Decoupling municipal solid waste generation and economic growth in the canton of Vaud, Switzerland

Resources, Conservation & Recycling 130 (2018) 260–266

Contents lists available at ScienceDirect

Resources, Conservation & Recycling journal homepage: www.elsevier.com/locate/resconrec

Full length article

Decoupling municipal solid waste generation and economic growth in the canton of Vaud, Switzerland

T

⁎

Rémi Jaligot , Jérôme Chenal Urban and Regional Planning Community, Swiss Federal Institute of Technology Lausanne, Switzerland

A R T I C L E I N F O

A B S T R A C T

Keywords: Solid waste management Socio-economic drivers Decoupling policies Switzerland

Switzerland is one of the largest producers of municipal solid waste (MSW) per capita. The principle of decoupling attempts to evaluate the relationship between consumption and production, and to measure the relationship between an activity and its impact on the environment. This paper uses the Environmental Kuznets Curve (EKC) hypothesis to understand the impact of three socio-economic drivers on MSW generation in the canton of Vaud in Switzerland. Vaud is a French-speaking canton that recently implemented several measures to limit MSW generation. We used time series of indicators for income, urbanisation and policy implementation in ten of the canton’s districts, which were set as independent variables, between 1996 and 2015. A panel data analysis was performed using a generalized least squares procedure to test for an EKC. Evidence shows that urbanization was slightly negatively associated with MSW generation, but without statistical significance. However, a direct policy mechanism such as the waste bag tax was significantly correlated with a decrease in waste generation. Overall, the presence of an EKC cannot be confirmed in the canton of Vaud, as waste generation tends to stabilize as income increases. It would be useful to perform a similar assessment in other cantons to fully inform decision-makers.

1. Introduction In recent decades, cities have been affected by a growing amount of municipal solid waste (MSW), which puts a strain on waste disposal capacities and on the environment (Cointreau, 2006). The total municipal solid waste (MSW) generated worldwide in 2012 was approximately 1.3 billion tonnes (Hoornweg and Bhada-Tata, 2012). Were all countries to continue to generate waste at the current rate of high income countries, total waste generation could reach 5.9 billion tonnes by 2025 (Scheinberg et al., 2010). The positive correlation between waste generation and income level is often demonstrated in the literature; as disposable income and living standards increase, consumption of goods tends to follow, and waste generation increases accordingly (Hoornweg and Bhada-Tata, 2012; Irwan et al., 2013; Keser et al., 2012; Wilson et al., 2012). Therefore, it is urgent to take appropriate measures to decouple economic growth from waste generation (Sjöström and Östblom, 2010; Unnisa and Rav, 2013). 1.1. The concept of decoupling The concept of decoupling, or delinking, has become a focus in economic studies, in order to understand the relationship between

⁎

consumption and production, and to measure the relationship between an activity and its impact on the environment. Is the elasticity of an environmental indicator relative to certain socio-economic drivers. This occurs when the value of the environmental indicator increases, but relatively less than the indicator of the driver (Mazzanti et al., 2008). Economist Simon Kuznets originally identified an inverted U-shaped relationship between income levels and inequality (Kuznets, 1955). He posited that income inequality would increase and then decrease as income grew within a country. Kuznets (1955) used time-series data from the United States, the UK and Germany for his analysis. The Environmental Kuznets Curve (EKC) hypothesis attempts to represent the decoupling of behaviours and to model a potential, inverted U-shaped relationship between an environmental indicator and indicators of socio-economic development. Household income’s impact on environmental degradation often has an inverted-U shape when plotted (Grossman and Krueger, 1995). A large body of literature focuses on the relationship between income and air pollution (Galeotti et al., 2006; Jalil and Mahmud, 2009; Selden and Song, 1994), whereas the impacts of municipal solid waste are less investigated (Arbulú et al., 2015). Some studies do not support the inverted U-shaped relationship (Chen, 2010; Johnstone and Labonne, 2004; Karousakis, 2006), while others find some evidence of a turning point for MSW generation

Corresponding author. E-mail addresses: remi.jaligot@epfl.ch, [email protected] (R. Jaligot).

https://doi.org/10.1016/j.resconrec.2017.12.014 Received 19 August 2017; Received in revised form 7 December 2017; Accepted 11 December 2017 Available online 15 December 2017 0921-3449/ © 2017 Elsevier B.V. All rights reserved.

Resources, Conservation & Recycling 130 (2018) 260–266

R. Jaligot, J. Chenal

Fig. 1. Study Area. Map (a) shows the location of the canton of Vaud in Switzerland. Map (b) shows the ten districts of interest in the canton of Vaud.

decentralized public political power and potential discrepancies with regard to waste management strategies. The canton of Vaud is located in the western French-speaking region of Switzerland. In Vaud, MSW is defined as the “waste produced by households, and those with similar a composition produced by industries, small-businesses, agriculture and tertiary activities” (Directorate General for the Environment (DGE, 2016). Incinerable waste and recyclable waste (i.e. paper, cardboard, glass, organic waste and metals) are MSW. The canton has designated areas for waste management within the canton that roughly corresponds to the district boundaries. Municipalities are required to evaluate the quantities of MSW, separated at source or not, and to communicate the results to the canton. The results are then cross-validated and aggregated by district. In this study, data on MSW generation are preferred to data on waste composition because MSW generation trends are more diverse across districts than is the evolution of waste composition. The canton has established several measures to limit MSW generation, which were applied differently across the various districts. For example, a ‘waste bag tax’ was officially introduced for the entire canton in 2013 to finance the treatment of MSW, but many municipalities in the northern districts had already introduced it as early as 2008. The aim of this research is to understand the relationship between three socio-economic drivers (i.e. income, urbanisation and policy implementation) and MSW generation in the canton of Vaud using four corresponding indicators (Section 2.2.1). The objectives are (i) to understand MSW generation trends in the canton at the district level, (ii) to test for the presence of an EKC, and (iii) to determine whether a correlation exists between MSW generation and important socio-economic drivers.

