Evaluating the Impact of Unit Based Pricing on Waste Production and Recycling in the Netherlands

(1)

Evaluating the Impact of Unit Based Pricing on Waste

Production and Recycling in the Netherlands

Student ID: s2049635

Student name: Lisa den Hamer

Master program: Public Administration Track: Economics & Governance Advisor: Kim Fairley

Second reader: Pierre Koning Date: 07-10-2020

(2)

2

Chapter 1: Introduction

In the fight against climate change, the world nations signed on the Paris Agreement on Climate Action to undertake real efforts to limit global warming. One of the most important goals of the Paris Agreement is to limit the rise of the global temperature below 2 degrees Celsius above pre-industrial levels. This can only be achieved by moving to a circular economy, which is a regenerative system in which resource inputs, waste, emissions, and energy leakage are minimized. The transition to a circular economy entails slowing, closing, and narrowing energy and material loops, which can be accomplished by repairing, reusing, remanufacturing, refurbishing, long-lasting designing and, more importantly, recycling (Wit et al., 2019).

The Circularity Gap Report 2019 released by Circle Economy, an organisation supported by the UN Environment and the Global Environment Facility, found that the global economy is only 9 percent circular with a negative trend. This means that only 9 percent of the more than 90 billion tonnes of fossil fuels and biomasses that enter the economy are reused every year. Even worse, the Circularity Gap is not closing as the upward trend in resource extraction and greenhouse gas emissions endures (Wit et al., 2019).

More and more countries are trying to deal with this issue. The Netherlands produces more waste per resident - household and industrial waste - compared to other European Union countries. The waste generated is processed in various ways. Waste can be deposited in landfills or incinerated to generate energy. Waste can also be recycled and reused to make new products. Recycling of waste is the most circular way as recycled waste can replace primary raw materials and thereby reduces the raw material requirement. According to the Central Bureau of Statistics (CBS) in the Netherlands, the amount of recycled waste generated in the Netherlands remained the same during the period between 2008 and 2016. Although recycling has hardly increased in the Netherlands, they have recycled substantially more compared to other countries. The statistical data shows that the Netherlands, together with Luxembourg and Belgium, recycle the largest amount of waste expressed in kilograms per resident in the European Union (Berkel & Schoenaker, 2020: 19).

Despite that, an important issue that municipalities are facing in the Netherlands, in order to meet national waste management objectives, is how to reduce the production of household waste or garbage (Bernstad, 2014). This is because of the growing quantity of garbage and the challenges involved in its collection, transportation, and disposal in ways that reduces the

(5)

5

impact on the environment. Garbage handling includes limiting production, increasing recycling diversion, encouraging composting of organic waste, and reducing the amount of garbage left for the municipality to collect (Abrashkin, 2015).

1.1. Research Background

In order to achieve the earlier mentioned environmental goals, the behaviour of individuals and nations need to change. Decisions regarding the environment are driven by multiple factors identified in behavioural science which include internal factors (such as personal motivation) external factors (such as financial incentives) and social norms (such as cultural norms and values). Therefore, municipalities use a wide range of policies focused on affecting environmental decisions of households, such as recycling behaviour. These measures might be a form of command-and-control stimuli, which sets specific laws and standards specifying behaviours, such as take-back schemes, as well as market-based incentives like landfill taxes, deposits, charges and fees, but it can also be indirect measures such as educational campaigns. A form of educational campaigns is primarily used, in contrast to direct measures such as financial incentives, to challenge the intrinsic motivation of individuals and build awareness. As a result, it is expected that attitudes of individuals towards, in this case the environment/recycling, change and therefore will be followed by a change in behavioural patterns (Remr, 2019). This thesis only focuses on forms of financial incentives and therefore, indirect measures will not be discussed any further.

Municipalities usually levy two types of economic instruments/financial incentives on households to reduce waste. This is in accordance with the principle proposed by Hahn and Stavins (1992) that economic instruments have more impact on the willingness of people to reduce polluting emissions compared to regulatory ‘command and control’ instruments. For example, Huang et al. (2019) concluded that the UBP policy was more effective than a mandatory recycling policy, because the unit-based pricing policy not only decreased the amount of unsorted waste but also increased biodegradable waste and recycling.

The objective of economic instruments is to assign an economic cost to garbage generation to trigger a change in the consumption patterns of households. This is in line with the ‘let the polluter pay’ principle (Ashworth et al., 2006). The first economic instrument charges a flat fee that households pay irrespective of the amount of garbage they produce (Ferrara & Missios, 2005). However, this amount is often too small to incentivise residents to reduce the amount of garbage produced or recycled. Sometimes this fee is included in property taxes and acts as

(6)

6

a hidden cost insulating households from the impact of the incremental amounts of garbage they generate (Skumatz, 2008). Thus, the fixed fee method has been unsuccessful in acting as a deterrent to excessive garbage generation.

The other economic instrument takes the form of unit-based pricing (UBP) also known as volume-based waste fee (VBWF) or Pay-As-You-Throw (PAYT) (Dijkgraaf & Gradus, 2008). This alternative to the fixed fee method charges households depending on the weight or volume of garbage they produce. That is, households pay a non-uniform rate based on the additional amounts of garbage they generate. The key benefits associated with UBP include a decline in the amount of garbage for disposal, a diversion of garbage to recycling purposes, a composting of organic waste, and a reduction in the total quantity of garbage (Abrashkin, 2015). Moreover, UBP is viewed as an additional funds generator for municipalities and validated as a mean for the just and fair allocation of ‘penalties’ on those who produce more garbage.

To achieve national waste management objectives, more and more municipalities are introducing the UBP mechanism in the Netherlands. In 2003, only 27 percent of all Dutch municipalities adopted some form of UBP (Allers & Hoeben, 2010). As of 2018, however, almost half of all Dutch municipalities (48 percent) have transitioned to the UBP system (Rijkswaterstaat, 2018). Several Dutch cities have adopted the PAYT form of UBP where households buy garbage disposal bags based on the amount of waste they expect to generate. The larger the estimated amount of garbage, the larger the bag and the higher the cost.

1.2 Research Problem

A concern regarding the implementation of UBP is related to the long-term benefits/effects of this scheme. Prior research suggests that the effectiveness of UBP diminishes in the long run as people might revert to their old behaviour once they get used to paying for the waste (Dijkgraaf & Gradus, 2014). On the contrary, people might gradually adapt their behaviour, in order to pay less for their waste. These opposing forces are known as “rebound effects”. Past research provides some evidence that a rebound effect can occur even a few years after the introduction of UBP (Usui & Takeuchi, 2012). Despite its significance, not much attention has been paid to this phenomenon. Therefore, it is crucial to study both short- and long-run effects of UBP to determine its long-term effectiveness.

(7)

7

1.3 Research Question

Given the prior mentioned developments regarding UBP, the main research question of this study is: What are the short- and long-term effects of different types of UBP systems on unsorted, recyclable and biodegradable waste generation in the Netherlands?

