Distances from Efficiency: A Territorial Assessment of the Performance of the Circular Economy in Italy

Abstract

This study investigates territorial disparities in transition toward circular economy within Italy, introducing an innovative methodological approach aimed at measuring regional efficiency and inequality. The research develops two complementary analytical tools: the Regional Circular Economy Index (ReCEI), a composite indicator designed for comparative evaluation of circular economy performance across regions, and the Regional Circular Economy Disparity Index (ReCED), inspired by the model of Sen which quantifies both the magnitude and spatial distribution of territorial inequalities. Applying this integrated framework to the 20 Italian regions reveals a pronounced heterogeneity: a select group of regions achieves or approaches efficiency benchmarks, whereas others exhibit persistent structural delays linked to infrastructural, institutional and innovative deficits. These findings, thus, confirm the persistence of a territorial dualism in the circular transition, only partially mitigated by instances of advanced governance and coordinated policies.

1. Introduction

Circular Economy (EC) has assumed a strategic role in global policy agendas, emerging as an integrated and innovative response to the environmental, economic and social challenges of our time [1,2]. In an era marked by increasing climate crises, instability in natural resource markets, and the need to strengthen territorial resilience and sustainability [3], the transition to circular economic models is no longer an option, but an ecological and socio-economic necessity. This transition represents a strategic drive to stimulate economic competitiveness, create quality employment and promote social cohesion at both national and regional levels [4]. However, its practical implementation is hindered by pronounced territorial heterogeneity—linked to structural, institutional, and cultural factors, including the availability and quality of infrastructure, the degree of technological innovation, capacities of multi-level governance, and the effectiveness of public–private partnerships [5,6,7,8]. In countries characterized by significant regional disparities, such as Italy, these conditions foster enduring fragmentation in both the extent of adoption and the efficiency of circular economy models [9].

From this perspective, the place-based approach, already affirmed in the debate on regional development [10,11], also appears crucial in the governance of circularity. Some scholars have highlighted that the adoption of green policies, if not carefully designed, may exacerbate existing disparities by disadvantaging regions with fewer resources, limited administrative capacities, or inadequate infrastructure [12,13]. Recent studies have analyzed these issues about circular economy practices implemented in specific territories. Some studies investigate the circular economy within the broader framework of business dynamics, assessing the direct implications for corporate sustainability performance [14], and exploring the governance mechanisms that affect firms’ strategic engagement with circular practices [15]. Increasing attention has been devoted to the territorial and institutional dimensions of circular economy. Ref. [16] employ a spatial analytical framework to assess the diffusion of circular practices across urban contexts, highlighting territorial specificities. Complementarily, Ref. [17] focus on cities and regions experiencing demographic decline, analyzing how local governance strategies enable territorial regeneration through circular models. At macro institutional scale, Ref. [18] provides a comparative assessment of regulatory and policy developments across European Union Member States, offering critical insights into the evolution of public policies supporting the circular economy. Adopting a data-driven perspective, Ref. [19] applies deep learning techniques to map the diffusion of circular entrepreneurship across Europe, emphasizing the phenomenon’s dynamic and innovative dimensions. Finally, Ref. [20], from an economic geography standpoint, examines the integration of circular practices within industrial networks in a national context, highlighting the importance of inter-firm collaboration and network configurations in fostering circular transformation.

Despite the variety and depth of these analyses, a notable gap persists in the availability of analytical tools that can rigorously and comparably measure circular performance at the regional level, as well as effectively support policymaking through territorially differentiated evidence. Furthermore, there is a lack of adequate tools to guide policy makers’ decisions, particularly concerning a fair and targeted distribution and allocation of financial resources.

This work aims to fill this gap, quantifying regional disparities in circular economy performance and providing analytical support to policymakers for the development of effective territorial programming strategy. This strategy should be designed to mitigate localized inequalities, foster the adoption of circular economy principles, and steer coordinated interventions that enhance environmental sustainability at the regional level. In doing so, we propose an innovative approach based on the combination of two newly developed tools: a Regional Circular Economy Index (ReCEI) has been built for the evaluation of the level of circular economy performance, in order to sup-port policy makers in carrying out cross-regional comparisons [21,22]; and a territorial inequality index, inspired by the Sen model [23], was adapted to capture both the extent of underperformance and its intensity and spatial distribution among the Italian regions [24,25]. The combined application of these tools enables the development of a dynamic mapping of territorial disparities in circular economy, offering a multidimensional and spatially detailed evaluation that directly integrates territorial inequalities into socio-economic planning and financial resources allocation.

The proposed approach is applied to the 20 Italian regions and provides a significant advancement to the ongoing debate on sustainable development, territorial cohesion, and public policy evaluation in the environmental field [24,26,27,28,29,30], by introducing a novel methodological framework tailored for circular economy planning and programming. Specifically, the main innovation of this study consists of its ability to rigorously quantify and compare the levels of circular economy performance and associated spatial disparities across regions. This analytical capacity allows for the generation of detailed and methodologically robust evidence, which is crucial for informing the design of more effective, context-specific, and territorially differentiated policy interventions. Thus, unlike traditional studies focused on static or descriptive assessments, this framework facilitates the rigorous, dynamic quantification of territorial disparities relative to predefined efficiency benchmark, enabling the precise prioritization of policy interventions. Furthermore, the novelty lies also in its methodological robustness and high replicability, integrating both aggregated and disaggregated indicators to deliver a multidimensional, spatially sensitive evaluation. Consequently, it provides policymakers with a sophisticated and comprehensive decision-support tool capable of design targeted, place-based strategies that enhance local specificities and strengthen the effectiveness of environmental governance. Furthermore, the integrated methodological framework enables the use of the territorial inequality index as a weighting vector for the allocation of financial resources, thereby promoting a more equitable and need-based distribution across local territories.

The study is structured as follows. Section 2 provides a literature review on territorial impact of circular economy. Section 3 describes the methodology applied and data collected. Section 4 presents empirical results, highlighting key patterns of inequality and identifying the highest- and lowest-performing regions. Section 5 presents robustness analyses. Finally, 6 discuss policy implications.

2. Literature Review

Territorial economic dynamics are crucial to understanding both the uneven progress of the circular economy transition and the broader success of the environmental ecological transition, which entails the adaptation and transformation of socio-economic systems at local and regional scales [31,32,33,34]. The ecological transition provides a comprehensive framework within which the circular economy serves as a pivotal mechanism for mitigating environmental impacts and fostering sustainable, resilient territorial development. This transformative process depends on the capacity of territories to integrate technological, infrastructural, and institutional innovations, while ensuring effective coordination across multiple levels of governance [5,28,31,32]. This framework is reinforced by objectives of the European Green Deal, which emphasize the need for territorially differentiated and integrated approaches to achieve decarbonization and sustainable transitions [4]. Consequently, a rigorous analysis of territorial dynamics is indispensable to identify disparities along ecological and circular transition pathways, and to design targeted, context-specific policy interventions. Focusing on the circular economy transition, empirical research by [35] highlights the role of multilevel governance and local actor engagement, at the regional level, for the effective implementation of circular economy strategies. Ref. [36], examining the Wielkopolska region in Poland, demonstrates that circular economy strategies must be adapted to regional specificities. Similarly, the comparative analysis carried out by [37] across three European contexts—the Alpine, transboundary, and urban regions—confirms that circular initiatives are more successful in areas with integrated ecosystems, adequate infrastructure, and well-developed administrative capacities.

