In transition to electric trucks

(1)

IN TRANSITION TO ELECTRIC TRUCKS

Rosa Schoenmaker

The electrification of trucks for regional

freight transport in the Netherlands to

mitigate climate change and reduce air

pollution; Insight in transition pathways

(2)

Master’s thesis by Rosa Schoenmaker

Student number VU: 2577324

Student number UvA: 11427868

E-mail: rosaschoenmaker@live.nl

MSc. Physics: Science, Business & Innovation

Vrije Universiteit Amsterdam (VU)

Universiteit van Amsterdam (UvA)

Commissioned by EVConsult

Supervisor: dr. M.L. Blankesteijn

External supervisor: R. van Sloten, MSc.

Co-assessor: dr. J.P. Dekker

Graduation period: January to July 2018

Amsterdam, 22-06-2018

(3)

Preface

This report represents the research conducted for the purpose of the master’s thesis project

for the master’s degree in Physics: ‘Science, Business & Innovation’. The research project was

supervised by Dr. Marie Louise Blankesteijn from the VU Amsterdam who was also the first

assessor. Ruud van Sloten, MSc, was the external coach who works at the internship company

EVConsult. The co-assessor of the research was Dr. Jan Dekker.

Acknowledgements

Almost six months ago I started at EVConsult and kicked off the final research project of my

study career at the university. Now that I am writing the final words of my thesis, I think it’s a

good time to thank the people who have supported me during the project. With their

contribution, the project not only resulted in this report but also in many teachable moments

on a personal level.

First of all, I would like to thank all experts who were willing to find an hour in their busy

schedules to cooperate with my investigation.

Furthermore, I would like to thank Ruud van Sloten and Marlous Blankesteijn for their valuable

supervision on the research project and their coaching during the entire process. Ruud, also

thank you for providing me with a lot of relevant information, contacts, and views on the

subject. Marlous, I would like to thank you especially for supporting me in getting back on the

right track at the moment I was losing the forest for the trees.

Next to my supervisors, several other persons have also made the effort to read my report

and to provide me with feedback. Thanks for this goes to my co-assessor Jan Dekker, to my

fellow graduate researcher at EVConsult, Tim van ‘t Wel, and to my brother Hugo

Schoenmaker who sacrificed his holidays for this.

Finally, I would like to thank my colleagues at EVConsult, my roommates, my friends and my

parents. You have made my internship period very enjoyable and besides, you supported me

with moral support when I needed it.

(4)

Executive summary

Introduction

To mitigate climate change and reduce air pollution in the freight sector, a transition to zero-emission vehicles is required with electric trucks being a serious option. The aim of this case study research is to generate insight in this transition by generating an answer to the following research question;

Which transition pathways will the transition towards electrification presumably follow, using insights from the multi-level perspective on socio-technical transitions and from factors influencing technology adoption?

The case analysed in this research is the regional freight transport system in the Netherlands concerning heavy-duty trucks driving medium-hauls (of 300 km on average).

Theory and methodology

The lens through which the research was conducted was grounded in theories on socio-technical system transitions. The multi-level perspective on socio-technical transitions as described by Geels (2004), and a further refinement of that theory explaining different transition pathways, formed the basis to the analyses.

The data for the analyses were collected through desk research (i.a. policy reports, consultancy reports, and expert blogs), ten qualitative expert interviews and one focus group. Several iterative analyses were performed on the collected data to distract relevant information on variables established from the literature. These variables are important to the determination of transition pathways.

Results

The findings reveal that at the supply side, the profit streams in the current system are primarily determined by combustion engines and diesel. At the demand side, the system is characterised by high competition and small profit margins. The trends on mitigating climate change and reducing air pollution are currently putting moderate pressure on the system which is mainly constituted by policy regulations and this pressure is expected to increase in the future. Battery-electric and fuel cell-electric trucks are candidates to enable this transition, but both show many challenges to adoption. Although hydrogen gets support from the Dutch government, the fuel cell-electric option requires more R&D to improve the fuel cell technology and the technologies to store hydrogen. Regarding battery-electric trucks, driving ranges over 100 km do not show a positive business case due to high costs of batteries. Improving those business cases, as well as the overall adoption of battery-electric trucks, faces multiple technical and organisational barriers. Among those, are the limited lifetime of the battery, the high weight and volume of the battery, and the reduction of the flexibility of the fleet.

Discussion & Conclusion

From the results, it became clear that both technological options (battery-electric and fuel cell-electric) are currently not sufficiently developed to meet the needs of most regional transport economically and practically. In combination with a currently moderate pressure on the stakeholders (i.e. original

(5)

activities. The pressure exerted by policy regulations will increase in the future whereby stakeholders will be forced to increasingly adopt electric trucks. This will only lead to direct substitution of diesel engines (i.e. that does not require additional changes to transport operations) when the electric trucks (battery- and/or fuel cell-electric) will have been improved sufficiently by that time. If not, no evident substitute to the internal combustion engine will be available whereby the increased pressure on the system will cause de-alignment of the stakeholders in the system. This unstable period will end when one of the two options gains support by the majority of the system, after which the system re-aligns.

Future research & policy recommendations

The research contributes to the existing literature by closing a gap in practice-oriented studies and generating suggestions for future research. Among others, future researchers should provide more insights on the potential economic performances of fuel cell-electric trucks and on the potential integration of biofuels in the transition to cover the period in which electric trucks are not ready yet. The research contributes to practice through the provision of policy recommendations. Policymakers should clearly articulate their visions and strategies on the transition in order to create trust and avoid resistance. Besides, policymakers should focus on stimulating market demands for battery-electric trucks with driving distances shorter than 100 km, and on facilitating R&D environments for developing electric trucks with longer ranges.

