\mathrm{it}} \end{gathered}$$

(3)

where ${\mathsf{\beta }}_0$ is the intercept term; ${\mathsf{\beta }}_1\text{→}{\mathsf{\beta }}_{17}$ are the estimated coefficients; $\mathrm{i}$ indicates the country (i = 1 to 27); $\mathrm{t}$ is the period (t = 2013 to 2021); ${\mathsf{\epsilon }}_{\mathrm{it}}$ is the error term.

3.2. Methodological Design

After selecting the variables and obtaining their statistical descriptions, a correlation analysis was carried out using the Pearson correlation coefficient. This stage is considered as a necessary and preliminary step to the application of regression methods with panel data, since it allows for the determination and verification, from a statistical point of view, as to whether the variables included in the model have association and linkage relationships. In the subsequent phase of the analysis, the stationarity assumption for each variable was tested. The Augmented Dickey–Fuller (ADF-Fisher) test proposed in [58], the Levin, Lin, and Chu [59] test, the Im, Pesaran, and Shin [60] test, and the Phillips–Perron (PP-Fisher) test proposed in [61] were all used to identify the autoregressive unit root.

Each unit root test was developed based on the following hypotheses: the null hypothesis (H₀) $\to$the existence of a unit root, or the alternative hypothesis (H₁)$\to$ the non-existence of a unit root, and the general equation applied was Equation (4):

$${\mathrm{X}}_{\mathrm{it}}={\mathsf{\alpha }}_{\mathrm{i}}+{\mathrm{Y}}_{\mathrm{it}}\mathsf{\beta }+{\mathsf{\omega }}_{\mathrm{it}}$$

(4)

where $\mathrm{i}$= 1 to N (number of countries); $\mathrm{t}$ = 1 to T (number of years); ${\mathsf{\alpha }}_{\mathrm{i}}$ is the individual constant coefficient;$\mathsf{\beta }$ is the slop parameter; ${\mathsf{\omega }}_{\mathrm{it}}$ indicates the stationary distribution; and ${\mathrm{X}}_{\mathrm{it}}$ and ${\mathrm{Y}}_{\mathrm{it}}$ are the I (1) variables.

After testing the stationarity hypothesis, the short-term impact of recyclable consumption was determined for the period and countries analysed. Following the steps outlined in [62,63], the analysis considered three possible specifications for estimating these parameters: the pooled-effects (POLS), fixed-effects (FE), and random-effects (RE) models. The general form for the FE model is shown in Equation (5), while Equation (6) presents the general form for the RE model. These formulas have been used in previous studies [64] and are standardised for the econometric analysis of the models used:

$${\mathrm{Y}}_{\mathrm{it}}={\mathsf{\beta }}_0+{{\mathsf{\beta }}^{\prime }{\mathrm{X}}_{\mathrm{i}\mathrm{t}}\mathsf{\alpha }}_{\mathrm{i}}+{\mathsf{\epsilon }}_{\mathrm{it}}$$

(5)

where ${\mathrm{Y}}_{\mathrm{it}}$ is the dependent variable (i.e., the natural logarithm of RECYCL); ${\mathsf{\beta }}_0$ is the intercept term; ${\mathsf{\beta }}^{\prime }$ indicates the estimated coefficients; ${\mathrm{X}}_{\mathrm{it}}$ is a vector including the independent and explanatory variables; ${\mathsf{\alpha }}_{\mathrm{i}}$ is the error component that reflects each cross-sectional dimension and shows the differences between the countries; ${\mathsf{\epsilon }}_{\mathrm{i}\mathrm{t}}$ is the error term; $\mathrm{i}$ indicates the cross-sectional dimension (i.e., the EU countries, t = 1 to 27); and$\mathrm{t}$ represents the temporal dimension (i.e., t = 2013 to 2021).

$${\mathrm{Y}}_{\mathrm{i}\mathrm{t}}={\mathsf{\beta }}_0+{\mathsf{\beta }}^{\prime }{\mathrm{X}}_{\mathrm{i}\mathrm{t}}+{\mathsf{\alpha }}_{\mathrm{i}}+{\mathrm{u}}_{\mathrm{i}}+{\mathsf{\epsilon }}_{\mathrm{i}\mathrm{t}}$$

(6)

where ${\mathrm{u}}_{\mathrm{i}}$ is the error term for each country, and the other terms are explained in Equation (5).

After applying the three regression models with panel data, the next step was to evaluate the predictive analysis of the consumption of recyclable materials for the whole period analysed. In this respect, the following indicators of the robustness of the prediction of the proposed two-stage model were applied: the Root Mean Square Error (RMSE), Mean Absolute Error (MAE), and Theil’s U statistic indicator, which are presented in Equations (7)–(9):

$$\mathrm{RMSE}=\sqrt{{\sum }_{\mathrm{i}=1}^{\mathrm{N}}{\displaystyle \frac{({\widehat{\mathrm{Y}}}_{\mathrm{i}\mathrm{t}}-{\mathrm{Y}}_{\mathrm{i}\mathrm{t}})}{\mathrm{n}}}}$$

(7)

$$\mathrm{MAE}={\displaystyle \frac{1}{\mathrm{n}}}{\sum }_{\mathrm{i}=1}^{\mathrm{n}}(\left|{\mathrm{Y}}_{\mathrm{i}\mathrm{t}}-{\widehat{\mathrm{Y}}}_{\mathrm{i}\mathrm{t}})\right|$$

(8)

$$\mathrm{Theil}\text{’}\mathrm{s}\mathrm{U}=\sqrt{{\displaystyle \frac{{{\sum }_{\mathrm{i}=1}^{\mathrm{n}}\left[\frac{{\widehat{\mathrm{Y}}}_{\mathrm{i}\mathrm{t}}-{\mathrm{Y}\mathrm{i}}_{\mathrm{t}}}{{\mathrm{Y}}_{\mathrm{i}\mathrm{t}-1}}\right]}_{\mathrm{i}}^2}{{{\sum }_{\mathrm{i}=1}^{\mathrm{n}}\left[\frac{{\mathrm{Y}}_{\mathrm{i}\mathrm{t}}-{\mathrm{Y}}_{\mathrm{i}\mathrm{t}-1}}{{\mathrm{Y}}_{\mathrm{i}\mathrm{t}-1}}\right]}_{\mathrm{i}}^2}}}$$

(9)

where ${\widehat{\mathrm{Y}}}_{\mathrm{i}\mathrm{t}}$ is the predicted value of the RECYCL; ${\mathrm{Y}}_{\mathrm{i}\mathrm{t}}$ is the observed value of the RECYCL; $\mathrm{n}$ is the number of observations; and $\mathrm{t}$ is the period.

The next step was to test for the existence of long-run relationships between the variables included in the model. This was tested using Kao and Johansen cointegration [65,66] tests, based on Equation (10) and the following assumptions: the null hypothesis (H₀) → there is no evidence of cointegration between the variables, and the alternative hypothesis (H₁) → there is evidence of cointegration between the variables. If this aspect was confirmed, i.e., the variables were cointegrated and thus had a long-run causal relationship, we chose to apply the cointegration regression method in a modified form, i.e., FMOLS regression:

$$\text{∆}{\mathrm{X}}_{\mathrm{it}}={\mathsf{\alpha }}_{\mathrm{i}}+\mathsf{\beta }{\mathrm{X}}_{\mathrm{i}\mathrm{t},\mathrm{t}-1}+{\sum }_{\mathrm{k}}^{{\mathrm{n}}_{\mathrm{i}}}{\mathrm{p}}_{\mathrm{ij}}\text{∆}{\mathrm{X}}_{\mathrm{it},\mathrm{t}-\mathrm{j}}+{\mathsf{\epsilon }}_{\mathrm{it}}$$

(10)

where $\mathrm{i}$ = 1 to N (number of countries); $\mathrm{t}$ = 1 to T (number of years); $\text{∆}$ is the first difference of the variables; ${\mathrm{X}}_{\mathrm{i}\mathrm{t},\mathrm{t}-1}$ is the series in the panel in the time ($\mathrm{t}$); ${\mathrm{n}}_{\mathrm{i}}$ is the no. of lags; and ${\mathsf{\epsilon }}_{\mathrm{it}}$ is the distributed random variables.