Given the relevance of territorializing circular strategies, the transition towards circular economy models risks exacerbating both intra- and interregional disparities. Economically advanced regions endowed with substantial structural and institutional resources are better positioned to capitalize on the opportunities presented by the circular economy, whereas less developed regions face the risk of further marginalization [16]. In this regard, Ref. [18] identifies persistent gaps at the national scale, with the most industrialized countries of Western and Northern Europe leading the transition. Even within these most advanced regions, significant variability remains in the capacity of municipalities to integrate circular economy principles into local policies [38,39]. Furthermore, Ref. [17] notes that regions characterized by demographic decline, economic fragility, or institutional weaknesses risk exclusion from the benefits of the circular transition, unless targeted place-based policy interventions are implemented. Likewise, Ref. [20] observes similar dynamics, emphasizing that a robust manufacturing fabric and well-established intersectoral networks facilitate the effective adoption of circular economy practices. In the Italian context, clear disparities have emerged between more industrialized and less developed regions in adopting circular economy models, underscoring the need for localized policies and interventions tailored to specific socio-economic settings [9,40].

From a methodological perspective, most of these studies tend to provide static and descriptive analyses, which do not allow for the evaluation of deviations from an efficiency benchmark nor the capture of evolutionary dynamics over time. Considering the evidence presented, this study aims to advance literature by introducing an innovative indicator of circular inequality, drawing on the theoretical framework of Amartya Sen [23]. This indicator enables a comprehensive assessment by capturing both the extent of territorial disparities and their intensity and spatial distribution, thereby providing a nuanced understanding of inequality within the circular economy. This approach constitutes a novel contribution within the circular economy field, carrying crucial implications for the evaluation and design of policies. It facilitates a dynamic characterization of territorial disparities in the circular economy context, thereby supporting the development of more effective, inclusive, and context-sensitive place-based strategies.

3. Materials and Methods

This section outlines the methodology employed to measure the territorial differences in circular economy performance. Specifically, we propose an approach combining two measurement tools as illustrated in Figure 1.

3.1. Construction of Regional Circular Economy Index (ReCEI)

We develop a novel metric of circular economy performance called Regional Circular Economy Index, ReCEI. It is a composite index designed to evaluate each region’s performance in implementing circular economy processes and policies. In line with the objectives of assessing circular transition progress, some pillars of ReCEI incorporate directionality that favors regions progressively divesting from high-impact environmental practices. This approach assigns higher scores to regions demonstrating significant reductions in unsustainable behavior (such as lower per capita waste production and less use of landfills and incineration), while advancing management methods more consistent with circular economy principles (notably elevated separate waste collection rate). Consequently, the ReCEI captures not only the absolute levels of the indicators, but also the quality of the improvement processes activated by the regions.

The construction of ReCEI follows the methodology proposed by [41] which includes three main steps: standardization, weighting and aggregation.

Prior to the standardization process, sporadic missing data at the level of individual observation (i.e., per region-year) were handled through a multiple imputation procedure [41] based on regional averages and temporal trends of each indicator. This approach ensured the completeness and internal consistency of the dataset, reducing potential distortions in the subsequent weighting and aggregation phases. Sporadic missing values are therefore fully accounted in the subsequent PCA-base weighting stage, which is performed on the imputed dataset to ensure that all indicators contribute consistently to the extraction of components. For variables exhibiting complete absence across specific regions throughout the observation period, detailed data validation analysis was carried out, and it is reported in Appendix C. These cases are treated separately from the PCA weighting procedure, as the systematic unavailability of variables reflects structural or infrastructural constraints rather than random missingness. This analysis examines the root causes of data unavailability and provides contextual insights for interpreting these gaps. This dual approach ensures that both sporadic and systematic missingness are transparently addressed: the former through imputation prior to component extraction, and the latter through dedicated validation rather than forced estimation. As a result, the robustness and interpretability of the composite index are enhanced.

Standardization is essential to make indicators measured in different units comparable [42]; for this purpose, z-score standardization was employed [43]. Thus, the original values are normalized according to the following formula:

$$z_i = \frac{x_i - \overline{x}}{s}$$

(1)

With $\overline{x}$ is the mean value of the observations considered, while $s$ represents the standard deviation. Based on Equation (1), the z-score $z_i$ measures the distance—expressed as the number of standard deviations—between the mean value and the regional rough score $x_i$.

The resulting standardized matrix is used as input for the weighting procedure, which aims to minimize bias and subjectivity in the definition of weights. Hence, the Principal Component Analysis (PCA) with varimax rotation [44] was applied, by using IBM SPSS Statistics software v21. This technique groups indicators into factors based on collinearity, maximizing shared variance among variables, and assigns weights proportional to the variance explained by each factor. Components were selected by following Kaiser’s criteria. According to this rule, the latent factors are identified according to two conditions. Firstly, only components with eigenvalues greater than 1 were retained, as those below this threshold explain less variance than a single standardized variable. Second, only solutions explaining at least 60% of cumulative variance were considered acceptable, thereby preserving substantial portion of the original dataset’s information. These characteristics suggest that additional components would contribute negligible incremental variance. Inspection of the eigenvalue distribution, the scree plot, and the observed communalities (reported in Appendix A for 2021 as a representative example) confirms that the retained components capture the dominant structure of the data. This empirical pattern aligns with the literature, which endorses the Kaiser criterion as appropriate for datasets of comparable dimensionality and complexity [45,46], ensuring an optimal balance dimensionality reduction and information preservation. In addition, the 60% cumulative variance threshold further guarantees that the retained components preserve a substantial portion of the original information. These criteria, aligned with [41] guidelines, ensure an appropriate balance between dimensionality reduction and information completeness.

Additionally, the varimax rotation was applied to compute the loadings of each variable, and to enhance factor interpretability by simplifying the association of indicators with the extracted components. This orthogonal rotation is justified by the assumption that are independence among latent dimensions and allows a clear distribution of indicators to components, facilitating the construction of a transparent and interpretable composite index.

The aggregation procedure combines the weighted scores obtained from the principal components into a single composite index value for each observation, according to the following formula:

$$S_j = ({\sum }_{i = 1}^n{x}_i \ast {\alpha }_i) \ast {\lambda }_{\alpha } + ({\sum }_{i = 1}^n{x}_i \ast {\beta }_i) \ast {\lambda }_{\beta } + \cdots + ({\sum }_{i = 1}^n{x}_i \ast {\omega }_i) \ast {\lambda }_{\omega }$$

(2)

Equation (2) is applied for computing each pillar composite index. Hence, $S_j$ represents the pillar values for each j-th region; $x_i$ stands for the 26 simple indicators; α, β and ω for the coordinates of each i-th variable on the components extracted through PCA; λ is the eigenvalue associated with each considered component. RCEI is the sum of $S_j$ values.

Additionally, to make the results comparable and interpretable, this aggregate score is normalized using Min-Max scaling within a 0 to 100 range to ensure comparability and ease of interpretability. This methodology yields a composite index that is transparent, statistically robust and aligned with best practices, thereby minimizing subjectivity in weighting process.

To define the weakness and straight of each observation, we replicated the prior procedure for each pillar of the circular economy, obtaining seven sub-indices that provide detailed insights into different dimensions. The methodology remains the same, except that in the final step we consider only the factors strongly associated with each pillar for the respective sub-index. After assigning weights to the factor scores, we retain only those scores with a pronounced relevance for individual dimension—namely Waste Production, Decoupling, Waste Management, Secondary Raw Materials, Competitiveness and Innovation, Regional Sustainability, and Regional Resilience. This approach aims to provide policymakers with a comprehensive and practical tool to better calibrate policy interventions across regions [41,42,43,44,45,46,47].