(Abrate and Ferraris, 2010; Arbulú et al., 2015; Khajuria et al., 2012; Mazzanti et al., 2008). An EKC may also arise only when residents show a diminishing marginal utility of consumption, but not when they consume more as their income increases (Swart and Groot, 2015). Economic growth is a recognised driver of urbanisation (Frenken et al., 2007; Moomaw and Shatter, 1996), and consequently a driver of change in terms of land-use. The impact of changes in land-use and the socio-economic variables associated therewith have been the focus of recent studies, wherein a positive statistical relationship was found between changes in land-use and MSW generation (George, 2015; Lei et al., 2016; Xiao et al., 2015), but a negative relationship was found between the density of dwellings and MSW generation (ChamizoGonzalez et al., 2016). 1.2. Waste management in Switzerland Switzerland is attempting to implement a strategy of avoidance, reuse, recycling similar to that of the European Union with the revised Waste Framework Directive (2008/98/EU), which introduced a new version of the “Waste Hierarchy”. It stipulates that waste prevention, reuse and recycling should be prioritized over other types of recovery and disposal into landfills unless proven otherwise. However, increasing MSW generation is a growing concern in the context of demographic growth and economic development, which trigger high levels of urban sprawl (Jaeger and Schwick, 2014). MSW generation increased from 603-kg per capita in 1990 to 729kg per capita in 2014 (Federal Office for Environment (FOE, 2016). This growth is largely associated with consumption habits, higher amounts of organic waste and the limited lifespan of electronic equipment. Despite high recycling rates (approx. 50%) and efficient waste disposal, waste generation reflects natural resource consumption, which is one of the greatest challenges for the country (Federal Office for Environment (FOE, 2016). For example, in order for the entire world population to match Swiss consumption levels, the amount of resources necessary would be equivalent to that of three Earths (Federal Office for Environment (FOE, 2016). Switzerland is a federation with twenty-six sovereign states called cantons. Each canton has its own government and parliament. Cantons are further divided into districts and municipalities, which results in

2. Study area and data 2.1. Study area The study area is the canton of Vaud, which, administratively speaking, is divided into ten districts: Aigle, Broye-Vully, Gros-de-Vaud, Jura-Nord Vaudois, Lausanne, Lavaux-Oron, Morges, Nyon, Ouest Lausannois and Riviera-Pays-d’Enhaut (Fig. 1). The canton had a population of 767,497 inhabitants in 2015. The least populated district 261

Resources, Conservation & Recycling 130 (2018) 260–266

R. Jaligot, J. Chenal

Table 1 Descriptive statistics of three socio-economic drivers and corresponding indicators. The socio-economic driver for each indicator is presented between parentheses in italics. A “Bag Tax” was widely implemented in the canton in 2013, which explains the low mean between 1996 and 2015.*CHF = Swiss Francs, where CHF 1 = $1.02 in August 2017. Indicator description (driver)

Unit

Abbreviation

Min.

Mean

Max.

-Municipal Waste Generation is the dependent variable -Tax Point Value (income) is an independent variable representing the wealth distribution per capita between districts. -Population density (urbanisation) may be positively or negatively correlated with the dependent variable as describe above. -Fixed waste tax (policy implementation) paid by all households may be positively correlated with lower MSW generation over the entire district -A “Bag Tax” (policy implementation) introduced in 2013 may be positively correlated with a sharp decrease in MSW generation when the tax is adopted by all municipalities

(kg/capita) %

MSW TPV

347 15.9

488 32.4

668 67

inhabitant/km2

DENS

75.6

617

2704

*CHF/capita

FTAX

0

49.7

169.1

% district population standardized

BTAX

0

0.113

1

important driver of MSW generation (Hong et al., 1993; Mazzanti et al., 2008; Spangenberg, 2001; Usui and Takeuchi, 2014). The principle of causality is the basis of the solid waste management (SWM) strategy in Switzerland. Articles 32 and 32a were incorporated in the Federal Law for Environmental Protection in 1995 and 1997 respectively. The articles stipulate that waste generators are responsible for post-generation treatment. Cantons are responsible for applying this policy through various mechanisms and/or taxes. The cost of post-generation treatment is based on (1) the type and quantity of waste and (2) the total costs related to the treatment mechanism, including maintenance of treatment facilities. The cantonal law in Vaud provides details with regard to this measure. It is the responsibility of municipalities to finance the SWM through tax mechanisms. 40% of the costs should be financed by a tax proportional to the quantity of waste generated. Two key approaches were implemented in Vaud: a variable tax based on the quantity or volume of waste generated, and a fixed tax. As such, two policy implementation indicators are presented in this research. A ‘waste bag tax’ (BTAX) was officially introduced in 2013 for the entire canton to finance the treatment of the bag. However, 55 municipalities in the northern part of the canton had introduced a local waste bag tax in 2008. In 2017, regional and local waste bag taxes were harmonized across the canton. The second tax mechanism was a fixed tax (FTAX), which is used to build and maintain costly waste management infrastructure such as incinerators and Material Recovery Facilities (MRF). The latter functions independently of the volume of waste generated. Table 1 provides a summary of the variables and corresponding descriptive statistics.