The methodology specific to multivariate autoregressive models (in this case, the VEC model) was also implemented to determine the series of short- and long-run interdependencies between the variables and their deviations from the long-run equilibrium condition, as well as their short-run adjustment power. Therefore, if I (1) variables were cointegrated, it meant that the variables had a long-run relationship, and we could successfully implement and run the VEC panel model to examine both the short-run and long-run dynamics. The VEC model is useful for estimating both the short-run and long-run effects of one time series on another. The term error correction (ECT_t−1) refers to the fact that the deviation of the last period from a long-run equilibrium, the error, affects its short-run dynamics. Thus, ECMs directly estimate the speed with which a dependent variable returns to equilibrium following a change in other variables. The generalised formula for the VEC model is illustrated in Equation (11):

$$\text{∆}{\mathrm{Y}}_{\mathrm{t}}={\mathsf{\sigma }}_{\mathrm{it}}+{\sum }_{\mathrm{i}=1}^{\mathrm{k}-1}{\mathsf{\gamma }}_{\mathrm{i}\mathrm{t}}\text{∆}{\mathrm{Y}}_{\mathrm{t}-\mathrm{i}}+{\sum }_{\mathrm{j}=1}^{\mathrm{k}-1}{\mathsf{\eta }}_{\mathrm{j}\mathrm{t}}\text{∆}{\mathrm{X}}_{\mathrm{t}-\mathrm{j}}+{\sum }_{\mathrm{m}=1}^{\mathrm{k}-1}{\mathsf{\xi }}_{\mathrm{m}\mathrm{t}}\text{∆}{\mathrm{R}}_{\mathrm{t}-\mathrm{m}}+{\mathsf{\lambda }}_{\mathrm{i}\mathrm{t}}{\mathrm{ECT}}_{\mathrm{i}\mathrm{t}-1}+{\mathsf{\mu }}_{\mathrm{i}\mathrm{t}-1}$$

(11)

where $\mathrm{k}-1$ is the lag length reduced by 1; ${\mathrm{Y}}_{\mathrm{i}\mathrm{t}}$ and ${\mathrm{X}}_{\mathrm{i}\mathrm{t}}$ designate the dependent and independent variables; ${\mathsf{\sigma }}_{\mathrm{i}\mathrm{t}}$ is the intercept term; ${\mathsf{\gamma }}_{\mathrm{i}\mathrm{t}}$, ${\mathsf{\eta }}_{\mathrm{j}\mathrm{t}}$, and ${\mathsf{\xi }}_{\mathrm{m}\mathrm{i}}$ represent the short-run dynamic coefficients of the model’s adjustment long-run equilibrium. ${\mathrm{ECT}}_{\mathrm{i}\mathrm{t}-1}$indicates the lagged OLS residual obtained from the long-run cointegration equation ${\mathrm{Y}}_{\mathrm{i}\mathrm{t}}=\mathsf{\sigma }+{\mathsf{\eta }}_{\mathrm{j}}{\mathrm{X}}_{\mathrm{i}\mathrm{t}}+{\mathsf{\xi }}_{\mathrm{m}\mathrm{i}}{\mathrm{R}}_{\mathrm{i}\mathrm{t}}+{\mathsf{\mu }}_{\mathrm{i}\mathrm{t}}$ and is expressed as ${\mathrm{ECT}}_{\mathrm{i}\mathrm{t}-1}=\left[{\mathrm{Y}}_{\mathrm{i}\mathrm{t}-1}-{\mathsf{\eta }}_1{\mathrm{X}}_{\mathrm{i}\mathrm{t}-1}-{\mathsf{\xi }}_1{\mathrm{R}}_{\mathrm{i}\mathrm{t}-1}\right]$, the cointegration equation. The ${\mathrm{E}\mathrm{C}\mathrm{T}}_{\mathrm{i}\mathrm{t}-1}$ of the previous period’s deviation from the long-run equilibrium (which is the error) influences the short-run movement in the dependent variable. ${\mathsf{\lambda }}_{\mathrm{i}\mathrm{t}}$ is the coefficient of the ECT and the speed of the adjustment. It measures the speed at which y returns to equilibrium after changes in ${\mathrm{X}}_{\mathrm{i}\mathrm{t}}$ and ${\mathrm{R}}_{\mathrm{i}\mathrm{t}}$. ${\mathsf{\mu }}_{\mathrm{i}\mathrm{t}-1}$ is the residuals.

The employment of panel regressions engenders superior estimation and prediction, a phenomenon attributable to the presence of a dual index for their variables. The i-index is instrumental for elucidating the cross-sectional dimension, whilst the t-index serves to illuminate the temporal dimension. Consequently, the utilisation of the panel methodology approach facilitates the estimation of the influence of a single coefficient on the independent variables specified over time and by entity, thereby enabling the determination and estimation of these regressors [62,63]. A further advantage of the panel methodology is that it can estimate the coefficients in a dynamic and modified way, where they can be controlled by individual fixed effects. Thus, we opted for panel regression methods, considering that we had 27 countries and a large number of variables.

The results were supplemented with impulse response functions (IRFs) and variance decompositions based on stochastic error terms. This methodological approach enabled the assessment of the response of recyclable consumption to shocks in the estimation of the additional variables incorporated into the systems represented by the model. Ten forward periods were employed for the IRF and variance decompositions, and one standard deviation was used for the Cholesky innovations and factorisation. Following the implementation of the VEC model, a Granger test was employed to assess the causal relationships between the variables. The methodological scheme utilised in this study is illustrated in Figure 1.

4. Results

This section presents and discusses the principal findings of the economic model’s application to the analysis of the short- and long-term relationships in the consumption of recyclables in the 27 EU countries over the period 2013–2021.

In order to achieve the aforementioned objective, the methodological steps delineated in Figure 1 were meticulously followed.

As a preliminary measure, an investigation was conducted into the correlations between the variables employed in the model. The results of this investigation are presented in

Table 3. Correlations between the variables.

	RECYCL	CMR	GHGE	INV	RMF	PPL	GDP_CAP	RENEW	TRADE
RECYCL	1.00
CMR	0.32 *	1.00
GHGE	$-$0.17 *	0.097	1.00
INV	0.29 *	0.53 *	$-$0.10	1.00
RMF	$-$0.19 *	$-$0.26 *	0.33 *	$-$0.23	1.00
PPL	0.46 *	0.41 *	$-$0.18 *	0.86 *	$-$0.29 *	1.00
GDP_CAP	0.002	0.26 *	0.53 *	0.20 *	0.26 *	0.017	1.00
RENEW	$-$0.06	$-$0.28 *	$-$0.22 *	$-$0.15 **	0.51 *	$-$0.19 *	$-$0.081	1.00
TRADE	0.27 *	0.62 *	0.023	0.62 *	$-$0.43 *	0.70 *	0.21 *	$-$0.25 *	1.00

Note: * indicates statistical significance at 1% level; ** indicates statistical significance at 5% level; Source: Authors’ work.

. In addition, the results of the multicollinearity test are included in Appendix A.

As the focus pertained to the utilisation of recyclable materials, this indicator exhibited a positive medium-intensity correlation with the rate of the use of recyclable materials. This finding suggests that EU countries consistently implemented a series of recycling and reuse programmes during the period 2013–2021. Recently, this has been increasingly practised, and the transition to a circular economy is becoming a priority and an effective industrial strategy for the European Union, as shown in the study [67]. In addition, a statistically significant negative relationship (−0.17) was found between the consumption of recycled materials and greenhouse gas emissions, which is essential at the European level for the successful implementation of policies aimed at a significant reduction in carbon dioxide emissions and those in the energy sector aimed at increasing the use of renewable energy sources [68].