This question will be answered through the following sub-questions

• What are the various forms of waste pricing (unit-based pricing, fixed fee etc.)? • To what extent are the forms of waste pricing implemented in the Netherlands?

• What is the overall effect of UBP systems on unsorted, recyclable and biodegradable waste quantities in the Netherlands?

• What are the effects of different types of UBP systems on unsorted, recyclable and biodegradable waste quantities in the Netherlands?

• What are the long-term effects of UBP programs on unsorted, recyclable and biodegradable waste quantities in the Netherlands?

• To what extent is there a rebound effect in the long run?

1.4 Purpose of the Study

The main objective of this research is to quantify the effect of different unit-based pricing programs on the levels of unsorted waste, biodegradable waste and recyclable waste compared to the municipalities that did not implement a unit-based pricing system (i.e. control group). In doing so, this study will provide some useful insights for policy makers regarding local waste management policies to improve environmental conditions in the Netherlands, which in turn will aid in achieving the objectives set by the Paris Agreement.

Given the long-term waste management objectives of the Dutch government, the gaps mentioned above can be very problematic. Whether the cost of implementing a UBP system yields adequate returns for municipalities both in the short and long run needs to be examined. Moreover, the current confusion in understanding the link between different types of UBP programs and the outcomes associated with them needs to be addressed. It is this gap – the effectiveness of different waste-pricing mechanisms – that sets the stage for this thesis.

1.5 Academic and Social Relevance

This paper adds to the literature on UBP by indicating the relationship between different types of UBP systems and the amount of unsorted, biodegradable, and recyclable waste.

(8)

8

Furthermore, this paper adopts a differences-in-differences approach, which avoids many of the disadvantages of the commonly used household-level studies. These disadvantages include self-selection bias, limited period of consideration, smaller number of jurisdictions, and limited types of UBP programs. These limitations in the prior research affect the representativeness and generalization of its findings. By using a methodology that eliminates most of these limitations, this thesis presents findings that are more representative of the problem of garbage production and recycling and can be applied to other studies as well.

Additionally, this thesis makes a social contribution of highlighting how to handle the fast-growing quantities of household waste in the Netherlands. In a recent report, the European Commission (2018) estimates that each person consumes 16 tons of materials in the EU every year, of which 6 tons are wasted. In this scenario, it becomes imperative to know whether or not UBP is an effective tool to handle garbage. Therefore, this study will open up an academic debate for solid waste managers or other government authorities who wish to use UBP as a municipal garbage management tool.

This thesis extends the study of Allers and Hoeben (2010) by incorporating the latest data available on UBP and garbage disposal trends in the Netherlands. In doing so, it identifies which forms of UBP have a larger effect as an economic incentive in reducing the problem of garbage. It is important to understand the effectiveness of different UBP programs in reducing garbage production, recycling of garbage, organic composting, and leaving lesser volumes of garbage for the municipality to collect.

1.6 Methodology Brief

Data is collected on a sample of 190 Dutch municipalities that existed between 2001 and 2018 with respect to per capita household garbage weight per municipality. The data include household garbage collected at the curb side by either the municipality or a private contractor and at drop-off locations. Data on these metrics is collected for the 18-year period 2001-2018. The time period of 18 years can be considered long enough to capture the long-term effects of waste pricing. This annual data is collected from the CBS and COELO. This data is analysed using the statistical suite of Stata through multiple regressions to establish the relation between (different forms of) UBP systems and waste quantities.

(9)

9

1.7 Research Structure

The remainder of this thesis is organized as follows. First, the literature review which identifies the key variables associated with garbage disposal is presented. Then, the key hypotheses of this research based on the gaps in prior literature are formulated. Next, the methodology is described which explains the type of primary data, sources of data, and the reliability and validity (internal and external). The fourth chapter presents the analysis which evaluates the impact of different UBP systems on the amount of waste generated. Then, the results are discussed with respect to the literature review and the hypotheses. Finally, we summarize the key findings of the research, answer the research question, identify research strengths and limitations, and formulate recommendations for future research.

(10)

10

Chapter 2: Literature Review

In an attempt to achieve the objectives of the Paris Agreement, the recycling and reduction of waste have become extremely important concerns for all nations. The unit-based pricing (UBP) system might be a successful solution as research shows it can change peoples’ behaviour effectively. This research intends to examine this.

This chapter begins by introducing the key concepts relevant to this thesis. A review of the related literature follows. The literature review addresses three areas related to the impact of unit-based pricing system on the waste quantities and recycling behaviour. The first section discusses research related to the traditional pricing system for municipal solid waste disposal. The second section focuses on studies about various forms of unit-based pricing systems. Finally, the third section discusses the literature regarding the intentional and unintentional outcomes of unit-based pricing.

2.1 Municipal Waste

Municipal waste is also known as municipal solid waste or MSW (Bernstad, 2014). It comprises of solid waste that can be broadly classified as rubbish, recyclable materials, and organic waste. MSW is produced by households, commercial, and institutional entities. Rubbish is also known as trash, waste or garbage. It is typically disposed of in landfills or waste-to-energy conversion facilities (Gallardo et al., 2016). Recyclables are those materials that need not be disposed but diverted to create new products and commodities (Kawai & Tasaki, 2016). Organic waste is defined as material that can be biodegraded or composted and includes food / green waste and soiled paper materials (Callan & Thomas, 2006). For the purposes of this research, MSW is defined in terms of garbage or any materials that get produced in Dutch households and must be discarded. See table 2.1 for an overview of the materials that can be classified under different waste categories.

(11)

11 Table 2.1: Waste streams defined

Recyclable waste Biodegradable waste Other household waste

Metallic waste (i.e.): aluminum cans, construction steel rods, old utensils

Kitchen waste (i.e.):

vegetable/fruit peels, egg shells, leftover / contaminated old food, tea leaves

Glass, plastic or metal packaging

Paper (i.e.): notebooks, books, newspaper, cardboards

Yard waste (i.e.): old leaves, grass, weeds, flowers, wood

Small chemical waste Plastic (i.e.): beverage bottles,

trash bags, toiletry, old toys, food packages, water bottles

Diapers

Glass (i.e.): used beer and alcohol bottles, jars

Old electrical / electronic equipment

Textiles (i.e.): old clothes Carpets

Rubber and leather (i.e.): slippers, shoes, belts, old tires

Mattresses

Ceramics Flat glass

Medicines Cast iron

Light bulbs Asbestos

Batteries Clean soil

Electronics Fire extinguishers

Cadavers

2.2 Pricing Systems

The pricing systems that have been discussed in the literature in terms of economic incentives to reduce the problem of waste include the traditional or flat fee model and the UBP model. The UPB model in itself consists of various forms. The flat fee model and the various UBP models are discussed as follows.