3.2. Measuring Regional Circular Economy Disparity (ReCED)

Once territorial circular economy performance level has been calculated, we have defined the thresholds allowing us to identify the regions not achieving acceptable levels. So, the scores of the ReCEI and its sub-indices were divided into quartiles to facilitate classification and comparative analysis.

Two scenarios were constructed for both the composite index and each sub-index:

• Scenario 1—Efficient Performance Threshold. It defines an “efficient value”—z₁ as the mean value of ReCEI scores (and of each pillar scores) within the interquartile range between the first and third quartile (Q1–Q3). This threshold captures a realistic benchmark for acceptable circular economy performance while mitigating distortions caused by extreme outliers. This approach aligns with methodologies employed in socioeconomic studies on synthetic index construction, which preferentially utilize robust statistics—such as the interquartile range (IQR) to mitigate the distorting effects of outliers [48,49]. In line with established practices in composite indicators design, robust central tendency benchmarks are widely adopted to ensure stability under distributional irregularities. The OECD methodological framework [41] highlights the importance of IQR-based thresholds to avoid distortions linked to high-leverage observations, and similar principles are applied in multidimensional poverty and vulnerability measures [50,51]. These contributions provide a theoretical framework for interpreting Scenario 1 as a justified threshold rather than an ad hoc statistical choice.
• Scenario 2—Ideal Performance Threshold. It defines an “ideal value”—z₂ as the maximum ReCEI value (and the maximum value of each pillar scores) observed within the fourth quartile (Q4). This threshold represents the ideal performance level and highlights the largest disparity in regional performance. The use of an upper-bound or frontier-type benchmark is consistent with the literature on performance gaps, polarization, and frontier evaluation. Studies on relative deprivation and polarization [52,53] employ extreme but empirically observed values to characterize the full span of disparity. This ensures that Scenario 2 captures the upper boundary of feasible performance without relying on hypothetical extreme.

These thresholds are theoretically grounded in Sen’s framework [23], which employs a poverty line to distinguish underperformers from adequately performing units. Scenario 1 establishes a pragmatic efficiency target, whereas Scenario 2 delineates the upper-bound idealistic goal. Within Sen’s conceptual architecture [23,54], evaluative spaces are often characterized using alternative cut-offs to explore how measurement properties vary when benchmarks shift from adequacy to aspiration. The dual threshold structure adopted here reflects this logic: a robust adequacy benchmark (Scenario 1) and a frontier-oriented aspirational benchmark (Scenario 2) jointly span the meaningful evaluative interval. Because the behavior of the index is monotonic across thresholds within the range z₁, −z₂ range, intermediate scenario is implicitly covered, supporting the sufficiency of these two theoretically grounded cases for the robustness assessment.

After identifying the observations requiring improvement in circular economy performance, territorial imbalances were quantified by calculating the distance between each observation’ current ReCEI (and its pillars) values from the efficient solution (Scenario 1) and the ideal one (Scenario 2).

These estimates were derived by applying different components of disparity, following the axiomatic formulation of poverty measurement introduced by Sen [23], which builds upon a refined version of the Headcount and Income Gap ratio.

The first component is the Circular Disparity Spread ($\mathit{ReCDS}$) computing the proportion of regions falling below the threshold. it is calculated as:

$$\mathit{ReCDS} = \frac{q}{N}$$

(3)

where $q$ denotes the number of regions with ReCEI (or one of pillars) score y_i below the threshold value—y_i < z₁ (or z₂) and $N$ is the total number of regions.

The second component is Circular Disparity Intensity ($\mathit{ReCDI}$), measuring the intensity of the disparity as the average relative distance of underperforming regions from the relevant thresholds:

$$\mathit{ReCDI} = \frac{1}{q}{\sum }_{i = i}^q\frac{z - y_i}{z}$$

(4)

This indicator can be interpreted as the quantum needed by each region to overcome or reach, at least, the threshold.

The third component is Circular Underperformance Inequality ($\mathit{ReCUI}$), measuring the degree of inequality among underperforming regions:

$$\mathit{ReCUI} = \frac{1}{2q^{2}m}{\sum }_{i = 1}^q{\sum }_{j = 1}^q|y_i - y_j|$$

(5)

where $m$ represents the maximum possible difference in ReCEI scores across all regions. This component captures the dispersion of deficits among regions below the threshold.

Combining the above components (Equations (3)–(5)), the Regional Circular Economy Disparity Index ($\mathit{ReCED}$) is defined as:

$$\mathit{ReCED} = \mathit{ReCDS}[I + (1 - \mathit{ReCDI}){\mathit{ReCUI}}_z]$$

(6)

This formulation allows the ReCED to capture simultaneously the extent, intensity, and inequality of circular economy underperformance across regions.

The values obtained must be interpreted as the relative weight that each region, among those present in the sample but below the threshold value, could assume as an objective to be maximized in an economic-financial planning process.

The methodology illustrated has been applied to the 20 Italian regions, corresponding to NUTS level 2.

3.3. Data

In this sub-section, we explain the variables used to realize the ReCEI. For its formulation we have followed the European Monitoring Framework [4], developed by the European Commission to assess progress in the transition towards the circular economy. Thus, ReCEI comprises a total of 26 indicators grouped into five dimensions:

• Production and consumption dimension measures the reduction in material and waste consumption, decoupling economic growth from resource use, and green public procurement [3,55,56].
• Waste management evaluates the effectiveness of minimizing waste and maximizing its recycling and reuse, in line with targets for reducing landfills and increasing separate collection [4,57].
• Secondary raw materials dimension reflects the regional capacity to re-use waste materials in new production processes, reducing dependence on virgin materials [1,58].
• Competitiveness and innovation dimension examines regional efforts to make the economy more competitive and innovative through circular practices, new models and technologies [8,59].
• Regional sustainability and resilience dimension evaluate resource efficiency to promote a self-sustaining local economy that is resilient to climate change and crises [2,60].

This multidimensional structure, as illustrated in Figure 2, allows a comprehensive analysis of regional progress towards circularity, by integrating key environmental, economic and social challenges [1,4,57,59,60]. All variables are collected for each Italian region, for the period 2015–2021, by using the following official databases: indicators for the Sustainable Development Goals (SDGs) developed by ISTAT; the waste register of ISPRA (Higher Institute for Environmental Protection and Research); Eurostat’s ARDECO database.

The choice of the period 2015–2021 for the analysis of variables is in line with international scientific literature, which adopts similar time frames to ensure the availability of harmonized data, methodological consistency and relevance concerning the main circular economy and sustainability policies introduced from 2015 onwards [60,61].

Table 1. Descriptive analysis of all variables used in ReCEI and its pillars.

