Open AccessArticle

Applying the Efficiency Analysis Tree Method for Enhanced Eco-Efficiency in Municipal Solid Waste Management: A Case Study of Chilean Municipalities

Departamento de Matemáticas para la Economía y la Empresa, Universidad de Valencia, Avd. Tarongers S/N, 46010 Valencia, Spain

Institute of Sustainable Processes, University of Valladolid, C/ Mergelina S/N, 47002 Valladolid, Spain

Department of Business, New York College, Leof. Vasilisis Amalias 38, 105 58 Athens, Greece

Author to whom correspondence should be addressed.

Clean Technol. 2024, 6(4), 1565-1578; https://doi.org/10.3390/cleantechnol6040075

Submission received: 2 July 2024 / Revised: 29 July 2024 / Accepted: 18 November 2024 / Published: 21 November 2024

(This article belongs to the Special Issue Towards Sustainable Consumption and Production Patterns: Strategies for Achieving SDG 12)

Download

Browse Figures

Versions Notes

Abstract

Enhancing the eco-efficiency of municipal solid waste (MSW) services is pivotal for the shift toward a circular economy. Although the Data Envelopment Analysis (DEA) method is widely used, it is susceptible to overfitting, potentially distorting eco-efficiency assessments. This study applies the efficiency analysis tree (EAT) method, which synergizes machine learning and linear programming, offering a more reliable framework for eco-efficiency evaluation in the MSW sector. This innovative approach provides deeper insights into the optimal levels of operational costs and unsorted waste. The research encompasses a case study of 98 Chilean municipalities from 2015 to 2019, uncovering significant disparities in optimal operational expenses and unsorted waste quantities, which underscores the necessity for customized waste management approaches. The average eco-efficiency scores for 2015–2019 range between 0.561 and 0.566. This means that assessed municipalities can reduce unsorted waste by amounts ranging from 1,632,409 tons/year (2016) to 1,822,663 tons/year (2018). Potential economic savings estimated are 105,973 USD/year (2019), which represents 44% of the total MSW management costs. Additionally, the investigation into the effects of external factors on eco-efficiency furnishes nuanced perspectives that can guide policymakers and municipal authorities in developing effective, context-specific waste management strategies. Beyond refining eco-efficiency evaluations, this study contributes to more informed decision-making processes, aiding the progression toward sustainable waste management practices.

Keywords:

eco-efficiency; efficiency analysis tree; municipal solid waste; unsorted waste; waste management; circular economy

1. Introduction

Waste management is of utmost importance for the health and well-being of the society [1]. It also helps to deal with universal challenges such as climate change and resource scarcity [2]. As the population grows and the consumption of goods increases as well, the collection, treatment, and recycling of municipal solid waste (MSW) would become even more challenging [3]. Recent European Union (EU) policy strategies such as the EU 7th Environment Action Programme highlight that EU countries should, among other things, aim to cut down the generation of waste and promote recycling and reuse of waste [4]. In an analogous manner, the United Nations (UN) underlined the importance of circular economy in achieving Goals 11 and 12 of the Sustainable Development Goals [5,6,7,8]. Policymakers and researchers are seeking for solutions to promote the efficient management of municipal solid waste from an economic and environmental perspective [3]. It can allow reducing operational costs and tariffs for citizens, improving the quality of service, safety, and health and complying with sustainable development policies [9]. Therefore, it is of great importance to study and evaluate the eco-efficiency of the waste management sector.

Eco-efficiency refers to the generation of more products employing less resources and minimizing environmental impacts [10,11]. In the framework of MSW management, eco-efficiency of units (municipalities, countries, service providers) involves increasing the amount of waste recycling and, at the same time, minimizing the generation of unsorted waste and operational costs [12,13]. The European Union’s Directives 2008/98/CE and 2018/851/UE set forth a regulatory framework that supports this eco-efficiency approach in waste management. These directives emphasize the importance of increasing the fraction of waste that is separately collected for recycling and recovery, aiming to transform waste into a valuable resource and drastically reduce the volume of waste sent to landfills [14]. Evaluating MSW services from an eco-efficiency perspective provides a comprehensive framework for assessing the environmental and economic performance of waste management systems, helping stakeholders make informed decisions that balance ecological considerations with service efficiency and cost-effectiveness [15,16].

According to the literature reviews conducted by Amaral et al. [3] and Lo Storto [14], data envelopment analysis (DEA) has been extensively utilized to evaluate the eco-efficiency of MSW service providers, offering a non-parametric approach that efficiently handles multiple inputs and outputs without the need for a predefined functional form of the production technology [17]. In the context of MSW management, DEA models typically treat operational costs as inputs while categorizing unsorted waste as an undesirable output and recycled waste as a desirable output [13]. Some studies have adopted advanced DEA models to refine their assessments. For instance, Romano and Molinos-Senante [9] utilized a metafrontier DEA model, while Delgado-Antequera et al. [18] employed the weighted Russell directional distance model. Despite its widespread application, DEA’s deterministic nature is a known limitation, as it makes the analysis sensitive to data quality issues, including outliers. Some researchers have addressed this by using partial frontier analysis, which provides a more robust efficiency estimation in the presence of data anomalies [1,3,12,19]. Additionally, Lo Storto [14] implemented a fuzzy DEA approach to tackle uncertainties and missing data issues. However, overfitting represents another critical challenge in DEA applications [20] that has not been addressed in MSW eco-efficiency literature. Overfitting arises because the efficiency score for each unit is calculated based on its deviation from the production possibility set’s frontier, reflecting each activity or production plan. This phenomenon manifests when the model excessively conforms to the particular dataset employed in the analysis, which may compromise the robustness of the efficiency scores [21].

To address overfitting in non-parametric methods, Esteve et al. [20] introduced an innovative methodology, i.e., efficiency analysis trees (EATs), which integrates machine learning and linear programming to assess efficiency. Specifically, the EAT method employs Classification and Regression Trees (CART) to predict the output variables’ values based on established rules (thresholds) for input variables. It incorporates the free disposability assumption to ensure that the predicted output variable value represents its maximum (optimal) level [22]. Subsequently, the EAT method applies linear programming to determine efficiency scores, ensuring their robustness and applicability in comparative efficiency analyses and decision-making processes [23,24]. In contrast to non-parametric methods such as DEA, the EAT method overcomes the well-known problem of overfitting by using cross-validation to prune back the deep tree obtained in a first stage [20].