2.2.1 Traditional Financing Model for MSW Disposal

In this model, a uniform fee is levied by all municipalities for all houses and establishments irrespective of the amount of garbage that is produced (Lebersorger & Beigl, 2011). The residents must place their garbage at the curbside once or twice a week, and this is then collected through municipalities’ owned and operated hauling services or through contracted private haulers. At times, the flat fee is included in the annual property taxes that must be paid and includes costs of collecting, disposing and recycling the garbage. Ferrara and Missios (2005) point out that in this method, the amount of the fee is not related to the actual amount of garbage produced. Therefore, this method is not associated with incentives for households to reduce waste. In cases where the fee is hidden in property taxes, households are disconnected from the actual costs of collecting and disposing garbage as well as from the incremental costs of garbage generated. Because householders are not aware of the costs of waste disposal, they are not incentivized financially to produce less garbage or to recycle waste. Anjum (2013) said

(12)

12

that this is because the flat fee levies do not act as financial incentives that trigger a behavioural change among households. This leads to a higher amount of waste and lower levels of recycling.

2.2.2 Unit-Based Pricing

The UBP waste fee system is an alternative to the traditional flat fee funding mechanism. In this system, households must pay a non-uniform fee based on the quantity of waste they generate (Kinnaman, 2006). This pricing system can be designed in different methods by municipalities in accordance with specific requirements of local communities. In the ‘Variable Collection in Size’ system, households can choose the number or dimensions of containers placed on the curbside for garbage collection (Mani & Singh, 2016). They can only dispose the quantity of garbage that fits in the containers or bins. This system is the most common UBP model because it is simple for residents and makes it easy for haulers to service residents living in single-family communities / dwellings (Mani & Singh, 2016). In the ‘Licensed Bag’ method, households must purchase licensed bags from the municipality and accommodate their waste in these bags only. The garbage placed in unlicensed bags is not collected by the municipality (Usui, 2008). The cost of the bags is included in their selling price and disposal charges resulting in additional tax revenues for the municipality. Moreover, as households use the bags being purchased, there are fewer issues related to invoicing or inventory in this type of system. In the ‘Sticker’ or ‘Tag’ system, households are required to buy stickers from the municipality and place them on bags that they fill with the garbage to be disposed (Davis & Kim, 2002). The sticker indicates how much garbage is in the bag and the cost of the stickers is incremental to the garbage volume left for disposal. In the hybrid system, households are charged a lower flat fee for garbage that fits into a standard-sized garbage disposal bin (Maskey et al., 2016). For additional garbage that does not fit into the bin, they must purchase licensed bags or stickers. In this method, households that produce extra garbage have to pay more for its disposal. This method is also known as a multi-tiered system and provides more security to the haulers as it ensures the coverage of fixed costs incurred (Nkansah et al., 2015).

Another method of classifying UBP programs is in terms of a metering system that determines costs of garbage disposal for residents. In the ‘Pay-Unit Service’, households buy bags, tags or stickers from municipal authorities depending on when they want to use them (Kipperberg, 2007). This method is a very simple billing mechanism as there is no need for municipal billing. However, it requires households to purchase / order bags on their own. In the ‘Subscription

(13)

13

Service’ method, households choose from different levels of fees in accordance with their garbage production (Ogwueleka, 2013). Based on the amount paid, they are allocated variably sized containers and bins. Thereafter, they are billed the same amount every month for the use of the same-sized bins. The larger the bin, the higher the price paid by the household. This method is simple as it involves payment of a fixed amount every month and eliminates pricing volatilities associated with pay-per-unit services. In the weight-based service, households are charged a fee based on the amount of garbage they generate as measured by the municipal crew that come for curbside collections (Taiwo, 2011). This method is less arbitrary than the other two methods mentioned above, and municipalities can alter or adjust the per-unit pricing more easily. However, it requires collection vehicles to be fitted with collection technologies that are either partly or fully automated as well as wireless communication systems (such as radio frequency identification or RFID) on both collection vehicles and garbage collection bins (Dijkgraaf & Gradus, 2008).

UBP programs are also broadly classified in terms of volume- and weight-based programs. In the former method, fees are levied according to the amount of garbage fitted into the bins whereas in the latter method, the amount of fee depends on the weight of the garbage (Nkansah et al., 2015). The volume-based method has the advantage of being easy to administer and providing some waste reduction incentives to households. However, the main disadvantage is that it fluctuates and does not provide any incentive beyond the lowest amount of subscription (such as one bag or container per week). The weight-based method provides a pricing signal that is transparent / continuous and similar to a weight-tipping fee (Trang, 2017). Its disadvantages include extended collection times, higher cost of equipment such as computerized billing systems and more chance of human error as it requires operators to hang bins or containers on the scale mechanism and scan barcodes fitted on the container. The weight-based system is more representative of the cost of waste disposal than the volume-based method as the fee is based on the weight. Furthermore, some technological developments such as the incorporation of RFID technologies reduce some of the disadvantages associated with weight-based pricing systems (Ferrara & Missios, 2005). Hauler education and testing weighing mechanisms more strictly for accuracy can reduce errors in levy of fee amounts. Moreover, the short-term investment costs of weight-based systems can be overcome by long-term savings from reduction in garbage production. Research on UBP indicates that it provides far more incentive for households to generate lesser amounts of MSW and to divert most of it

(14)

14

for recycling, and thereby minimizes the volume of garbage left on the curbside for collection (Kipperberg, 2007; Roy & Deb, 2013).

The above views summarize various forms of UBP systems. Their impact on garbage production and disposal is examined in the following section.

2.3 Outcomes of UBP Programs

Past research found various outcomes related to the impact of UBP programs. See also table 2.2 for a schematic overview.

2.3.1 Lower unsorted waste production

Wertz (1976) studied a pricing system for waste services that was implanted in San Francisco for different levels of household income. The author compared the average production of garbage in San Francisco in 1970 and the average amount of waste in other comparable US cities which had not adopted incentive pricing at that time. Wertz (1976) found that the quantity of garbage generated decreased when the waste tax increased. The author estimated the price elasticity of -0.15. This means that a 1 percent rise in incentive pricing results in a 15 percent decrease in the amount of waste produced. This finding was corroborated by Miranda et al. (1994) who observed an average waste reduction level of 30 percent in communities where UBP was introduced. Such a reduction could not be explained solely through recycling, composting, illegal disposal, or errors in measurement. It was found that most communities reduced production of waste in the first place.

The findings of Miranda et al. (1994) were supported by Adamec (1991) according to which the introduction of UBP in Illinois, lead to a total decrease of 31 percent in the amount of garbage produced by households. Together, these findings imply that one of the most important outcomes of UBP is a decline in garbage production or reduction of garbage produced by the source, i.e. households.

Linderhof et al. (2001) studied the effects of the implementation of the first weight-pricing system in the Netherlands in Oostzaan and measured similarities and dissimilarities between the behaviour of households before and after the fee was adopted. This is the first study which examined both short- and long-run effects of UBP. Their results indicated an increase in the price elasticity by 30 percent in the long term. This indicates that the effects of weight-based pricing are lasting. Moreover, the authors found that 3 years after the introduction of

(15)

weight-15

pricing system, annual waste collection was reduced by 42 percent. The amount of non-recycled waste was declined by 56 percent.