Variable	U.M.	Obs	Mean	Std. Dev.	Min	Max
Hazard Waste Generation	tonnes per capita	140	0.15	0.07	0.06	0.34
Waste generation (Urban waste + Special waste)	tonnes per capita	140	2.95	0.92	1.41	5.48
Domestic material consumption per capita	tonnes per capita	140	8.98	3.29	3.70	16.90
Waste generation per unit of Value Added	tonnes per thousand euros	140	122.68	173.72	7.90	729.22
Separate collection of Municipal Waste	kg per capita	160	284.42	84.37	59.74	468.93
Urban waste treated in composting plants	kg per capita	140	63.99	41.27	0.00	214.74
Incinerated Urban Waste	kg per capita	140	70.44	76.54	0.00	299.73
Landfilled Industrial Special Waste	kg per capita	140	225.78	202.79	0.00	862.04
Landfilled urban Waste	kg per capita	140	238.49	179.91	0.00	654.91
Incinerated Industrial Special Waste	kg per capita	140	16.34	21.76	0.00	83.17
Urban waste treated in aerobic and anaerobic plants	kg per capita	140	42.18	66.46	0.00	287.77
Special Waste Recovery	tonnes per 1000 inhabitants	140	80.20	19.18	41.00	149.79
Waste reused as a source of energy	tonnes per 1000 inhabitants	140	1.62	1.68	0.00	6.92
Firms with Energy management system Certification	Number per 1000 inhabitants	140	0.03	0.02	0.00	0.13
Firms with Environmental management system Certification	Number per 1000 inhabitants	140	0.39	0.17	0.14	0.94
Organization/entrerprises with EMAS registration	Number per 1000 inhabitants	140	0.02	0.02	0.00	0.09
Greeb Purchases or Green Public Procurement	%	140	37.62	18.20	10.70	69.90
GHG emission—Industry sector	kt CO2 eq per capita	140	1.55	0.85	0.38	3.65
GHG emission—Transport sector	kt CO2 eq per capita	140	2.11	0.92	0.78	5.05
GHG emission—Agriculture sector	kt CO2 eq per capita	140	0.51	0.35	0.06	1.92
GHG emission—Waste sector	kt CO2 eq per capita	140	0.25	0.11	0.06	0.49
Air quality—PM2.5	%	140	78.96	20.04	6.10	100.00
Waste produced from tourism sector	kg per capita	140	11.20	12.32	1.13	59.60
Electricity from Renewable sources	%	140	57.12	61.44	7.30	323.10
Renewable energy share	%	140	24.44	18.93	0.00	106.30
Energy Efficiency Certificates (TEE)	Number per 1000 inhabitants	140	613.77	667.64	12.59	3429.17

contains statistics of all variables used, whereas more details on data validation are illustrated in Appendix B.

4. Results and Discussion

4.1. Regional Circular Economy Performance

Following the procedure outlined in Section 3.1, the Regional Circular Economy Index (ReCEI) and its pillars (subindexes) were constructed for all Italian regions for each year of the period 2015–2021. To enhance transparency and reproducibility, the factor loading matrix and scree plot for 2021 are reported in Appendix A as an illustrative example. Similar patterns were observed for all other years (2015–2020), confirming the stability of the weighting procedure over time.

Figure 3 presents the average values for ReCEI and each pillar, while annual values are illustrated in the Supplementary Materials.

The results of ReCEI reveal a significant territorial heterogeneity in circular economy performance across Italian regions from 2015 to 2020, that diverges partly from the traditional Italian dualism characterized by a more developed North and a less developed South. Campania (98.11) and Lazio (94.00) stand out as top performers, driven by strong outcomes in the pillars of Waste Production, Waste Management, and Secondary Raw Materials. These areas reflect efficient waste handling, effective recovery and transformation of secondary materials, and a notable reduction in waste generation—results aligned with focused regional policies and dynamic public–private collaborations [62,63,64].

Campania’s leading position in the composite indicator may initially seem paradoxical, given the well-known regional challenges in waste management over time [65,66]. Nevertheless, the index methodology explicitly prioritizes the reduction in waste generation as well as the discontinuation of the most environmentally damaging disposal methods. Viewed through this lens, Campania’s performance trajectory over the past decade aligns coherently with the resultant scores.

Quantitatively, the region exhibits one of the lowest levels of per capita waste production rates (1.54 t/inhabitant) and provides clear evidence of a decoupling between economic growth and waste generation. Moreover, waste management dynamics have also undergone substantial reconfiguration: the region’s limited reliance on landfills, combined with a steady increase in separate waste collection—exceeding 340 kg per capita in 2021—has significantly strengthened waste management performance. These trends represent a structural break compared to the patterns that characterized previous phases of crisis [66,67,68,69]. During 2015–2016, the average volume of municipal waste disposed of in landfills amounted to 448 kg per capita, whereas after 2016 this value declined to 8 kg per capita, indicating a profound improvement in waste management practices.

These achievements are partially attributable to targeted regulatory interventions. Campania Regional Law No. 14/2016, updated by Regional Law n. 2/2018, formalized the region’s commitment to advancing a circular economy model focused on the integrated reduction, reuse and recycling of materials. This legislation introduced measures to support research and innovation, and to promote the development of facilities for energy recovery from waste: the special waste recycling rate increased on average by 4% after 2016, with peak values recorded in 2018 (+7.15%) and 2021 (+6.48%). Though these comprehensive initiatives, Campania has developed a virtuous model of circularity at the regional scale, positioning itself among leading Italian territories in this domain [9]. This model also reflects the region’s economic structure, which is characterized by a predominance of service sectors and light industries, as highlighted by recent studies [62,69]. Far from suggesting the absence of residual challenges, this outcome illustrates how targeted policy interventions and inherent structural characteristics can synergistically reverse historically adverse trajectories.

The Lazio Region has systematically incorporated the ecological transition within its strategic planning framework, with particular emphasis on the Regional Waste Management Plan (PRGR), which serves as a comprehensive and integrated reference framework. It is designed to promote waste prevention, enhance recycling efforts, and maximize recovery of secondary raw materials through the deployment of advanced technologies and innovative processes within environmental management systems. Such policies reinforce the pivotal role played by technologically advanced firms and cooperatives in disseminating circular economy models, as illustrated in several recent scientific studies [61,70,71,72,73].

Conversely, regions such as Sardinia, Trentino-Alto Adige, Veneto, and Piedmont hold intermediate positions, consistently performing in Waste Management and Material Recovery. By contrast, Lombardy and Emilia-Romagna—despite their roles as economic powerhouses—fall below the national average. These results can be explained by their regional policies, which, although updated and ambitious, face structural challenges. Lombardy’s Regional Waste Management Program (PRGR) adopts an integrated strategy that emphasizes waste prevention, recycling, and landfill reduction. However, the region’s substantial reliance on incineration and the complexity of its treatment infrastructure limit full compliance with circular economy principles. Similarly, Emilia-Romagna’s Regional Waste Management and Remediation Plan (PRRB) sets rigorous targets, including 80% separate collection and 70% recycling rates, yet its efforts contend with the need to modernize existing plants and processes, which remain anchored to traditional disposal methods [74].

Molise and Valle d’Aosta face deeper structural constraints due to limited infrastructure and dependence on land, reflected in their low scores for Waste Management and Regional Resilience. These findings highlight ongoing structural challenges in low-density regions with limited plant infrastructure [75,76], as well as a marked delay in political commitment to the circular economy transition. In 2023, the Molise region introduced a strategic regional development law aimed at orienting the production system toward regenerative and sustainable models inspired by natural cycles. Similarly, in 2022, the Valle d’Aosta region updates its regulatory framework through Regional Law No. 4, which revises the Regional Waste Management Plan for the period 2022–2026. Both legislative measures, aligned with European waste regulations and circular economy principles, aim to foster waste reduction, recovery, and reintegration into production cycles, thereby conserving primary resources and mitigating environmental impacts. The temporal lag in policy development and implementation of both regions, coupled with their limited infrastructural capacity, thus accounts for the comparatively lower performance of these territories relative to more advanced regions.

Despite this territorial diversity, only the Competitiveness and Innovation pillar reveals a clear North–South divide: Emilia-Romagna, Lombardy, and Lazio excel with advanced industrial ecosystems, a high concentration of innovative firms, and a strong ability to attract R&D investment, whereas southern regions lag significantly—with Campania as a notable exception [5,11,77]. In addition, within the Regional Sustainability pillar, southern regions such as Campania and Sicily achieve high scores driven by effective environmental policies targeting emission reductions. This indicates the potential for a more rapid and effective sustainable transition compared to certain industrialized regions burdened by greater environmental pressures [78,79].

