The impact of recycling on the productivity of Municipal Solid Waste: a comparison of the Malmquist and Malmquist-Luenberger Productivity Indexes

ABSTRACT

Enhancing the economic and environmental efficiency of municipal solid waste (MSW) management is crucial for transitioning to a circular economy and achieving cost-efficient and sustainable waste practices. In this study, we introduce an innovative approach to examine the influence of MSW recycling on waste performance over multiple years. We employ a non-parametric method to assess and compare the Malmquist Productivity Index (MPI) and the Malmquist-Luenberger Productivity Index (MLPI) across 143 municipalities in Chile from 2015 to 2019. The findings reveal a positive trend in both MPI and MLPI. The average eco-productivity change (MLPI) is relatively lower than the average productivity change (MPI) throughout the evaluated years. The aggregated scores for MPI and MLPI between 2015 and 2019 stand at 2.73 and 1.33, respectively. The primary driver of productivity change is attributed to technical advancements, leading to a shift in the efficient frontier, while eco-productivity change predominantly arises from efficiency enhancements. This implies that municipalities have employed diverse strategies to enhance their economic and environmental performance. The results and implications of this study hold significance for policymakers at regional and local levels, offering valuable insights for optimising MSW management within the context of a circular economy.

KEYWORDS:

• Waste management
• productivity change
• eco-productivity change
• Malmquist Productivity Index (MPI)
• Malmquist-luenberger Productivity index (MPI)
• circular economy

1. Introduction

The alarming and escalating rise in the generation of municipal solid waste (MSW) has emerged as a significant global concern, primarily attributed to its detrimental impacts on the environment and public health (Guerrini et al. 2017; Soni et al. 2022). Recognising the gravity of this issue, the United Nations has acknowledged and emphasised the significance of enhancing MSW management through Goals 11 and 12 of the Sustainable Development Goals (Al-Dailami et al. 2022; Di Foggia and Beccarello 2018).

In December 2019, the European Green Pact was introduced as a roadmap to modernise the European economy, enhancing its efficiency and competitiveness in resource utilisation (The European Green Deal 2019; Romano et al. 2020). As part of the European Green Pact (adopted in 2020), the new Action Plan for the circular economy was approved, encompassing measures to encourage businesses, public administrations, and consumers to adopt a sustainable model. Consequently, European countries have made progress in implementing the circular economy. There exists a notable disparity in other regions, such as Latin America and the Caribbean, regarding the adoption of the circular economy (Mihai et al. 2022). This highlights the significance of contributing to this study, specifically focusing on waste management in Latin American countries.

Literature highlights the relevance of the efficiency assessment of MSW management in the transition towards a circular economy and the fulfilment of its objectives (Cerciello, Agovino, and Garofalo 2019; Llanquileo-Melgarejo and Molinos-Senante 2021). Evaluating the performance in the provision of MSW services allows municipalities to determine the most cost-effective strategies for handling solid waste. It provided valuable data for policy development and decision-making. Municipalities can use this information to formulate effective waste management policies, set goals and targets, allocate resources, and track progress towards sustainability objectives (Llanquileo-Melgarejo and Molinos-Senante 2021; Romano, Masserini, and Lombardi 2021).

Some studies have evaluated the efficiency of units (countries, regions, and municipalities) in providing MSW services (Gastaldi et al. 2020; Song et al. 2012; Struk and Boda 2022). Most research on this topic has used economic variables as inputs and waste collected as outputs to assess the performance of units and economic efficiency (Halkos and Petrou 2019; Romano et al. 2020). Considering the relevance of moving towards a circular economy, the concept of eco-efficiency is being used incipiently to evaluate the performance of municipalities in the provision of MSW services (Exposito and Velasco 2018; Romano, Rapposelli, and Marrucci 2019). Eco-efficiency is defined as the production of more goods (products) and services with fewer resources (inputs) and less environmental impact (Beltrán-Esteve, Reig-Martínez, and Estruch-Guitart 2017). The prefix”eco” represents environmental and economic issues; the eco-efficiency assessment provides relevant information from an economic and environmental perspective (Lo Storto 2021).

Eco-efficiency provides information about the performance of the units analysed for a specific time, i.e. it is a static evaluation that cannot account for changes in performance over time (Gémar et al. 2018). This assessment might be extended using methods and models that allow evaluating the performance of units over time, i.e. evaluating the eco-productivity change of units (Mahlberg, Luptacik, and Sahoo 2011). The assessment of eco-productivity change involves extending the notion of eco-efficiency to a temporal setting. Assessing changes in eco-productivity allows one to compute the eco-efficiency of a unit for any given period and helps to compare the eco-efficiency among units. Obtaining information about the temporal dynamics of eco-efficiency, i.e. eco-productivity change, is essential to support decision-making and enhance the circular economy (Romano, Masserini, and Lombardi 2021).