(6)

List of abbreviations

EC European Commission EU European Union FHWA Federal Highway Administration ICE Internal combustion engine LSP Logistics service provider MLP Multi-level perspective OEM Original equipment manufacturer R&D Research and development ST Socio-technical TTW Tank-to-wheel WTT Well-to-tank WTW Well-to-wheel

List of figures and tables

Figure 1. The basic elements and resources of socio-technical-systems ... 13

Figure 2. Co-evolution of technology and user environment ... 13

Figure 3. Sub-regimes in the ST-system adapted from Geels (2004) ... 13

Figure 4. Multi-level perspective on socio-technical transitions ... 16

Figure 5. Transformation pathway ... 17

Figure 6. De-alignment and re-alignment pathway ... 18

Figure 7. Technological substitution pathway ... 19

Figure 8. Reconfiguration pathway ... 20

Figure 9. FHWA Vehicle classifications ... 26

Figure 10. Tractor-unit on the left and a tractor-trailer-combi on the right ... 27

Figure 11. Parallel hybrid-electric drivetrain confirguation ... 39

Figure 12. Representation of a rechargeable Lithium battery ... 40

Figure 13. Energy density of different battery chemistries ... 41

Figure 14. Representation of basic principles of a fuel cell ... 42

Table 1. Examples of institutions in the sub-regimes ... 14

Table 2. Conceptual model for analysing data and answering sub-questions ... 22

Table 3. Conceptual model for analysing results and answering main question ... 22

Table 4. Interview sample... 24

Table 5. Trip distances and freight weights for national transport ... 30

Table 6. Landscape developments impacting the system ... 36

Table 7. Battery-electric and fuel cell-electric compared ... 44

Table 8. Business case of electric trucks relative to a diesel truck by 2020 ... 48

(8)

1. Introduction

24% of the global anthropogenic carbon dioxide (CO2) emissions causing climate change is due to the

transport sector, of which road transport is responsible for 75% (IEA, 2017a). Additionally, these emissions cause high levels of air pollution which is specifically a problem in densely populated areas such as cities (Ostro, 2004). According to the World Health Organization, around three million deaths per year are linked to exposure to outdoor air pollution (WHO, 2016). Besides the effects on climate change and air pollution, combustion engines in transport are a major cause of oil depletion since they account for 55% of the global oil consumption (Kodjak, 2015). Reducing and eventually eliminating this diesel consumption and corresponding emissions requires a huge transition. This transition must lead to substitution of polluting fossil fuels by sustainable and clean energy carriers such as electricity (generated from renewables).

First and foremost, the transport sector can be divided into two categories; passenger transport and freight transport. When considering road transport only, passenger transport is dominated by private ‘smaller’ vehicles and among freight transport more larger and heavier vehicles are predominant (CBS, 2016a). Although solely 11% of all road vehicles are heavy-duty vehicles, they nevertheless account globally for 46% of the CO2 and 71% of the particulate emissions by road transport (Kodjak, 2015). But,

whereas the adoption of private and other ‘smaller’ passenger electric vehicles is increasing every day and the diffusion now occurs at different levels of the market, the widespread introduction of full electric vehicles for (heavy-duty) freight transport stays out (Moultak, Lutsey & Hall, 2017). However, Tesla recently announced the first all-electric semi-truck which attracted a lot of publicity (The Guardian, 2017). This announcement will probably act as an important trigger for the market to start researching, developing and implementing this or other sustainable and clean technologies more in the freight transport sector.

Besides the publicity about the Tesla Semi, also other stakeholders have acknowledged the relevance of this sector in the ongoing energy transition. Multiple consultancy studies and policy reports take up the subject, argue for the importance of it and try to come up with possible solutions (Schiller, Maier & Büchle, 2016; McKinsey Energy Insights, 2017; Roland Berger, 2012). This has already resulted in governmental bodies setting goals and announcing targets aiming at reducing the emissions of the freight transport sector (European Commission, 2011; SER, 2013).

1.1. Problem outline

Although different stakeholders are on the same page in their desire to reduce emissions, many challenges remain for clean vehicles to successfully diffuse among the entire freight transport sector (OECD/IEA, 2017). The freight transport sector lies in a robust and complex socio-technical system with many embedded stakeholders having many different characteristics, objectives, interests and activities (Klitkou, Bolwig, Hansen & Wessberg, 2015). This sector has a long history and according to therefore, it is likely that the socio-technical system is characterised by a great path dependency, sunk cost, many fixed routines, and several lock-ins (Klitkou et al., 2015). Similar to most changing socio-technical regimes, this transition requires, next to efficient novel technologies, also changes in all involved domains (e.g. institutions, manufacturers, logistics, road infrastructure, charging infrastructure, user

(9)

demand actors (Bos & Brown, 2012). Since these actors are mainly for-profit organisations, required collaborations between competing actors are not very common and financial investments that are uncertain in their outcomes are undesired as well (Faems, Janssens & van Looy, 2010). Because of this, it is presumably that, without proper support, the transition will take off too slow. Therefore, it is important that the required changes are supported and accelerated sufficiently, and governmental bodies could play an important role in this (Rotmans, Kemp & van Asselt, 2001). To develop this support, more insight must be gained on the transition and its corresponding required changes.

Literature gap

Many studies have been conducted on sustainable transitions concerning the electrification of the automobile industry (e.g. van Bree, Verbong & Kramer, 2010; Greene, Park & Liu, 2014; Tate, Harpster & Savagian, 2008;Sovacool & Hirsh, 2009). Although they have diverse aims, one thing most academic studies have in common is their focus on ‘smaller’ vehicles for passenger transport rather than freight vehicles (trucks). However, for the transition towards electric trucks, analysing studies on smaller passenger vehicles will not provide many relevant insights because freight transport is significantly different from passenger transport. Since freight transport concerns other stakeholders and other market needs, entirely other barriers and opportunities will be faced during this transition. To provide insight in these barriers and opportunities, research on the entire freight transport sector is required that can eventually serve as input for policy recommendations to support the transition in this sector (Geels, 2002).