The Regional Resilience pillar highlights Liguria and Sicily for their effective efforts in energy efficiency and diversification. Liguria, despite low renewable electricity production (7–9%), demonstrates strong resilience through efficient consumption management and widespread use of Energy Efficiency Certificates (TEEs). This performance is largely attributable to the adoption of regulatory measures such as Regional Law no. 22 of 29 May 2007, which regulates the energy planning and interventions of the Liguria Region. The law promotes improvements in energy efficiency, energy saving, and the development of renewable sources compatible with territorial specificities, including support for initiatives aimed at enhancing local resources through mechanisms such as TEEs. This regulatory framework exemplifies the adaptive strategies of regions overcoming territorial and infrastructural limitations [80]. Sicily shows steady growth in renewable energy generation (from 23% to 28%) alongside strong adoption of efficiency measures, with resilience reinforced by energy diversification, a key factor in reducing vulnerability to shocks [9]. This progress stems from actions supported by Cohesion Funds (EFRD 2014–2020), which provided financial incentives for the construction and modernization of plants, thereby encouraging the diffusion of clean energy sources such as solar and wind. These policies, together with the updating of the Regional Environmental and Energy Plan, promote a diversified and resilient energy model aligned with the principles of sustainable transition. Conversely, Molise and Valle d’Aosta exhibit low resilience due to reliance on less diversified, traditional energy systems [2,80], as well as a comprehensive delay in the adoption of environmental policies.

Overall, the RCEI and its pillars demonstrate that regional circularity depends not only on economic strength but also on the adequacy of infrastructure and innovation capacity. The emerging picture is nuanced: some southern regions—Campania and Sicily— achieve notable success, likely due to targeted waste management interventions and a growing focus on closing material supply chains. In contrast, industrially advanced regions such as Lombardy and Emilia-Romagna, despite larger production volumes, face difficulties fully embracing circular economy principles, highlighting the need for stronger policies to promote innovation and waste reduction.

4.2. Regional Circular Economy Disparity

The second phase of the methodology employed in this study focuses on quantifying the extent of regional disparities in circular economy performance. Specifically, the deviation of each Regional Circular Economy Index (ReCEI) value—and its underlying pillars—is assessed relative to carefully defined benchmark thresholds. As detailed in Section 3.2, two distinct analytical scenarios were developed:

• Scenario 1 (efficiency threshold, z1): establishes a pragmatic benchmark reflecting a satisfactory level of performance, calculated as an “interquartile” average between the first and third quartiles of the distribution.
• Scenario 2 (ideal threshold, z2): sets an aspirational benchmark representing best practice, corresponding to values within the fourth quartile.

The distances from these benchmarks were quantified using the Regional Circular Economy Disparity Index (ReCED), computed for each region, by applying Equation (6) (see Section 3.2). Lower ReCED values indicate closer proximity to the efficiency threshold, reflecting better performance, while higher values reveal inefficiencies or structural lags in the circular transition. Results for Scenario 1 are illustrated in Figure 4a–h, whereas those for Scenario 2 are provided in Supplementary Materials. The graphs illustrate, with gray bars, the regional distribution of the Regional Circular Economy Index (ReCEI) and of its pillars, as well as the distance (shown with blue bar) from the reference threshold value (indicated by the red line), thereby highlighting the degree of territorial heterogeneity captured by the ReCED Index.

The interregional comparison in Scenario 1 highlights a marked polarization between a limited group of high-performing regions and a broader set of territories that remain distant from the efficiency threshold, thereby confirming the presence of structural territorial dualism in the circular economy transition. These spatial disparities are aligned with existing empirical research on the regional green transition gaps and underscore the critical role of institutional quality in promoting resilience and sustainable growth [81,82].

Specifically, among the twenty regions analyzed, only five—Campania, Lazio, Tuscany, Apulia and Sicily—successfully achieve the efficiency thresholds for the composite index ReCEI (Figure 4a) and for at least six of the seven sub-indexes (Figure 4b–h). This indicates a level of circular maturity and institutional resilience that is crucial for effectively advancing sustainable transition processes [82].

Particularly, Lazio and Campania demonstrate full efficiency in all dimensions, supported by well-established innovative ecosystems and consistent environmental policy frameworks. For example, the Lazio region has embedded ecological transition within key planning instruments—including the Regional Waste Management Plan (PRGR)—complemented by targeted support policies for SMEs focused on eco-innovation and digitalization. A notable initiative is the 2018 public funding call aimed at financing innovative projects in energy and environmental sectors. Such interventions corroborate empirical findings on the pivotal role of technologically advanced firms and cooperatives in driving the diffusion of circular economy models [70,71,72,73,83]. In Campania, the implementation of the new regional waste plan (Regional Law 29/2018), which sets targets of 65% separate waste collection and 70% recovery, has steered the region toward innovative waste management strategies that avoid new landfill development and encourage investments in aerobic treatment facilities. Concurrently, awareness-raising campaigns have contributed to the diffusion of a circular economy culture, positioning Campania as a green innovation hub for the Southern Italy [84,85].

By contrast, Tuscany, Apulia and Sicily are positioned close to the efficiency threshold, exhibiting only minor residual inefficiencies. This reflects the ability of medium-scale regions to offset lower industrial density through strong territorial cohesion and multilevel governance coordination. Thus, Tuscany region has formally recognized the circular economy as a guide principle through the Regional Statute (Regional Law 9/2018) and Regional Law 48/2018, effectively translating these guidelines into integrated strategies that encompass innovation, sustainability and territorial planning. Notably, supply chain agreements and circular industrial districts play a crucial role in this framework by strengthening advanced waste management and material recycling practices [79,86]. Both Apulia and Sicily have promoted a transition to a circular economy by a regional law proposal, developed through an inclusive process with local actors [87,88].

Jointly, these five regions exhibit a sophisticated profile of institutional and territorial capacity, reinforced by proactive policies, mature innovative ecosystems and well-established multilevel governance frameworks. This aligns with the broader literature demonstrating that regions characterized by resilient institutions and sustained long-term policy commitments consistently achieve higher sustainable efficiency levels [2,58,81,83,89,90].

The lagging regions exhibit significantly higher ReCED values, indicating pronounced structural delays and systemic misalignments relative to the efficiency threshold. Regions such as Molise, Umbria, Valle d’Aosta, Friuli-Venezia Giulia and Basilicata, face infrastructural, institutional, and productive vulnerabilities that constrain their capacity to deploy effective circular strategies [91].

Molise exhibits systemic inefficiencies driven by low rates of separate waste collection, insufficient infrastructure, and limited adoption of innovative technologies. This is a dynamic typical of territories with low administrative capacity [2,83,85,92]. Similarly, Basilicata shows moderately high ReCED across several dimensions. The most pronounced inefficiencies are evident in waste production, secondary raw materials use, and in the decoupling of economic growth from resource consumption, indicating a persistent dependence on material-intensive production processes. This is consistent with findings by [81] regarding barriers to eco-innovation. Thus, the lack of a well-established innovation ecosystem constrains Basilicata’s capacity to progress toward more advanced circular economy models [75].

ReCED values for Umbria indicate widespread inefficiencies in all domains, particularly in secondary raw material use and waste production. However, its regional sustainability and resilience indicator suggests a latent capacity for adaptation and response, aligning with the smoother transition patterns identified by [93]. Conversely, Valle d’Aosta shows substantial imbalances in regional sustainability decoupling, reflecting vulnerabilities typical of small-scale economies affected by administrative fragmentation [91]. Finally, Friuli-Venezia Giulia presents a heterogeneous yet broadly critical profile, with high deficiencies in waste generation, alongside moderate inefficiencies regional sustainability. Despite its advanced industrial development, the region’s overall efficiency is hindered by the lack of comprehensive circular governance strategies [91,94].