Despite the crucial implications of assessing the change in eco-productivity for significant public decisions, to our knowledge, only Romano et al., (2021) and Lo Storto (2021) have conducted evaluations on the eco-productivity change in a sample of large cities regarding solid waste services. On one hand, Romano et al. (2021) focused on comparing the eco-productivity change among municipalities with publicly, privately, and mixed-owned entrusted waste utilities, utilising the Metafrontier Malmquist-Luenberger productivity index. On the other hand, Lo Storto (2021) applied the Global Malmquist Productivity Index to assess the eco-productivity change in the provision of MSW services in a sample of Italian municipalities from 2010 to 2017. In the context of a circular economy, it is imperative to evaluate the impact of changes in MSW recycling rates on the performance of municipalities. In other words, it is essential to compare the metrics of productivity change (which do not integrate environmental variables) and eco-productivity change (which differentiates between recycled waste and unsorted waste) in the provision of MSW services by municipalities.

Against this background, the main objective is to evaluate the impact of MSW recycling on the dynamic performance of municipalities in the provision of MSW services by estimating and comparing their productivity and eco-productivity change. Both metrics of dynamic performance can be decomposed into technical change and efficiency change, which is why the approach in this study allowed us to identify the main drivers of both productivity and eco-productivity change. This information is essential for designing specific policies to enhance MSW recycling and improve the performance of municipalities.

The significance of this study to the existing body of literature can be summarised as follows: This paper stands out as one of the limited number of studies that examine the dynamics of performance in MSW management. As far as our knowledge extends, this study is the first to compare estimations of productivity and eco-productivity. It introduces a novel approach by examining the impact of recycling rates on the productivity of MSW management. Previous research on this subject has predominantly focused on developed countries, where circular economy concepts and MSW recycling policies are well-established. In contrast, our case study centres on Chile, a middle-income country with emerging regulations to promote MSW recycling. Consequently, the insights gained from our research in Chile could offer valuable lessons for other developing nations.

2. Material and methods

This analysis was divided into two steps: First, the productivity change in the provision of MSW services in Chilean municipalities was evaluated through the Malmquist Productivity Index (MPI) (Caves, Christensen, and Diewert 1982). In the second step, environmental variables were included in the assessment by estimating the eco-productivity change using the Malmquist-Luenberger Productivity Index (MLPI) (Chung, Färe, and Grosskopf 1997).

2.1 Productivity change estimation: Malmquist Productivity Index (MPI)

First, a set of production possibilities that describes feasible relationships between inputs and outputs is presented. Assuming there are $k=1,\dots ,K$ municipalities during $t=1,\dots ,T$ periods, the production technology for municipalities producing $M$ desirable outputs, $y\in R_{+}^M$, by using $N$ inputs, $X\in R_{+}^N$, is represented by the possibility set $T\left(x\right)$, which is as follows:

(1) $T\left(x\right)=\left\{\left(x,y\right):xcanproducey\right\}$(1)

The output distance function is defined as follows:

(2) $D_0\left(x,y\right)=min\left\{\theta :x,\frac{y}{\theta }\in T\left(x\right)\right\}$(2)

The productivity change for each unit (municipality) is estimated based on the MPI, which is defined as follows (Caves, Christensen, and Diewert 1982):

(3) $MPI\left(y_t,x_{t,}{y}_{t+1,}{x}_{t+1}\right)=\sqrt{\frac{D_0^t\left(y_{t+1,}{x}_{t+1}\right)+D_0^{t+1}\left(y_{t+1,}{x}_{t+1}\right)}{D_0^t\left(y_{t,}{x}_t\right)+D_0^{t+1}\left(y_{t,}{x}_t\right)}}$(3)

Productivity change is decomposed into efficiency change ($MECH$) and technical change ($MTCH$) as follows:

(4) $MPI\left(y_t,x_{t,}{y}_{t+1,}{x}_{t+1}\right)=\frac{D_0^t\left(y_{t+1,}{x}_{t+1}\right)}{D_0^t\left(y_{t,}{x}_t\right)}\sqrt{\frac{D_0^t\left(y_{t+1,}{x}_{t+1}\right)+D_0^{t+1}\left(y_{t+1,}{x}_{t+1}\right)}{D_0^t\left(y_{t,}{x}_t\right)+D_0^{t+1}\left(y_{t,}{x}_t\right)}}=MECH\ast MTCH$(4)

The advantage of this decomposition is that the drivers that contribute the most to each municipality’s productivity change (positive or negative) are identified. Efficiency change component informs about the movement of units (municipalities) towards the best practice frontier (Kumar 2006). This component is related to the ability of units to be managed by the best operating practices. On the other hand, the technical change component measures the change in the efficient frontier between the two periods (Kumar 2006). The measurement of these components is essential for strategic and effective long-term planning.