The positive, moderate, and statistically significant correlation between the consumption of recyclable materials and investments in specific sectors of circular economy activities is another aspect that emerges from the analysis addressed in this study. Thus, the more the European Union provides support and financial assistance to the circular economy sectors, the more positive externalities will be generated, attracting the young population to engage in these sectors and increasing the level of responsibility of European citizens towards circular economy practices.

The unit root tests were then used to test the stationarity hypothesis for each variable included in the panel. The results of these tests are presented in

Table 4. Panel unit root test results.

Variable	Levin, Lin, and Chu		Im, Pesaran, and Shin		ADF-Fisher		PP-Fisher
	Level	First Difference	Level	First Difference	Level	First Difference	Level	First Difference
RECYCL	1.63	$-$10.33 *	4.07	$-$4.23 *	23.41	115.83 *	20.28	136.83 *
RMF	$-$5.56 *	$-$18.33 *	$-$1.70	$-$7.92 *	81.33 *	175.40 *	76.54	187.61 *
TRADE	$-$4.24 *	$-$13.22 *	$-$1.40	$-$6.26 *	82.22 *	149.79 *	105.70 *	216.97 *
GHGE	$-$2.56 *	$-$11.81 *	0.09	$-$4.83 *	56.98	127.44 *	49.07	116.95 *
INV	$-$4.00 *	$-$13.88 *	$-$2.10	$-$6.03 *	84.26 *	146.21 *	125.41 *	201.94 *
GDP_CAP	$-$4.24 *	$-$13.71 *	0.30	$-$6.65 *	50.70	159.12 *	39.18	205.02 *
RENEW	3.43	$-$10.11 *	5.23	$-$4.19 *	21.60	117.96 *	25.24	138.06 *
CMR	$-$3.57 *	$-$12.27 *	$-$0.71	$-$6.43 *	68.00	151.14 *	66.80	154.58 *
PPL	$-$7.94 *	$-$5.62 *	0.23	$-$0.97	75.68	67.38	98.40 *	80.98 *
Result	I (1)

Note: * indicates statistical significance at 1% level. Source: Authors’ work.

. The results of the endogeneity of the regressors are included in the Appendix B. As a result, the variables were stationary at the first difference, with the applied tests indicating an extremely high level of significance of 1%, and the econometric methods for studying the impact on the consumption of recyclable materials could be successfully applied and implemented.

Table 5. Panel regression results for the first-stage model (with the RECYL as the dependent variable).

Estimated Coefficients
Independent Variables	Common-Effects Model	Fixed-Effects Model	Random-Effects Model
b0	2.04	8.55	$-$14.85 **
RMF	$-$0.32	$-$0.16	$-$0.069
TRADE	$-$0.24 **	0.096	0.12 ***
GHGE	$-$0.70 ***	0.12	0.15
INV	0.59 **	0.21 *	0.23 *
GDP_CAP	$-$0.70 **	0.32	0.16
RENEW	0.88 *	0.73 *	0.61 *
CMR	0.35 ***	$-$0.069	$-$0.048
PPL	0.87 *	$-$0.70	0.84 *
Diagnostic and Robustness Tests
R2	0.55	0.99	0.32
Adj. R2	0.54	0.98	0.30
S.E. of regression	1.84	0.26	0.26
F-statistic	36.54	783.86	13.72
F-statistic (prob.)	0.00	0.00	0.00
Forecast Evaluation Indicators
RMSE	1.80	0.28	0.24
MAE	1.26	0.14	0.14
Bias Proportion	0.00	0.00	0.00
U statistic indicator	0.14	0.018	0.018

Note: * indicates statistical significance at 1% level; ** indicates statistical significance at 5% level; *** indicates statistical significance at 10% level. RMSE: Root Mean Squared Error; MAE: Mean Absolute Error. Source: Authors’ work.

presents the results obtained for the model in the first stage using the pooled OLS model, the fixed-effects model, and the random-effects model to analyse the interdependencies between the variables and the short-term influence on the consumption of recyclable materials in the analysed European countries in the period 2013–2021.

The first observation is that both circular investment and renewable energy sources have a positive impact on the consumption of recyclables, while the development of circular trade has a different impact on the consumption of recyclables, as the results show a positive impact from the random-effects estimation method and a negative impact from the common-effects method. Specifically, a 10 p.p. increase in the trade of recyclable materials in the short term increases the consumption of recyclable materials by about 1.20 p.p., which is a positive and direct effect in the short term.

The results also showed an indirect and negative interdependence between the consumption of recyclable materials and greenhouse gas emissions, meaning that for every 10 p.p. increase in the consumption of recyclable materials, GHG emissions are reduced by about of 7 p.p. in the short run. This negative influence is justified by the need to significantly reduce greenhouse gas emissions at the level of EU countries to ensure the most practical, plausible, and beneficial integration of the components of the circular economy for European citizens.

Another interesting finding is the positive and statistically significant impact of the population on the consumption of recyclable materials. For example, in the short term, the consumption of recyclable materials tended to increase by approximately 8.70 p.p. in the common-effects model and by 8.40 p.p. in the random-effects model for a 10 p.p. change in the population. The results of the common-effects regression model show a direct and positive relationship between the consumption of recyclable materials and the recycling rate; i.e., a 10 p.p. increase in the recycling rate changes (i.e., increases) this consumption by approximately 3.50 p.p. In the model, the results of the regression methods used show that renewable energy sources have a statistically significant impact on the consumption of recyclable materials. For instance, the consumption of recyclable materials will increase by 7.40 p.p. in the short run for a 10 p.p. change in the renewable energy sources.

The consumption of recyclable materials showed an increase of about 5.90 p.p. according to the common-effects model, an increase of 2.10 p.p. according to the fixed-effects model, and an increase of about 2.30 p.p. according to the random-effects model when the investments in the circular sectors were modified by 10 p.p. in the short run. This is due to higher investments in those sectors of activity that generate specific circular economy actions and strategies. All these results are adequate and appropriate to the analysis performed. They highlight the importance of using the consumption of recyclable materials in economic activities to achieve an important step towards sustainable growth, attracting sustainable investment, instilling a sense of social responsibility, and educating the population to protect and care for nature, all of which contribute consistently to achieving the objectives of the circular economy [69,70,71].

Table 6. Panel regression methods results for the second-stage model (with the RECYL as the dependent variable).

Estimated Coefficients
Independent Variable	Common-Effects Model	Fixed-Effects Model	Random-Effects Model
b0	1.28 **	$-$2.54	1.37
RMF	$-$0.15	$-$0.07	$-$0.11
TRADE	0.073	0.080	0.075
GHGE	0.11	0.28	0.14
INV	0.09 ***	0.11 ***	0.09 ***
GDP_CAP	0.59 ***	0.39	0.48
RENEW	0.30 ***	0.29 ***	0.24 ***
CMR	0.008	0.0029	0.024
PPL	0.65	2.98	1.60
RECYCL (−1)	1.00 *	0.80 *	0.99 *
RMF (−1)	0.20	0.21	0.18
TRADE (−1)	$-$0.065	$-$0.048	$-$0.063
GHGE (−1)	$-$0.17	$-$0.26	$-$0.21
INV (−1)	$-$0.081	$-$0.030	$-$0.069
GDP_CAP (−1)	$-$0.63 ***	$-$0.42	$-$0.54 ***
RENEW (−1)	$-$0.36 **	$-$0.33 **	$-$0.33 **
CMR (−1)	0.022	$-$0.001	0.001
PPL (−1)	$-$0.67	$-$2.81	$-$1.62
Diagnostic and Robustness Tests
R2	0.99	0.99	0.99
Adj. R2	0.98	0.98	0.98
S.E. of regression	0.17	0.16	0.16
F-statistic	3108.62	1452.61	1329.08
F-statistic (prob.)	0.00	0.00	0.00
Forecast Evaluation Indicators
RMSE	0.37	0.16	0.20
MAE	0.23	0.09	0.13
Bias proportion	0.0028	0.00014	0.0041
U statistic indicator	0.028	0.012	0.015

Note: * indicates statistical significance at 1% level; ** indicates statistical significance at 5% level; *** indicates statistical significance at 10% level. RMSE: Root Mean Squared Error; MAE: Mean Absolute Error. Source: Authors’ work.

shows the results for the second stage of the model, including the control variables that were expressed in the first lag. The findings of this study provide evidence that circular investments have a positive and statistically significant impact, indicating that for a 10 p.p. change in such investments, there is an observed average increase in the consumption of recyclable materials of 1 p.p. For the analysed EU countries, it can be observed that for a 10 p.p. change in the real GDP per capita (i.e., an increase), the impact is in the same direction: the consumption of recyclable materials increases by about 5.90 p.p. in the short term, which was captured by the common-effects regression model. However, the effect undergoes a shift when a 10 p.p. change in the GDP per capita is considered, with a lag (i.e., lag 1 of GDP per capita). This is evidenced by the decline in the consumption of recyclable materials, averaging 5.85 p.p. (this decrease was deemed statistically significant when examining the fixed- and random-effects models).