Only one academic study was found on transitioning to sustainable freight transport (Mattila & Antikainen, 2011). In this study, Mattila & Antikainen (2011) used a back-casting technique to identify technological options for reducing greenhouse gas emissions. However, due to the high variability of identified back-casts and the different preferences among stakeholders, the results of this approach did not allow for any specific policy recommendations on guiding technological development. Among non-academic reports, more studies exist on this subject that do provide more insights that are relevant to policymakers (OECD/IEA, 2017; ICCT, 2007; McKinsey Energy Insights, 2017). These studies had a more practice-oriented and empirical approach. Since the aim of this research is to provide insights in the transition towards a sustainable freight transport sector that should be able to serve as a basis for policy recommendations, a more practice-oriented approach is taken here as well.

Because this research is investigating the transition towards electric trucks based on an academically sound method, it closes the gap between previous academic and non-academic literature. Besides, the previous practical studies were mainly internationally oriented whereas this one investigates a national-level case. Next to the fact that no previous studies were found with this level of analysis, it is also relevant because especially national policies are considered as important mechanisms to support sustainable transitions (Rotmans et al., 2001).

1.2. Research scope

For this case-study research, the country of the Netherlands is chosen as the unit-of analysis because it is a frontrunner in the adoption and diffusion of electric vehicles (Ministry of Economic Affairs, 2017; Netherlands Enterprise Agency, 2017). Besides, it is the base of the researcher who also speaks the language which eases the data collection and analysis.

(10)

The relevance to focus on heavy-duty vehicles for freight transport instead of passenger transport was already stated in the second paragraph of chapter 1. Within the freight transport sector, however, much variation still exists and therefore more scoping is needed before a feasible research question can be formulated. First, in this study the focus is on vehicles for freight transport by road whereby transport by rail and water is left out of consideration. This is because most freight rail transport is already electrified in the Netherlands, and transport by water is considered as an entirely different from road transport (Koninklijk Nederlands Vervoer, 2014; European Commission, 2011).

Then, even within the road freight transport sector many different heavy-duty vehicles exist and also different purposes. This research focuses on weight-based class eight (> 15 tons) regional transport (medium hauls) only. This means that tractor-trailer combinations are considered that barely enter cities and thereby city transport and international long-haul transport, as well as, vans and other single-unit trucks are ignored. This choice was made because long-haul transport is aimed by the European Commission (EC) to get shifted to water- and railways and electrification of city transport is already taking off since that is easier than electrification of regional transport (European Commission, 2011; McKinsey Energy Insights, 2017; Werner, 2018).

A final aspect that must be discussed is the choice for electrification only, instead of also taking into account biofuel which is also considered as a sustainable option for (road) transport (OECD/IEA, 2017). In this research, biofuel was ignored because of two major disadvantages compared to electricity. One being that the combustion of biofuel still emits polluting PM10 and NOx which are harmful to health

(Scharlemann, & Laurance, 2008). The other being that a widescale production of biofuel will inherently compete with food production, which is regarding the growing global population and the forecasted food scarcity not very convenient (Godfray et al., 2010).

Altogether, the scope of this research is the transition of the system of heavy-duty (class eight) freight transport by road for regional distances in the Netherlands towards electric trucks. When later in this report ‘freight transport system’ is mentioned, it refers to this scoped system.

Research objective and research questions

Now that the research is introduced, the problem is outlined, a gap in empirical literature is appointed and the subject is scoped, the following research objective can be formulated;

This research aims to generate insight in the transition towards electrification of the regional freight transport system in the NLs that can serve as a basis for supporting governmental policy.

This objective must be achieved by finding an answer to the following research question;

Which (of the five) transition pathways will the transition of the regional freight transport system in the NLs towards electrification presumably follow, using insights from the multi-level perspective on socio-technical transitions and from factors influencing technology adoption?

This will be done through first thoroughly studying the current freight transport system including its stakeholders. Then, the developments influencing that system and the novel technologies that are candidate to realising the electrification will be identified. Both desk research and qualitative expert interviews must allow for establishing this overview of the freight transport system as well as for

(11)

research and structure to the results, the following four sub-questions are formulated that are based on the problem outline in section 1.1. and on the theories discussed in chapter 2;

1. What is the composition of the current freight system in the Netherlands? 2. What trends are influencing the system and triggering a transition?

3. What are novel technologies that are candidate to the electrification of trucks capable of? 4. What are the main barriers and opportunities regarding the adoption of the technology?

1.3. Report outlook

To find an answer to this research question, first a clear understanding of the relevant concepts and relating theories that underlie this question must be established. These underlying theories are ‘the multi-level perspective on socio-technical system transitions’ by Geels (2002), the consequently emerged ‘transition pathways’ by Geels & Schot (2007) and ‘technology adoption and diffusion’ by Rogers (2003) and by Frambach & Schillewaert (2002). They are discussed in chapter two and the relevant subjects, concepts and variables are culminated into conceptual models presented in section 2.4. at the end the chapter.

The third chapter describes the methodology used in this case-study research and elaborates on the data collection, the data analysis and the quality assurance.