A municipality has improved its productivity over time if MPI > 1; whereas it has suffered retardation in productivity if MPI < 1. An MPI = 1 means that the productivity is constant over time. For the components of the MPI, i.e. MECH and MLTCH, the same interpretation applies.

The MPI can be calculated using a non-parametric methodology, such as the data envelopment analysis (DEA). The following models should be solved for each unit evaluated to compute productivity indicators:

${\bar{D}}_0^t\left(x_{k^{\prime }}^t,y_{k^{\prime }}^t\right),\hspace{0.17em}{\bar{D}}_0^t\left(x_{k^{\prime }}^{t+1},y_{k^{\prime }}^{t+1}\right),\hspace{0.17em}{\bar{D}}_0^{t+1}\left(x_{k^{\prime }}^{t+1},y_{k^{\prime }}^{t+1}\right)\mathrm{a}\mathrm{n}\mathrm{d}{\bar{D}}_0^{t+1}\left(x_{k^{\prime }}^t,y_{k^{\prime }}^t\right):$

(5) ${\bar{D}}_0^t\left(x_{k^{\prime }}^t,y_{k^{\prime }}^t\right)=Max\beta$(5)

$s.t.:$

${\sum }_{k=1}^K{\lambda }_k^t{y}_{km}^t\ge \left(\beta \right)y_{k^{\prime }m}^t,m=1,2,\dots ,M$

${\sum }_{k=1}^K{\lambda }_k^t{x}_{kn}^t\le x_{k^{\prime }n}^t,n=1,2,\dots ,N$

${\lambda }_k^t\ge 0,k=1,2,\dots ,K$

(6) ${\bar{D}}_0^t\left(x_{k^{\prime }}^{t+1},y_{k^{\prime }}^{t+1}\right)=Max\beta$(6)

$s.t.:$

${\sum }_{k=1}^K{\lambda }_k^t{y}_{km}^t\ge \left(\beta \right)y_{k^{\prime }m}^{t+1},m=1,2,\dots ,M$

${\sum }_{k=1}^K{\lambda }_k^t{x}_{kn}^t\le x_{k^{\prime }n}^{t+1},n=1,2,\dots ,N$

${\lambda }_k^t\ge 0,k=1,2,\dots ,K$

(7) ${\bar{D}}_0^{t+1}\left(x_{k^{\prime }}^{t+1},y_{k^{\prime }}^{t+1}\right)=Max\beta$(7)

$s.t.:$

${\sum }_{k=1}^K{\lambda }_k^{t+1}{y}_{km}^{t+1}\ge \left(1+\beta \right)y_{k^{\prime }m}^{t+1},m=1,2,\dots ,M$

${\sum }_{k=1}^K{\lambda }_k^{t+1}{x}_{kn}^{t+1}\le x_{k^{\prime }n}^{t+1},n=1,2,\dots ,N$

${\lambda }_k^{t+1}\ge 0,k=1,2,\dots ,K$

(8) ${\bar{D}}_0^{t+1}\left(x_{k^{\prime }}^t,y_{k^{\prime }}^t\right)=Max\beta$(8)

$s.t.:$

${\sum }_{k=1}^K{\lambda }_k^{t+1}{y}_{km}^{t+1}\ge \left(1+\beta \right)y_{k^{\prime }m}^t,m=1,2,\dots ,M$

${\sum }_{k=1}^K{\lambda }_k^{t+1}{x}_{kn}^{t+1}\le x_{k^{\prime }n}^t,n=1,2,\dots ,N$

${\lambda }_k^{t+1}\ge 0,k=1,2,\dots ,K$

where $M$ is the number of outputs generated, $N$ is the number of inputs used, $K$ is the number of units evaluated and ${\lambda }_k$ is a set of intensity variables.

2.2 Eco-productivity change estimation: Malmquist-Luenberger Productivity Index (MLPI)

The evaluation of the eco-productivity change of the evaluated units was based on the estimation of the MLPI by (Chung, Färe, and Grosskopf 1997). This approach used the directional distance function, which does the following: i) seeks to increase the desirable outputs while decreasing the undesirable outputs and ii) avoids computational problems associated with the calculation of the output efficiency as a solution to non-linear programming problems.

The notation used to estimate the change in the eco-productivity of the municipalities is similar to that described in Section 2.1 for the MPI. The main difference is that in the MLPI methodological approach, the vector of outputs is decomposed into desirable outputs, $y\in N_{+}^M$ (recycled waste), and undesirable outputs, $b\in N_{+}^M,$ (unsorted waste).

The directional distance function, including undesirable outputs, is defined as:

(9) $\overrightarrow{D_o}\left(x,y,b;g\right)=sup\left\{\beta :\left(y,b\right)+\beta g\in P\left(x\right)\right\}$(9)

where $g$ is the vector of directions in which outputs are scaled, with $g=\left(g^y,-g^b\right)$ being $g^y$∈ $R_{+}^M$ and $g^b$∈ $R_{+}^H$. In this case, $g=\left(1,-1\right)$ assumes that the desired outputs are increased and undesirable outputs are decreased.