was Broye-Vully with 40,207 inhabitants; the most populated was Lausanne with 160,446 inhabitants in 2015 (Statistique Vaud, 2017). 2.2. Data 2.2.1. Data collection The panel data have a cross-sectional and a temporal dimension. For this study, the former includes the ten districts. The timeframe spans from 1996 to 2015. Data were gathered from the Cantonal Statistical Office and the National Office for Spatial Data for the 1996–2015 period. The following variables were calculated for each district: yearly waste generation per capita, population density (inhabitants/hectare), tax point value per capita, the payable amount of fixed annual tax for waste management per capita and the percentage of the population that pays the bag tax. The choice of control variables was based on the literature and relevance to the local context, as the densification of urban areas and waste policies are two key topics of debates in the canton. The data are described in more details below. Spatial statistical data were gathered using the official GIS tool provided by the cantonal office. As complete datasets were available, population density was the preferred indicator of urbanisation, versus sporadic data on the proportion of built areas. Other non-spatial demographic, economic and environmental data were gathered from the Cantonal Statistical Office for the 20-year period. The datasets were balanced, as all variables are observed for the entire period. The Tax Point Value (TPV) is the amount of money collected from local taxes divided by the local coefficient applied for the same year. The results are displayed per district and per capita. The total amount from local taxes in a given district is correlated to the demographic and social structure of its population (e.g. rural municipalities, urban municipalities, suburban municipalities, percentage share of single households, single-parent households, etc.). The median revenue varies between municipalities, so the TPV is the covariate used as a proxy of wealth distribution or income per capita. Data on a more direct measure of income per capita (e.g. revenue per capita) were incomplete for all districts for the study period. TPV was therefore the next best option as an income per capita proxy. Many studies show that wealth is positively correlated with MSW generation (Dyson and Chang, 2005; Fargier, 2015; Sudhir et al., 1997; Wilson et al., 2012). Population density (DENS) is an indicator of both urbanisation and the current trend of densification of urban centres (Frenken et al. 2007; Stankowski, 1972), and was used as a control variable in the study of EKC in the waste sector (Chen, 2010; Mazzanti et al., 2008). It remains unclear how MSW generation reacts to fluctuating densities. One can infer that MSW generation decreases as density increases, due to decreased land availability. However, one could also argue that MSW generation increases as people have less space to source separation of waste (Chen, 2010; Mazzanti et al., 2008). DENS is included as an independent variable and an indicator of urbanisation. The implementation of direct and indirect policies may be an

2.2.2. Data visualization The waste generated for all districts ranges from 347-kg/capita to 668-kg/capita (Fig. 2). The evolution of MSW generation varied between districts, with some being erratic. Most districts showed an increase in MSW generation from 1996 to about 2007, followed by a gradual decrease. Waste generation in the remaining districts, namely Aigle and Ouest Lausannois, was more erratic. 55 municipalities of Jura-Nord-Vaudois (ID 4) introduced the BTAX in 2008. This was reflected in a slight drop (35-kg/capita in 2009) in MSW generation. All districts showed a decrease in MSW generation around 2012–2013. A sharp drop was observed in Gros-de-Vaud, Lausanne, Lavaux-Oron, Morges, Nyon, Ouest Lausannois and Riviera-Paysd’Enhaut. The districts of Aigle, Broye-Vully and Jura-Nord-Vaudois showed the most even continuous decrease in MSW generation. The BTAX came into effect in 2013 and appears to have had a strong impact on waste generation across six of the ten districts (Fig. 2). It is important to note that MSW generation followed an inverted Ushape over time in all districts, despite some outliers.

262

Resources, Conservation & Recycling 130 (2018) 260–266

R. Jaligot, J. Chenal

Fig. 2. Evolution of municipal solid waste generation in ten districts in the canton of Vaud, Switzerland. District ID – 1. Aigle, 2. Broye-Vully, 3. Gros-de-Vaud, 4. Jura-Nord-Vaudois, 5. Lausanne, 6. Lavaux-Oron, 7. Morges, 8. Nyon, 9. Ouest Lausannois, 10. Riviera-Pays-d’Enhaut). The vertical line represents the introduction of the BTAX in 2013. It was plotted in Year 2012 to show the impact on the following year. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

3. Methods

(ii) β1 > 1 and β2 =β3 =0. This implies a linear relationship between MSW and TPV. (iii) β1 < 0 and β2 =β3 =0. This implies a monotonic decreasing relationship between MSW and TPV. (iv) β1 > 0 , β2 < 0 and β3 =0. This implies an inverted U-shaped relationship such as an EKC. (v) β1 < 0 , β2 > 0 and β3 =0. This implies a U-shaped relationship. (vi) β1 > 0 , β2 < 0 and β3 > 0 . This implies a cubic polynomial or Nshaped figure. (vii) β1 < 0 , β2 > 0 and β3 < 0 . This implies the opposite of (vi).