In the fourth chapter, the results are presented structured by the four sub-questions. Section 4.1. lists all results of the analysis on the composition of the current freight system including its stakeholders and institutions. Then, the trends that are influencing the system and triggering or supporting the transition are listed in section 4.2. Section 4.3. discusses the technologies that are candidate to the electrification of trucks (battery-electric and fuel cell-electric) and presents the results of an economic analysis of battery-electric trucks in section 4.3.2. Finally, in the identified barriers and opportunities to the adoption of battery-electric trucks are presented in section 4.4. Also, for each sub-question the findings are summarised and presented in sections 4.1.5. (sub-question 1), 4.2.2. (sub-question 2), 4.3.3. (sub-question 3), and 4.4.3. (sub-question 4).

Through re-interpretation and re-analysis of the results according to the theories described in chapter two, the results are discussed in chapter five. From this feedback to the theory, the possible transition pathways are examined. Also, limitations of the research are discussed in this chapter (section 5.2.). Eventually, in chapter six, all findings are summarised, an answer to the research question is formulated, and recommendations on future research and policy development are listed.

2. Theory

As stated above, a clear understanding on the relevant theoretical concepts is required to find an answer to the research questions. When analysing the research question, it can be noticed that the theoretical concepts were already processed in it. These concepts are the building blocks of the question and thereby also the guiding topics in this research.

Which (of the five) transition pathways will the transition of the freight transport system in the NLs towards electrification presumably follow, using insights from the multi-level perspective on socio-technical transitions and from factors influencing technology adoption?

(12)

In this chapter, the theories behind these concepts will be explained and discussed. The main building block and overarching theory is the multi-level perspective (MLP) on socio-technical system transitions. This MLP is the basis for the first three sub-questions in this research and will be discussed first. Besides, the MLP is also the basis for the different theoretical transition pathways on which insight is required to answer the main research question. These are outlined in section 2.2. Finally, the input for answering the fourth research question is covered by discussing multiple theories on innovation adoption. At the end of this chapter, the most relevant aspects of all theories are summarised and presented in a conceptual model.

2.1. Multi-level perspective on socio-technical system transitions

When considering the transition pathway of a system, an understanding of that system must be established first. This understanding must also serve as a framework for answering the first sub-question. The electrification of the freight transport system concerns technological innovation which can have effects on different stakeholders. Therefore, the understanding of this system is sought in theories explaining systems in which technologies play an important role, and the innovation of them touches multiple domains. This is the theory on socio-technical (ST) systems.

Socio-technical systems

The theory on socio-technical systems (ST-systems) is drawn upon similar theoretical roots as the theory on systems of innovation (Markard & Truffer, 2008; Weber & Rohracher, 2012). Especially the definition of the system of innovation can be extended to serve as a basis for the ST-systems theory. Edquist (2004) defines the overall system of innovation as “the determinants of innovation processes, i.e. all important economic, social, political, organizational, institutional, and other factors that influence the development, diffusion, and use of innovations” (pp. 182). He furthermore states organisations and institutions as the main components of systems of innovation with organisations being the players and institutions being the rules of the game. Organisations are defined as “formal structures that are consciously created and have an explicit purpose” and institutions as “sets of common norms, habits, established practices, routines, rules and laws that regulate the relations and interactions between individual groups and organisations” (Edquist, 2004, pp. 182).

The strengths of this theory on systems of innovation are, among others, the holistic and interdisciplinary perspective; the interdependence and non-linearity; and the emphasis on the role of institutions (Edquist, 2004). However, according to Geels (2004), the theory is mainly focussed on the production side where innovations emerge whereby, he argues, it lacks focus on the user side. Instead, theory on ST-systems provide a unit of analysis which encompass production, diffusion and use of technology. Geels (2004) defines ST-systems as “the linkages between elements necessary to fulfil societal functions such as transport, communication and nutrition” (pp. 900). A distinction is made between the production, distribution and use of technologies characterised as sub-functions and the elements necessary to fulfil these sub-functions characterised as resources. Figure 1 illustrates these basic elements of a ST-system. This approach allows for a wide perspective where, besides industries, also room is for other actors such as users.

(13)

Figure 1. The basic elements and resources of socio-technical-systems (Geels, 2004)

In a ST-system, technologies, which are developed at the production side, are aligned with their user environment, or user side. When incremental innovations to the technologies occur at the production side, the user environment co-evolves accordingly (this is depicted in figure 2). As time goes by, the alignment is getting stronger; the actors and the institutions (i.e. habit, routines, established practices, rules and laws) are joining the system and increasingly getting fixed to the present technology. This eventually leads to the efficiency and stability of the ST-system.

Next to the technology production side and the user environment, more social groups can be distinguished that co-constitute a ST-system; science actors develop new knowledge required for technological innovations, policy actors develop and maintain laws and regulations that shape the activities, and socio-cultural actors, such as media organisations, may influence opinions and attitudes of all actors in the system which in turn, might also affect the activities. Because these social groups have their own institutions, Geels (2004) refers to the groups as regimes. He defines the regimes as “semi-coherent sets of rules, which are linked together and difficult to change without altering others” (Geels, 2004, pp. 904) (depicted in figure 3). All interests of the different actors are covered by their institutions and the alignment between these institutions gives a regime stability, and ‘strength’ to coordinate activities.

Figure 2. Co-evolution of technology and user

(14)

Among the institutions that act in the regimes, a distinction can be made between regulative (or formal), normative and cognitive institutions. Regulative institutions include formal rules, laws, sanctions, incentive structures, etc. Normative institutions include, among others, values, norms, role expectations and codes of conduct. Cognitive institutions are, for example, problem agendas, beliefs, bodies of knowledge (paradigms), classifications and search heuristics. For each sub-regime Geels (2004) provided examples of institutions which he derived from many different relevant literatures (e.g. innovation studies, business studies, sociology of technology, cultural studies and evolutionary economics). The examples that are most relevant to the freight transport system are distracted from the examples listed by Geels (2004) and are presented in table 1.