Considering $t=1,\dots ,T$ periods, the MLPI is defined as follows:

(10) $MLP{I}_t^{t+1}={\left[\frac{\left(1+{\underset{-}{D}}_0^t\left(x^t,y^t,b^t;g^t\right)\right)}{\left(1+{\underset{-}{D}}_0^t\left(x^{t+1},y^{t+1},b^{t+1};g^{t+1}\right)\right)}\ast \text{break}\frac{\left(1+{\underset{-}{D}}_0^{t+1}\left(x^t,y^t,b^t;g^t\right)\right)}{\left(1+{\underset{-}{D}}_0^{t+1}\left(x^{t+1},y^{t+1},b^{t+1};g^{t+1}\right)\right)}\right]}^{\frac{1}{2}}$(10)

The MLPI can be decomposed into two variables: the efficiency change (MLECH) and the technical change (MLTCH) (Ananda 2018):

(11) $MLEC{H}_t^{t+1}=\left[\frac{(1+D_0^t\left(x^t+y^t{+}_{\hspace{0.17em}}{b}^t;g^t\right)}{(1+D_0^t\left(x^{t+1}+y^{t+1}{+}_{\hspace{0.17em}}{b}^{t+1};g^{t+1}\right)}\ast \frac{(1+D_0^{t+1}\left(x^t+y^t{+}_{\hspace{0.17em}}{b}^t;g^t\right)}{(1+D_0^{t+1}\left(x^{t+1}+y^{t+1}{+}_{\hspace{0.17em}}{b}^{t+1};g^{t+1}\right)}\right]$(11)

(12) $MLTC{H}_t^{t+1}=\left[\frac{(1+D_0^{t+1}\left(x^t+y^t{+}_{\hspace{0.17em}}{b}^t;g^t\right)}{(1+D_0^t\left(x^t+y^t{+}_{\hspace{0.17em}}{b}^t;g^t\right)}\ast \frac{(1+D_0^{t+1}\left(x^{t+1}+y^{t+1}{+}_{\hspace{0.17em}}{b}^{t+1};g^{t+1}\right)}{(1+D_0^{t+1}\left(x^{t+1}+y^{t+1}{+}_{\hspace{0.17em}}{b}^{t+1};g^{t+1}\right)}\right]$(12)

A municipality has improved its eco-productivity in the provision of MSW services if MLPI > 1, whereas an MLPI < 1 means worsening of the eco-productivity; an MLPI = 1 means that the productivity is unchanged. For the components of the MLPI, i.e. MLECH and MLTCH, the same interpretation applies (Maziotis, Molinos-Senante, and Sala-Garrido 2017).

Based on the definition of the MLPI (Eq. 10), four linear-programming problems must be solved for each unit (municipality) to compute the MLPI. The problems are shown in Eqs. (13—16).

(13) ${\bar{D}}_0^t\left(x_{k^{\prime }}^t,y_k^t,b_{k^{\prime }}^t;g_{k^{\prime }}^t\right)=Max\beta$(13)

$s.t.:$

${\sum }_{k=1}^K{\lambda }_k^t{y}_{km}^t\ge \left(1+\beta \right)y_{k^{\prime }m}^t,m=1,2,\dots ,M$

${\sum }_{k=1}^K{\lambda }_k^t{b}_{ki}^t=\left(1-\beta \right)b_{k^{\prime }i}^t,i=1,2,\dots ,I$

${\sum }_{k=1}^K{\lambda }_k^t{x}_{kn}^t\le x_{k^{\prime }n}^t,n=1,2,\dots ,N$

${\lambda }_k^t\ge 0,k=1,2,\dots ,K$

(14) ${\bar{D}}_0^{t+1}\left(x_{k^{\prime }}^{t+1},y_{k^{\prime }}^{t+1},b_{k^{\prime }}^{t+1};g_{k^{\prime }}^t\right)=Max\beta$(14)

$s.t.:$

${\sum }_{k=1}^K{\lambda }_k^{t+1}{y}_{km}^{t+1}\ge \left(1+\beta \right)y_{k^{\prime }m}^{t+1},m=1,2,\dots ,M$

${\sum }_{k=1}^K{\lambda }_k^{t+1}{b}_{ki}^{t+1}=\left(1-\beta \right)b_{k^{\prime }i}^{t+1},i=1,2,\dots ,I$

${\sum }_{k=1}^K{\lambda }_k^{t+1}{x}_{kn}^{t+1}\le x_{k^{\prime }n}^{t+1},n=1,2,\dots ,N$