3.1. Regression functional relationship The first step of our analysis was to define the regression functional relationship. The aim is to understand whether the correlation between MSW generation and socio-economic variable is consistent with the EKC hypothesis. Different EKC regression models are presented in various works (Choi and Cho, 2010; Lindmark, 2002; Mazzanti et al., 2008; Stern, 2004). The quadratic equation was consistently used in studies that tested for the inverted U-shaped relationship between a dependent environmental variable and socio-economic covariates. The use of the quadratic equation implies that the environmental variable tends to zero or infinity as the independent variable increases. As such, certain authors also use higher-order polynomials. This implies a more complex relationship between income and environmental degradation. We tested for this complexity by including a cubic term, as MSW is unlikely to tend to zero when wealth increases. Eq. (1) represents the functional relationship used in this research:

The dependent variable was combined with each independent variable. Eq. (1) was tested with different specifications in order to assess the suitability of the EKC. The use of a logarithmic scale was not advantageous in this study. The analysis therefore uses a natural scale. Model (1) is the simple regression of MSW on TPV. Model (2) tests the quadratic form of Eq. (1) with TPV2 to test the Environmental Kuznets Curve hypothesis. Model (3) adds each explanatory variable to model (2) to capture differences between districts. Finally, model (4) fits Eq. (1) with the independent variables and cubed TPV values. Cross-country data have been used to assess the relationship between socio-economic variables and environmental degradation (Dasgupta et al., 2001; Raleigh and Urdal, 2007; Shafik and Bandyopadhyay, 1992; Shi, 2003). The underlying assumption of a common structure across countries, e.g. the economic development trend at a given time, is often made in such studies but cannot be guaranteed due to cross-country variations (Arbulú et al., 2015; Dinda, 2004). The use of panel data methodology is relevant in this study, versus cross-section and time series analysis. Panel data have several advantages over cross-section or time-series data (Arbulú et al., 2015; Hsiao, 2014; Woolridge, 2002).

MSWit = α + β1 (Tax point value )it + β2 (Tax point value )it2 + β3 (Tax point value )3it +

∑ β4 (Xit ) + εit

(1)

where α is the intercept parameter and X is a vector referring to one of the socio-economic independent variables, the subscript i refers to the district, the subscript t to the year and ε to the error term. Eq. (1) allowed us to test different relationships between the environmental and socio-economic variables (Dinda, 2004). (i) β1+ β2 + β3=0. This implies no relationship between MSW and TPV. 263

Resources, Conservation & Recycling 130 (2018) 260–266

R. Jaligot, J. Chenal

Table 2 Generalized least squares estimates of four municipal solid waste generation models in the canton of Vaud, Switzerland. Correlation is significant when p-value < 0.05. Variable

Intercept TPV TPV2 TPV3 DENS FTAX BTAX Pseudo Rsquared (i)

Model 1 Coeff.

SE(i)

t-value

p-value

460.02 0.83

24.27 0.44

18.95 1.91

0.000 0.057

0.0415

Model 2 Coeff.

SE

t-value

p-value

Model 3 Coeff.

SE

t-value

p-value

334.83 8.08 −0.09

50.70 2.33 2.89e-02

6.06 3.47 −3.21

0.000 0.000 0.002

303.61 9.22 −0.09

64.81 2.64 3.08e-02

4.68 3.49 −3.02

0.000 0.000 0.003

−4.15e-03 0.15 −60.89 0.536

6.12e-03 0.13 16.34

0.58 1.15 −3.73

0.561 0.251 0.000

0.0755

Model 4 Coeff. 205.66 16.82 −0.28 0.001 0.005 0.15 −64.40 0.536

SE

t-value

p-value

120.29 7.93 0.18 1.38e-03 6.15e-03 0.13 15.92

1.71 2.12 −1.54 1.08 0.78 1.19 −4.05

0.089 0.035 0.125 0.280 0.435 0.235 0.000

SE = Standard Error.

4. Results and discussion

(i) Higher sample variability and more degrees of freedom allowing for more accurate inference of model parameters. (ii) Better understanding of the complexity of socio-economic drivers. It allows for a more precise analysis of the dynamics of adjustment of socio-economic variables. (iii) Panel data considers all countries or regions as heterogeneous, versus cross-sectional data. It allows us to consider more information and potentially biased estimators.

4.1. Preliminary testing Plots of residuals for all FE specifications show some autocorrelation. The Breusch-Pagan statistic was used to test for heteroskedasticity, where variance in the error terms of a variable are unequal across the range of a second variable (often time) that predicts it. It was detected at a 5% statistical significance level. Autocorrelation and heteroskedasticity are addressed using the procedure presented in Section 3.2. After implementing the GLS procedure, insignificant autocorrelation in the residuals was observed for model (1) and model (2). A slightly significant negative autocorrelation was observed at lag 1 for models (3) and (4). The impact of small autocorrelation on slow changing variables, such as MSW generation, can be ignored (Greene, 2000). The structure of our model is therefore relevant to the system.

