اندازه گیری و تبیین بهره وری هزینه جمع آوری زباله های جامد شهری و خدمات پردازش

This paper proposes an adjusted version of the popular efficiency measurement technique data envelopment analysis (DEA) that makes it possible (1) to evaluate the cost efficiency of municipalities in the collection and processing of multiple household waste fractions, (2) robustifying the cost efficiency evaluations for the impact of measurement errors in the data or municipalities with outlying and atypical performances (if present in the sample set), and (3) to correct the evaluations for differences in the operating environments of municipalities (e.g., factors such as demography and median income of the municipality population). The paper illustrates the usefulness of the methodology by carrying out and analyzing a cost efficiency evaluation using data on 293 municipalities in Flanders, Belgium, for the year 2008.

The growing focus on municipal solid waste (MSW) reduction, re-use and recycling by international, national and regional governments has prompted important reforms in local/municipal waste collection and processing policy. As it is often at the local level that higher order policy initiatives are operationalized, municipalities are increasingly subjected to intricate waste collection and processing systems. In addition, municipalities are ever more challenged to create a sound budget without raising the tax burden or cutting the provision of these vital services to the residents. Fuelled by this evolution and supported by methodological innovations, a growing number of studies aim at assessing the efficiency of local MSW collection and processing services. Nevertheless, capturing the full complexity of MSW services when analysing the cost efficiency of municipalities proves to be a tortuous matter, as important conceptual and methodological difficulties are often neglected in the existing literature. In this paper we draw upon recent developments in the efficiency literature in order to provide an answer to three of those conceptual and methodological issues. The first important issue is related to the peculiar role that the input ‘waste cost’ is playing in the municipalities’ collection and processing efforts of the different MSW fractions. Municipalities use public resources to manage and finance the collection and treatment processes of the various fractions of MSW. It is therefore particularly interesting for the municipalities to have an idea not only about aggregate cost efficiency, but also about the cost efficiencies realized in the collection and treatment of the each waste fraction separately (i.e., partial cost efficiencies). Of course, the estimation of partial cost efficiency measures requires detailed information on how costs are precisely shared between the different MSW fractions. However, such detailed information is typically lacking. Previous studies that focused on the cost efficiency of municipalities used the basic DEA-model that does not explicitly account for this peculiar role of the input ‘waste cost’. Consequently, these studies were only able to compute a global cost efficiency measure for the municipalities and not detailed partial cost efficiency measures. A second issue in the cost efficiency evaluations of the municipalities involves correcting for the impact of the operating environment. The basic DEA-model presumes that all municipalities are homogeneous in the sense that they perform similar activities under comparable operating circumstances. Critical readers would remark that this assumption of “operating under comparable circumstances” is hard to maintain in practical settings. Both the empirical results in the literature (for an overview, see Section 2) as well as the practical experiences of the people that are active in this business (for instance, policy makers at the municipal level, administrators in government institutions-like OVAM in Flanders, etc.) suggest that some operating environments are more beneficial to cost efficiency in the collection and treatment of MSW waste, while other operating environments make it more difficult for a municipality to operate cost efficiently. A third difficulty is related to the DEA-model itself, and more particular to the relative perspective that is used by this model in the estimation of the cost efficiencies. The DEA-model relates the cost and waste quantity data of each evaluated municipality in a relative perspective to the cost and waste data of the other municipalities in order to evaluate the assessed municipality's aggregate and partial cost efficiencies. This relative perspective means that the presence merely of one municipality in the data set with outlying and/or atypical cost and waste data suffices considerably to impact the cost efficiencies of all municipalities in the dataset. That is to say, the basic DEA-based estimates are non-robust in the sense that they are sensitive to the impact of such municipalities. A correction of the DEA-model such that cost efficiency estimates are less sensitive to such influences would clearly benefit the reliability and validity of the DEA-model in evaluating municipality cost efficiency. The main objective of this paper is to adjust the DEA-model so that it provides answers to the three conceptual and methodological problems raised. In particular, this paper develops and advocates a robust and conditional version of the shared input DEA-model in evaluating the cost efficiencies of Flemish municipalities in the collection and treatment of MSW. This model is based on recent additions to the DEA-literature (i.e., the ideas of the shared input DEA-model follow Beasley [2], Cook et al. [10], and Cook and Green [9], whereas the additions to the robust and conditional DEA-model follow, among others, Cazals et al. [7], Daraio and Simar [11], [12] and [13], and De Witte and Kortelainen [18]). As we will argue, introducing this format of DEA when analyzing the cost efficiency of MSW related services, enables us to (1) compute partial cost efficiency measures for each waste fraction; (2) mitigate the impact of municipalities with outlying or atypical cost and quantity data; (3) correct for the impact of the operating environment and policy variables which are only controllable to the municipalities in the long term, and (4) analyze the way in which the characteristics of the operating environment are precisely related to the municipalities’ cost efficiency in MSW collection and processing. No doubt, such knowledge is informative and useful for both municipal and regional policy makers. This paper unfolds as follows. The next section provides a brief literature review of previous DEA-studies in the waste management literature. The focus is on studies that examined the efficiency measurement of municipalities as well as on studies that discussed the question of how the operating environment may affect the municipalities’ efficiency. In a third section, we discuss the household waste cost and quantity data for Flemish municipalities for the year 2008. Section 4 presents the robust and conditional version of the shared input DEA-model. In Section 5, we present and analyze the cost efficiency evaluations for the Flemish municipalities. In particular, we focus on the impact of the operating environment and long-term policy variables on the outcomes of our efficiency evaluation. In a final section, we make some concluding remarks and provide some directions for further research.