Similar to the results in Table 5, an interrelationship was found between the consumption of recyclables and renewable energy sources: if the latter increased by 10 p.p., then consumption also increased by 3 p.p. in the common-effects regression, by 2.90 p.p. in the fixed-effects regression, and by 2.40 p.p. in the random-effects regression. A 10 p.p. change in the lag values of renewable energy sources led to an average reduction in the consumption of recyclable materials of around 3.40 p.p. in the short run. The observed variability can be attributed to the challenges faced by certain European nations in augmenting the contribution of renewable energy sources to their aggregate energy production [72,73].

Another interesting aspect found in Table 6 relates to the statistically increased impact of the past value of the consumption of recyclable materials on the current value, which allows us to assume that European countries have a strong inclination and willingness to achieve maximum recycling, reuse, and waste reduction in the present and soon.

The accuracy and diagnostic results for the applied two-stage model are presented in Table 5 and Table 6 according to the regression methods used. In particular, the proposed model was statistically validated according to the two specifications, with the convergence of high F-test values and probabilities to 0 confirming this evidence. The high values of the adj. R² also emphasise and qualify the statistical power of the proposed model, showing that the independent variables included in the model had a significant impact on the dependent variable, which, in this case, was the consumption of recyclable materials.

Analysing the results of the RMSE, MAE, bias indicator, and Theil’s U statistic coefficient indicators, it was concluded that the best prediction was given by the random-effects regression model, as can be seen in Figure 2. In this sense, these indicators show the degree of performance in predicting the consumption of recyclable materials. The regression method with random effects shows the minimum values for the MAE (0.14 and 0.13), while the U statistic is close to zero, indicating a very good prediction result. Also, the model proposed and applied in this study is homoscedastic, and there is no autocorrelation or serial correlation in the series of the error term. In addition, robustness tests were carried out on the predictive and forecasting ability of the economic model analysed.

The present study also examined cointegration relationships, i.e., whether there were long-term association relationships between the variables included in the model. Following the research direction of other studies [74,75], the verification of the cointegration hypothesis was tested by means of the Johansen cointegration test and the Kao residual cointegration test, and the results are presented in

Table 7. Panel cointegration test results.

Johansen Cointegration Test
Trace
Equation	Trace Statistic	0.05 Critical Value	Prob. **
None *	265.98	197.37	0.0000
At most 1 *	197.88	159.52	0.0001
At most 2 *	137.31	125.61	0.0079
At most 3	94.39	95.75	0.0618
At most 4	55.45	69.81	0.4001
At most 5	25.50	47.85	0,9045
At most 6	11.63	29.79	0.9436
A most 7	1.27	15.49	0.9997
At most 8	0.11	3.84	0.7293
Maximum Eigenvalue
Equation	Max. Eigen Statistic	0.05 Critical Value	Prob.**
None *	68.10	58.43	0.0043
At most 1 *	60.57	52.36	0.0059
At most 2	42.91	46.23	0.1088
At most 3	38.94	40.07	0.0668
At most 4	29.94	33.87	0.1372
At most 5	13.86	27.58	0.8317
At most 6	10.86	21.13	0.7094
A most 7	1.15	14.26	0.9996
At most 8	0.11	3.84	0.7293
Kao Residual Cointegration Test
ADF
T-Statistic		$-$12.044 *
Prob.		0.0000

Note: * indicates rejection of the null hypothesis (H₀ $\to$ no cointegration) at the 5% level. ** p-values. Source: Authors’ work.

. Thus, the values obtained by the Johansen and Kao cointegration tests indicate the presence of at least two cointegrating equations, while the values suggested by the trace and maximum eigenvalue tests are higher than the critical values at the 5% significance level. In this context, the long-run interrelationships between the variables included in the panel were analysed by applying cointegration regression in a modified form, namely, panel FMOLS cointegration regression, and the specific results are shown in

Table 8. Results of the panel FMOLS regression (with the RECYL as the dependent variable).

Independent Variables	RMF	TRADE	GHGE	INV	GDP_CAP	RENEW	CMR	PPL
Estimated coefficients	−0.1647 *	0.1587 *	0.1895 *	0.2492 *	0.2343 *	0.6835 *	−0.0370 *	−0.7785 *
R2	0.9931
Adj. R2	0.9918

Note: * indicates statistical significance at 1% level. Source: Authors’ work.

.

The first point raised is suggested by the fact that all the estimated coefficients are statistically significant at a significance level of at least 5%. It can be observed that investments in the circular economy have a positive effect on the consumption of recyclable materials in the long term. A 10 p.p. increase in investments results in an increase in the consumption of recyclable materials by approximately 2.50 p.p. Another significant long-term positive impact comes from the trade in recyclable materials, which generates an increase in the consumption of recyclable materials of around 1.59 p.p.

The positive sign of the renewable energy coefficient indicates a positive impact on the consumption of recyclable materials. For instance, a 10 p.p. increase in this indicator will generate an increase of more than 6.80 p.p. of the consumption of recyclable materials in the long term. From this standpoint, renewable energy resources appear to be one of the most efficient and effective solutions to the environmental problems currently being experienced.

These problems (i.e., air pollution, waste management, plastic pollution, and food) require long-term actions for sustainable development in the 27 EU countries. Another important finding of the FMOLS regression is the long-term relationship between the consumption of recyclable materials and the real GDP per capita. It was demonstrated that when the GDP per capita increases by 10 p.p., the consumption of recyclable materials shows a growth of 2.35 p.p. in the long term, which shows that EU countries are starting to apply specific circular economy practices and policies in their industries and economic activities. It is confirmed that the transition to a greener economy generates positive effects and externalities at the economic level.

This is due to the simple fact that there must always be an interdependence between the environment and the economy, which facilitates growth and sustainable development, aspects that are captured in the studies carried out by the authors of [76,77,78]. Furthermore, the specific results of the regression method used were statistically validated, and the estimated coefficients confirm that there is a long-term relationship between the variables, and that the independent variables explain more than 95% of the impact on the consumption of recyclable materials for the European countries.

Alternatively, multivariate and autoregressive methods were used to provide additional details on the short- and long-term relationships between the consumption of recyclable materials and the rest of the explanatory variables at the European level and over the period analysed. Several indicators (i.e., the final prediction error, Akaike information criterion, Schwarz information criterion, and Hannan–Quinn information criterion) were used to determine the optimal number of lags in the implementation of the VEC model in the constituted panel. In addition, the panel VEC model was implemented with three lags, which is shown by the statistical indicators of accuracy in the choice of the correct and optimal number of lags.

Table 9. Results for the panel VEC model long-run coefficients.