Dijkgraaf and Gradus (2004) extended the work of Linderhof et al., (2001) by measuring the effects of four different unit-pricing systems (based on weight, volume, bags and collection frequency) on the production of total, unsorted, biodegradable, and recyclable waste. In the case of unsorted waste, the authors found that there was a decrease in the quantity of waste of 50 percent for pricing based on weight or bags, a reduction of 27 percent if the pricing was based on collection frequency and 6 percent if it was based on waste volume. This means that, according to their research, the pricing systems based on weight and bags are the most effective.

Dijkgraaf and Gradus (2009) extended their 2004 study by using a larger panel data set and controlling for the differences in citizens’ attitude towards environmental concerns (i.e. environmental activism). The authors found sizeable and significant effects of different unit-based pricing systems, even after correcting for environmental activism. They found that the weight-, the bag- and the frequency-based system reduces the amount of unsorted waste by respectively 39 percent, 31 percent and 23 percent. They found no significant effects for the volume-based system. Also, their results indicate 9 percent less unsorted waste in municipalities with a high level of environmental activism. They conclude that not including environmental activism does overestimate the unit-based pricing effect, but environmental activism does not dominate this effect and therefore the effects of the different unit-based pricing systems are still sizeable and thus effective. Interestingly, they found that the effect of environmental activism decreased over time.

(16)

16 Table 2.2: An overview of earlier research on the outcomes of UBP

Reference Results

Lower unsorted waste production

Wertz (1976) A 1% increase in incentive pricing results in a 15% decrease in the amount of waste produced. Miranda et al. (1994) Waste reduction of 30% after implementation of UBP.

Adamec (1991) Waste reduction of 31% after implementation of UBP. Linderhof et al. (2001) Waste reduction of 30% as a result of UBP in the long run.

Dijkgraaf & Gradus (2004) A decrease in the quantity of waste of 50% for weight- and bag-based systems, a reduction of 27% for frequency-based system and a reduction of 6% for the volume-frequency-based system.

Dijkgraaf & Gradus (2009) The weight-, bag- and frequency-based system reduces the amount of unsorted waste by respectively 39%, 31% and 23%. No significant results for the volume-based system.

Dijkgraaf & Gradus (2014) For the weight-based system, there was a waste reduction of 45% in 1998 and of 34% in 2010. For the bag-based system, there was a waste reduction of 30% in 1998 and 41% in 2010. For the frequency-based system, there was a waste reduction of 6% in 1998 and of 22% in 2010. For the volume-based system, the waste reduction was between 3% and 6% in the period 1998-2010.

Dijkgraaf & Gradus (2016) The effects of the bag-, weight- and frequency system for unsorted waste are respectively 14,3 and 9 percent points. Allers & Hoeben (2010) For the weight-, bag- and frequency-based system, there was a waste reduction effect of respectively 39%, 28% and

21%.

Huang et al. (2019) They found that the UBP policy reduced the quantity of unsorted waste by 6.78 kg per capita per month. Higher recycling rates

Fullerton & Kinnaman (1996) The volume and weight of recyclable materials increased by respectively 37% and 16%. Kourtik (1990) Unit based pricing results in an increase of recycling tonnage by 60%.

(17)

17

Dijkgraaf & Gradus (2004) The amount of waste increased by 21% for the weight-based system and by 10% for the frequency-based system. Dijkgraaf & Gradus (2009) For the weight-based system and the bag-based system, recyclable waste increased by respectively 12% and 5%. Dijkgraaf & Gradus (2014) They found that the effect on recycling varies from 5% to 10% for the weight-based system, 5% to 14% for the

bag-based system and -1% to 10% for the frequency-bag-based system during the period 1998-2010.

Dijkgraaf & Gradus (2016) They found significant and positive effects of UBP on four types of recyclable waste (paper, textile, plastic and glass).

Huang et al. (2019) They found an increase of 7% of recyclable waste quantities due to the implementation of UBP.

Kim & Kim (2012) They found that UBP results in households adopting waste reduction techniques as well as increasing garbage recycling.

Palatnik (2014) The author found that households that are participating in a garbage recycling program are less likely to reduce the source production of waste.

Lower biodegradable waste rates

Dijkgraaf & Gradus (2009) The weight-, bag- and frequency-based system diminishes the amount of biodegradable waste by respectively 59%, 31% and 44%.

Allers & Hoeben (2010) The weight- and frequency-based system decreased biodegradable waste by respectively 51% and 43%.

Dijkgraaf & Gradus (2014) The weight-based system reduced biodegradable waste by 56% in 1998 and 53% in 2010. The frequency-based system, reduced biodegradable waste by 16% in 1998 and 38% in 2010. No significant effects were found on bag- and volume-based systems.

(18)

18

In 2014, Dijkgraaf and Gradus (2014) conducted another research using panel data for 1998-2010. In this study, the authors investigated both short- and long-run effects of unit-based pricing. For unsorted waste, they found that the weight-based system has a highly significant and negative effect of 45 percent in 1998 and 34 percent in 2010. However, the effect of the bag-based system on waste reduction increased over time from -30 percent in 1998 to -41 percent in 2010. The most interesting result is that the effect of the frequency-based system on unsorted waste was more than tripled from -6 percent in 1998 to -22 percent in 2010. The effect of the volume-based system was small (between -3 and -6 percent) which is not surprising since this system is less refined than the other systems. The results of Dijkgraaf and Gradus (2016) also show that the bag-based system effectively reduces unsorted waste rates with 14.3 percent points. However, the effects of the weight-based and frequency-based pricing systems are much smaller (no larger than 9 percent).

Allers and Hoeben (2010) used a 10-year dataset of all Dutch municipalities and found evidence that for unsorted waste, the weight system has the greatest price effect and reduces waste quantities by 39 percent. The bag system and the frequency system reduced garbage quantities by respectively 28 percent and 21 percent. This is because the bag- and frequency programs give an incentive to reduce garbage volumes, which can be accomplished by reducing weight, more likely, by compressing waste. As a result, it can be expected that a volume-based system has less effect on waste weights than a weight-based program. Similarly, Allers and Hoeben (2010) found that the impact of the weight-based program is almost twice of that of volume-based programs. Although, their results differ from the results of Dijkgraaf and Gradus (2009) who found that the weight program and bag program have similar impacts.

Furthermore, Huang et al. (2019) studied the comparison between a nationwide command and control practice that is essentially a mandatory recycling policy and the UBP policy in Taiwan. Under the mandatory recycling policy, people who dispose of recyclable waste along with ordinary waste will be fined and the waste collectors can refuse to collect mixed wastes. The authors found that, in contrast to the UBP policy that reduced the quantity of unsorted waste by 6.78 kg per capita per month, the mandatory recycling policy had no statistically significant effect on diminishing unsorted waste.