3.2. Selection of model estimator Fixed-effects (FE) is often used in panel data analysis (Baltagi, 2008; Hsiao, 2014; Petersen, 2009). Different waste management companies operate in the canton, which can potentially lead to measurement errors and differences in the amount of waste generation calculated. The use of FE captures some of this variation. Moreover, the coefficients for the districts cannot be considered equal. The choice of an FE estimator over a pooled OLS estimator is confirmed by the results of an F-test (Croissant and Millo, 2008; Woolridge, 2002; Yaffee, 2003). In this respect, the analysis was first performed with FE. According to some authors, FE is suitable for testing for the presence of an EKC (Grossman and Krueger, 1991; Karousakis, 2006). This research used the FE-type model to obtain the preliminary results. FE estimations were specified for each model presented in Section 3.1. The results are not presented in details here for the sake of conciseness (c.f. Section 4.1). Data was collected over time, so it was important to include preliminary tests on serial correlations and heteroscedasticity, which we used to check for randomness in our data and to identify underlying trends at varying time-lags. Innocent et al. (2016) demonstrated that the sample autocorrelation persists at high lags when measures are taken monthly due to the seasonality of waste generation, particularly in developing countries. In this study we considered that a one-year lag is relevant when measures are yearly and seasonality is not of interest. If autocorrelation and/or heteroskedasticity are found, it is important to modify the estimator so as to address both issues, as either may lead to an important loss of efficiency in the model approach. The generalized least squares (GLS) procedure makes it possible to estimate panel data when heteroskedasticty and autocorrelation are present (Cole et al., 1997; Hill and Magnani, 2002; Johnstone and Labonne, 2004; Woolridge, 2002). Using a GLS estimator to correct for autocorrelation allowed us to adjust the FE regression by changing the model so that errors are independent, in order to get efficient parameter estimates. The error structure (εit ) is modelled as a first-order autoregressive process. The estimator derived from this procedure is consistent and asymptotically normally distributed (Ajinkya et al., 1991; Parks, 1967). This procedure was used in previous waste management studies, wherein the model corrects for both autocorrelation and heteroskedasticity (Arbulú et al., 2015; Johnstone and Labonne, 2004; Mazzanti et al., 2008).

4.2. Empirical findings This study provides empirical evidence of the relationship between municipal solid waste (MSW) generation and four socio-economic indicators in the ten districts of the canton of Vaud. It analyses a panel data set over a 20-year period, between 1996 and 2015. Tax point value (TPV) is a proxy of income. The results confirm the previous observation that high-income households tend to generate more waste (Table 2). It should be noted that household size and/or awareness regarding waste management vary between income groups and have been found to be negatively correlated with MSW generation (Ogwueleka, 2013; Pfeffer, 1992). Hence, certain variables may moderate this result. The standard error of the estimate, which provides a measure of accuracy of the prediction, is the square root of the average squared deviation. The quadratic equation fitted by model (2) suggests the emergence of a waste Kuznets curve. The inclusion of socio-economic variables in model (3) significantly increases the elasticity of the TPV variable; an EKC thereby emerges. The low elasticity of TPV2 (-0.09-kg for model (3)) suggests the existence of an EKC in this context. However, we concluded that MSW generation stabilizes as TPV increases (Table 2). Key differences in the level of urbanisation in the canton have been identified. As such, it remains unclear how MSW generation behaves with fluctuating levels of urbanisation. The introduction of the density factor (DENS) aimed to test the relationship between urbanisation and MSW generation. It shows slightly negative elasticity with poor significance (Table 2), suggesting that more densely populated areas produce relatively less waste. Despite the poor significance, this result contradicts the outcomes of an earlier analysis, which show that population density positively affects waste generation (Chen, 2010; Mazzanti et al. 2008). Certain factors might explain this variation, such as differences in waste management awareness in urban (more densely populated) and rural (less densely populated) areas in Switzerland 264

Resources, Conservation & Recycling 130 (2018) 260–266

R. Jaligot, J. Chenal

stabilize as wealth increases. The concept of EKC is subject to debate, as it implies that MSW generation tends to zero as income increases. This contradicts other studies that support the EKC hypothesis for MSW generation. Population density (DENS), fixed tax measures (FTAX) and bag tax (BTAX) measures were the other indicators used to test relationship between MSW generation and socio-economic drivers. Only the BTAX was found to be negatively and significantly correlated with MSW generation across specifications. When residents must pay for bin bags, it triggers a reduction of waste generation and an increase in waste separation. Interestingly, the effect of the FTAX mechanism is not as significant. There are two possible reasons for this. First, the FTAX is paid indirectly, along with other local taxes. This indirect payment may, as such, trigger changes in waste management practices. The second relates to age and income level. Residents aged under 18 or 20 and those with low incomes are exempt from this tax. The positive correlation between FTAX and income could imply similar behaviours for both indicators as regards MSW generation. DENS is used as an indicator of urbanisation. Its relationship to MSW generation is ambiguous, as detailed in previous studies. Here, however, evidence shows that population density is slightly negatively associated with MSW generation. This suggests that an increase in population density, and hence the level of urbanisation, is correlated with a decrease in waste generation. However, the low significance level tempers this correlation. This study introduces the relationship between socio-economic drivers and MSW generation in the canton of Vaud in Switzerland. It demonstrates that the most efficient way to reduce waste generation is a dissuasive policy approach. Further research is needed to test whether similar trends are observable in other Swiss cantons that use different policy measures. A similar exercise might be conducted at the canton level to inform decision-making at the federal level.