${\lambda }_k^{t+1}\ge 0,k=1,2,\dots ,K$

(15) ${\bar{D}}_0^{t+1}\left(x_{k^{\prime }}^t,y_{k^{\prime }}^t,b_{k^{\prime }}^t;g_{k^{\prime }}^t\right)=Max\beta$(15)

$s.t.:$

${\sum }_{k=1}^K{\lambda }_k^{t+1}{y}_{km}^{t+1}\ge \left(1+\beta \right)y_{k^{\prime }m}^t,m=1,2,\dots ,M$

${\sum }_{k=1}^K{\lambda }_k^{t+1}{b}_{ki}^{t+1}=\left(1-\beta \right)b_{k^{\prime }i}^t,i=1,2,\dots ,I$

${\sum }_{k=1}^K{\lambda }_k^{t+1}{x}_{kn}^{t+1}\le x_{k^{\prime }n}^t,n=1,2,\dots ,N$

${\lambda }_k^{t+1}\ge 0,k=1,2,\dots ,K$

(16) ${\bar{D}}_0^t\left(x_k^{t+1},y_k^{t+1},b_k^{t+1};g_k^t\right)=Max\beta$(16)

$s.t.:$

${\sum }_{k=1}^K{\lambda }_k^t{y}_{km}^t\ge \left(1+\beta \right)y_{k^{\prime }m}^{t+1},m=1,2,\dots ,M$

${\sum }_{k=1}^K{\lambda }_k^t{b}_{ki}^t=\left(1-\beta \right)b_{k^{\prime }i}^{t+1},i=1,2,\dots ,I$

${\sum }_{k=1}^K{\lambda }_k^t{x}_{kn}^t\le x_{k^{\prime }n}^{t+1},n=1,2,\dots ,N$

${\lambda }_k^t\ge 0,k=1,2,\dots ,K$

where $M$ is the number of desirable generated outputs, $I$ is the number of undesirable outputs produced, $N$ is the number of used inputs, $K$ is the number of evaluated units and ${\lambda }_k$ is a set of intensity variables.

2.3 Case analysis

The empirical application in this study utilised a sample of 143 out of 345 Chilean municipalities (41.4%) during the period of 2015 to 2019. 22 out of the 143 analysed municipalities can be categorised as small because its population is lower than 10,000 people. In recent years, Chile has implemented several policies aimed at enhancing the environmental management of MSW. For instance, the National Plan for Sustainable Consumption and Production 2017–2022 emphasises the development, implementation, and strengthening of mechanisms to prevent waste generation and promote the valorisation of waste across all sectors of the economy through financial and educational tools that incorporate concepts like eco-design and circular economy. Additionally, the National Waste Policy 2015–2025 has been instrumental in shaping waste management practices.

From an institutional standpoint, the Ministry of the Environment has established a Circular Economy Office, which focuses on research and innovation programmes (such as Eco-design) and material recovery in collaboration with research centres, universities, and the Agency for Sustainability and Climate Change. Concerning the management of solid waste in Chile, Law 20,920 was approved in 2016. This legislation established a framework for waste management, extended producer responsibility, and the promotion of recycling. Its objectives include reducing waste generation, increasing recovery, reuse, and recycling, as well as safeguarding human and environmental health. The law requires producers to take responsibility for the processing and valorisation of products throughout their lifecycle. The implementation of this law was initiated in mid-2020. As noted by (Valenzuela-Levi et al. 2021), the rates of separate collection and recycling of MSW in Chile still remain relatively low.

The evaluation of the productivity and eco-productivity change is based on the previous research (Marques and Simões 2009; Simões, De Witte, and Marques 2010). The input is the total annual cost of the provision of MSW services. It is defined as the expenditure by each municipality in the provision of MSW services, including cleaning services, waste collection and waste treatment or disposal. The total annual cost is expressed in Chilean pesos per year.¹ This variable was collected from the National Municipal Information System (SINIM) for 2015—2019.

Regarding outputs, in the case of productivity change, only the volume of generated MSW was integrated into the model and expressed in metric tons per year. By contrast, to evaluate eco-productivity change, both undesirable and desirable outputs were integrated into the assessment. Per the existing studies on the eco-efficiency of MSW management (Llanquileo-Melgarejo and Molinos-Senante 2021; Sarra, Mazzocchitti, and Rapposelli 2017), unsorted waste (tons/year) was considered undesirable output. Additionally, a set of recyclable wastes (tons/year) was considered as desirable output: (i) paper and paperboard, (ii) glass, (iii) plastic, (iv) organic waste and (v) other inorganic waste. Data about recyclable and unsorted waste were collected from the National System of Waste Declaration (SINADER in Spanish).

Due to many units analysed in this study (143 municipalities during 2015ꟷ2019), atypical values were detected based on the average values and the standard deviation of the sample (Wilson 1993). The detection of outliers is essential to evaluate the performance of municipalities using non-parametric methods such as the DEA since it is a deterministic method. Therefore, outliers can affect productivity and eco-productivity results since they act as pairs of other units (Ananda 2019). Table 1 presents the average of the variables used to assess the productivity and eco-productivity change in terms of the provision of MSW services among 143 Chilean municipalities.