(19)

19 2.3.2 Higher recycling rates

Jenkins (1991) study in five American cities that had adopted incentive pricing system found that households began to actively consider recycling as an option to reduce waste. Households were particularly incentivized if they received money for recyclable materials. The study concluded that incentive pricing was a greater stimulus for recycling than charging a flat rate fee. Fullerton and Kinnaman (1996) measured quantities of garbage generated by 75 households in Charlottesville, before and after an incentive pricing system had been adopted. This was a unit-pricing system for weight with stickers that indicate 0.8 dollars for a 120-L bag and 0.4 dollars for a 60-L bag collected at the curbside. The authors found that the weight of the collected garbage decreased by 14 percent and the volume and weight of recyclable materials increased by 37 percent and 16 percent respectively. Similar findings were reported by Kourtik (1990) who found that the implementation of UBP systems in Seattle resulted in an increase of recycling tonnage by 60 percent, with the overall recycling participation rate in the city increasing by 80 percent.

Dijkgraaf and Gradus (2004) found, for recyclable waste, that the amount of waste increased by 21 percent for a weight-based pricing and by 10 percent when the system was based on frequency. The pricing system based on volume did not have a significant effect on the quantity of recyclable waste. Dijkgraaf and Gradus (2009) found that the weight-based system leads to higher effort of 12 percent in recycling of glass, paper and textiles. The system based on frequency increased the recyclable waste only by 5 percent. Furthermore, they found that recyclable waste in municipalities with a high level of environmental activism (‘green’ municipalities) is 6 percent higher which means that households in municipalities with a UBP program are more active in sorting their waste regardless of the presence of such a UBP system. This means that municipalities with high waste levels do not necessarily have more incentives to introduce a UBP system.

Suwa and Usui (2015) examined the garbage reduction effect and the substitution effect between garbage and recyclables due to UBP. They collected data of garbage generation and amount of recyclables from 1726 municipalities in Japan in 2010. Suwa and Usui (2015) found that higher levels of UBP for garbage lead to higher amounts of PET bottles, plastic containers, and paper containers for recycling. Furthermore, they found that the amount of garbage diminished. Their results suggest that higher levels of UBP leads to an increase in some type of recyclables collected. Usui and Takeuchi (2014), who also studied cities in Japan, examined

(20)

20

the effect of a UBP policy by measuring at least 12 years after the introduction of UBP, and found a long-lasting effect of unit-based pricing on recycling.

More recent studies including Dijkgraaf and Gradus (2014, 2016) and Huang et al. (2019) also demonstrated that incentive pricing systems based on weight, volume or bag had a negative effect on the reduction of garbage and a positive effect on the amount of recycled waste. For recyclable waste, Dijkgraaf and Gradus (2014) found a significant and positive effect of weight-, bag-, and frequency-based pricing. Specifically, they found that the effect on recycling varies from 5 to 11 percent for weight-based pricing, 5 to 14 percent for bag-based pricing, and -1 to 10 percent for frequency-based pricing system during the period 1998-2010. Dijkgraaf and Gradus (2016) found that the UBP coefficients are significant and positive for all four types of recycling rates (paper, textile, plastic, and glass). Also Huang et al. (2019) found significant and positive effects of unit-based pricing of an increase of 7 percent of the quantity of recycling. From these findings it may be concluded that UBP encourages participation in garbage recycling.

Kim and Kim (2012) evaluated the interactions between source reduction and participation in recycling and found that both programs are complementary. UBP stimulates greater awareness about waste and appropriate ways to handle it. This induces households to adopt waste reduction techniques as well as increase garbage recycling. However, Palatnik (2014) observed that households that are participating in a garbage recycling program are less likely to reduce the source production of waste. This can occur when recycling services are provided free of charge. To encourage both forms of garbage handling, it was recommended that UBP for recycling services should be set at a lower rate than for garbage collection.

2.3.3 Lower biodegradable waste rates

Another effect of UBP is a decrease in biodegradable waste rates. Dijkgraaf and Gradus (2009) found that pricing waste on the basis of weight diminishes biodegradable waste by 59 percent. The bag based and frequency-based systems diminishes the amount of biodegradable waste by respectively 31 percent and 44 percent. They state that Dutch households probably use home composting to reduce this type of waste. Allers and Hoeben (2010) reported that the effect of the weight system (51 percent) is larger than under the frequency system (43 percent). For the bag system, they found no significant effect in relation to the biodegradable waste. Dijkgraaf and Gradus (2014) reported that the weight-based pricing reduced biodegradable waste by 56 percent in 1998 and 53 percent in 2010 representing a quite stable effect on composting over

(21)

21

time. More importantly, they found that the frequency-based pricing reduces biodegradable waste by 16 percent in 1998 and 38 percent in 2010. This might be because of an improvement in the waste collection mechanism by municipalities. But the authors did not find any significant effects for the bag- and volume-based pricing on biodegradable waste. However, Tuladhar and Spuhler (2016) observed that the adoption of home-composting depends on awareness and training among residents and the availability of gardens or flowerpots in which to do so.

2.4 Negative side effects of UBP programs

2.4.1 Higher waste compaction

Another phenomenon associated with the UBP is waste compaction. This is the packing of garbage as densely as possible to minimize per-container fees (Cooper, 2003). The study of Fullerton and Kinnaman (1996) in Charlottesville, Virginia, found that the volume of garbage from households reduced by about 37 percent whereas the total reduction in weight was only 14 percent. It was concluded that this was due to garbage compaction through stomping by residents. In other words, they compressed waste to reduce volume (the so-called Seattle Stomp). This suggests that the weight-based system may provide a more accurate pricing mechanism than the volume-based system. Dijkgraaf and Gradus (2016) found that the bag-based system, besides that it is an effective system which reduced unsorted waste with 14.3 percent points, has an incentive for households to put as much waste as possible in each bag which makes it more difficult to handle these bags.

2.4.2 Illegal dumping

Yet, another negative effect of UBP is undesirable diversion of waste through dumping in neighbours’ bins, burning, and dumping of waste in the workplace. Fullerton and Kinnaman (1996), in their study of Charlottesville, observed that up to 28 percent of all reduction in waste production post implementation of UBP was due to undesirable diversion. Huang et al. (2019) also found some evidence for illegal dumping. After the unit-based pricing program was implemented, the government of Taipei encouraged residents to be involved in monitoring any illegal dumping by providing a unique online internet platform where residents can upload evidence of violators in a form of videos or pictures. Their results showed that the number of violations for illegal dumping increased immediately after the implementation of UBP policy but dropped dramatically a few years after the policy became effective. This might imply that the implementation of a form of UBP could increase illegal dumping but has no long-term

(22)

22

effect (as the violations drop over time). Kim et al. (2008) confirmed these findings since they estimate that a one percent increase in the unit price of a trash bag resulted in a 3 percent increase in the number of reports of illegal dumping in South Korea. By contrast, Dijkgraaf and Gradus (2004) and Allers and Hoeben (2010) found no evidence for illegal dumping as there is no indication that surrounding municipalities without UBP collect part of the garbage produced by municipalities with a UBP policy. In their opinion, one would expect that many municipalities would have terminated their UBP policies if illegal dumping would occur. According to them, this has not happened. Furthermore, the results in the study by Carattini et al. (2018) indicate that four years after the implementation of a UBP program, illegal dumping remained a minor issue.