Table 1. Examples of institutions in the sub-regimes (Geels, 2004)

Sub-regimes Institutions/rules

Technological and product regimes

(research, development, production)

Formal rules:

• Technical standards,

• Product specifications (e.g. emissions, weight), Functional requirements (articulated by customers or marketing departments),

• Expected capital return rate for investments,

• R&D-subsidies

Science regimes Formal rules:

• Formal research programmes (in research groups, governments),

• Professional boundaries,

• Rules for government subsidies

Policy regimes Formal rules:

• Administrative regulations and procedures which structure the legislative process,

• Formal regulations of technology (e.g. safety standards, emission norms),

• Subsidy programmes,

• Procurement programmes Normative rules:

➢ Policy goals,

➢ Interaction patterns between industry and government (e.g. corporatism),

➢ Institutional commitment to existing systems,

➢ Role perceptions of government Cognitive rules:

• Ideas about the effectiveness of instruments,

• Guiding principles (e.g. liberalisation),

• Problem-agendas

Socio-cultural regimes (societal groups, media)

Formal rules:

• Rules which structure the spread of information production of cultural symbols (e.g. media laws)

Users, markets and distribution networks

Formal rules:

(15)

• Market subsidies,

• Tax credits to users

• Competition rules • Safety requirements Cognitive rules: • User practices • User preferences • User competencies

• Interpretation of functionalities of technologies

• Beliefs about the efficiency of (free)markets, perceptions of what ‘the market’ wants (i.e. selection criteria, user preferences)

Multi-level perspective on socio-technical transitions

It has thus become clear that technologies do not arise, distribute and live on their own, but that they are embedded in a ST-system. This ST-system is characterised by stability and path dependency due to increasing alignment between its technologies and its actors and institutions. However, when a core technology of a ST-system is being debated and incremental innovations do not suffice to this debate, this stability is being jeopardised. Because all regimes, including its actors and institutions, are interlinked, changes in one regime will inevitably affect the other regimes. When these changes in one regime undermine institutions in other regimes, whereby the interests of the different actors are no longer covered, this may lead to distrust and eventually to de-alignment of the regimes in the system. This could be the start of a transition process of the entire system.

To get an understanding of how such a transition process actually arises and develops, the multi-level perspective (MLP) as described by Geels (2002) could provide comprehensive insights. Therefore, second and third sub-questions in this research are based on the MLP. This framework provides an overall view of the multi-dimensional complexity of changes in socio-technical systems and is especially helpful for understanding transitions to sustainability. Three analytical levels are distinguished in the MLP: niches (which are the locus for radical innovations), socio-technical regimes (the interlinked sub-regimes that are locked in and stabilized on several dimensions), and an exogenous socio-technical landscape. The MLP proposes that transitions, which are also defined as regime shifts, are caused by interactions between and within these levels. Within the regimes innovation occurs incrementally along predictable trajectories and radical innovations emerge in niches. In these niches, dedicated actors can develop novel technologies and nurture alignment on multiple dimensions in order to construct ‘configurations that work’. If external landscape developments create pressure on the ST-regimes that leads to cracks (e.g. debates, distrust, market failures), windows of opportunity arise for novel niche-technologies to break through in the regime. Examples of these landscape developments are climate change, urbanisation, and digitalization. Besides, Geels (2004) lists also four reasons that can cause mis-alignment and tensions in the regime that come from inside that regime. Among these, are internal technical problems, negative externalities, changing user preferences and strategic games between firms. The MLP is depicted in figure 4.

(16)

Figure 4. Multi-level perspective on socio-technical transitions (Geels & Schot, 2007)

2.2. Transition pathways

As with many societal or policy approaches, the MLP as described by Geels (2002) has not been without criticism. Geels & Schot (2007) outlined some general points on which they reacted. This point states that the approach emphasizes the role of technological niches as the principal locus for regime change too much. Berkhout et al. (2004) claim that MLP-approaches are emphasizing too much the processes of regime changes that begin within niches and then work up and breakthrough in the regime. As a result, the processes that directly address the various dimensions of the sociotechnical regime or those which operate ‘downwards’ from general features of the sociotechnical landscape are underexposed. To counter this third criticism, Geels and Schot (2007) further refined the MLP and developed four transition pathways based on different multi-level interactions.

Two criteria determining the pathway

The distinction of these pathways is made based on two criteria; the timing of interactions and the

nature of interaction. The timing of landscape pressure on the regime is particularly relevant in

combination with the state of niche-developments. Whether niche-technologies are sufficiently developed to meet the needs in the regime or whether they are still in their early development stages when landscape pressure occurs, will affect the transition path. When landscape pressure occurs, a window of opportunity is created in the regime that allows novel technologies to climb up and enter the regime. However, a technology that is not yet fully developed cannot take advantage of this window, which will then close again.

(17)

have disruptive or reinforcing relationships with the current the regime. Reinforcing landscape developments do not result in a transition but rather in a reinforcement of the robustness of the current regime, whereas disruptive landscape developments exert pressure and create incentives for change in the regime. Similarly, novel niche-technologies that are competing with the regime have the aim to replace its technologies and can thereby encourage change and enable a transition. On the other hand, technologies that have symbiotic relationships with the regime, can be adopted as competence enhancing additions in the regime.

The four transition pathways are developed based on combinations of the described two criteria and can be labelled as transformation; reconfiguration; technological substitution; and de-alignment and re-alignment.

1. Transformation

The transition pathway of transformation occurs when moderate landscape pressure causes a need for changes in the regime, but the niche-technologies have not yet been sufficiently developed to meet the needs and enter the regime. In this case, regime players will react to the landscape pressure by reorienting their practices and modifying the direction of development paths and innovation activities. They will adopt symbiotic elements of niche technologies.