Table 1. Descriptive statistics of the variables used for 2015 and 2019 for 143 Chilean municipalities.

Year	Total anual costs (CLP/year)	Paper and paperboard (tons/year)	Glass (tons/year)	Plastics (tons/year)	Organic waste (tons/year)	Other inorganic waste (tons/year)	Unsorted waste (tons/year)
2015	1,039,630	182	75	18	817	712	26,205
2019	2,299,943	3,853	411	77	1,470	6,304	45,202
	Total anual costs (CLP per capita)	Paper and paperboard (ton per capita)	Glass (ton per capita)	Plastics(ton per capita) (ton per capita)	Organic waste (ton per capita)	Other inorganic waste (ton per capita)	Unsorted waste (ton per capita)
2015	12762	2.23	0.92	0.22	10.03	8.74	321.69
2019	17250	6.02	2.30	0.28	3.70	14.04	350.20

3. Results and discussion

This section reports the estimates of productivity and eco-productivity change based on the MPI and MLPI metrics, which were calculated using the MAX-DEA software. In other words, it shows the impact of changes in recycling rates over time on the performance of Chilean municipalities in terms of the provision of MSW services.

Figure 1 shows the average values of MPI and MLPI for a sample of 143 Chilean municipalities. The average MPI was larger than the MLPI for all evaluated years. It means that when environmental variables are included in the assessment, municipalities are penalised in terms of their performance. Since the average eco-productivity change was found to be lower than the average productivity change, it suggests that recycling MSW harms the economics of municipalities. In the case of Chile, where around 50% of households do not pay for MSW services (CSP 2020), municipalities have to make additional economic efforts for recycling MSW services. It is essential to implement policies to increase the municipal revenue associated with MSW services by implementing alternative tariff schemes, such as pay-as-you-throw or volume-based waste fees as a part of the application of the ‘polluter pays principle’ (Drosi et al. 2020; Ukkonen and Sahimaa 2021). Alternatively, for the 22 small municipalities evaluated in this study, which would present difficulties in the implementation of tariff schemes, additional funds from regional and/or national entities should be received to increase the MSW recycling rates.

Figure 1. Average evolution of the productivity change (MPI) and eco-productivity change (MLPI) and its drivers efficiency change (MECH and MLECH) and technical change (MTCH and MLTCH) of the Chilean municipalities evaluated.

Figure 1 illustrates that both performance metrics, MPI and MLPI, are consistently higher than the index evaluated for all years. On average, Chilean municipalities have demonstrated improvement in their performance over the years from both an economic and environmental standpoint, reflecting the significant efforts made to enhance MSW management in recent times. This improvement is particularly notable in the most recent evaluated year (2018–2019), suggesting that two years after the implementation of the National Law for promoting MSW recycling (Law 20,920), Chilean municipalities have made strides in waste management, including recycling initiatives.

Further analysis of the results presented in Table 2 reveals substantial divergences among the evaluated municipalities. Specifically focusing on productivity change, there are significant variations between the minimum and maximum MPI scores for all the years examined. None of the evaluated Chilean municipalities have shown improvement in productivity over time. For instance, between 2018 and 2019, 29.4% of the municipalities (42 out of 143) did not witness any enhancement in their economic performance. The differences among municipalities become less significant when environmental variables are integrated into the assessment, i.e. when eco-productivity change is evaluated. Nevertheless, a noteworthy percentage of municipalities (32.9%) did not experience an improvement in their eco-productivity between 2018 and 2019, as well as in previous years. This indicates that certain municipalities need to reduce operational costs and/or increase their MSW recycling rates in order to match the efficiency levels of their peers.

Table 2. Main statistics of productivity change and eco-productivity change of Chilean municipalities evaluated.

	Malmquist Productivity Index				Malmquist-Luenberger Productivity Index
	Average	Minimum	Maximum	% municipalities improved	Average	Minimum	Maximum	% municipalities improved
2015/16	1.254	0.296	7.423	55.9	1.010	0.341	1.925	33.6
2016/17	1.161	0.086	6.109	58.0	1.029	0.234	5.446	42.0
2017/18	1.295	0.109	8.109	53.1	1.021	0.050	3.821	45.5
2018/19	3.018	0.009	9.039	70.6	2.750	0.585	4.466	67.1