versus other countries. This result may have implications in terms of forecasting waste trends, as the canton of Vaud is aiming to densify urban centres. The role of two policies (tax measures) was also tested. The first the FTAX, a fixed measure - concerned the operation and maintenance of waste management infrastructures. FTAX is not related to the quantity of MSW generated; rather, it complements the second tax measure based on the quantity of MSW generated (BTAX). The FTAX is calculated based on the number of household members, excluding residents aged under 18 or 20 and those with low incomes. Municipalities set the amount of the tax, which cannot exceed CHF 200. Hence, age and income are considered in the calculation, implying that only household members with average to high income will pay the FTAX. Empirical evidence shows that income often positively affects waste generation (Hockett et al., 1995; Ojeda-Benítez et al., 2008; Sudhir et al., 1997). Therefore, the positive correlation may show the relationship between the FTAX and income rather than MSW generation directly. The second tax mechanism, or bag tax, is indirectly paid by residents when they purchase bin bags and, as such, is related to the quantity of MSW generated. The BTAX is negatively and significantly correlated with MSW generation, which shows that specific charge policies are effective to reduce MSW generation (Priefer et al., 2016; Meng et al., 2016). When all the residents in a district are subject to the BTAX, MSW generation per capita decreases by approximately 61-kg per capita (Table 2). The p-values obtained in model (3) suggest that population density and the FTAX do not have a significant impact on waste generation, whilst the impact of TPV, TPV2 and BTAX are statistically significant. We also noted that, for Aigle, none of the specifications significantly fit the variables. Fig. 2 shows that waste generation is erratic in the district. One explanation for this may be the high rate of tourism in the Alps relative to the district’s population. A positive correlation between tourism and waste generation was established in previous studies (Arbulú et al., 2015; Mateu-Sbert et al., 2013; Mazzanti et al., 2008). Most of the canton’s alpine leisure areas are located in Aigle. The number of hotel nights is stable at around 480,000 per annum in the Alps, whereas in Lausanne and Ouest-Lausannois this figure reaches 1,000,000. However, the population is almost four times higher in both districts than in Aigle (Statistique Vaud, 2017). Model (4) includes the cubic TPV term. The coefficient is almost null, which confirms an earlier observation that waste generation stabilizes as income increases. The p-value obtained from model (4) shows that a change in most predictors is not associated with a change in the response. Only TPV and BTAX have low p-values, suggesting that they are meaningful additions to our model (Table 2).

Declaration of conflicting interests The authors declare no potential conflicts of interest with respect to the research, authorship and/or publication of this article. Acknowledgements This research would not have been possible without the help of those who provided assistance and data, including the staff of the Statistical Office of Vaud. We are also grateful to Professor Anthony Davidson for his input on statistical models. References

5. Conclusion Abrate, G., Ferraris, M., 2010. The environmental Kuznets curve in the municipal solid waste sector. HERMES Working Paper, 1. (Accessed 2 October 2017). https://pdfs. semanticscholar.org/a239/f6ea0d6fa38410100039f68c5a28b5353d57.pdf. Ajinkya, B.B., Atiase, R.K., Gift, M.J., 1991. Volume of trading and the dispersion in financial analysts' earnings forecasts. Account. Rev. 389–401. Arbulú, I., Lozano, J., Rey-Maquieira, J., 2015. Tourism and solid waste generation in Europe: a panel data assessment of the environmental Kuznets curve. Waste Manage. 46, 628–636. Baltagi, B., 2008. Econometric Analysis of Panel Data. John Wiley & Sons. Chamizo-Gonzalez, J., Cano-Montero, E.I., Muñoz-Colomina, C.I., 2016. Municipal solid waste management services and its funding in Spain. Resour. Conserv. Recycl. 107, 65–72. Chen, C.C., 2010. Spatial inequality in municipal solid waste disposal across regions in developing countries. Int. J. Environ. Sci. Technol. 7 (3), 447–456. Choi, A., Cho, Y., 2010. An Empirical Study of the Relationships between CO2 Emissions, Economic Growth and Openness. IZA DP No-5304. Cointreau, S., 2006. Occupational and Environmental Health Issues of Solid Waste Management: Special Emphasis on Middle-and Lower-Income Countries. World Bank, Washington, DC. Cole, M.A., Rayner, A.J., Bates, J.M., 1997. The environmental Kuznets curve: an empirical analysis. Environ. Dev. Econ. 2 (04), 401–416. Croissant, Y., Millo, G., 2008. Panel data econometrics in R: the plm package. J. Stat. Softw. 27 (2), 1–43. Dasgupta, S., Mody, A., Roy, S., Wheeler, D., 2001. Environmental regulation and