Estimated Long-Run Coefficient	Value
RECYCL t−1	1.00 *
CMRt−1	76.95 *
GHGEt−1	2.67 *
INVt−1	−2.44 *
RMFt−1	449.45 *
PPLt−1	7.53 × 10−5 *
GDP_CAPt−1	−0.25 *
RENEWt−1	32.18 *
TRADEt−1	−0.003 *

Note: The signs of the coefficients are reversed in the long run. In the case of the present model, the GHGE had a negative impact on the RECYCL, the INV had a positive impact on the RECYCL, while the GDP_CAP had a positive impact on the RECYCL, on average, caeteris paribus. * The estimated long-run coefficients are statistically significant at the 1% level. Source: Authors’ contribution.

and

Table 10. Results for the VEC panel model short-run coefficients.

Δ (Dependent Variables)
Δ (Independent Variables)	RECYCLt	CMRt	GHGEt	INVt	RMFt	PPLt	GDP_CAPt	RENEWt	TARDEt
ECTt−1	−0.014 *	$-9.75\times {10}^{-7}$	−0.003	0.023 *	$2.85\times {10}^{-7}$	−0.764 *	−0.0001	$3.02\times {10}^{-6}$	−0.133
RECYCLt−1	−0.080	$9.02\times {10}^{-5}$	−0.091	0.138	$-9.68\times {10}^{-5}$	12.280	−0.098	$2.67\times {10}^{-5}$	−6.245
RECYCLt−2	−0.133 **	−0.0001	0.086	0.235 **	0.0002	28.555 *	0.306 ***	−0.0001	80.483 *
RECYCLt−3	0.035	$6.75\times {10}^{-5}$	0.015	−0.089	−0.0001	−16.286 **	−0.105	$-7.31\times {10}^{-5}$	−31.746 ***
CMRt−1	−1.992	0.235 **	11.810	60.206	−0.127	−3557.026	−63.540	−0.001	−7741.093
CMRt−2	−1.278	−0.298 **	−19.923	−41.973	0.021	4734.505	−60.119	−0.294 ***	−925.504
CMRt−3	−29.566	−0.048	−44.164	24.413	−0.074	1204.022	−48.219	0.285 ***	−3416.642
GHGEt−1	−0.016	$-1.07\times {10}^{-5}$	0.288 *	−0.021	−0.0002	5.125	−0.593 *	0.0004	−10.640
GHGEt−2	0.008	$-1.84\times {10}^{-5}$	−0.278 **	0.064	0.0001	−2.751	0.071	−0.0009 *	36.023
GHGEt−3	0.027	−0.0004 ***	−0.299 **	−0.080	−0.0003	7.579	−0.450 ***	0.0003	−9.619
INVt−1	−0.074	$1.38\times {10}^{-5}$	0.022	0.104	−0.0001	−11.297	0.029	$-5.25\times {10}^{-5}$	24.432
INVt−2	−0.031	$-5.68\times {10}^{-5}$	−0.023	−0.241 *	$6.93\times {10}^{-5}$	7.567	0.048	−0.0001	1.561
INVt−3	−0.006	$9.66\times {10}^{-5}$	−0.002	0.185 *	−0.0001 ***	−19.404 *	−0.134	$-6.20\times {10}^{-5}$	0.213
RMFt−1	30.901	−0.030	−95.359 **	−28.337	−0.144	549.265	−16.710	0.096	−2798.767
RMFt−2	3.981	0.046	−72.722 ***	0.575	−0.050	−2697.182	−106.618	0.030	−6883.586
RMFt−3	17.440	−0.011	10.831	14.872	−0.096	−1676.384	−105.46	−0.047	3326.290
PPLt−1	−0.001 *	$-1.46\times {10}^{-6}$	−0.0004	0.011 *	$-6.24\times {10}^{-7}$	0.152 **	0.0008	$1.45\times {10}^{-7}$	0.352 ***
PPLt−2	0.002 **	$5.13\times {10}^{-7}$	−0.0007	0.0008	$2.62\times {10}^{-7}$	0.391 *	−0.001	$1.93\times {10}^{-6}$	−0.776 **
PPLt−3	0.001 ***	$9.00\times {10}^{-7}$	−0.0001 **	−0.003 *	$-1.09\times {10}^{-6}$	−0.035	−0.001	$1.52\times {10}^{-6}$	−0.257
GDP_CAPt−1	0.111 **	0.0003 **	−0.149 **	0.077	−0.0001	7.286	−0.216 ***	$9.46\times {10}^{-5}$	−10.773
GDP_CAPt−2	−0.028	$-5.06\times {10}^{-5}$	0.011	−0.027	$8.22\times {10}^{-6}$	3.421	0.445 *	$9.92\times {10}^{-5}$	−21.473
GDP_CAPt−3	−0.064	$-4.21\times {10}^{-5}$	0.070	−0.045	−0.0001	0.004	0.502 *	$-5.54\times {10}^{-5}$	53.688 *
RENEWt−1	−35.924	−0.067	5.746	−69.565	0.018	1165.866	76.147	−0.168	1263.726
RENEWt−2	−12.890	0.020	−66.945	−52.302	−0.176	−5866.790	−36.150	0.035	16391.08
RENEWt−3	37.814	−0.128	−53.716	−21.994	0.152	6180.600	15.133	−0.148	−8655.599
TRADEt−1	−0.0002	$1.25\times {10}^{-6}$ ***	$9.15\times {10}^{-5}$	0.0006 ***	$8.22\times {10}^{-9}$	−0.028	−0.0008	$-1.24\times {10}^{-7}$	−0.451 *
TRADEt−2	−0.001	$1.06\times {10}^{-6}$ ***	−0.0001	0.0007 ***	$-7.66\times {10}^{-7}$	−0.041	−0.0006	$-2.96\times {10}^{-7}$	0.072
TRADEt−3	−0.0003	$4.93\times {10}^{-7}$	0.0001	0.0002	$1.46\times {10}^{-7}$	0.051 ***	0.0004	$-1.26\times {10}^{-7}$	−0.234 *

Note: * indicates statistical significance at 1% level; ** indicates statistical significance at 5% level; *** indicates statistical significance at 10% level. Source: Authors’ work.

show the estimated long-run and short-run coefficients and how they may be adjusted in the next period.

First, the long-term coefficient (ECT_t−1) is negative and statistically significant at the 1% significance level for the consumption of recyclable materials, which explains that the previous year’s standard deviation from the long-run equilibrium was corrected in the next year at an adjustment rate of 0.14 p.p. for the dependent variable. This is also reflected for circular investment and population, where the long-term equilibrium adjustment coefficients indicate corrections to the next year’s values of around 0.23 p.p. and 7.60 p.p., respectively. Using the value of the estimated short-term coefficients, it was found that the population has a negative effect on the consumption of recyclable materials, while the real level of GDP per capita has a positive effect on the consumption of recyclable materials within the EU countries. For example, a 1 p.p. change in consumption is associated with a short-term increase in the GDP per capita of about 0.11 p.p.

Another interesting finding captured by the VEC model results is that the present value of the consumption of recyclables has an impact on the future value of the consumption of recyclables, so that a decrease in the present consumption of recyclables leads to a decrease in the future consumption of recyclables even for over 2 years, indicating the need for European countries to implement programmes aimed at encouraging the population to recycle and reduce waste. Again, these mathematical results have a logical connotation, namely, that the more the citizens of European countries are attracted by and convinced of the benefits of recycling, the more the European Union will achieve its environmental goals and make significant progress in supporting the circular economy and achieving the much desired goal of climate neutrality by 2050.

Furthermore, the financial support provided by the European Union is essential and welcome for the development of sectors specific to the circular economy, with net positive effects on employment rates, productivity growth, and gross value-added growth (i.e., all contributing to higher GDP levels). This is confirmed, for example, by the short-term interaction between circular investment and the consumption of recyclable materials; i.e., a 1 p.p. change in circular investment is associated with a 0.23 p.p. increase in the previous value of recyclable consumption. In addition, the results showed that for a 1 p.p. change in the GDP, the consumption of recyclables is associated with an increase of about 0.31 p.p.

Table 11. Robustness results for the panel VEC model.