2.5 Rebound effect

In the context of this study, a rebound effect refers to an increase or decrease in the effectiveness of a unit-based pricing system in the long run. A rebound effect can be defined as an effect of unit-based pricing in the long run that can be attributed to the unit-based pricing, not to some other cause. The rebound effect is illustrated in Figure 2.1 (see also Usui & Takeuchi, 2012: 247). The vertical axis shows the amount of waste per capita and the horizontal axis shows time. In this case, it is assumed the waste generation will increase in time. When unit-based pricing is implemented at time t1, the amount of waste reduced from r0 to r1. A few

years after the implementation of unit-based pricing, waste generation might grow and approach the level before the implementation. There might not necessarily be a rebound effect as there might be a positive time trend in waste generation. If the predicted waste generation (the dashed line) and realized waste generation (the solid line) run parallel, there is no rebound effect, because the increase in waste generation cannot be attributed to the effect of unit-based pricing. But if the slope of the latter is greater than the former, there is a rebound effect. There is also a possibility that waste generation decreases below the dashed line (see Figure 2.1); this is called an ‘inverse rebound effect’.

(23)

23 Figure 2.1: Illustration of rebound effect

Depending on the positive or negative change in citizens’ attitude towards waste reduction, different rebound effects can occur. On the one hand, a few years after the introduction of UBP, residents might get used to the UBP system and revert to their old behaviour. This represents a negative rebound effect known as ‘awareness erosion’. The awareness erosion effect implies that the effectiveness of UBP erodes over time as people become less sensitive to paying for waste. In other words, households get used to pricing waste and, as a result, go back to their old behaviour and the awareness effect of unit-based pricing might disappear over time (Dijkgraaf & Gradus, 2008; 2014). On the other hand, residents might gradually adapt their behaviour by learning how to sort out and separate recyclable waste. Households need time to learn how to change their behaviour. This represents a positive rebound effect known as ‘the learning effect’. The learning effect implies that the effectiveness of UBP increases over time as people learn to change their behaviour in an optimal direction (Dijkgraaf & Gradus, 2014).

Earlier research mainly focuses on the short-run effects of UBP on waste reduction. In this regard, researchers have examined the impact of UBP on different types of waste collected, such as total, unsorted, sorted, biodegradable, and recyclable waste (e.g. Hong et al., 1993; Callan & Thomas, 2006; Kinnaman & Fullerton, 2000; Jenkins, 1993; Allers and Hoeben, 2010). However, there have been fewer studies that examine the long-run effect of UBP on waste reduction. Some studies that do investigate the long-run effect of UBP include Linderhof

(24)

24

et al. (2001), Dijkgraaf and Gradus (2009, 2014), and Usui and Takeuchi (2012). Linderhof et al. (2001) is the first study that examines the long-run effect of UBP using a panel data set. In this study, the authors empirically showed that the long-run price elasticity of weight-based pricing is 30 percent higher than the short-run price elasticity.

The studies by Dijkgraaf and Gradus (2009, 2014) investigated the long-run effect of UBP using a panel data set on waste collection by several municipalities in the Netherlands. Dijkgraaf and Gradus (2014) study found some evidence for awareness erosion and the learning effect. For unsorted waste, the authors found a decrease in the effectiveness of pricing based on weight and bag in the long run. Furthermore, they found a significant and positive effect for the frequency-based system on the amount of biodegradable waste collected in the long run. One possible explanation for this effect is that people gradually learn to avoid fee by separating unsorted waste. Dijkgraaf and Gradus (2009) found evidence for a learning effect in this study, although the effect was not sizeable. However, they found nearly no evidence for an awareness erosion effect. They conclude that neither learning effects nor awareness erosion play a dominant role, which means that their results implicate that the effect of unit-based pricing does not decrease over time.

Using panel data on 665 cities in Japan, Usui and Takeuchi (2014) investigated the long-term effect of UBP on waste reduction and recycling. In contrast to Dijkgraaf and Gradus (2009), the authors found evidence for the presence of rebound effects, although the size of the effects were negligible. With respect to waste reduction, they found that the effect of UBP decreases as more years elapse, after the implementation of UBP. Moreover, they found that the effect of UBP on waste recycling first decreases for up to 16 years and then gradually started increasing. Furthermore, the authors concluded that the learning effect on waste recycling becomes stronger than the awareness erosion effect on waste reduction in the long run. This implies UBP is more effective in increasing recycling than reducing waste generation in the long run.

2.6 Description of systems

For the purpose of this study, the following descriptions / definitions of the flat fee system and the different UBP programs will be used (see tables 2.3 and 2.4).

(25)

25 Table 2.3: Description of flat fee system

Flat fee model Description

Flat rate system Residents pay a fixed amount irrespective of the amount of garbage that they produced

Table 2.4: Description of UBP programs

UBP program Description

Frequency-based system Residents pay a fee every time their container is emptied at the curbside

Bag-based system Residents buy special garbage bags or labels to put on their own bags. The more waste a household produces, the more bags a household needs

Weight-based system Residents pay a fee depending on the weight of the waste offered at the curbside

2.7 Hypotheses

The findings presented in the literature review imply that the impact of UBP systems is to reduce unsorted waste produced, to enhance waste recycling, to promote waste composting, and to minimize the amount of waste left on the curbside for municipalities to collect. The negative side effects include waste compacting and undesirable diversion. In the Netherlands, the initial research on UBP programs indicated that the weight-based pricing system is more of an incentive to reduce waste than the frequency-based pricing system. Furthermore, prior research provides some evidence regarding the awareness erosion and learning effect in the long run. Based on these findings, the key hypotheses are constructed below.

H1: There is a negative and significant relationship between UBP programs (in general), as compared to the flat fee model, and the quantity of unsorted waste produced by Dutch households.

H2: There is a positive and significant relationship between UBP programs (in general), as compared to the flat fee model, and the quantity of recyclable waste (glass, paper and textile) produced by Dutch households.

H3: There is a negative and significant relationship between UBP programs (in general), as compared to the flat fee model, and the quantity of biodegradable waste produced by Dutch households.

(26)

26

H4: Compared to the flat fee model, the individual UBP programs – weight-, bag- and frequency-system – significantly reduce the quantity of unsorted waste produced by Dutch households.

H5: Compared to the flat fee model, the individual UBP programs – weight-, bag- and frequency-system – significantly increase the quantity of recyclable waste (glass, paper and textile) produced by Dutch households.

H6: Compared to the flat fee model, the individual UBP programs – weight-, bag- and frequency-system – significantly reduce the quantity of biodegradable waste produced by Dutch households.

H7: After the introduction of UBP programs, the waste reduction effect regarding the unsorted and biodegradable waste erodes over time.

H8: After the introduction of UBP programs, the waste separation or recycling effect regarding recyclable waste erodes over time.

(27)

27

2.8 Conceptual Model

This paragraph translates the hypotheses into the conceptual model of this thesis.