Figure 5. Transformation pathway (Geels & Schot, 2007)

The main important actors in this pathway are the regime players and outside groups. Landscape changes only exert pressure if they are perceived by the regime players and actors outside the regime can articulate the changes and thereby thus increase the pressure. Regime players that must adapt to the landscape pressure are not always doing so without a fight, but this usually involves conflicts, power struggles or contestations. Hence, social-institutional dynamics with social groups acting to alter regime rules directly are important in this pathway. These outsiders can be societal pressure groups, social movements, outside professional scientists or engineers, outsider firms, entrepreneurs or activists. When they succeed to demonstrate viable alternatives, regime players may change their perceptions and will reorient and rethink their existing technologies or activities. In this path (depicted in figure 5), new regimes arise out of old regimes through cumulative alteration.

(18)

2. De-alignment and re-alignment

The transition pathway of de-alignment and re-alignment occurs when the landscape developments are large, sudden and divergent which increase problems within the regime and cause regime players to lose their faith and legitimacy. This leads to de-alignment of the regime and if niche-technologies are not sufficiently developed to step in, then there is no evident substitute. This path allows multiple niche-technologies to co-exist, further develop and compete for resources and attention of the regime players. Eventually, one of these technologies will gain support and become dominant. That technology will form the core for re-alignment of the new regime.

The main important actors in this transition path (depicted in figure 6) are the new niche actors. Due to a lack of institutional stability, multiple innovation trajectories and directions are being explored. Besides, the co-existence of multiple niche actors leads to increased uncertainty, because actors of winning products make contrasting claims. Thus, before the new regime is re-aligned, there is a prolonged period of co-existence, experimentation, competition and uncertainty among these niche actors.

Figure 6. De-alignment and re-alignment pathway (Geels & Schot, 2007)

3. Technological substitution

The transition pathway of technological substitution occurs when the landscape pressure is very high at a moment when niche-technologies have sufficiently developed to meet regime needs. The novel technologies will (directly) break through and substitute the regime technologies.

In this scenario, it is assumed that radical technologies are being developed in niches while the regime is stable and shutting out the novel technologies. Without influential landscape ‘shock’, ‘avalanche change’ or ‘disruptive change’, no pressure on the regime poses and the regime keeps in its reproduction process. Thus, for the transition to happen, an important role is played by the landscape

(19)

The pressure causes tensions in the regime which create windows of opportunity for the niche-technologies to enter the regime. Because these niche-technologies have sufficiently developed in their niches already, they can easily enter the regime markets. However, the diffusion often occurs through niche accumulation whereby the technology enters greater (niche) markets successively. Eventually, the novel technology will enter the mainstream markets and then regime actors will go in defence by investing in improving their products. A competition fight emerges between the established incumbents and the new companies and if the newcomers win, this leads to a domino effect for wider changes and eventually an entire transition of the regime (depicted in figure 7).

Figure 7. Technological substitution pathway (Geels & Schot, 2007)

4. Reconfiguration

Reconfiguration is a transition pathway which occurs when local problems in the regime are solved by adopting symbiotic innovations that subsequently trigger more changes in the structure of the system. The radical technologies are developed in niches and as they have symbiotic relationships with the practices in the regime, they can, driven by economic and functional reasons, easily be adopted as addition or as component replacement. When the regime players start to learn more about the adopted new technologies and are going to explore new technological configurations, it may lead to technical changes and changes in user practices, perceptions, and search heuristics. Eventually, under pressure of landscape developments, the accumulation of innovations can thus add up to major technological reconfigurations and changes in interpretation and user practices, indicating a transition of the regime.

The main important actors in this pathway (depicted in figure 8) are the regime players, and specifically the technology suppliers. The adopted novel component-technologies are developed by new suppliers which causes competition between the old and the new suppliers. In the reconfiguration pathway, the new regime arises from the old regime where regime actors survive, and component suppliers undergo the competition and tensions.

(20)

Figure 8. Reconfiguration pathway (Geels & Schot, 2007)

Combination of transition pathways

Besides following one single pathway out of the four described ones, a transition is also likely to follow a sequence of different pathways. This sequence begins with transformation which leads to reconfiguration and possibly followed by either de-alignment and re-alignment or by substitution. This combination of transition pathways occurs when landscape pressure is caused by a ‘disruptive change’ in landscape developments.

In initial stages, this development seems a moderate change due to its slow rate. The regime actors address the emerged problems by changing the direction of their activities and innovation trajectories using their own resources, indicating a transformation pathway. When, however, the pressure continues and increases, the landscape change becomes more disruptive. Regime players may become willing to adopt symbiotic niche-technologies and implement component innovations. If these adoptions trigger structural system changes, this indicates a reconfiguration path and if problems are solved, regime players will survive. If problems are not solved and pressure keeps continuing, incumbents will lose faith in their new configurations and niche technologies will develop further. Then a path of either substitution or de-alignment and re-alignment occurs, depending on whether niche-technologies are sufficiently developed to meet regime practices.