This paper advocated the DEA-model and, more precisely, the conditional and robust shared input version of this model, as a very appealing approach for evaluating the cost efficiency of municipalities in the collection and processing of MSW. In particular, the non-parametric nature of the DEA-model seems to be a key advantage in the study of the cost efficiency of the Flemish municipalities where there is uncertainty not only concerning the importance weights of the MSW fractions but also concerning the exact allocation of the costs between the different MSW fractions. It was shown that the DEA-framework provided robust and environment-adjusted estimations of the aggregate cost efficiency of the evaluated municipality as well as of the partial cost efficiencies for the separate MSW fractions. Addressing the three conceptual and methodological issues raised in the introduction enables us to accommodate some recent evolutions in local MSW management. In particular, the growing diversification in waste fractions justifies the use of a shared input version of the DEA method. In addition, by accounting properly for different and non-controllable background conditions, we are able to present a stronger argument in favour of the reliability of our results. Nevertheless, as in any parametric or non-parametric performance evaluation, the result is only as strong as the assumptions behind the method. In this sense, the nonparametric nature of the DEA method poses an important advantage. However, other methodological postulates might not correspond to reality. For instance allowing that the resource splits are determined in such a way that they are optimal for the municipality under consideration, has implications on the interpretation of the results. Attaining cost efficiency under the DEA-based evaluations is an indication that the municipality is doing well, but using true instead of optimal cost shares might entail a different result. In this sense, a DEA-based cost inefficient score can be seen as a much stronger result, as even with the optimal weights and resource split, the municipality is still outperformed by other municipalities. Even so the results of a shared input DEA-model could be of prime interest to local incumbents themselves. Nevertheless, as noted by García-Sánchez [22] municipal politics can have different goals with respect to MSW management. Depending on the political ideology, municipalities might consider maximizing efficiency to be subordinate to, for instance, maximizing the number of jobs in waste collection. However, this is not the only conceivable link between political motives and the performance of local waste management in terms of cost efficiency. As argued by De Jaeger et al. [17], the implementation of the polluter-pays principle – i.e., the principle that costs of collection and processing of waste are to be borne by those who produce the waste – discloses critical information about the municipalities’ performance to the electorate. Although the costs linked to waste management services generally represent a marginal fraction of total household expenses, residents seem to be sensitive to the pricing policy. For instance, the results of De Jaeger [15] show that the popularity of local incumbents in Flemish municipalities depends on the level of the price for waste collection and processing. Moreover, the decision to increase the price is mainly taken in the first four years of the six-year term served by the municipal council. This is at least an indication that local incumbents believe that voters can hold them responsible for waste bag prices. Finally, environmental concerns might be more important to local policy makers than cost efficiency gains themselves. Even so De Jaeger et al. [17] show that the implementation of policies aiming at MSW reduction does not necessarily reduce the municipalities’ level of performance in terms of cost efficiency. Notwithstanding our explicit focus on an assessment of cost efficiency estimates which captures the complex nature of MSW collection and processing services, further adjustments could prove valuable. For instance, the nature of our data only allows for a cross-sectional analysis, which evidently implies that we have refrained from considering the evolution over time of cost efficiency estimates. Nevertheless a time dependent approach could be valuable. For instance, Simões and Marques [40] are, due to their time dependent analysis, able to observe a productivity decline in the Portuguese urban waste utilities in spite of the introduction of quality of service regulation. Also, given the variation in the number of separately collected waste fractions, exploring potential benefits of scope would be an interesting avenue for further research. Finally, alternative approaches to DEA such as the stochastic frontier analysis (see, for instance, [39] for a recent application in waste management), or the use of performance indicators (examples in waste management include [23], [26] and [32]) prove also to be effective instruments when assessing efficiency of urban waste services.