The EAT method offers enhanced flexibility and accuracy in handling data variations and model specifications. Unlike DEA, which assumes a fixed form of the efficiency frontier and may overestimate inefficiencies due to noise, the EAT method integrates a more adaptive approach that accounts for data variability [23]. Furthermore, unlike stochastic frontier analysis (SFA), which requires a predefined distributional form affecting the robustness of the results, the EAT method allows for a more direct and intuitive analysis of efficiency and its determinants without imposing strict distributional assumptions [25]. This approach not only broadens the analytical scope but also increases the reliability of the results, providing clearer insights into the eco-efficiency of MSW management systems.

Although various studies have applied DEA to evaluate the eco-efficiency of the MSW sector, to the best of our knowledge, none of them have specifically dealt with overfitting limitations within this context. Our study aims to fill this gap in the literature.

This study aims to advance the evaluation of eco-efficiency in the MSW sector by addressing the issue of overfitting through the application of the EAT method. This method enables the determination of the maximum level of operating expenses and unsorted waste based on the total amount of waste collected and recycled. By leveraging the eco-efficiency scores calculated for each evaluated municipality, the potential reductions in costs and unsorted waste were deduced. Additionally, the influence of certain external variables on the eco-efficiency of municipalities in providing MSW services was explored.

Our study has the following contributions to existing knowledge. By utilizing the EAT method, this study pioneers in addressing the overfitting issue in eco-efficiency assessments within the MSW sector. Moreover, it allows not only for estimating eco-efficiency scores but also for determining the optimal levels of operational costs and undesirable outputs (unsorted waste) relative to the quantities of waste collected and recycled. This study further explores the impact of selected exogenous variables on the eco-efficiency of municipalities, providing insights into how factors such as population size, municipality area, population density, and waste generation per capita influence MSW management eco-efficiency. By addressing these aspects, this study significantly contributes to the body of knowledge on eco-efficiency in MSW management, offering methodological advancements and practical insights that can aid municipalities in improving their waste management practices and achieving greater environmental and economic sustainability.

2. Methodology

2.1. Methodology for Eco-Efficiency Estimation

Eco-efficiency of municipalities in the provision of MSW services is estimated using the EAT method [23]. It involves two main stages, which are defined as follows. Firstly, the entire sample of municipalities is broken up into several nodes based on a set of thresholds of the predictor variables (collected and recycled waste, operational costs, and unsorted waste). In particular, it uses regression trees to visualize the maximum value of the response variable (eco-efficiency) based on a set of thresholds of the predicted variables [22,26]. In doing so, the EAT approach imposes the assumption of free disposability based on the concept of efficiency analysis [20], which involves that if a specific pair of input and output is producible, any pairs of more input and less output for the specific one are also producible [27].

Let us assume that we have a vector of response variables

y, \dots, y_{n}

, where

y_{i} \in R^{n}

, and a vector of predictor variables

x_{1}, \dots, x_{m}

, where

x_{i} \in R^{m}

. In our case study, response factors are operational costs and unsorted waste, whereas the predictor factor is the recycled MSW. The EAT approach chooses a predictor

j

and a threshold

s_{j} \in S_{j}

, where

S_{j}

is the set of likely rules (or thresholds) for

j

to split the units into two nodes depicted at the right and left sides of the decision tree,

t_{R}

and

t_{L}

, respectively [22,23,24]. The separation of the units into a right and a left node is carried out by employing the sum of the mean squared error (MSE). The EAT approach is mathematically presented as follows:

R (t_{L}) + R (t_{R}) = \frac{1}{n} \sum_{(x_{i}, y_{i}) \in t_{L}} {(y_{i} - y (t_{L}))}^{2} + \frac{1}{n} \sum_{(x_{i}, y_{i}) \in t_{R}} {(y_{i} - y (t_{R}))}^{2}

(1)

In Equation (1), the size of the sample and the nodes of the decision tree are defined as

n

and

t

, respectively. Moreover,

R (t_{L})

and

R (t_{R})

present the MSE for the left and right nodes, respectively [28]. Also,

y (t_{L})

and

y (t_{R})

are the estimated values of the response variable for the left and right nodes, respectively [20]. These are provided below:

\begin{array}{l} y (t_{L}) = m a x \{m a x \{y_{i} : (x_{i}, y_{i}) \in t_{L}\}, y (I_{T (k | t^{*} \to t_{L}, t_{R})} (t_{L}))\} \\ y (t_{R}) = m a x \{m a x \{y_{i} : (x_{i}, y_{i}) \in t_{R}\}, y (I_{T (k | t^{*} \to t_{L}, t_{R})} (t_{R}))\} \end{array}

(2)

In Equation (2),

T

is the subtree that the EAT method generates,

k

is the number of breaks, and

y (I_{T (k | t^{*} \to t_{L}, t_{R})} (t_{L}))

and

y (I_{T (k | t^{*} \to t_{L}, t_{R})} (t_{R}))

are the leaf nodes of the regression tree after the completion of the k-th break based on Pareto rules at nodes

t_{L}

and

t_{R}

, respectively (for more details, please see Esteve et al. [20,23,24]).

Subsequently, a machine-learning approach incorporates V-fold cross-validation techniques to determine the optimal number of regression trees, effectively addressing potential overfitting problems [20]. In this cross-validation, the learning sample is randomly divided into

V

disjoint subsamples with the same sample size or as close as possible. Then, the process to obtain a score for each subtree is based on repeating the tree growing and pruning procedure commented above using each subsample instead of the complete sample [20]. By utilizing this approach, the EAT method estimates the underlying production technology, which can provide a more accurate and robust understanding of the data structure and the relationships within it. In doing so, the EAT approach estimates the following production technology:

\hat{{P T}_{T_{k}}} = \{(x, y) \in R_{+}^{m + 1} : y \leq d_{T_{k}} (x)\}

(3)

where

d_{T_{k}} (x)

is the estimated value of the predictor variable regarding the subtree

T_{k}

Finally, the EAT method employs the following linear programming model to measure the eco-efficiency of each unit (municipality in our study) evaluated:

\begin{matrix} ψ (x_{k}, y_{k}) = \max ψ \\ s u b j e c t t o : \\ \sum_{t \in \tilde{T^{*}}} λ_{t} a_{j}^{t} \leq x_{j k}, j = 1, \dots, m \\ \sum_{t \in \tilde{T^{*}}} λ_{t} d_{r T^{*}}^{t} (a^{t}) \geq {ψ y}_{j k}, r = 1, \dots, p \\ \sum_{t \in \tilde{T^{*}}} λ_{t} = 1 \\ λ_{t} \in \{0,1\}, i = 1, \dots, n \end{matrix}

(4)

In Equation (4), the eco-efficiency of each municipality is defined by

ψ

. The terms

(a^{t}, d_{T^{*}} (a^{t}))

are points in the input–output vector for all

t \in T^{*}

, where * presents the final subtree [29]. The

λ

variables in the constraints of the linear programming model are employed to make the efficient frontier [28]. The eco-efficiency score, ranging from zero to one, serves as a metric to evaluate the environmental performance of a municipality in relation to its waste management practices. A score below one indicates eco-inefficiency, suggesting that there are opportunities for the municipality to enhance its environmental performance. On the other hand, a score of one denotes full eco-efficiency, implying that the municipality is optimizing its environmental performance.

The derivation of eco-efficiency scores further allows the quantification of savings from an economic and environmental perspective. These are calculated as follows:

{O p e r a t i n g c o s t s}_{s} = {O p e r a t i n g c o s t s}_{c} * (1 - ψ)

(5)

{U n s o r t e d W a s t e}_{s} = {U n s o r t e d w a s t e}_{c} * (1 - ψ)

(6)

where

{O p e r a t i n g c o s t s}_{s}

and

{U n s o r t e d W a s t e}_{s}

present the potential savings in operational costs and unsorted waste, respectively, that a municipality can achieve if it operated on the efficient frontier. Finally,

{O p e r a t i n g c o s t s}_{c}

and

{U n s o r t e d W a s t e}_{c}

present the actual (observed) levels of operating costs and unsorted waste for each municipality that is included in the assessment exercise.

2.2. Potential Exogenous Variables Influencing Eco-Efficiency

To explore how external factors influence the eco-efficiency of municipalities in managing MSW services, the Kruskal–Wallis test, a non-parametric method, is applied to compare eco-efficiency scores across different groups of municipalities defined by exogenous variables [12,30] (Equation (6)). This test is particularly useful when the data do not meet the normal distribution assumption required for ANOVA [9]. The null hypothesis posits that there are no differences in the eco-efficiency scores across the groups—in other words, all groups are from the same distribution or population in terms of eco-efficiency. The test provides a p-value that is used to determine the statistical significance of the observed differences. If the p-value is less than or equal to the chosen significance level (commonly 0.05 for 95% confidence), the null hypothesis is rejected. This outcome would indicate that there are statistically significant differences in eco-efficiency scores among the groups, suggesting that the exogenous variable in question has a discernible impact on eco-efficiency.

H = \frac{12}{N (N + 1)} \sum_{i = 1}^{g} n_{i} {(\bar{r_{i}} - \bar{r})}^{2}

(7)

where

N

is the total number of observations from all groups evaluated;

n_{i}

is the number of observations for group

i

;

\bar{r_{i}} = \frac{\sum_{j = 1}^{n_{i}} r_{i j}}{n_{i}}

;

r_{i j}

is the range (among all observations) of observation

j

in the group

i

; and

\bar{r} = \frac{(N + 1)}{2}

Table 1 shows the groups (

i

) considered for each variable analyzed:

3. Data and Sample Selection

This study assesses the eco-efficiency of 98 out of 345 municipalities in Chile over the 2015–2019 period, focusing on their management of solid waste and on the availability of data for specified inputs and outputs. Initially, 211 municipalities were excluded for reporting zero values for at least one of the variables critical for estimating eco-efficiency scores, reducing our sample to 134 municipalities. To ensure data robustness, we applied Tukey’s method (1977) to detect outliers based on average and standard deviation values. This process identified 36 out of the 134 municipalities as outliers due to anomalies in at least one of the seven variables assessed for eco-efficiency. After accounting for both data availability and the impact of outliers, the final sample was refined to 98 municipalities.

In the Chilean context, local municipalities are tasked with solid waste management, typically outsourcing waste collection to private companies that perform door-to-door service, as public waste containers are not a common practice. These companies are also responsible for the waste’s subsequent transportation and disposal. It has been reported that 55% of Chilean municipalities had introduced some recycling services by 2019 [31], indicating a growing awareness and initiative toward sustainable waste management. Nevertheless, data from the Chilean Environment Ministry [32] indicate that, at present, a mere 12.5% of packaging waste is subjected to recycling. In response, the government has actively sought to improve this scenario via legislative changes. Specifically, the 2016 Law for Promoting Recycling and Extended Producer Responsibility (REP) is designed to elevate waste recycling rates, targeting a 70% recovery for paper and cardboard, 45% for plastics, and 65% for glass by the year 2030.

The data from the SINIM database [33] illustrate a significant rise in the operational expenses (OPEX) for MSW management in Chile, escalating from USD 245 million in 2012 to USD 528 million by the referenced year. It involves the annual expenses of the municipality in collecting, transporting, and disposing the MSW generated in its territory. This substantial increase in OPEX and the low percentage of MSW recycling rates underscore the urgent need for municipalities to enhance their eco-efficiency in MSW service provision.

In assessing the eco-efficiency of the waste sector using the EAT method, the selection of appropriate response and predictor variables is crucial, guided by data availability and insights from prior research [9,14,18,34,35]. This study incorporates two response variables: i) the OPEX associated with the collection, transportation, and disposal of waste, expressed in United States dollars (USD) per year—this variable reflects the financial aspect of waste management services; and ii) the quantity of unsorted waste, measured in tons per year, representing the physical scale of waste that is not segregated for recycling or recovery. The predictor variable is the quantity of waste collected and recycled, measured in tons per year. This variable serves as an indicator of the effectiveness of waste recycling and recovery efforts within the municipality. The data for this study are sourced from two authoritative databases: MSW collection and disposal data from the Chilean National Waste Declaration System (SINADER), and OPEX data for MSW services from the Chilean National Municipal Information System (SINIM).