2.3. Factors influencing technology adoption

When analysing a socio-technical transition, several barriers and opportunities to this transition can be revealed. Opportunities are factors that could stimulate the transition and barriers are factors that could hinder the transition. For example, a high degree of landscape pressure on the regime can stimulate a transition and can therefore be seen as an opportunity, and a low degree of pressure as a barrier. However, as described earlier, this pressure needs a right timing which means that it should occur at the moment when the novel technologies are sufficiently developed to meet the needs of the regime. Whether this is the case, could be analysed based on the likelihood that an actor would adopt

(21)

From literature on the adoption and diffusion of technological innovations, different factors can be listed that can influence the adoption of a novel radical technology. Rogers (2003, pp. 222) distinguishes five perceived attributes of an innovation that determine an adoption-decision. These are; relative advantage, compatibility, complexity, trialability and observability. The relative advantage of an innovation “is the degree to which an innovation is perceived as being better than the idea it supersedes” (Rogers, 2003, pp. 229). ‘Being better’ can be interpreted in multiple ways (e.g. being less costly, being more useful, being more sustainable) whereby measuring this attribute can be difficult (Tornatzky & Klein, 1982). However, when considering for-profit organisations only, this relative advantage mainly refers to the perceived cost and the ability to increase profits when implementing the technology; i.e. economic profitability of the innovation (Rogers, 2003, pp. 291-230). Since increasing profits is generally the main objective of such an organisation, this attribute is the most important one in determining the adoption-decision of an innovation. Moreover, Frambach & Schillewaert (2002) studied the variables influencing innovation adoption-decisions specifically for organisations and added the uncertainty of a novel technology to the five attributes of Rogers. It is theorised that a high degree of perceived uncertainty decreases the rate of adoption and thus forms a barrier to the transition.

Next to the attributes of the innovation itself, certain characteristics of the potential adopter (i.e. regime actor) can also influence the adoption-decision. An important one is the extent of innovativeness of the adopter (Frambach & Schillewaert, 2002). This innovativeness can be reflected in degree to which an organisation is receptive to change. When organisations are highly receptive to change, they are likely to have a positive attitude towards novel technologies. This generally leads to a higher willingness and a higher perceived ability to adopt the innovation.

When the attributes of the innovation are not perceived positive and/or the actors are not willing or do not perceive themselves as being able to adopt the innovation, a rejection to of the adoption-decision will generally be the result. However, when concerning innovations that serve a common or societal good, such as sustainable technologies, the need for adoption is greater. In these cases, governmental policies may be deployed to stimulate or regulate adoption-decisions (Kemp, 1994).

2.4. Conceptual models

In this section, two conceptual models are presented in tables 2 and 3 (on the next page) that will be used to answer the research (sub-)questions. Both models list the relevant concepts or premises distracted from the described theories. Table 2 was mainly used to analyse the collected data, to translate it into relevant results whereby answering the sub-questions of the research. Besides, tables 2 also provided input to the questions asked in the expert-interviews and structuration to the presentation of the results.

Table 3 is based on the theory of the different transition pathways and summarises the main relations and premises that are relevant to determining those pathways. When the answers to the sub-questions are found, they can be interpreted according to this table whereby an answer to the main research question will be formed. The model in table 2 will thus mainly be used to analyse the raw data to form and structure the results and the model in table 3 is mainly used to discuss those results and form a conclusion to the research.

(22)

Table 2. Conceptual model for analysing data and answering sub-questions

Subject Theoretical concepts Variables

System’s composition Socio-technical system Sub-regimes Actors

Institutions (incl. market needs) Developments affecting the

system

Factors putting pressure on the regime

Changes on landscape level Negative externalities Internal technical problems Changing user preferences Novel technologies candidate to

adoption by the system

Niche developments Technological performances Economic performances Experimentation/pilot projects Adoption barriers and

opportunities

Factors influencing innovation adoption

(Perceived) economic profitability of novel technology (Perceived) uncertainty of novel technology

(Perceived) willingness/ability of actors to adopt technology

(Perceived) role of government to influence technology adoption

Table 3. Conceptual model for analysing results and answering main question

Circumstance Pressure effect Action Pathway

Technologies are sufficiently developed

Tensions from high landscape pressure

Novel technologies break through directly and substitute regime technologies.

Technological substitution First local problems only,

later landscape pressure

First symbiotic innovations are adopted that eventually trigger more structural changes. Besides technical reconfigurations, also changes in user practices can be a result.

Reconfiguration

Technologies are underdeveloped

Need for changes due to moderate landscape pressure

Regime players start reorienting and modifying innovation trajectories.

Transformation

Loss of faith and legitimacy due to high landscape pressure

Regime de-aligns; niche technologies compete for attention and resources; one technology wins; regime re-aligns.

De-alignment & re-alignment

(23)

3. Methodology

To find an answer to the research question, an exploratory case study was performed in which the analyses are mainly based on qualitative data (except for the analysis of the economic performance which was based on quantitative data). Because the aim of this study is rather practice-oriented and case study research allows for generating results directly relevant to the people concerned, this method was chosen. Besides, an exploratory case study allows for an in-depth study of a complex situation which matches the research aim best (Baxtar & Jack, 2008).

To collect the required data, an extensive desk research, a focus group and 10 expert interviews were conducted. The data obtained from the desk research and the focus group were mainly used for answering the first three sub-questions and the data obtained from the interviews mainly for the fourth sub-question.

3.1. Data collection

Several types of data sources were used in this research. The majority of the data was obtained through desk research including policy reports, consultancy reports, technical academic papers, news articles, expert blog posts and digital expert interviews.

Sample and interview protocol

In order to find an answer to the research (sub-)question(s), it is important to collect the right information. Therefore, both the digital documents in the desk research and the interviewed respondents were strategically sampled.

The interview respondents were sampled within three categories which represented the most important regimes to the transition; markets, policy and science/technology (see table 4 and appendix A). For privacy reasons, the names of the respondents cannot be revealed. This anonymity must also have a positive influence on receiving unbiased answers.

The interview questions were drawn upon the variables listed in the conceptual model. Although the interviews were mainly focussed on answering sub-question four, also other variables were discussed; for example, institutions such as market needs. In table 4, presented on the next page, the third column lists the topics that were addressed during the interviews per respondent category. Because the experts have different backgrounds and thereby also different knowledge bases and levels, the interviews were semi-structured. This allowed for adapting the questioning to the answers the respondents gave and also to insights gained from desk research.