A fixed-effects model was initially applied to the panel dataset to test the relationship between MSW generation in the canton of Vaud at the district level and four socio-economic indicators. MSW generation varied between districts. Most districts show an increase in waste generation between 1996 and approximately 2007, followed by a gradual decrease until 2015. The sharpest decline was observed around 2012–2013. Four model specifications were tested using fixed-effects, based on an Environmental Kuznets Curve (EKC) functional relationship for the canton’s ten districts. We then checked for the existence of autocorrelation in the residuals and for heteroskedasticity before assessing underlying trends at varying time-lags. Both autocorrelation and heteroskedasticity were significant. A generalised least squares (GLS) approach was used to correct for both factors. Insignificant autocorrelation at varying time-lags confirms that the structure is appropriate. The existence of an EKC was tested with the tax point value (TPV), a proxy of income. The presence of an EKC cannot be confirmed. The low value of the TPV2 parameter shows that MSW generation tends to 265

Resources, Conservation & Recycling 130 (2018) 260–266

R. Jaligot, J. Chenal

333–347. Mateu-Sbert, J., Ricci-Cabello, I., Villalonga-Olives, E., Cabeza-Irigoyen, E., 2013. The impact of tourism on municipal solid waste generation: the case of Menorca Island (Spain). Waste Manage. 33 (12), 2589–2593. Mazzanti, M., Montini, A., Zoboli, R., 2008. Municipal waste generation and socioeconomic drivers: evidence from comparing Northern and Southern Italy. J. Environ. Dev. 17 (1). Meng, X., Wen, Z., Qian, Y., 2016. Multi-agent based simulation for household solid waste recycling behavior. Resour. Conserv. Recycl. 128, 535–545. Moomaw, R.L., Shatter, A.M., 1996. Urbanization and economic development: a bias toward large cities? J. Urban Econ. 40 (1), 13–37. Ogwueleka, T.C., 2013. Survey of household waste composition and quantities in Abuja, Nigeria. Resour. Conserv. Recycl. 77, 52–60. Ojeda-Benítez, S., Armijo-de Vega, C., Marquez-Montenegro, M.Y., 2008. Household solid waste characterization by family socioeconomic profile as unit of analysis. Resour. Conserv. Recycl. 52 (7), 992–999. Parks, R.W., 1967. Efficient estimation of a system of regression equations when disturbances are both serially and contemporaneously correlated. J. Am. Stat. Assoc. 62 (318), 500–509. Petersen, M.A., 2009. Estimating standard errors in finance panel data sets: comparing approaches. Rev. Financial Stud. 22 (1), 435–480. Pfeffer, J.T., 1992. Solid waste management engineering. Sources and Characteristics of Urban Solid Wastes. Prentice Hall, New Jersey, pp. 47–70. Priefer, C., Jörissen, J., Bräutigam, K.R., 2016. Food waste prevention in Europe–A causedriven approach to identify the most relevant leverage points for action. Resour. Conserv. Recycl. 109, 155–165. Raleigh, C., Urdal, H., 2007. Climate change, environmental degradation and armed conflict. Pol. Geogr. 26 (6), 674–694. Scheinberg, A., Wilson, D., Rodic, L., 2010. Solid waste management in the world’s cities. UN-Habitat’S State of Water and Sanitation in the World’S Cities Series. Earthscan for UN-Habitat. Selden, T.M., Song, D., 1994. Environmental quality and development: is there a Kuznets curve for air pollution emissions? J. Environ. Econ. Manage. 27 (2), 147–162. Shafik, N., Bandyopadhyay, S., 1992. Economic Growth and Environmental Quality: Time-Series and Cross-Country Evidence Vol. 904 World Bank Publications. Shi, A., 2003. The impact of population pressure on global carbon dioxide emissions, 1975–1996: evidence from pooled cross-country data. Ecol. Econ. 44 (1), 29–42. Sjöström, M., Östblom, G., 2010. Decoupling waste generation from economic growth–a CGE analysis of the Swedish case. Ecol. Econ. 69 (7), 1545–1552. Spangenberg, J.H., 2001. The environmental Kuznets curve: a methodological artefact? Popul. Environ. 23 (2), 175–191. Stankowski, S.J., 1972. Population density as an indirect indicator of urban and suburban land-surface modifications. US Geological Survey Professional Paper, 800. pp. 219–224. Statistique Vaud, 2017. Districts et Communes: Autres Tableaux. Accessed 1 June 2017. http://www.scris.vd.ch/Default.aspx?DomId=33. Stern, D.I., 2004. The rise and fall of the environmental Kuznets curve. World Development 32 (8), 1419–1439. Swart, J., Groot, L., 2015. Waste management alternatives: (Dis) economies of scale in recovery and decoupling. Resour. Conserv. Recycl. 94, 43–55. Sudhir, V., Srinivasan, G., Muraleedharan, V.R., 1997. Planning for sustainable solid waste management in urban India. Syst. Dyn. Rev. 13 (3), 223–246. Unnisa, S.A., Rav, S.B., 2013. Sustainable Solid Waste Management. CRC Press p 123. Usui, T., Takeuchi, K., 2014. Evaluating unit-based pricing of residential solid waste: a panel data analysis. Environ. Resour. Econ. 58 (2), 245–271. Wilson, D.C., Rodic, L., Scheinberg, A., Velis, C.A., Alabaster, G., 2012. Comparative analysis of solid waste management in 20 cities. Waste Manage. Res.: J. Int. Solid Wastes Public Clean. Assoc., ISWA 30 (3), 237–254. Woolridge, J., 2002. Econometrics Analysis of Cross Section and Panel Data. MIT Press, Cambridge, Massachusetts. Xiao, L., Lin, T., Chen, S., Zhang, G., Ye, Z., Yu, Z., 2015. Characterizing urban household waste generation and metabolism considering community stratification in a rapid urbanizing area of China. PloS One 10 (12), e0145405. Yaffee, R., 2003. A primer for panel data analysis. Connect: Information Technology at NYU. pp. 1–11.