The comparison between high- and low-performing regions confirms the presence of a territorial gradient of sustainable efficiency, aligning with international literature on regional disparities in the green transition [81]. Campania, Lazio, Tuscany, Apulia and Sicily distinguish themselves through a synergistic combination of institutional maturity, stable environmental policies, and well-established innovative ecosystems. Conversely, Molise and Umbria, Aosta Valley, Basilicata and Friuli-Venezia Giulia exhibit multidimensional vulnerabilities encompassing both material and energetic domains as well as institutional and innovative capacities.

The most notable disparities are evident in waste production, decoupling, and secondary raw materials indicators, which measure regional capacities to decouple economic growth from resource consumption, as well as to promote production models based on reuse and recycling [81,92]. Overall, these fundings underline the need for place-based policies that enhance administrative capacity and integrate sustainable innovation within regional development strategies, thereby promoting territorial convergence toward the efficiency threshold.

Scenario 2 reveals a marked deterioration in performance, since only a limited number of regions exhibit low disparity index (ReCED) values that demonstrate significant proximity to the ideal threshold for the ReCEI and its pillars. Notably, Campania emerges as the most efficient region, achieving the ideal threshold, followed by Lazio and Sicily, which display relatively low overall disparity index, indicative of performance closely approaching, yet not fully attaining, to the ideal objective.

The territorial analysis corroborates the presence of marked regional disparities, consistent with findings from Scenario 1. Northern regions generally exhibit higher levels of disparities in the pillars associated with waste generation and waste management, reflecting a persistent dependence on incineration and less circular waste generation practices [74]. In the Central regions, Umbria exhibits widespread critical vulnerabilities, especially in second raw materials use and in waste production, whereas Tuscany and Lazio show lower disparity values, reflecting more solid performances close to the ideal threshold. The Southern regions present a heterogeneous landscape: Campania maintains low disparity levels across all assessed pillars, highlighting well-established circular economy practices; on conversely, Molise and Basilicata display high disparity values across multiple dimensions, reporting significant structural challenges.

These results underscore the need to design territorially differentiated policy interventions, aimed at overcoming the structural barriers adversely affecting regions with higher disparities. Such targeted approaches are essential to facilitate the broad adoption of circular economy models that are both highly efficient and sustainable.

5. Robustness Analyses

5.1. Conceptual Validation of ReCEI

The assessment of the conceptual validation of the ReCEI is based on theoretical considerations relating to the validity and consistency of composite indicators. The goal is twofold: to verify whether the index designed to measure the circular sustainability of Italian regions aligns with the ASviS SDG12 synthetic indicator, and to explore the effects of adding extra pillars in the full version of the index.

To carry out this assessment, a comparable version of the index, ReCEI_comp, was developed based on four pillars closely associated with the circular economy: “Waste Production”, “Decoupling”, “Waste Management”, and “Secondary Raw Material”. For each pillar, a corresponding variable was selected from the Asvis SDG12 synthetic indicator (see

Table 2. List of variables of SDG 12 used in ReCEI_comp.

Variables	SDG12	ReCEI_comp	ReCEI Pillars
Domestic material consumption per capita	✓	✓	Decoupling
Urban wasteproduction	✓	✓	Waste production
Separate collection of municipal waste	✓	✓	Waste management
Circular material rate	✓	✓	Secondary Raw Material

), ensuring maximal comparability with official metrics. This methodological approach enables the isolation of the index’s core components, thereby providing a robust foundation for rigorous comparisons and minimizing the risk of distortion caused by additional variables absent from the official indicators. Consequently, the construction of ReCEI_comp serves as a clear methodological benchmark for verifying the internal validity of the indicator.

As illustrated by the two maps in Figure 5a,b, the rankings of the two indicators exhibit a high degree of similarity, with only minor exceptions. Notably, Lazio and Campania consistently emerge as leading regions in both classifications. In contrast, Emilia-Romagna, Sardinia and Basilicata show discrepancies, positioning themselves differently in the two rankings.

Three types of correlation were employed to evaluate the consistency between the indicators, capturing both linear relationships and concordance in the regional ranks. Pearson correlations assess the linear relationships between the indicator values, while Spearman and Kendall correlations analyze the consistency of regional rankings, with Kendall being more appropriate for small samples and tie data. This approach ensures a rigorous evaluation of internal consistency and alignment with the official SDG12 indicators.

The correlations analysis between ReCEI_comp and SDG12 reveals a high level of consistency in regional rankings. The results shown in

Table 3. Correlation tests between ReCEI_comp and Sdg12 Asvis.

ReCEI_comp—SDG12	Pearson (r)	Spearman ρ	Kendall τ-b
Coeffiecient	0.7529	0.8932	0.7263
p-value	0.0001	0.0000	0.0000
N	20	20	20

indicate that ReCEI_comp effectively captures the key dimensions of sustainable consumption and production, thereby validating the choice of pillars and confirming the comparable index’s capacity to replicate the conceptual framework of the official indicators. Consequently, the comparable index (ReCEI_comp) is both robust and coherent, providing a reliable basis for interregional comparisons.

Overall, the robustness analysis confirms that the ReCEI is methodologically sound and maintains a high degree of consistency with the official SDG12 indicators. Simultaneously, the inclusion of additional pillars and variables in the full version enhances the indicator, providing a broader and more nuanced understanding of regional performances in circular economy and sustainability. From a policy perspective, the robustness of the index ensures that regional evaluations are both reliable and interpretable, providing decision-makers with a rigorous and adaptable tool to guide policies aimed at advancing the circular economy and sustainable development through targeted, evidence-based interventions.

5.2. Robustness Analysis of ReCEI

To verify the robustness of the ReCEI with respect to economic and institutional contextual factors, a panel regression with fixed effects was estimated, encompassing the 20 Italian regions for the period 2015–2021. The primary objective was to evaluate whether the composite circular economy indicator sustains both conceptual validity and empirical stability when accounting for structural determinants such as economic development, territorial density and quality of governance. Specifically, three models were estimated (see

Table 4. Fixed-effect regression (Dipendent variable—ReCEI).

Dipendent Variable: ReCEI	M1	M2	M3
GDP pro capite	−0.7039 (0.8642)		−0.9447 (0.7671)
Industrial density	−6.4438 *** (1.6679)		−9.3755 *** (2.8296)
EQI		0.1842 (0.1123)	0.0694 (0.1230)
RIS		−2.3458 ** (1.0591)	−0.0768 (0.9561)
Public investment in R&D		−0.2817 (0.2716)	0.2014 (0.2565)
Private investment in R&D		−0.2074 ** (0.0955)	0.3836 * (0.1891)
Industrial GVA			0.0870 (0.5337)
Population density			−7.7414 * (4.3550)
_cons	11.8615 (8.8439)	1.8885 * (0.9548)	53.3333 ** (25.1026)
N	133	133	133
pseudo R2
Log Likelihood	−15.79	−27.82	−11.87
Chi squared

Time variables not reported. Robust standard error sin parentheses. *** p < 0.01; ** p < 0.05; * p < 0.1. Fixed Effects (FE) Model with Multiple Fixed Effects; Dependent variable: ReCEI.

). Model 1 (M1) incorporates GDP per capita (lgdp_pc) and industrial density (ldens_ind) as explanatory variables. Model 2 (M2) extends the specification by including institutional quality (EQI), the Regional Innovation Scoreboard (RIS) and measures of public and private investment in R&D (lpubinvrd and lprivinvrd), thereby examining the interplay between innovative performance, institutional quality and circularity outcomes. Finally, Model 3 (M3) further expands the framework by integrating additional structural variables such as population density (ldens_pop) and industrial value added (lgvaind), thereby capturing broader territorial and economic dynamics that might influence regional circular economy performance.