In selecting the potential exogenous variables impacting on eco-efficiency scores, the extant literature [14,36,37] and statistical data availability for Chilean municipalities were considered. The municipalities were grouped based on average values for each selected exogenous variable, which include the following: i) population served defined as the number of inhabitants of the municipality—it provides insights into the scale of MSW management needed and the potential economies of scale influencing eco-efficiency; ii) municipality size in km²—it can influence the logistics and costs associated with waste collection and transportation, and thereby eco-efficiency scores; iii) population density expressed as the number of people per km²; and iv) the annual amount of MSW generated per capita (kg/year·person). This metric can help gauge the waste generation patterns and their impact on the efficiency of waste management systems.

Table 2 in this study presents the descriptive statistics for these variables. R software 4.4.2 was used for the statistical analysis.

4. Results and Discussion

4.1. Optimal Operational Expenditure and Unsorted Waste Estimation

The regression tree results from the EAT approach, based on 2015–2019 variables, as depicted in Figure 1, offer insightful categorizations of municipalities based on their MSW collection and recycling volumes, elucidating the associated OPEX and the quantity of unsorted waste. Third groups of municipalities based on the quantity of MSW collected and recycled are observed. A first group of municipalities includes those that collect and recycle more than 317 tons per year of MSW, representing 36% of the sample (178 out of 490 observations). For these municipalities, the maximum estimated OPEX is USD 17,371 per year, translating to approximately USD 54.79 per ton. A second group of municipalities comprises 40% of the observations (197 out of 490) and is characterized by handling between 41 and 317 tons of MSW per year. The maximum OPEX for this group is predicted to be USD 8870 per year, with the cost per ton ranging between USD 239.74 and USD 27.98. The final group accounts for 24% of the sample (115 out of 490), which includes municipalities that collect and recycle less than 41 tons/year of MSW. The maximum OPEX for this group is estimated at USD 2525 per year or USD 61.59 per ton.

These results underscore the substantial variability in OPEX per ton across different municipality groups. This study’s results also shed light on the economies of scale in MSW management. The findings from Llanquileo-Melgarejo and Molinos-Senante (2021) align with the observed OPEX variability, indicating diverse returns to scale across municipalities. Specifically, they demonstrated that 6 out of 142 municipalities (4.2%) present constant returns to scale, 55 out of 142 (38.7%) demonstrated increasing returns to scale, and 81 out of 142 (57.1%) demonstrated decreasing returns to scale, which is consistent with the findings from this study.

The results of the analysis via the EAT method indicate that the maximum quantity of unsorted waste across the three identified groups of municipalities substantially varies, ranging from 56,399 tons/year to 466,772 tons/year (Figure 1). Despite this wide range in the absolute quantities of unsorted waste, when these figures are contextualized against the total amount of MSW generated, they reveal a striking similarity: the proportion of unsorted waste hovers around 99% across all groups. This consistency points to a pervasive challenge in the Chilean waste management landscape—the extremely low rate of recycling. Such a high percentage of unsorted waste underscores the pressing need for enhanced waste segregation and recycling initiatives within the country. The near uniformity in the proportion of unsorted waste, irrespective of the municipality group’s size or waste management capacity, highlights that the issue of low recycling rates is widespread, affecting municipalities of various scales and capacities.

4.2. Eco-Efficiency Estimation

The results of the eco-efficiency analysis of the municipalities over the years present a consistent trend, as depicted in Figure 2. The average eco-efficiency scores exhibit minimal variation, ranging from 0.561 in 2015 and 2019 to a slightly higher 0.566 in 2016, with an overall interannual average of 0.564. This indicates that, on average, there is a potential for improving economic and environmental performance by up to 43.6%. A notable observation from the results is the absence of any municipality achieving an eco-efficiency score of one, involving that all the municipalities assessed have the potential to enhance their eco-efficiency. This uniformity in scores, with the lowest observed eco-efficiency at 0.514 and the highest at 0.535, indicates a remarkable homogeneity in the economic and environmental performance across the 98 municipalities studied. The findings underscore a broader opportunity for systemic improvements in waste management practices across the assessed Chilean municipalities. The uniformity in eco-efficiency scores suggests that similar strategies for enhancement can be applicable across various municipalities, focusing on increasing recycling rates and optimizing waste collection and transportation. By addressing these areas, municipalities can work toward achieving higher eco-efficiency, leading to economic benefits and reduced environmental impacts.

The findings from our study, indicating average eco-efficiency scores between 0.561 and 0.566 for the assessed Chilean municipalities in MSW services, align with previous research, albeit with some variations due to the methodologies employed. The studies by Llanquileo-Melgarejo et al. [38] and Llanquileo-Melgarejo and Molinos-Senante [30], which utilized the DEA method, reported average eco-efficiency scores of 0.540 and 0.580, respectively, for different sets of municipalities. Molinos-Senante et al. [29] observed a higher average eco-efficiency of 0.689 using a double-bootstrap DEA method. Conversely, Molinos-Senante et al. [39] found a significantly lower average eco-efficiency score of 0.332 employing the semi-parametric envelopment of data (StoNED) method for 140 municipalities. This variability highlights the importance of method selection in performance assessment and the implications it can have on decision-making and policy development in municipal waste management. Understanding the methodological underpinnings and contextual appropriateness of each approach is crucial for deriving meaningful insights and making informed decisions aimed at enhancing the eco-efficiency of MSW services.

The inefficiency in MSW recycling among the 98 evaluated Chilean municipalities not only impacts environmental sustainability but also has significant economic implications. The potential reductions in unsorted waste, as depicted in Figure 3, indicate that if these municipalities achieved eco-efficiency, they can reduce unsorted waste by amounts ranging from 1,632,409 tons/year (2016) to 1,822,663 tons/year (2018). This reduction translates to 43.3% (2018) to 44.1% (2016) of the total MSW generated and collected, showcasing a substantial opportunity for waste diversion from landfills through enhanced recycling efforts. Moreover, the potential cost savings, as depicted in Figure 4, show an increasing trend from 2015 to 2019, peaking at a potential saving of 105,973 USD/year in 2019. This figure represents approximately 44% of the total MSW management costs for the 98 municipalities. Such substantial potential savings underscore the economic benefits of improving eco-efficiency in MSW management, alongside the environmental benefits of reducing unsorted waste.