From a geographical perspective, it is interesting to analyse the difference between the aggregated MPI and MLPI (values for the 143 Chilean municipalities calculated between 2015 and 2019 (Figures 2(a, b). The findings reveal that in 101 out of the 143 municipalities (70.6%), the aggregated values of MPI from 2015 to 2019 were higher than those for the MLPI metric. Over two-thirds of the analysed Chilean municipalities demonstrated better dynamic performance from an economic standpoint compared to an environmental perspective. These results differ from the conclusions drawn by (Llanquileo-Melgarejo and Molinos-Senante 2021), who stated that most Chilean municipalities exhibit higher eco-efficiency scores than efficiency scores. It should be noted that their study focused solely on 2018 data. Thus, our study highlights the importance of incorporating the temporal dimension when evaluating the performance of municipalities in terms of the provision of MSW services. Figures 2(a, b) demonstrate that there was no discernible geographical pattern observed in the comparison of the MPI and MLPI metrics. This finding aligns with the results reported by (Llanquileo-Melgarejo and Molinos-Senante 2021), indicating that the geographical factor does not determine the performance of Chilean municipalities in terms of MSW service provision.

Figure 2. (a and b) Difference between aggregated MPI and MLPI values for 2015/19 years for the Chilean municipalities evaluated.

Not a clear pattern is observed across Chilean municipalities in terms of productivity change and eco-productivity change.

Regarding the drivers of productivity and eco-productivity change estimations, the average values are depicted in Figure 1. In terms of economic dynamic performance metrics, specifically MPI, technological change (MTCH) was the primary factor contributing to the improvement in productivity change among Chilean municipalities, except for the period of 2017–2018 when it exceeded one. Throughout most of the evaluated years, there was a positive shift in the efficient frontier, particularly noticeable in the last evaluated year where MTCH improved by 4.215. Important to mention is the adoption of the Chilean Law 20,920 concerning waste management in 2016. Despite its gradual implementation over the years, it has led to significant changes in the primary waste management approaches in Chile. These changes are evident in the considerable positive shift observed in the estimated efficient frontier during the recent evaluated years.

This significant increase counterbalanced the considerable decrease in the efficiency change (MECH) driver, which experienced a decline of 29% during the 2018–2019 period. These results highlight notable variations among Chilean municipalities in terms of economic performance (MPI). While some municipalities showcased excellent performance, leading to a positive movement of the production frontier (MTCH >1), many others did not improve their efficiency and remained further away from the efficient frontier, resulting in a negative efficiency change.

The trend observed in the environmental dynamic performance metric, MLPI, was slightly different (refer to Figure 1). The driver of efficiency change (MLECH) had a negative impact on the eco-productivity change of municipalities during the first two evaluated years (2015–2017). Following the implementation of Law 20,920, which promotes MSW recycling in Chilean municipalities, the average MLECH became positive. Between 2017 and 2019, the overall performance of municipalities improved, bringing them closer to the eco-efficient production frontier. Similarly, to MPI, the driver of technical change (MLTCH) was positive for all years except for 2017–2018. This suggests that, on average, the eco-efficiency frontier of municipalities, in terms of MSW service provision, has undergone a positive shift, indicating that municipalities are making improvements in their daily operations. The average values of the MLTCH component were closer to one, indicating that these improvements are gradual and likely related to the availability of funds for implementing policies to promote MSW recycling at the household level. This could be attributed to initiatives such as the Recycling Fund (FPR), one of the competitive funds established by the Chilean Ministry of the Environment. The FPR aims to support Extended Producer Responsibility under Law 20,920 by financing projects that prevent waste generation, promote reuse, recycling, and other types of recovery. These funds primarily target municipalities and associations.

Figures 3(a, b) illustrate the difference between the aggregated values of MECH and MLECH from 2015 to 2019 at the municipal level. It was observed that 61 out of 143 Chilean municipalities (42.7%) exhibited higher aggregate MECH values than MLECH for the 2015–2019 period. This indicates that the majority of evaluated Chilean municipalities (57.3%) have prioritised improving their environmental performance rather than focusing on economic aspects. Any observed regional trends could potentially be attributed to the low recycling rates in Chile (Valenzuela-Levi et al. 2021).

Figure 3. (a and b) Difference between aggregated MECH and MLECH values for 2015/19 years for the Chilean municipalities evaluated.

Most of the municipalities located in the south of the country suffered aretardation of the efficiency change between 2015 and 2019.

Similarly, Figures 4(a, b) present the difference between the aggregated values of MTCH and MLTCH from 2015 to 2019 as drivers of (eco)productivity change. In this case, the green colour dominates the map, indicating that aggregate MTCH values were larger than MLTCH. Specifically, 139 out of 143 municipalities (97.2%) demonstrated higher scores in technological change when the assessment did not incorporate environmental variables. This implies that while municipalities have made efforts to adopt improved practices for enhancing MSW management from an environmental perspective, the most significant shift in the efficient frontier is driven by economic changes. It is worth noting that MSW management in Chile is funded through the municipal budget (SUBDERE 2020) necessitating that the poorest municipalities concentrate on improving MSW management not only from an environmental standpoint but also from an economic standpoint.