Dependent Variable	RECYCL	CMR	GHGE	INV	RMF	PPL	GDP_CAP	RENEW	TRADE
R2	0.71	0.22	0.31	0.76	0.17	0.74	0.42	0.23	0.57
Adj. R2	0.64	0.01	0.13	0.70	−0.044	0.67	0.27	0.03	0.45
S.E.	467.21	1.24	607.31	804.02	1.55	65,071.60	1216.96	1.69	152,241.00
F-Statistic	9.44 *	1.07	1.72 *	12.21 *	0.79	10.51 *	2.69 *	1.15	4.95 *
DW stat	1.61	1.84	2.13	1.59	1.94	2.17	2.37	2.15	2.14
LM test	1.24 (0.21)	10.28 (0.00)	1.58 (0.11)	3.60 (0.063)	7.36 (0.00)	3.76 (0.00)	28.26 (0.071)	10.39 (0.00)	0.32 (0.74)

Note: * indicates statistical significance at 5% level. DW: Durbin−Watson; S.E.: standard error. Source: Authors’ work.

shows the accuracy and diagnostic results for the VEC panel model. The results presented show that the VEC model does not suffer from autocorrelation and has a constant variance value; i.e., it is homoscedastic, which is supported by the Durbin–Waston and Breusch–Pagan LM tests. Looking strictly at the situation of this key variable (i.e., the consumption of recyclable materials), it was observed that the regression model is valid and correctly specified, as suggested by the high and statistically significant value of the F-test (the value is 9.44), while the coefficient of determination shows that the level of the consumption of recyclable materials is explained and influenced by 71% of the independent variables used in the proposed model. Furthermore, it was performed a sensitivity analysis to assess how the results change under different assumptions or model specifications and it is included in the Appendix C.

Another relevant aspect that can be derived from the use of the VEC panel model is its ability to predict the variance in and evaluate the response of an endogenous/dependent variable to the impulse of a group of exogenous or independent variables. Consequently, this study was able to analyse the response of the consumption of recyclable materials and predict the variance in this indicator by applying impulse response functions and variance decomposition analysis over a 10-year period. The results are shown in Figure 3 and Figure 4.

For example, the impact of the GDP per capita on the consumption of recyclable materials is positive throughout the 10-year forecast horizon and is particularly stable between the sixth and ninth years of the forecast horizon. In addition, circular investments have a positive impact on the consumption of recyclable materials, along with the material and population footprints. This confirms that the 27 European countries are making remarkable progress towards a circular economy and sustainable growth. At the same time, renewable energy is seen as a key determinant with a positive impact on the consumption of recyclable materials over the 10-year forecast period, while the negative impact of greenhouse gas emissions on the consumption of recyclable materials is highlighted.

The findings are consistent with the results of previous studies, including those referenced in [79,80,81]. These studies indicate that the implementation of more efficient and circular industrial material use, as outlined in circular economy strategies, could result in a 40% reduction in global greenhouse gas (GHG) emissions by 2050. This reduction would be achieved by utilising just four key industrial materials—cement, steel, plastics, and aluminium—in a more sustainable mode. The implementation of circular economy solutions is a crucial element in the pursuit of decarbonisation scenarios. In the context of the Paris Agreement, the absence of the explicit identification of circular actions in the nationally determined contributions is a notable limitation. However, achieving the targets outlined in the agreement is contingent upon the incorporation of these actions. The extraction and processing of raw materials, encompassing fossil fuels and agricultural sources, contributes to approximately 50% of the global greenhouse gas emissions, with the related carbon footprint being a significant element in this statistic.

According to the results of the variance decomposition for the consumption of recyclable materials (Figure 4), it is suggested that this indicator itself has the greatest impact, but its degree of impact gradually decreases from 100% in the first year of the forecast to 73.83% in the last year of the forecast. The major influence on the consumption of recyclable materials is underlined by the levels of population, trade in recyclable materials, and greenhouse gas emissions, which together contribute about 17% to the shocks on the consumption of recyclable materials, indicating the increased and significant influence of these exogenous variables in forecasting the future level of the consumption of recyclable materials for the EU countries. It should also be noted that investment, renewable energy, and the trade in recyclable materials are other relevant variables whose impact on the consumption of recyclable materials was 3.10%, while the impact of the GDP per capita increased from 5.20% in the mentioned prediction interval.

Another important aspect of this study is the implementation of Granger causality tests, which provided a better explanation of how the variables interacted and produced long-term effects on the consumption of recyclable materials within European countries. The results of the applied causality test are presented in

Table 12. Panel causality test results.

Null Hypothesis (H0) → Variable on the Column Does Not Cause Variable on the Line.
Variable	RECYCL	CMR	GHGE	INV	RMF	PPL	GDP_CAP	RENEW	TRADE
RECYCL	-	0.40	0.51	4.39 *	0.20	11.20 *	0.70	0.30	2.73 **
CMR	1.51	-	0.97	0.06	1.37	0.08	1.78	1.001	0.01
GHGE	0.06	0.86	-	0.29	1.01	0.41	2.30 ***	2.77 **	0.73
INV	8.00 *	0.59	0.82	-	0.83	0.79	1.42	0.40	3.42 **
RMF	0.38	2.41 ***	5.18 *	0.33	-	0.07	0.69	0.95	1.08
PPL	30.88 *	0.99	0.49	36.65 *	0.67	-	1.14	0.85	3.73 *
GDP_CAP	0.33	6.04 *	0.06	0.18	3.90 *	1.14	-	1.99 ***	0.12
RENEW	0.84	1.90 ***	2.15 ***	0.36	2.43 ***	0.11	2.54 **	-	0.80
TRADE	2.43 ***	0.87	0.28	2.37 ***	0.45	0.96	0.72	0.16	-

Note: The Granger causality test was performed using three lags. * Indicates statistical significance at 1% level; ** indicates statistical significance at 5% level; *** indicates statistical significance at 10% level. Source: Authors’ work.

.

The bidirectional relationship between the consumption of recyclable materials and investments in specific circular economy sectors was identified in this study. This aspect has been validated and confirmed by previous expert studies that have sought to explain the interplay between circular material consumption and investment dynamics [4,10,32,33]. The findings of these studies demonstrated that, in the context of short-term causality, an enhancement in material recycling results in a reduction in waste generation. This suggests that the promotion of circular economy-related innovation and investment, as well as the stimulation of material recycling to encourage the secondary raw material market, can facilitate the achievement of zero-waste targets.

At the same time, the existence of a bidirectional causal relationship between the consumption of recyclable materials and the trade in these materials has been identified. In the context of EU practices, one of the most visible actions in terms of the consumption and trade of recyclable materials is the introduction of new requirements and conditions for the packaging of recyclables. This directive is in line with the objective proposed by the European Commission [3]. The overarching aim of this directive is to reduce the negative impact of plastic packaging by up to 5% by 2030.

Nonetheless, the findings of numerous pieces of research have indicated that the utilisation and recuperation of recyclable materials is not wholly efficacious, as it has been demonstrated to give rise to issues with the quality of the new materials produced. Simultaneously, it has been acknowledged that these product recycling practices represent a significant step towards reducing waste and conserving resources [10,11,53]. Conversely, the extant literature has placed an emphasis on the reduction in contamination and the enhancement of infrastructure to streamline the processes of the use and recuperation of recyclable materials [18,24,55]. Consequently, the impact of sustainability and the circular economy on economic growth is not augmented by a straightforward transition to renewable resources or materials.

The authors emphasise that recycling rates and eco-innovation play significant roles in sustainable economic development and growth. However, other studies [82,83] contend that environmental taxes are among the most crucial drivers of economic growth, while they also validate the findings that recycling rates and trade in recyclable materials demonstrate a positive and substantial impact on economic growth. The study further indicates that an impetus for the consumption of recyclable materials is the mounting use of renewable energy sources.

The present findings are consistent with the extant literature [82,83,84,85] on the subject of the EU, which also demonstrates that renewable energy contributes almost half as much to greenhouse gas emissions as fossil energy. However, these findings are concomitant with results that indicate a more significant relationship between economic growth and renewable electricity consumption as the share of the renewable energy sector in the economy increases. Therefore, the consumption of recyclable materials is becoming increasingly important across the 27 EU countries.