The consistency in potential waste reduction and cost savings across the five years studied reflects the nearly uniform eco-efficiency scores observed year-over-year (Figure 2), suggesting that the municipalities’ performance in waste management has remained relatively stable over the period. The year-to-year differences in potential waste reduction and cost savings are primarily attributable to fluctuations in the total volumes of MSW collected rather than significant changes in eco-efficiency levels. These findings underscore the critical need for targeted interventions to improve recycling rates and waste management practices in these municipalities. Enhancing eco-efficiency not only aligns with environmental sustainability goals but also offers economic benefits by potentially reducing operational expenditures associated with waste collection and disposal. Implementing best practices, investing in recycling infrastructure, and fostering community engagement in waste reduction and recycling can drive progress toward achieving these potential reductions in unsorted waste and OPEX.

4.3. Exogenous Variables Influencing Eco-Efficiency

The results of the analysis exploring the impact of exogenous variables on the eco-efficiency scores of Chilean municipalities provide valuable insights into how different factors influence waste management efficiency. For the year 2019, the Kruskal–Wallis test was employed to determine the statistical significance of eco-efficiency score differences across groups of municipalities according to the exogenous variables explored (Table 2). For the population size, the categorization was as follows: (i) fewer than 30,000 inhabitants; (ii) between 30,000 and 96,000 inhabitants; and (iii) more than 96,000 inhabitants. The p-value obtained from the Kruskal–Wallis test being less than 0.05 indicates that the differences in eco-efficiency scores among these population-based groups are statistically significant. The finding that smaller municipalities tend to be more eco-efficient than larger ones suggests the presence of diseconomies of scale in the provision of MSW services among the evaluated Chilean municipalities. This observation contrasts with several other studies [9,40], which found economies of scale in MSW management. This discrepancy highlights the complexity of waste management systems and suggests that the relationship between municipality size and eco-efficiency can vary based on local conditions, management practices, and other contextual factors. The result aligns with Guerrini et al. [12], who observed that smaller municipalities can have more efficient operational environments.

The examination of the impact of municipality size on the eco-efficiency of MSW service provision, categorized into three size groups (<200 km², 200–1000 km², and >1000 km²), utilized the Kruskal–Wallis test to assess statistical significance. The estimated p-value (Table 3) did not lead us to reject the hypothesis of equality of means for eco-efficiency with 95% significance based on the three ranges. It indicates that, within the assessed municipalities, size does not have a significant impact on eco-efficiency in MSW management. This outcome aligns with the findings of Llanquileo-Melgarejo and Molinos-Senante [30], suggesting that the physical size of a municipality may not be a critical determinant of eco-efficiency in MSW services for the studied sample. However, this conclusion contrasts with other research [18,34], who identified municipality size as a significant exogenous variable affecting the eco-efficiency of MSW service providers.

The assessment of the influence of population density on the eco-efficiency of municipalities involved grouping the municipalities into three categories based on their population density: (i) fewer than 50 inhabitants/km²; (ii) between 50 and 1000 inhabitants/km²; and (iii) more than 1000 inhabitants/km². The relationship between population density and economies of density in MSW management is a topic of ongoing debate in the literature, with mixed findings regarding whether municipalities benefit from economies of density [17] or not [3]. In the context of the current case study, the results of the analysis revealed that population density does not significantly impact the eco-efficiency of municipalities in managing MSW. This outcome might be attributed to the specific waste collection method employed across the assessed municipalities—door-to-door collection—which may not be as influenced by population density as more centralized or communal waste collection systems.

The investigation into the impact of per-capita MSW generation on municipal eco-efficiency categorized municipalities into three groups based on their annual MSW generation per inhabitant: (i) less than 350 kg/inhabitant·year, (ii) between 350 and 500 kg/inhabitant·year, and (iii) more than 500 kg/inhabitant·year. The results of the analysis revealed that municipalities with the lowest per-capita MSW generation exhibited the lowest average eco-efficiency, with these differences being statistically significant across the three groups. This finding aligns with the research by Llanquileo-Melgarejo and Molinos-Senante [30], which underscores the significance of fixed costs in the MSW management sector in Chile.

Understanding the influence of exogenous variables on the economic and environmental performance of MSW service providers is crucial for policymakers and municipal planners as it suggests that tailored strategies that consider the specific characteristics and needs of municipalities are essential for improving the overall eco-efficiency of MSW services. The discrepancies observed between studies highlight the complex interplay of factors that influence eco-efficiency, indicating that the impact of variables such as municipality size and population density can differ based on local conditions, including geographical, economic, and operational characteristics. This variation underscores the necessity for further research to delineate the circumstances under which specific municipal characteristics may influence eco-efficiency in MSW management. For policymakers and waste management professionals, these insights emphasize the importance of considering a broad range of factors when designing and implementing strategies to enhance the eco-efficiency of MSW services, rather than relying on single-variable explanations.

Given the moderate eco-efficiency of the 98 municipalities evaluated, policymakers should implement several actions to reduce OPEX and simultaneously increase the rate of MSW recycled to enhance eco-efficiency. Based on the current practices for managing MSW in Chile, recommended actions include the following: (i) implement advanced routing software to optimize waste collection routes, reducing fuel consumption and operational hours, thereby decreasing OPEX; (ii) expand and promote community recycling programs to increase MSW separation and recycling rates—this can involve more extensive public education campaigns and improved recycling facilities; (iii) adopt state-of-the-art technologies for MSW sorting and processing that increase efficiency and the capacity to handle recyclable materials; (iv) encourage private sector investment in MSW management solutions through subsidies or tax incentives, particularly for the development of recycling and waste-to-energy projects; and (v) introduce metered waste collection (pay-as-you-throw) to incentivize MSW reduction and segregation at the source, leading to lower collection and processing costs. These strategic actions can lead to significant improvements in the eco-efficiency of waste management across Chile’s municipalities, aligning with sustainable environmental practices and cost-effective management strategies.