_ + _

UBP:

- Weight system - Frequency system - Bag system

Biodegradable

waste

Unsorted

waste

Recyclable

waste

(28)

28

Chapter 3: Methods and Data

The main objective in this thesis is to investigate the short- and long-term effects of different UBP systems on the total waste quantities and waste separation in the Netherlands. This study follows a quantitative research design, using a differences-in-differences analysis. In this chapter, the data and method used is discussed. This chapter begins by presenting the reasons for choosing a quantitative approach. Next, an extensive description of the data follows containing a description of the population and of the sample. Then, the method of data collection and the method of analysis are described. Finally, the reliability and validity of the measures is presented.

3.1 Why quantitative research?

The reason for choosing a quantitative research design is twofold. First, the number of observed subjects (population), which are municipalities (553) that existed in the Netherlands between 2001 and 2018, can be considered large enough to resort to a quantitative approach. Based on the population (553) and the sample size (190), it would be difficult to execute qualitative measures, such as a case study or small N comparisons. As a qualitative method requires deep understanding of the underlying reasons of the phenomenon studied, it would be an extremely time- and energy- consuming task to apply a qualitative method to the large number of research subjects in this study that goes beyond the author’s capability given the limited amount of time and resources. However, quantitative methods are perfectly capable of processing large amounts of data through statistical techniques.

Second, the essence of the research question in this study is more suitable to a quantitative design than a qualitative design. The research question in this study enquires a confirmation or denial of the existence of relationships between the independent (UBP programs) and dependent variables (different waste streams), which can only be accomplished through a statistical analysis. In a quantitative design, the results should be, to some extent, generalizable to the population. Observing the ‘overall pattern’ is the advantage of quantitative methods, which is also the main purpose of this study. Qualitative methods are better at providing more in-depth analyses of small sample size cases, but due to the limited amount of cases analyzed in such studies, the generalizability of the results is much weaker than of the quantitative results.

(29)

29

3.2 Data description

The quantities of unsorted, recyclable, and biodegradable waste per capita (in kilograms) are the dependent variables for this research. Particularly, this study distinguishes between these different waste streams as data is collected for biodegradable waste (such as vegetable, food and garden waste), recyclable waste (glass, paper and textiles), and unsorted waste. Data on the dependent variables – the quantities collected of unsorted, biodegradable, and recyclable waste in kilograms per inhabitant – come from the CBS (the Dutch Central Bureau of Statistics) who sends an annual inquiry to the waste collection units of all Dutch municipalities (CBS, 2019a). The figures for the quantity of waste collected provided by these waste collection units are reliable as the amount they have to pay depends on the quantity of waste supplied to waste treatment firms.

The independent variables are the different types of UBP programs. Data on the independent variables come from COELO, an independent research institute affiliated with the University of Groningen. This institute focuses on research of local and regional taxes, for instance property tax, waste tax etc. They keep track of what is levied where and investigate the backgrounds. Furthermore, since prior similar research (see Allers & Hoeben, 2010) has also controlled for several demographic characteristics including population density, average household size, percentage of young children, percentage of the elderly, the number of inhabitants of a municipality, and percentage of immigrants, this study also controlled for these demographic characteristics. The sample of municipalities is divided into two groups: the treatment group and the control group. The treatment group includes municipalities with a UBP system and the control group includes municipalities without a UBP system (flat fee system). These groups will be discussed in more detail in paragraph 3.3. The data was collected from the period 2001 to 2018.

Dutch municipalities are free to choose the financing mechanism for the collection of unsorted and biodegradable waste. In most municipalities (63.6 percent in 2001 and 46.8 percent in 2018), garbage fees depend on household size. This is a flat fee, which means that the price is independent of waste quantities. Another flat fee policy used by Dutch municipalities is a pricing program with an annual fixed cost per household. In 2001, 11.6 percent of the municipalities used this fixed cost per household program compared to 6.2 percent in 2018. Furthermore, a volume-based pricing system is a program where residents pay a fee depending on the volume / size of their bin or bag. In 2001, 34 municipalities used such a volume-pricing system. In 2018, this number declined to 18 municipalities. This volume-based program is not

(30)

30

considered as a UBP system because households pay a fee according to the bin size they choose and are therefore pre-committed to a particular volume and are charged regardless of their actual waste quantities. Therefore, the costs of a marginal increase in waste will be zero under this system in most cases. These three different types of flat fee policies are considered as the ‘flat rate / fee model’ (i.e. the control group) in this study.

In the Netherlands, there are six different types of UBP. Most common is a frequency system where residents pay a fee every time their container is emptied at the curbside. In 2001, 9.2 percent of the municipalities used such a system, and this increased up to 29.2 percent in 2018. Another system is the bag program where residents buy special garbage bags or labels to put on their own bags. The more waste a household produces, the more bags a household needs. The bag system was used by 12 municipalities in 2001, which later increased to 21 municipalities in 2018. This bag system is also used in combination with a fixed fee system which depends on household size (mentioned above) by 10 municipalities in 2001 and 8 municipalities in 2018 (‘bag + household size’). The use of a weight-based system declined from 17 municipalities in 2001 to 10 municipalities in 2018. Under a weight-based system residents pay a fee depending on the weight of the waste offered. Municipalities with such a system use garbage trucks which automatically weigh every container before and after emptying and combines this information with the identity of the owner stored in a chip integrated in the collection bin. The reason for the decline of this system is probably the high administrative costs for the respective system.

Furthermore, a few municipalities (4 in 2001 and 11 in 2018) use a combination of the weight and frequency systems (‘weight + frequency’). Here, residents pay for every time their bin is emptied but also for the weight of their garbage. In 2016, one municipality started to combine the bag system with the frequency-based system (‘bag + frequency’). Table 3.1 gives an overview of the unit-based pricing systems used in the period 2001-2018 for the population. The percentage of municipalities using a flat rate (i.e. flat fee model) decreased from 82 percent in 2001 to 58 percent in 2018.

(31)

31

Table 3.1: Use of UBP by year: proportion of municipalities (population)

Table 3.2 shows the proportion of unit-based pricing systems weighted by the number of inhabitants from 2001 to 2018 for the population. In the last column, the weighted proportion of municipalities using a flat rate is given; it decreased from 88 percent in 2001 to 70 percent in 2018. This indicates that UBP programs are applied more in small municipalities whereas flat rate systems are used more in large cities. The development of different unit-based pricing systems over time is also illustrated by Figure 3.1, which shows that the use of the frequency system has especially increased and more than tripled between 2001 and 2018. Prevalence to the bag systems especially increased after 2015. However, the weight system declined after 2010 while the weight system in combination with the frequency system increased.