The research suggests that multilateral policies should be implemented to foster economic growth and the expansion of the circular economy. These policies have the potential to have a positive impact on sustainable growth and development and are a key component of the circular economy. The findings underscore the significance of augmenting the utilisation of recyclable materials as a pivotal strategy to sustaining productive economic activity. They further underscore the pivotal role of the proactive promotion of the circular economy as a crucial instrument in the pursuit of global sustainable development.

The principal benefit of adopting a sustainable circular economy model is evident from the standpoint of human life [83]. By establishing a parallel between the longevity of products and human life, it becomes apparent how environmentally sustainable factors are inextricably linked with quality of life. In this regard, the European Commission’s Environmental Policy Report emphasises the importance of resource reuse at the EU level, underscoring the necessity for Member States to allocate substantial resources to the preservation of environmental infrastructure and the development of effective instruments to achieve their environmental objectives [86,87,88].

5. Discussion

The model proposed in this research showed a statistically significant and positive impact of the recycling rate on the consumption of recyclable materials in the 27 EU Member States. It is important to note that, in this paper, we have always used the ceteris paribus (all else being equal) hypothesis when analysing the effects of the explanatory variables.

The results confirm and support hypothesis 1 (H₁): the consumption of recyclable materials, raw material footprint, and circular material use rate are expected to be positively correlated. In this sense, the model was able to demonstrate that the consumption of recyclable materials is influenced by the recycling rate, showing an increase of about 3.50 p.p. in the short term when the recycling rate was changed by 10 p.p.

Conversely, the findings of the dynamic model (i.e., the second stage of the model) indicate that the recycling rate exerts a direct and positive influence on the consumption of recyclable materials. Furthermore, it is observed that in the long term, a negative interdependence exists between the consumption of recyclable materials and the recycling rate. This indicates that a 10 p.p. change in the recycling rate will result in a decrease in the consumption of recyclable materials of approximately 0.37 p.p.

This can be explained by the fact that the European countries have different and distinct orientations towards the targets related to recycling and the use of recyclable materials, an aspect that is explained in the studies [42,56,81]. These authors argue that countries can minimise their demand and use of raw materials in the long term by promoting eco-innovation and recycling. These studies show that the recycling rates vary considerably across Europe, with high rates in Northern and Western European countries and a common factor being the level of economic development and advancement. The view is that the key to more sustainable development is the recovery of as much material value from waste as possible through effective recycling.

The following hypothesis (H₂) can be also confirmed: The consumption of recyclable materials is directly influenced by the trade in recyclable materials, and this direct relationship is statistically significant at the 5% significance level.

There is scientific support for hypothesis 2. In this regard, attention is drawn to the study by [89]. The results of this study suggest that a highly important and relevant indicator that positively affects circular economy indicators is the trade in recyclable materials. Furthermore, the authors applied the dynamic form of the regression method to the 27 EU countries and found that private investment in the circular economy sectors shows an increase in the trade in recycled raw materials. The transition to a circular economy therefore involves the development of the trade in recyclable raw materials, which could have far-reaching effects. Investment represents a key relevant factor for the benefits of the circular economy.

As far as hypothesis 3 (H₃): The consumption of recyclable materials is positively influenced by renewable energy sources is concerned, this has a statistically positive impact on the consumption of recyclable materials in the long run. Specifically, according to the results of the FMOLS cointegration regression method, a 1 p.p. increase in the share of renewable energy sources increases the consumption of recyclable materials by up to 0.70 p.p. in the long run.

From this standpoint, hypothesis 3 (H₃) is substantiated, and the subsequent notion, as posited in the extant literature [90,91,92], can be endorsed. This assertion pertains to the significance of renewable energy utilisation in the context of circular product and resource generation, encompassing the design, manufacturing, construction, and management of renewable assets, in addition to the management of their post-use phase.

Furthermore, one of the main reasons for the transition from fossil fuels to renewable energy is climate change. The energy transition is defined as a long process. It involves replacing current systems that rely on fossil fuels with clean energy from renewable sources.

The results demonstrated a negative correlation between the consumption of recyclable materials and greenhouse gas emissions, thereby validating hypothesis 4 (H₄): The consumption of recyclable materials is negatively influenced by greenhouse gas emissions. The model was able to demonstrate the negative impact of greenhouse gas emissions on the consumption of recyclable materials, with an approximate decrease of 0.70 p.p. in consumption when these emissions increase by about to 10 p.p. in the long term. According to the results of subsequent studies [27,68,72,93], the transition to a circular economy in European countries is predicated on a considerable reduction in greenhouse gas emissions. This reduction will be accompanied by corresponding increases in the consumption of recyclable materials and the recycling rates in the long term.

The model was able to demonstrate that investment in the circular sectors has a positive and statistically significant impact on the consumption of recyclable materials in the short and long terms, suggesting a two-way causal relationship in the long term. These findings lead to the validation and positive evaluation of hypothesis 5 (H₅), which is formulated as follows: The consumption of recyclable materials will be positively affected by investments in circular economy sectors. Investment is a variable that explains the dynamics and evolution of the consumption of recyclable materials in the 27 European countries studied.

The evidence shows that a 10 p.p. increase in circular investments leads to an increase in the consumption of recyclable materials of around 0.35 p.p. This is also confirmed in the long run, where a 10 p.p. increase in circular investment increases the consumption of recyclable materials by about 2.50 p.p. Accordingly, it was confirmed that innovation and investment significantly attenuate environmental degradation; however, higher investments have no impact on resource efficiency.

Investment is a critical factor in the formulation of strategies for the implementation of the circular economy in various sectors and companies. As the authors of [94] emphasise, the government plays a pivotal role in fostering the growth of the circular economy and enhancing small and medium-sized enterprises’ access to financial resources. These enterprises, which are active in the recycling, repair, and innovation sectors, adhere to the principles of the circular economy.

European Union (EU) institutions and agencies are progressively enhancing their efforts to promote the circular economy agenda. A notable example is the European Green Deal Investment Plan (EIP), which presently provides support to sectors associated with sustainable energy provision, energy efficiency, sustainable cities, and sustainable agricultural practices, among others [95,96]. These entities possess the capacity to offer technical assistance, mobilise financial resources, and catalyse positive impact investments in circular economy systems.

Hypothesis 6 (H₆): The consumption of recyclable materials is positively influenced by the level of GDP per capita is confirmed by the results of the economic model proposed in this study. In this respect, the GDP per capita has an impact on the consumption of recyclable materials; if the GDP increases by 10 p.p., then the consumption of recyclable materials increases by approximately 1.60 p.p.in the short term. This effect is also confirmed in the long term: when the GDP per capita increases by 10 p.p., the consumption of recyclable materials changes in the same direction and increases by about 2.45 p.p.

More specifically, the results of the model show that there is a positive and statistically significant long-term impact between the consumption of recyclables and GDP per capita. For example, a 1 p.p. change in the consumption of recyclables changes the GDP per capita by about 0.11 p.p. This evidence is also found in other studies [97,98,99], where the economic growth and the environmental performance must be considered in tandem. The environment provides the key resources necessary to produce services and goods, as well as to process and absorb waste and pollution. It also underpins economic activity and growth.

The results achieved through the implementation of the proposed economic model confirm hypothesis 7 (H₇), which is formulated as follows: The consumption of recyclable materials is positively influenced by the population in the EU countries. As a result, during the period under study, most European countries experienced an increase in this indicator, leading to an intensification and accentuation of actions to protect the environment and increase social responsibility. This is also confirmed by the results of the model applied, where the impact of the consumption of recyclable materials is statistically significant at the 1% level and contributes to an increase of up to 9 p.p. of the current value in the short run.

Th estimated short-run coefficients showed that a 1 p.p. increase in the population is associated with an average 0.80 p.p. increase in the consumption of recyclable materials. The application of the Granger test revealed a bidirectional causal relationship between the consumption of recyclable materials and the population [93,94,100].