The results demonstrated that the ReCEI indicator exhibits a consist and coherent response to the key explanatory variables, thereby affirming its conceptual validity. Industrial density emerges as a statistically significant negative determinant of circular performance. This finding suggests that regions characterized by higher industrial concentration experience increased pressures on material flows, aligning with the literature on resource management and the environmental impacts associated with production systems [61]. Similarly, population density presents a significant negative effect on circulatory outcomes, highlighting the infrastructural and environmental challenges faced by densely populated urban areas in managing circular material flows. This result is consistent with studies highlighting the complexities and pressures inherent in high-density contexts [95].

On the contrary, neither GDP per capita nor institutional quality (EQI) are statistically significant determinants of the ReCEI. This outcome suggests that ReCEI captures circular economy performance independently, without confounding influences from regional wealth or institutional capabilities, corroborating prior findings [96,97]. The Regional Innovation Scoreboard (RIS) exhibits a significant negative association in M2, indicating that higher overall innovation performance does not inherently translate into greater circularity. This underscores an important conceptual distinction between profit-oriented innovation and circular economy-specific innovation [98]. This result is also confirmed by private investment in R&D which shows a negative and statistically significant coefficient in M2 but loses significance in the M3 model. In contrast, empirical evidence suggests that public investment in R&D is positively associated with enhanced circularity outcomes, reflecting the strategic role of targeted public funding in fostering sustainable transitions.

Overall, this robustness analysis confirms that the ReCEI represents a methodologically sound and conceptually coherent tool: the most relevant structural variables influence the indicator in a predictable manner, while the economic and institutional context variables do not distort its regional ranking. Thus, empirical evidence supports the ability of ReCEI to capture the operational dimensions of circularity, providing a reliable tool for the comparative analysis of regional performances in the circular economy.

5.3. Sensitivity Analysis of ReCEI

To assess the robustness of the effect of public investment in R&D on the Regional Circular Economy Index (ReCEI), we conducted a sensitivity analysis employing the sensemakr technique. This approach, developed by [99], 2020, extends traditional omitted variable bias techniques by incorporating formalized quantitative robustness metrics and intuitive visual diagnostics. A key feature of the sensemakr framework is its ability to benchmark the explanatory power of hypothetical unobserved confounders against that of observed covariates, thereby enabling a contextualized and data-driven interpretation of sensitivity results. This enhances both transparency and methodological rigor in evaluating the potential influence of hidden biases on casual inferences derived from regression analyses.

As reported in

Table 5. Sensitivity analysis. General results.

Outcome Regional Circular Economic Index (ReCEI)
Treatment	Estimate	S.E.	T-Value	Partial R2 of the Treatment (R2yd.x)	Robustness (RV_q = 1)	Robustness of t-Value (RV_q = 1 a = 0.05)
Public Invstment in R&D	18.418	2.9458	6.2522	0.2298	0.4171	0.3097

, the estimated coefficient for public investments in R&D (lpubinvrd) is positive and statistically significant (β = 18,418, standard error = 2946, t = 6.25), demonstrating a robust and favorable impact of such investments on regional circular economy performance. Sensitivity analysis reveals that an extreme confounder, orthogonal to the covariates included in the model, would have to explain at least 22.98% of the residual variance of the treatment to entirely nullify the observed effect (partial R_(yd.x)² = 0.2298). The Robustness Value (RV) quantifies the minimum strength that an unobserved confounder must possess to alter the estimated casual effect. In this context, an RV of 0.4171 for a total bias (q = 1.00) indicates that an unobserved confounder should explain more than 41.7% of the residual variance in both the treatment and the dependent variable in order to reduce the estimated coefficient to zero. Additionally, the RV_qa value of 0.3097 at the 5% significance level of 5% (α = 0.05), implies that a confounder would have to explain at least 30.97% of the residual joint variance to make the observed effect statistically significant. These findings indicate that the estimated effect is robust, suggesting that it is not readily nullifiable by realistic confounders.

Complementary to this, the Bounds on Omitted Variable Bias analysis (see

Table 6. Sensitivity analysis. Bounds on Omitted Variable Bias.

Bound	R2dz.x	R2yz.dx	Coef.	S.E.	t(H0)	Lower CI	Upper CI
1.00× Institutional quality	0.0066	0.0033	18.2604	2.9621	6.1647	12.4002	24.1206
2.00× Institutional quality	0.0132	0.0066	18.1017	2.9671	6.1007	12.2316	23.9718
3.00× Institutional quality	0.0199	0.0098	17.9419	2.9722	6.0366	12.0618	23.8221

), using the European Quality of Government Index (EQI) as a benchmark, confirms the stability of the estimated coefficient. Specifically, an unobserved confounder with explanatory power equivalent to twice or three times that of the EQI induces only marginal changes in the estimated coefficient (18.26, 18.10 and 17.94, respectively), preserving both the sign and substantive significance of the effect. Notably, the effect remains positive even under extreme conditions wherein the confounder accounts for the entirety of the residual variance (ranging between 13.62 and 15.67), thereby underscoring the robustness and practical relevance of the casual inference.

In conclusion, the sensitivity analysis supports the robustness and credibility of the positive effect of public R&D investments on ReCEI, reinforcing the hypothesis that public investment policies in R&D represent an effective mechanism to enhance regional performances in circular economy. The combination of a statistically significant coefficient, high robustness values, and narrow bounds confirms the rigorous methodology employed and substantially reduces the likelihood that omitted confounders could compromise the validity of the findings.

5.4. Sensitivity Analysis of Regional Circular Economy Disparity Index (ReCED)

To assess the robustness of the ReCED index in capturing regional disparities in circular economy performance—as measured by the ReCEI—this section presents a sensitivity analysis based on an alternative benchmark threshold. Specifically, we define a new threshold, z₃, as the median of the ReCEI distribution. This measure is widely employed in territorial welfare studies, institutional benchmarking, and comparative evaluations of public service quality, owing to its property as a central point that divides the distribution into equal halves, rendering it insensitive to skewness or outliers [63,100]. Employing the median as an alternative baseline serves two key objectives: first, it tests the index’s stability against a distinct measure of central tendency, distinct from the interquartile mean; second, it aligns with established methodologies in performance indicator construction.

Table 7. ReCED values and absolute distance from z3. ReCEI and Waste Production pillar.

Regions	ReCED ReCEI	Distance (ReCEI-z3)	Regions	ReCED WasteProd	Distance (WasteProd-z3)
Trentino-Alto Adige	0.003	−0.11	Sardinia	0.007	−2.23
Veneto	0.004	−1.49	Aosta Valley	0.013	−10.61
Piedmont	0.004	−2.71	Trentino-Alto Adige	0.017	−16.45
Emilia-Romagna	0.008	−7.04	Umbria	0.026	−28.20
Lombardy	0.008	−8.19	Piedmont	0.027	−28.61
Friuli-Venezia Giulia	0.014	−15.71	Basilicata	0.031	−34.22
Basilicata	0.016	−17.96	Veneto	0.034	−37.95
Umbria	0.023	−27.92	Emilia-Romagna	0.034	−38.17
Molise	0.031	−38.38	Friuli-Venezia Giulia	0.036	−40.37
Aosta Valley	0.031	−38.69	Lombardy	0.044	−51.56

,

Table 8. ReCED values and absolute distance from z3. Decoupling pillar and Waste Management pillar.