Frequency Bag Bag + household size Bag + frequency Weight Weight + frequency Flat rate 2001 0.09 0.02 0.02 0 0.03 0.01 0.82 2002 0.11 0.02 0.02 0 0.04 0.01 0.80 2003 0.11 0.03 0.02 0 0.04 0.01 0.80 2004 0.13 0.02 0.02 0 0.04 0.01 0.78 2005 0.14 0.02 0.02 0 0.04 0.01 0.76 2006 0.15 0.02 0.02 0 0.04 0.01 0.75 2007 0.15 0.02 0.02 0 0.04 0.01 0.75 2008 0.16 0.02 0.02 0 0.04 0.01 0.74 2009 0.17 0.02 0.02 0 0.04 0.01 0.73 2010 0.17 0.02 0.02 0 0.04 0.01 0.73 2011 0.18 0.02 0.02 0 0.03 0.02 0.72 2012 0.19 0.03 0.02 0 0.03 0.02 0.71 2013 0.21 0.03 0.02 0 0.03 0.03 0.69 2014 0.25 0.02 0.02 0 0.03 0.03 0.65 2015 0.25 0.03 0.02 0 0.03 0.03 0.65 2016 0.28 0.03 0.02 0 0.03 0.03 0.61 2017 0.28 0.05 0.02 0.002 0.03 0.03 0.60 2018 0.29 0.05 0.02 0.002 0.03 0.03 0.58

(32)

32

Table 3.2: Use of UBP by year: weighted by inhabitants (population)

Frequency Bag Bag + household size Bag + frequency Weight Weight + frequency Flat rate 2001 0.06 0.03 0.01 0 0.02 0.01 0.88 2002 0.07 0.03 0.01 0 0.03 0.01 0.86 2003 0.07 0.03 0.01 0 0.02 0.01 0.86 2004 0.09 0.03 0.01 0 0.03 0.01 0.84 2005 0.10 0.03 0.01 0 0.03 0.01 0.83 2006 0.10 0.02 0.02 0 0.03 0.01 0.83 2007 0.11 0.02 0.02 0 0.02 0.01 0.82 2008 0.11 0.02 0.02 0 0.03 0.01 0.82 2009 0.12 0.02 0.02 0 0.03 0.01 0.81 2010 0.12 0.02 0.02 0 0.03 0.01 0.81 2011 0.13 0.02 0.02 0 0.02 0.02 0.81 2012 0.14 0.02 0.02 0 0.01 0.02 0.80 2013 0.16 0.02 0.02 0 0.01 0.02 0.78 2014 0.18 0.02 0.02 0 0.01 0.02 0.75 2015 0.18 0.02 0.02 0 0.01 0.02 0.75 2016 0.19 0.02 0.02 0.001 0.01 0.02 0.74 2017 0.20 0.04 0.01 0.001 0.01 0.02 0.71 2018 0.21 0.04 0.01 0.001 0.01 0.02 0.70

Figure 3.1: Use of UBP by year: percentage of municipalities (population)

% o f m u n icip alities Years Frequency/bin Bag

Bag + Household size Bag + frequency Weight Weight + frequency

(33)

33

3.3 Sample

In this paragraph, the process of selecting cases for this study, also known as the sampling plan, is described. Specifically, how the sample of municipalities is selected from the broader population will be explained. In the context of this thesis, the broader population consists of 553 municipalities that existed between 2001 and 2018. This includes municipalities that merged with other municipalities due to municipal re-divisions. Therefore, the number of municipalities declined every year from 504 in 2001 to 380 in 2018 (CBS, 2019b). From this population of municipalities, a sample based on the years a municipality ‘existed’ and their geographical location is selected. Every municipality that started to use a UBP program between 2001 and 2018 (95 in total) was assigned to the treatment group. Except for Loppersum, which had too many missing values (42) to be useful for the model and was therefore excluded. Because of this, the combination of the bag and frequency system (‘bag + frequency’) is not part of the sample and therefore not included in this study.

Then, these municipalities are matched to the municipalities with a flat rate policy (i.e. the control group) that existed in the same period and were closest geographically. For instance, the municipality Bellingwedde, which existed from 2001 until 2017 and started to use a UBP program in 2010, was assigned to the treatment group. Bellingwedde was matched to the municipality Vlagtwedde, which was assigned to the control group, as it used a flat fee policy and also existed from 2001 until 2017 and is a neighboring municipality of Bellingwedde. The reason for this approach was to correct for the possibility that people have become more environmentally conscious over the years and that sustainability norms might depend on the region in which a municipality is located. This might influence the amount of sorted and unsorted waste generated independent from the introduction of UBP. Table 3.3 gives an overview of the unit-based pricing systems used in the period 2001-2018 for the sample. Furthermore, table 3.4 shows the proportion of unit-based pricing systems weighted by the number of inhabitants from 2001 to 2018 for the sample selection.

(34)

34

Table 3.3: Use of UBP by year: proportion of municipalities (sample)

Table 3.4: Use of UBP by year: weighted by inhabitants (sample)

3.4 Procedure (method of data collection)

In this section, how the data was collected and the procedures that were followed throughout the study, is explained. The data was collected from the Central Bureau of Statistics (CBS) in the Netherlands and the COELO for the years 2001 until 2018. This annual data collected from the CBS consists of per capita household garbage weight per municipality for recyclable

Frequency Bag Bag + household size Weight Weight + frequency Flat rate 2001 0 0 0 0 0 1.00 2002 0.017 0 0 0.017 0 0.97 2003 0.033 0 0 0.011 0 0.96 2004 0.065 0 0 0.016 0 0.92 2005 0.087 0 0 0.016 0.005 0.89 2006 0.092 0 0 0.016 0.005 0.89 2007 0.11 0 0 0.011 0.006 0.87 2008 0.13 0 0 0.022 0.006 0.84 2009 0.15 0 0 0.022 0.006 0.83 2010 0.15 0 0 0.022 0.006 0.82 2011 0.16 0 0 0.017 0.011 0.81 2012 0.17 0.006 0 0.017 0.011 0.80 2013 0.23 0 0.006 0.017 0.022 0.73 2014 0.3 0 0.006 0.017 0.022 0.66 2015 0.31 0 0.006 0.017 0.017 0.65 2016 0.37 0.006 0.006 0.017 0.017 0.59 2017 0.39 0.022 0.011 0.017 0.017 0.55 2018 0.41 0.034 0.011 0.017 0.017 0.51

Frequency Bag Bag + household size Weight Weight + frequency Flat rate 2001 0 0 0 0 0 1.00 2002 0.009 0 0 0.018 0 0.97 2003 0.015 0 0 0.015 0 0.97 2004 0.055 0 0 0.017 0 0.93 2005 0.064 0 0 0.017 0.002 0.92 2006 0.072 0 0 0.017 0.001 0.91 2007 0.079 0 0 0.015 0.001 0.91 2008 0.09 0 0 0.018 0.001 0.89 2009 0.1 0 0 0.017 0.001 0.88 2010 0.11 0 0 0.017 0.001 0.87 2011 0.12 0 0 0.005 0.013 0.86 2012 0.13 0.005 0 0.005 0.013 0.84 2013 0.18 0 0.005 0.005 0.021 0.79 2014 0.22 0 0.005 0.005 0.021 0.74 2015 0.23 0 0.005 0.005 0.02 0.74 2016 0.26 0.004 0.005 0.005 0.02 0.71 2017 0.29 0.012 0.008 0.005 0.02 0.67 2018 0.3 0.021 0.008 0.005 0.02 0.64

Evaluating the Impact of Unit Based Pricing on Waste Production and Recycling in the Netherlands