This should teach the population, irrespective of its numbers and growth rate, that recycling is not merely an obligation but rather a duty by addressing behaviours that recognise the consequences of non-recycling. In concrete terms, the rapid growth of the global population, which has reached a point where it is now being driven by industrialisation and intensifying consumerism, will generate a new paradigm. This new paradigm will see recycling play a key role in reducing the pressure on raw material extraction while also reducing energy consumption and greenhouse gas emissions. It is therefore evident that the trends today show that recycling is not just an option but rather a deciding factor for maintaining a sustainable environment [7,14,26,81].

The circular economy is a concept of significant importance, as it possesses the capacity to attract both businesses and policymakers to the domain of sustainability. However, it is imperative to recognise the necessity for scientific research in order to ascertain the actual environmental impacts of the circular economy and ensure that it contributes to the pursuit of sustainability.

The conclusion of this study indicates that the shift to a circular economy has the potential to positively impact economic growth and development in European countries. However, this transition is predicated on the development and implementation of concrete projects, actions, plans, and policies by these countries. These initiatives must be designed to encourage population engagement in recycling, the utilisation of recyclable materials, the reduction in resource waste, and the adoption of rational, fair, and responsible attitudes and behaviours.

The Member States of the European Union are engaged in the independent development of national circular economy indicators, which are aligned with the strategies implemented and the related policies and activities particular to each country. The EU must encourage market actors to adopt sustainable production and consumption patterns by reducing, reusing, recovering, refurbishing, and recycling resources at all stages of their value chains, and by raising the volume of trade in these materials. The transition to a circular economy can probably be made easier if EU Member States and their respective governments create a favourable climate for actors to get involved [101,102,103].

6. Conclusions

This study was conducted with the intention of investigating the impacts of different variables on the consumption of recyclable materials. This investigation focused on a period spanning almost a decade, during which time the study encompassed 27 EU Member States. The main empirical evidence reported stems from the model’s ability to capture and identify causal and dependent relationships between the selected variables aimed at measuring and quantifying the circular economy [4,5,6]. In this sense, the consumption of recyclable materials is directly and positively influenced by the level of GDP per capita, with an increasing trend of more than 1.50 p.p. Another aspect found in the analysis carried out in this study concerns the positive and direct relationship between the consumption of recyclable materials, renewable energy sources, and investments in the circular economy sectors.

From this point of view, it is found that the transition to a circular economy at the European level can be ensured and achieved since the formulation of strategic actions and policies at the national level by government authorities are aimed at generating positive and beneficial externalities soon [1,9,11]. For this reason, the adoption of various awareness-raising and empowerment campaigns is advocated to help the population understand the new concept of the circular economy [15,56].

It is also considered that educating the population in the spirit of the circular economy objectives of recycling, reducing, and reusing is essential for the sustainability of recyclable consumption [19,22,23]. At the same time, the consumption of recyclable materials is positively and statistically significantly correlated with the recycling rate. This evidence is supported by the propensity and motivation of European countries to not only increase their recycling rates consistently and sustainably, but also to significantly reduce the generation of industrial waste.

The main conclusion of this study is that Europe cannot achieve sustainable development without a circular economy. Achieving a sustainable development economy is more easily realised in the presence of coherent national policies for the more efficient use of European funds and for the creation of synergies with other public funds. Moreover, the adequate education of individuals in the spirit of the circular economy is imperative. The implementation of these policies can be achieved much more effectively at a national level.

At the microeconomic level, when making strategic decisions, companies should consider the bidirectional relationship between the consumption of recyclable materials and investments in specific sectors of the circular economy, suggesting a mutually potentiating relationship between the two. That is to say, the consumption of recyclable materials is likely to increase for firms that invest in the circular economy, which, in turn, will lead to innovation in production processes and investment in more efficient technologies. This, in turn, will create a virtuous circle, resulting in an increased demand for recycled materials. In the long term, companies that invest in circular economy technologies can gain a competitive advantage in the market, especially in the context of increasing EU sustainability regulations.

At the meso level, as investment in the circular economy sectors increases, so does the demand for specialists in waste management, recycling engineering, sustainable design, and circular supply chain management.

To adapt to the demands of the circular economy, companies will need to make significant investments, but these should not be seen as expenses that will affect their short-term profitability, but rather as investments that will generate sustainable growth and much higher profits in the long term. It is imperative for companies to acknowledge the significance of renewable energy in the production of circular products and resources. This encompasses the design, manufacturing, construction, and management of components within renewable assets. Additionally, the role of government in promoting the circular economy and enhancing financial access for small and medium-sized enterprises engaged in recycling, repair, and innovation that adhere to circular economy principles is emphasised.

The following policy measures are proposed to facilitate the transition to a circular economy and reduce environmental impacts, based on the results:

-

The introduction and extension of the carbon tax: a progressive tax on carbon emissions can act as a disincentive to the use of non-renewable resources and encourage investment in recyclable materials and renewable energy sources;

-

Economic mechanisms to support investments in green technologies: the creation of special funds and tax incentives for companies that invest in recycling, emission reduction technologies, and renewable energy sources;

-

The establishment of legally binding targets for the circular material utilisation rate (CMUR) at the national and sectoral level to ensure that businesses prioritise recycled materials over virgin resources;

-

The imposition of a levy on industries heavily dependent on virgin raw materials to incentivise the transition to recyclable materials;

-

The increased use of renewable energy should be facilitated through the development of EU policies for the accelerated integration of renewables into energy grids, including subsidies for green energy producers and investment in sustainable energy infrastructure;

-

Education and social programs to increase recycling rates should be launched, including awareness campaigns and environmental education programs for the general public, along with initiatives such as efficient separate waste collection systems and incentives for recycling;

-

The regulation of the trade in recyclable materials is also recommended, with a view to standardising and facilitating the market for recyclable materials in the EU, with a view to encouraging a steady and efficient flow of reusable resources.

The implementation of these policies by European countries has the potential to reduce their carbon footprints, stimulate investment in the circular economy, and improve long-term environmental sustainability.

The external validity of the findings varies according to the degree of development of the countries in question. In high-income regions (e.g., North America, Japan, Australia), similar circular economy policies could be successfully adopted because they already have strong regulatory frameworks, technological capacities, and investment capabilities.

In developing economies (e.g., parts of Africa, South Asia, Latin America), barriers such as limited infrastructure, weak enforcement, and lack of investment could hinder the effectiveness of circular economy policies.

As in any quantitative research, some limitations were encountered. Among these, data availability is a suggested limitation in this case. In most cases, data collection is an essential step to build and apply the economic model, so it is important to have data for all the variables analysed and for the whole period of investigation [64]. A further limitation that should be noted is the processing of the data and the utilisation of the analytical methods. In this instance, particular care was exercised in determining the most appropriate econometric methods and tools for the analysis of the panel data.

In practice, the primary constraint encountered during the execution of this study pertained to the selection of the variables, with the objective being to accentuate the role of recyclable resource consumption as a catalyst for the circular economy. Concurrently, while the analysis encompassed European Union countries, a limitation of the empirical investigation was the restriction to the period under scrutiny (i.e., 2013–2021) and to these 27 EU countries, with the possibility of analysing other categories of countries (developed, developing, underdeveloped, etc.) in the future. Consequently, it is recommended that in future research, the period under analysis be extended to encompass a more extensive timeframe, accompanied by a thorough examination of the results.

Given these limitations, the research suggests the following future directions for analysis. The present study proposes an extension of the economic model on the interdependencies of recyclable material consumption by including other explanatory variables (i.e., the municipal waste rate, employment in circular economy sectors) and for other time periods and countries analysed. Furthermore, the study puts forward the application of other econometric machine learning methods designed to study and predict the consumption of recyclable materials. Finally, it is acknowledged that the concept of a circular economy is a nascent field that lends itself to analysis and debate from multiple perspectives. One potentially fruitful avenue of research would be the development of a composite index to estimate and quantify the development of the circular economy in a comparative context.