Regions	ReCED Decoup	Distance (Decoup-z3)	Regions	ReCED WasteManag	Distance (WasteManag-z3)
Marche	0.002	−0.16	Liguria	0.004	−3.07
Sardinia	0.003	−0.91	Veneto	0.008	−8.34
Friuli-Venezia Giulia	0.003	−1.16	Marche	0.009	−8.89
Emilia-Romagna	0.003	−2.06	Lombardia	0.012	−13.14
Abruzzo	0.004	−2.83	Emilia-Romagna	0.013	−14.85
Trentino-Alto Adige	0.004	−3.48	Sardinia	0.014	−16.30
Umbria	0.009	−11.14	Umbria	0.014	−16.73
Basilicata	0.015	−20.68	Piedmont	0.018	−21.37
Molise	0.042	−63.59	Aosta Valley	0.018	−21.92
Aosta Valley	0.042	−64.76	Molise	0.021	−26.32

,

Table 9. ReCED values and absolute distance from z3. Waste recovery level pillar and Sustainable Innovation pillar.

Regions	ReCED WasteRecov	Distance (WasteRecov-z3)	Regions	ReCED SustInnov	Distance (SustInn-z3)
Trentino-Alto Adige	0.003	−0.63	Umbria	0.003	−0.10
Veneto	0.005	−3.56	Sardinia	0.003	−0.25
Emilia-Romagna	0.010	−11.02	Marche	0.006	−2.83
Piedmont	0.011	−12.58	Aosta Valley	0.007	−3.99
Liguria	0.012	−13.62	Abruzzo	0.008	−4.76
Basilicata	0.013	−15.55	Apulia	0.014	−9.69
Calabria	0.016	−20.12	Calabria	0.015	−10.17
Aosta Valley	0.019	−24.75	Basilicata	0.018	−12.62
Molise	0.039	−53.93	Sicili	0.029	−21.25
Umbria	0.042	−58.55	Molise	0.041	−31.56

and

Table 10. ReCED values and absolute distance from z3. Environmental Sustainability pillar and Environmental Resilience pillar.

Regions	ReCED EnvSust	Distance (EnvSust-z3)	Regions	ReCED EnvRes	Distance (EnvRes-z3)
Toscana	0.004	−1.95	Piedmont	0.002	−0.29
Umbria	0.004	−2.26	Friuli-Venezia Giulia	0.002	−0.59
Sardinia	0.005	−3.05	Basilicata	0.002	−1.22
Emilia-Romagna	0.007	−5.99	Abruzzo	0.003	−1.91
Veneto	0.007	−6.36	Calabria	0.003	−2.51
Basilicata	0.008	−7.04	Trentino-Alto Adige	0.004	−3.39
Friuli-Venezia Giulia	0.009	−8.44	Umbria	0.005	−4.93
Molise	0.019	−20.88	Lombardy	0.015	−16.53
Trentino-Alto Adige	0.030	−34.65	Molise	0.016	−17.70
Aosta Valley	0.044	−52.68	Aosta Valley	0.027	−30.85

report the results of the sensitivity analysis, displaying ReCED values—computed for the composite ReCEI indicator and its pillars—alongside the absolute distance from the alternative threshold z₃ (median). Lower ReCED values, approaching zero, denote proximity to z₃; under this metric, the relative structure of regional performance remains substantially unaltered. The regions exhibiting the greatest deviation from z₃ mirror those identified under the baseline threshold z₁ (interquartile mean), with no substantive shifts in the overall ranking.

This invariance implies that ReCED’s measurement of territorial disparities is insensitive to threshold specification, robustly capturing authentic distributional differences in circular economy performance. Thus, incorporating the median threshold validates that empirical findings are parameter-independent, confirming the solidity of the methodological framework adopted [41,101].

6. Conclusions

In recent years, Italy has consolidated a leading position in Europe in the circular economy domain, standing out for efficiency in the use of resources, high recycling rates, and use of secondary raw materials. Reports from the European Commission [102] and the Circularity Gap Report Italy [103], place Italy among the European leaders in municipal waste recycling—over 54% in 2023, well above the EU average—and in resource productivity and recycled material use.

However, this national excellence is accompanied by significant regional disparities. To address these, the present study introduces a combined methodological approach aimed at quantifying disparities in circular economy performance, thereby providing detailed insights critical for policy design. The initial step involves developing a composite index, called the Regional Circular Economy Index (ReCEI), designed for adaptability across territorial contexts that enables standardized assessment of circular economy performance at the Italian NUTS 2 region level. The index provides a granular perspective on regional progress towards circularity, distinguishing leaders implementing sustainable practices from laggards, in order to provide targeted insights for policy intervention. ReCEI also identifies regional strengths and weaknesses, fostering a better understanding of where additional resources and actions are needed.

The development of ReCEI acknowledges ongoing debates on composite indicators’ (CIs) statistical validity, including methodological biases, subjectivity in indicator selection, normalization, weighting, and aggregation, alongside sensitivity to data quality and measurement errors [104,105,106]. These limitations risk distorting empirical insights and policy recommendations due to opacity or inadequate statistical rigor [107,108,109]. Comprehensive sensitivity and validation analyses (see Section 5) address these concerns, confirming ReCEI’s robustness for regional policymakers to monitor circular economy progress, pinpoint interventions, and enable benchmarking.

The second methodological step deals with the construction of indicators measuring the dissimilarity of territorial areas in circular economy performance, as distance from threshold values. The Regional Circular Economy Disparity Index (ReCED), inspired by Sen’s Theory [23], is designed to capture not only the magnitude of underperformance but also its intensity and spatial distribution [24,25,110]. The integrated application of these indices enables the creation of a dynamic and multidimensional mapping of territorial disparities in the circular economy, providing a spatially detailed evaluation framework that directly incorporates territorial inequalities into socio-economic planning processes.

Empirical findings confirm a heterogeneity in circular economy performance across Italian regions. While some regions –such as Campania, Lazio, Tuscany, Apulia and Sicily—have developed integrated strategies and effective tools to advance the circular economy transition, others—namely Molise, Umbria, Valle d’Aosta, Friuli-Venezia Giulia and Basilicata—still struggle to achieve the desired standards, as evidenced by their regional composite indices. The persistent North–South divide is further reflected in the South’s lower self-sufficiency in waste processing facilities, constraining closure of material cycles and maximizing recycled input. Thus, the success of the circular transition will depend on the ability to bridge these regional gaps through an integrated strategic vision. Embedding circularity into regional planning processes necessitates setting explicit objectives and challenging investments into infrastructure, innovation and skills development—with particular emphasis on supporting SMEs through tailored incentives and facilitation of widespread adoption of circular technologies [14,17,18]. Moreover, fostering active participation from citizens and companies, coupled with transparent and adaptive monitoring systems, is essential for ensuring responsiveness to emerging needs [15,19,20].

Despite the robust insights generated, this study’s methodology bears inherent limitations. First, annual variation in factor structures and weights derived from principal component analysis (PCA) induce variability in indicator interrelationships, thereby impacting the construction of the composite index (CI). Consequently, the index is well-suited for cross-sectional comparisons of statistical units within a given period but presents challenges for longitudinal or temporal analyses. Additionally, while data availability at the national level is comprehensive, regional-level granularity remains limited, constraining finer-scale comparative analyses. Despite these limitations, the proposed methodological approach constitutes a valuable decision-support instrument for regional policymakers, facilitating targeted strategy development aimed at bridging existing circular economy gaps. By interpreting territorial inequality indices as weighting factors, the framework further supports the design of equitable and efficient resource allocation mechanisms, establishing a quantitative amount for prioritizing interventions within territorial planning processes.