5. Conclusions

In the pursuit of a circular economy, enhancing the eco-efficiency of MSW services is crucial. Reliable and robust eco-efficiency estimations are essential for informed decision-making to enhance waste management practices in line with sustainability goals. While DEA has been extensively utilized to evaluate the eco-efficiency of the MSW sector, its susceptibility to overfitting can potentially compromise the accuracy and reliability of the efficiency estimations. To address this limitation, this study introduces an innovative approach by employing the EAT method for eco-efficiency assessment. The EAT method combines elements of machine learning and linear programming, offering a novel and sophisticated analytical framework that enhances the robustness of eco-efficiency evaluations. The application of the EAT method in assessing the eco-efficiency of a sample of municipalities represents a significant methodological advancement in the field. It not only provides more reliable eco-efficiency scores but also offers insights into the optimal levels of operational costs and unsorted waste outputs, considering the varying levels of waste collection and recycling. This approach enables municipalities to identify more accurately where improvements can be made and how resources can be allocated more effectively to enhance their waste management systems within the context of a circular economy.

The case study underscores the variability in optimal levels of operational costs and unsorted waste across municipalities, contingent upon the volume of recycled waste. This finding suggests that municipal targets and policies for waste management should be bespoke, reflecting the unique conditions and capacities of each municipality rather than adopting a one-size-fits-all approach. From 2015 to 2019, the average eco-efficiency scores for the municipalities assessed ranged from 0.561 to 0.566. This indicates that these municipalities have the potential to significantly reduce their amounts of unsorted waste, with reductions ranging from 1,632,409 tons per year in 2016 to 1,822,663 tons per year in 2018. Moreover, the potential economic savings from such reductions are substantial. In 2019, it was estimated that municipalities can save approximately USD 105,973 annually, which represents 44% of the total MSW management costs. The absence of any municipality reaching full eco-efficiency within the sample highlights a significant potential for enhancing both economic and environmental performance in MSW management. This gap underscores the need for environmental and local authorities to intensify efforts to boost MSW recycling. Such endeavors should engage a broad spectrum of stakeholders, including citizens, to foster a collective and concerted push toward improved recycling practices. Concerning the influence of exogenous variables on eco-efficiency scores, this study adds nuance to the mixed findings in existing literature. By analyzing these variables, the research provides insights into how factors external to the waste management process itself can impact the eco-efficiency of municipalities. These insights are valuable for policymakers and municipal managers, offering a more informed basis for designing and implementing waste management strategies that are not only effective but also context-sensitive.

Author Contributions

Conceptualization, A.M. and M.M.-S.; methodology, R.S.-G.; software, M.M.-A.; validation, M.M.-S. and A.M.; formal analysis, M.M.-S.; investigation, M.M.-A.; resources, A.M.; data curation, R.S.-G.; writing—original draft preparation, M.M.-S. and A.M.; writing—review and editing, R.S.-G.; visualization, A.M.; supervision, M.M.-S.; project administration, M.M.-S.; funding acquisition, M.M.-S. All authors have read and agreed to the published version of the manuscript.

Funding

Centro de Desarrollo Urbano Sustentable ANID/FONDAP/15110020, Pontificia Universidad Católica de Chile, Santiago 4860, Chile.

Institutional Review Board Statement

Not applicable.

Informed Consent Statement

Not applicable.

Data Availability Statement

Data will be available upon to a reasonable request.

Conflicts of Interest

The authors declare no conflict of interest.

Abbreviation List

CART	Classification and Regression Trees
DEA	Data envelopment analysis
EAT	Efficiency analysis tree
MSW	Municipal solid waste
OPEX	Operational expenditure
REP	Law for Promoting Recycling and Extended Producer Responsibility
SFA	Stochastic frontier analysis
SINADER	Chilean National Waste Declaration System
SINIM	Chilean National Municipal Information System
StoNED	Semi-parametric envelopment of data
UN	United Nations