Figure 4. (a and b) Difference between aggregated MTCH and MLTCH values for 2015/19 years for the Chilean municipalities evaluated.

Most of the municipalities assessed experienced apositive growth of the technical change between 2015 and 2019.

The information provided in this study has implications for policymakers aiming to enhance productivity and eco-productivity in the provision of MSW services. Policymakers can use the findings to inform the development and implementation of targeted actions and policies to improve MSW management. One important aspect highlighted by the study is the identification of municipalities that have demonstrated eco-productive practices. These exemplary municipalities can serve as models for others to learn from and replicate their successful strategies. Policymakers can monitor and compile the actions and policies implemented by these eco-productive municipalities and disseminate this information to other municipalities, facilitating knowledge sharing and promoting best practices. Forging alliances with the private sector can be an effective strategy to finance the necessary infrastructure for improving MSW management. Collaboration with private entities can bring in resources and expertise, enabling municipalities to implement initiatives that enhance eco-productivity. Additionally, implementing specific measures for MSW recycling can significantly contribute to improving eco-productivity. Policymakers can introduce policies and incentives that encourage and support increased recycling rates, such as public awareness campaigns, infrastructure development, and extended producer responsibility programmes.

4. Conclusions

In the context of a circular economy, the economic and environmental performance of municipalities must be improved in the provision of MSW services. Assessing productivity change and eco-productivity change in MSW management is essential for achieving cost-effective, environmentally sustainable, and socially responsible waste management practices. This study analysed the impact of MSW recycling on the performance of municipal waste management by evaluating and comparing the productivity change and eco-productivity scores of a sample of 143 Chilean municipalities between 2015 and 2019.

The empirical findings revealed that the recycling rates of MSW in Chile did not undergo significant changes over the evaluated years, as environmental policies promoting MSW recycling were still in their early stages. Correspondingly, the average productivity change estimations (MPI) were higher than the eco-productivity change estimations (MLPI) for all the years examined. This indicates that Chilean municipalities have placed more emphasis on improving their performance in MSW management from an economic perspective rather than an environmental one. Substantial variations were observed among municipalities, highlighting disparities in the available funds for implementing measures to promote MSW recycling. It also suggests a lack of collaboration in MSW management among neighbouring municipalities. Therefore, to leverage potential economies of scale, the implementation of a regional policy is necessary to enhance the management of MSW services in Chile.

While this study makes a valuable contribution to the existing knowledge regarding the assessment of MSW services’ performance, it also paves the way for further research in several areas. Firstly, the current assessment focuses on specific types of solid waste, excluding electronic waste, batteries, and used oil from consideration. Thus, future research could explore the performance of municipalities in providing MSW services for these waste types as well. Secondly, our study primarily examines the recycling of waste as desirable outputs, without delving into the recycling technologies employed or the sub-products generated. Therefore, an extension of our analysis would involve incorporating the evaluation of the stages undertaken by municipalities for MSW valorisation. Lastly, conducting a second stage of analysis would help identify potential environmental factors that influence the dynamic performance of municipalities in MSW management.

Nomenclature

DEA	=	Data Envelopment Analysis
DMUs	=	Decision Making Unit
MLECH	=	Malmquist-Luenberger Efficiency Change
MLTCH	=	Malmquist-Luenberger Technical Change
MPI	=	Malmquist Productivity Index
MLPI	=	Malmquist-Luenberger Productivity Index
MSW	=	Municipal Solid Waste
SFA	=	Stochastic Frontier Analysis
SINADER	=	National System of Waste Declaration
SINIM	=	National Municipal Information System

Highlights

• The dynamic performance of municipal solid waste management was assessed.

• Productivity and eco-productivity metrics were compared for Chilean municipalities.

• Economic and environmental performance of municipalities improved across years.

Acknowledgments

The authors are grateful for support from the following sources: CEDEUS, ANID/FONDAP 15110020 (María Molinos-Senante), ANID Fondecyt Regular 1210077 (María Molinos-Senante), ANID Beca Doctorado Nacional 21182091 (Paula Llanquileo-Melgarejo).

Disclosure statement

No potential conflict of interest was reported by the authors.

Data availability statement

The data that support this research and other findings of this paper are available from the corresponding author upon reasonable request of each reader.

Additional information

Funding

The work was supported by the Agencia Nacional de Investigación y Desarrollo [Fondecyt 1210077].

Notes on contributors

Paula Llanquileo-Melgarejo

Paula Llanquileo Melgarejo is bachelor in chemical engineering and PhD candidate on hydraulic and environmental engineering at Pontificia Universidad Católica de Chile.

María Molinos-Senante

María Molinos-Senante is PhD on Local Development and researcher at Centre for Sustainable Development.

Notes

1. On 9th February, the conversion rate was 1€ $\cong$ 942 CLP and 1 US$ $\cong$ 824 CLP