development: a cross-country empirical analysis. Oxf. Dev. Stud. 29 (2), 173–187. Dinda, S., 2004. Environmental Kuznets curve hypothesis: a survey. Ecol. Econ. 49 (4), 431–455. Directorate General for the Environment (DGE), 2016. Plan de gestion des déchets 2016. Accessed 2 October 2017 http://www.vd.ch/fileadmin/user_upload/themes/environnement/dechets/fichiers_pdf/ DIRNA_GEODE_PGD__version_finale_novembre_2016.pdf. Dyson, B., Chang, N.B., 2005. Forecasting municipal solid waste generation in a fastgrowing urban region with system dynamics modelling. Waste Manage. 25 (7), 669–679. Fargier, M., 2015. Statistical Comparison of Solid Waste Management Performance in 40 Cities. MSc Thesis. Imperial College, London. Federal Office for Environment (FOE), 2016. Elimination des déchets. (Accessed 4 April 2017). https://www.bafu.admin.ch/bafu/fr/home/themes/dechets/publicationsetudes/publications/elimination-des-dechets.html. Frenken, K., Van Oort, F., Verburg, T., 2007. Related variety, unrelated variety and regional economic growth. Reg. Stud. 41 (5), 685–697. Galeotti, M., Lanza, A., Pauli, F., 2006. Reassessing the environmental Kuznets curve for CO 2 emissions: a robustness exercise. Ecol. Econ. 57 (1), 152–163. George, E.T., 2015. The effects of communal land sub-division and land use changes on household waste production in the Southern rangelands of Kenya. Int. J. Peace Dev. Stud. 6 (2), 21–29. Greene, W., 2000. Econometric Analysis, fourth edition. McGraw-Hill, New York. Grossman, G.M., Krueger, A.B., 1991. Environmental Impacts of a North American Free Trade Agreement (No. w3914). National Bureau of Economic Research. Grossman, G.M., Krueger, A.B., 1995. Economic growth and the environment. Q. J. Econ. 110 (2), 353–377. Hill, R.J., Magnani, E., 2002. An exploration of the conceptual and empirical basis of the environmental Kuznets curve. Aust. Econ. Pap. 41 (2), 239–254. Hockett, D., Lober, D.J., Pilgrim, K., 1995. Determinants of per capita municipal solid waste generation in the Southeastern United States. J. Environ. Manage. 45 (3), 205–217. Hong, S., Adams, R.M., Love, H.A., 1993. An economic analysis of household recycling of solid wastes: the case of Portland, Oregon. J. Environ. Econ. Manage. 25 (2), 136–146. Hoornweg, D., Bhada-Tata, P., 2012. What a Waste: a Global Review of Solid Waste Management. World Bank, Washington, DC. Hsiao, C., 2014. Analysis of Panel Data (No. 54). Cambridge University Press. Innocent, N.M., Dieudonné, B., Jose, S.N., 2016. Modelling the temporal variations of municipal solid waste generation for future projection in the Douala Municipality, Cameroon. Transportation 3 (7). Irwan, D., Basri, N., Watanabe, K., Abushammala, M., 2013. Influence of income level and age on per capita household solid waste generation in Putrajaya, Malaysia. Jurnal Teknologi 65 (2). Jaeger, J.A., Schwick, C., 2014. Improving the measurement of urban sprawl: weighted urban proliferation (WUP) and its application to Switzerland. Ecol. Indic. 38, 294–308. Jalil, A., Mahmud, S.F., 2009. Environment Kuznets curve for CO 2 emissions: a cointegration analysis for China. Energy Policy 37 (12), 5167–5172. Johnstone, N., Labonne, J., 2004. Generation of household solid waste in OECD countries: an empirical analysis using macroeconomic data. Land. Econ. 80 (4), 529–538. Karousakis, K., 2006. Municipal solid waste generation, disposal and recycling: a note on OECD inter-country differences. In: 24th March. Applied Environmental Economics Conference. Keser, S., Duzgun, S., Aksoy, A., 2012. Application of spatial and non-spatial data analysis in determination of the factors that impact municipal solid waste generation rates in Turkey. Waste Manage. 32 (3), 359–371. Khajuria, A., Matsui, T., Machimura, T., Morioka, T., 2012. Decoupling and environmental Kuznets curve for municipal solid waste generation: evidence from India. Int. J. Environ. Sci. 2 (3), 1670–1674. Kuznets, S., 1955. Economic growth and income inequality. Am. Econ. Rev. 45 (1), 1–28. Lei, K., Liu, L., Lou, I., 2016. An evaluation of the urban metabolism of Macao from 2003 to 2013. Resour. Conserv. Recycl. 128, 479–488. Lindmark, M., 2002. An EKC-pattern in historical perspective: carbon dioxide emissions, technology, fuel prices and growth in Sweden 1870–1997. Ecol. Econ. 42 (1),

266