References

Romano, G.; Rapposelli, A.; Marrucci, L. Improving waste production and recycling through zero-waste strategy and privatization: An empirical investigation. Resour. Conserv. Recycl. 2019, 146, 256–263. [Google Scholar] [CrossRef]
Campitelli, A.; Aryoug, O.; Ouazzani, N.; Bockreis, A.; Schebek, L. Assessing the performance of a waste management system towards a circular economy in the Global South: The case of Marrakech (Morocco). Waste Manag. 2023, 166, 259–269. [Google Scholar] [CrossRef] [PubMed]
Amaral, C.; Pedro, M.I.; Ferreira, D.C.; Marques, R.C. Performance and its determinants in the Portuguese municipal solid waste utilities. Waste Manag. 2022, 139, 70–84. [Google Scholar] [CrossRef] [PubMed]
Dragomir, V.D.; Dumitru, M. The state of the research on circular economy in the European Union: A bibliometric review. Clean. Waste Syst. 2024, 7, 100127. [Google Scholar] [CrossRef]
United Nations. Sustainable Development Goals. 2015. Available online: https://www.un.org/sustainabledevelopment/sustainable-development-goals/ (accessed on 11 November 2024).
Di Foggia, G.D.; Beccarello, M. Improving efficiency in the MSW collection and disposal service combining price cap and yardstick regulation: The Italian case. Waste Manag. 2018, 79, 223–231. [Google Scholar] [CrossRef]
Lombardi, G.V.; Gastaldi, M.; Rapposelli, A.; Romano, G. Assessing efficiency of urban waste services and the role of tariff in a circular economy perspective: An empirical application for Italian municipalities. J. Clean. Prod. 2021, 323, 1029097. [Google Scholar] [CrossRef]
Zhao, J.; Li, X.; Chen, L.; Liu, W.; Wang, M. Scenario analysis of the eco-efficiency for municipal solid waste management: A case study of 211 cities in western China. Sci. Total Environ. 2024, 919, 170536. [Google Scholar] [CrossRef] [PubMed]
Romano, G.; Molinos-Senante, M. Factors affecting eco-efficiency of municipal waste services in Tuscan municipalities: An empirical investigation of different management models. Waste Manag. 2020, 105, 384–394. [Google Scholar] [CrossRef]
Beltran-Esteve, M.; Reig-Martínez, E.; Estruch-Guitart, V. Assessing eco-efficiency: A metafrontier directional distance function approach using life cycle analysis. Environ. Impact Assess. Rev. 2017, 63, 116–127. [Google Scholar] [CrossRef]
Alizadeh, S.; Vali, F.; Vatani, Z.; Avami, A. Sustainable analysis of Waste-to-Energy systems in cities by eco-efficiency assessment using DEA approach: A case study of Iran’s municipalities. Sustain. Cities Soc. 2023, 98, 104825. [Google Scholar] [CrossRef]
Guerrini, A.; Carvalho, P.; Romano, G.; Marques, R.C.; Leardini, C. Assessing efficiency drivers in municipal solid waste collection services through a nonparametric method. J. Clean. Prod. 2017, 147, 431–441. [Google Scholar] [CrossRef]
Exposito, A.; Velasco, F. Municipal solid-waste recycling market and the European 2020 Horizon Strategy: A regional efficiency analysis in Spain. J. Clean. Prod. 2018, 172, 938–948. [Google Scholar] [CrossRef]
lo Storto, C. Measuring ecoefficiency of municipal solid waste management in Apulia to account for governance heterogeneities. Clean. Waste Syst. 2024, 7, 100131. [Google Scholar] [CrossRef]
Díaz-Villavicencio, G.; Didonet, S.R.; Dodd, A. Influencing factors of eco-efficient urban waste management: Evidence from Spanish municipalities. J. Clean. Prod. 2017, 164, 1486–1496. [Google Scholar] [CrossRef]
Sala-Garrido, R.; Mocholi-Arce, M.; Molinos-Senante, M.; Maziotis, A. Measuring technical, environmental and eco-efficiency in municipal solid waste management in Chile. Int. J. Sustain. Eng. 2022, 15, 71–85. [Google Scholar] [CrossRef]
Salazar-Adams, A. The efficiency of municipal solid waste collection in Mexico. Waste Manag. 2021, 133, 71–79. [Google Scholar] [CrossRef]
Delgado-Antequera, L.; Gémar, G.; Molinos-Senante, M.; Gómez, T.; Caballero, R.; Sala-Garrido, R. Eco-efficiency assessment of municipal solid waste services: Influence of exogenous variables. Waste Manag. 2021, 130, 136–146. [Google Scholar] [CrossRef]
Pérez-López, G.; Prior, D.; Zafra-Gómez, J.L.; Plata-Díaz, A.M. Cost efficiency in municipal solid waste service delivery. Alternative management forms in relation to local population size. Eur. J. Oper. Res. 2016, 255, 583–592. [Google Scholar] [CrossRef]
Esteve, M.; Aparicio, J.; Rabasa, A.; Rodriguez-Sala, J.J. Efficiency analysis trees: A new methodology for estimating production frontiers through decision trees. Expert Syst. Appl. 2020, 162, 113783. [Google Scholar] [CrossRef]
Maziotis, A.; Molinos-Senante, M. Understanding energy performance in drinking water treatment plants using the efficiency analysis tree approach. npj Clean Water 2024, 7, 13. [Google Scholar] [CrossRef]
Maziotis, A.; Molinos-Senante, M.; Sala-Garrido, R.; Mocholi-Arce, M. Evaluation of dynamic eco-efficiency of water companies: The influence of non-revenue water and water supply interruptions. npj Clean Water 2023, 6, 20. [Google Scholar] [CrossRef]
Esteve, M.; Rodriguez-Sala, J.J.; Lopez-Espin, J.J.; Aparicio, J. Heuristic and Backtracking Algorithms for Improving the Performance of Efficiency Analysis Trees. IEEE Access 2021, 9, 17421–17428. [Google Scholar] [CrossRef]
Esteve, M.; Aparicio, J.; Rodriguez-Sala, J.J.; Zhu, J. Random forests and the measurement of super efficiency in the context of Free Disposal Hull. Eur. J. Oper. Res. 2022, 304, 729–744. [Google Scholar] [CrossRef]
Maziotis, A.; Sala-Garrido, R.; Mocholi-Arce, M.; Molinos-Senante, M. A comprehensive assessment of energy efficiency of wastewater treatment plants: An efficiency analysis tree approach. Sci. Total Environ. 2023, 885, 163539. [Google Scholar] [CrossRef]
Maziotis, A.; Sala-Garrido, R.; Mocholi-Arce, M.; Molinos-Senante, M. Carbon efficiency analysis in the provision of drinking water: Estimation of optimal greenhouse gas emissions. J. Clean. Prod. 2023, 392, 136304. [Google Scholar] [CrossRef]
Mehdiloo, M.; Podinovski, V.V. Selective strong and weak disposability in efficiency analysis. Eur. J. Oper. Res. 2019, 276, 1154–1169. [Google Scholar] [CrossRef]
Maziotis, A.; Molinos-Senante, M. A Comprehensive eco-efficiency analysis of wastewater treatment plants: Estimation of optimal operational costs and greenhouse gas emissions. Water Res. 2023, 243, 120354. [Google Scholar] [CrossRef]
Molinos-Senante, M.; Maziotis, A.; Sala-Garrido, R.; Mocholí-Arce, M. Factors influencing eco-efficiency of municipal solid waste management in Chile: A double-bootstrap approach. Waste Manag. Res. 2023, 41, 457–466. [Google Scholar] [CrossRef]
Llanquileo-Melgarejo, P.; Molinos-Senante, M. Evaluation of economies of scale in eco-efficiency of municipal waste management: An empirical approach for Chile. Environ. Sci. Pollut. Res. 2021, 28, 28337–28348. [Google Scholar] [CrossRef]
Valenzuela-Levi, N. Factors influencing municipal recycling in the global south: The case of chile. Resour. Conserv. Recycl. 2019, 150, 104441. [Google Scholar] [CrossRef]
MMA. Waste in Chile. 2023. Available online: https://sinia.mma.gob.cl/wp-content/uploads/2022/01/C10-residuos-rema-2021.pdf (accessed on 20 November 2024).
SINIM. National System of Municipal Information. 2024. Available online: https://www.sinim.gov.cl/ (accessed on 20 November 2024).
Sarra, A.; Mazzocchitti, M.; Rapposelli, A. Evaluating joint environmental and cost performance in municipal waste management systems through data envelopment analysis: Scale effects and policy implications. Ecol. Indic. 2017, 73, 756–771. [Google Scholar] [CrossRef]
Ríos, A.-M.; Picazo-Tadeo, A.J. Measuring environmental performance in the treatment of municipal solid waste: The case of the European Union-28. Ecol. Indic. 2021, 123, 107328. [Google Scholar] [CrossRef]
Doussoulin, J.P.; Colther, C. Evaluating the Efficiency of Municipal Solid Waste Collection Services in Developing Countries: The Case of Chile. Sustainability 2022, 14, 15887. [Google Scholar] [CrossRef]
das Mercês Costa, I.; Dias, M.F.; Robaina, M. Evaluation of the efficiency of urban solid waste management in Brazil by data envelopment analysis and possible variables of influence. Waste Dispos. Sustain. Energy 2024, 6, 283–295. [Google Scholar] [CrossRef]
Llanquileo-Melgarejo, P.; Molinos-Senante, M.; Romano, G.; Carosi, L. Evaluation of the Impact of Separative Collection and Recycling of Municipal Solid Waste on Performance: An Empirical Application for Chile. Sustainability 2021, 13, 2022. [Google Scholar] [CrossRef]
Molinos-Senante, M.; Maziotis, A.; Sala-Garrido, R.; Mocholí-Arce, M. The eco-efficiency of municipalities in the recycling of solid waste: A stochastic semi-parametric envelopment of data approach. Waste Manag. Res. 2023, 41, 1036–1045. [Google Scholar] [CrossRef]
Pérez-López, G.; Prior, D.; Zafra-Gómez, J.L. Temporal scale efficiency in DEA panel data estimations. An application to the solid waste disposal service in Spain. Omega 2018, 76, 18–27. [Google Scholar] [CrossRef]

Figure 1. Efficiency analysis tree (EAT) for estimating optimal operational costs and unsorted waste, where Id denotes the node; n(t) shows the number of observations; y₁ is the maximum operating costs in USD per year; and y₂ is the maximum level of unsorted waste in tons per year.

Figure 2. Statistics of eco-efficiency scores in the provision of municipal solid waste services for assessed municipalities.

Figure 3. Potential reduction of unsorted waste expressed in tons per year and percentage in relation to total municipal solid waste generated.

Figure 4. Potential economic savings in managing MSW expressed in USD per year and percentage in relation to total operational costs.

Table 1. Groups for Kruskal–Wallis test.

Variable	Group 1	Group 2	Group 3
Population served	<30,000 people	30,000–96,000 people	>96,000 people
Municipality size	<200 km²	200–1000 km²	>1000 km²
Population density	<50 people/km²	50–1000 people/km²	>1000 people/km²
MSW per capita	<350 kg/inhabitant·year	350–500 kg/inhabitant·year	>500 kg/inhabitant·year

Table 2. Descriptive statistics of variables for eco-efficiency assessment. Source: SINADER and SINIM databases.

Variables	Unit of Measurement	Mean	St. Dev.	Minimum	Maximum
Collected and recycled waste	Tons/year	3696	11,434	0.01	97,290
OPEX	USD/year	2108	2701	0	17,371
Unsorted waste	Tons/year	40,285	56,634	0.05	466,777
Population served	People	96,939	116,137	4507	633,021
Municipality size	km²	986	1549	7	10,863
Population density	People/km²	1743	3938	3	18,221
MSW per capita	kg/year·person	452	214	107	1495

Total observations for 2015–2019: 490.

Table 3. Average eco-efficiency scores and Kruskal–Wallis test results according to assessed exogenous variables.

Variables and Groups	Number of Municipalities	Average Eco-Efficiency	p-Value Kruskal–Wallis Test
Number of inhabitants
<30,000	37	0.564	0.025
30,000–96,000	26	0.560
>96,000	35	0.558
Municipality size (km²)
<200	33	0.562	0.090
200–1000	34	0.563
>1000	31	0.557
Population density (inhabitants/km²)
<50	38	0.561	0.120
50–1000	37	0.562
>1000	23	0.557
Generation of MSW (Kg/inhabitant·year)
<350	29	0.557	0.040
350–500	42	0.561
>500	27	0.563

Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

© 2024 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/).

Share and Cite

MDPI and ACS Style

Sala-Garrido, R.; Mocholi-Arce, M.; Molinos-Senante, M.; Maziotis, A. Applying the Efficiency Analysis Tree Method for Enhanced Eco-Efficiency in Municipal Solid Waste Management: A Case Study of Chilean Municipalities. Clean Technol. 2024, 6, 1565-1578. https://doi.org/10.3390/cleantechnol6040075

AMA Style

Sala-Garrido R, Mocholi-Arce M, Molinos-Senante M, Maziotis A. Applying the Efficiency Analysis Tree Method for Enhanced Eco-Efficiency in Municipal Solid Waste Management: A Case Study of Chilean Municipalities. Clean Technologies. 2024; 6(4):1565-1578. https://doi.org/10.3390/cleantechnol6040075

Chicago/Turabian Style

Sala-Garrido, Ramon, Manuel Mocholi-Arce, Maria Molinos-Senante, and Alexandros Maziotis. 2024. "Applying the Efficiency Analysis Tree Method for Enhanced Eco-Efficiency in Municipal Solid Waste Management: A Case Study of Chilean Municipalities" Clean Technologies 6, no. 4: 1565-1578. https://doi.org/10.3390/cleantechnol6040075

APA Style

Sala-Garrido, R., Mocholi-Arce, M., Molinos-Senante, M., & Maziotis, A. (2024). Applying the Efficiency Analysis Tree Method for Enhanced Eco-Efficiency in Municipal Solid Waste Management: A Case Study of Chilean Municipalities. Clean Technologies, 6(4), 1565-1578. https://doi.org/10.3390/cleantechnol6040075

Article Menu

Applying the Efficiency Analysis Tree Method for Enhanced Eco-Efficiency in Municipal Solid Waste Management: A Case Study of Chilean Municipalities

Abstract

1. Introduction

2. Methodology

2.1. Methodology for Eco-Efficiency Estimation

2.2. Potential Exogenous Variables Influencing Eco-Efficiency

3. Data and Sample Selection

4. Results and Discussion

4.1. Optimal Operational Expenditure and Unsorted Waste Estimation

4.2. Eco-Efficiency Estimation

4.3. Exogenous Variables Influencing Eco-Efficiency

5. Conclusions

Author Contributions

Funding

Institutional Review Board Statement

Informed Consent Statement

Data Availability Statement

Conflicts of Interest

Abbreviation List

References

Share and Cite

Article Metrics

Article Access Statistics

Further Information

Guidelines

MDPI Initiatives

Follow MDPI