3.2. Data analyses

The data collected from both the desk research and the interviews were analysed based on whether they implicated things on the variables outlined in the conceptual model (section 2.4.). The digital documents were examined, and relevant information was extracted and collected in documents that were structured according to the variables to be analysed for answering the sub-questions (see table 2). Eventually, this information was merged and resulted in a starting overview of the current freight transport system, the landscape developments and the novel technologies, presented in sections 4.1., 4.2. and 4.3.

(24)

The data obtained from the interviews required some preparation before they could be analysed on relevant information. After the interviews were conducted, they were transcribed first. The next step was the coding of the data; fragments that revealed information on the relevant concepts were identified, highlighted and labelled. After this coding was performed, the labelled fragments were connected to the information topics from table 2 and were reviewed again within the contexts of these topics and the contexts of the entire interview. This was done to confirm the labels were interpreted right and actually represent the expressions of the respondent. The structuration of the information behind the labels was done in a similar way the information from the desk research was structured. Table 4. Interview sample

Category Respondents Information topics

Markets Supply side:

Three experts of two OEMs Two transition/niche actors Demand side:

Two logistics providers (LSPs)

One employers’ organisation for transport and logistics companies

Market needs

Perceived ability and willingness to adopt and implement novel technologies Perceived ability and willingness to change practices in favour of novel technology Perceived barriers and opportunities to implementation

Policy Two nationally operating governmental organisations that provide advice to the ministry of infrastructure and water and the to the ministry of economic affairs

Perceived role of government in transition Perceived ability and willingness to stimulate implementation of novel technologies

Perceived ability, willingness and likeliness to steer developments in certain directions Preferences for certain technologies Perceived barriers and opportunities to implementation

Science/technology Focus group with technical researchers:

One senior researcher One junior researcher Three MSc researchers

Performances of novel technologies Economic performances

(25)

3.3. Quality assurance

For quantitative research, researchers generally agree on the quality criteria of internal and external validity, reliability and objectivity of the research. However, whether these specific criteria also apply to qualitative research is sometimes subject to discussion (Seale, 1999). Case study research is sometimes even more so being debated on its scientific quality because some argue that is lacks rigour (Kyburz-Graber, 2004). However, Kyburz-Graber argues that when certain quality criteria (following from the objectivity, reliability and validity criteria) are taken into account, this quality should be assured. These criteria for a scientifically acknowledged case study are derived from Yin (1994) and can be listed as;

a) a theoretical basis including research questions is described; b) triangulation is ensured by using multiple sources;

c) a chain of evidence is designed with traceable reasons and arguments; d) the case-study research is fully documented; and

e) the case-study report is compiled through an iterative review and rewriting process. The first criterion is already guaranteed in section 1.2. and chapter 3. Triangulation was established by the mix of desk research and interviews. As stated earlier, the desk research was mainly focussed on sub-questions one to three and the interviews on sub-question four. However, to ensure reliability, the information obtained from the desk research was also validated during the interviews and vice versa; the information obtained in the interviews was checked by desk research. Thereby, various data sources were consulted for the same topics and, especially because interviews allowed for obtaining information on other dimensions (e.g. perceptions and opinions), this assured triangulation of the data. According to the third criterion, the results are described with a chain of evidence based on clear arguments and other information from the data with references to the sources. The case study research is fully documented in a way that all non-digital data sources (i.e. focus group and interviews) were transformed into transcripts. Finally, the entire report was constructed in an iterative way; it was restructured multiple times and besides, it was reviewed by fellow academics and entirely rewritten once.

4. Results

In this chapter, the results of the data analyses are presented. The results are structured based on the four sub-research questions; in section 4.1. the composition of the current freight transport system is described, section 4.2. lists the landscape developments that influence that system, thereafter the novel technologies that are a candidate to the transition are specified in section 4.3., and finally section 4.4. elaborates on the barriers and opportunities to the adoption of the novel technologies.

4.1. The current freight transport system

In this section, the results of the analysis of the composition of the current freight transport system are presented. At the end of the section, the main findings are summarised to generate an answer to the first sub-question of the research. For this analysis, the five regimes that were provided by the theory in section 2.1 served as a basis. For each regime in the system, the relevant stakeholder groups

(26)

and relating institutions are identified and presented in this chapter. The main findings of this section are summarised on section 4.1.5.

The production and application domains of the system are mainly constituted by the technological and product regime (supply) and the regime of users and markets (demand). Besides, the science regime can mainly be assigned to the production domain because it especially generates input for the development of (technological) innovations to the product. The policy and socio-cultural regimes are active at the interplay between the two domains.

4.1.1. Technological and product regime

The main stakeholder group in this regime consists of the original equipment manufacturers (OEMs). Multiple OEMs are active in the system by supplying their products to the Dutch market (i.e. trucks). Among the OEMs represented by the largest truck fleet in the Netherlands are; DAF, Volvo, Mercedes-Benz, Scania, MAN and Iveco (Luman, 2015).

When analysing the institutions such as technical standards, it is relevant to first get insight in different product types.

Within the sector of freight transport by road, several typical categorisations can be made. These categorisations exist because of the high extent of diversity between vehicle types. The most commonly used categorisations are the weight-based classification and the Federal Highway Administration (FHWA) vehicle classification (AFCD, 2018). As stated earlier, the focus in this research is on weight-class eight trucks. This class is the highest in the classification and contains all heavy-duty trucks that weigh 15 tons or more, based on the gross vehicle weight (GVW) (i.e. the weight of the vehicle, including driver and cargo, excluding trailer). However, because many trucks weigh much more than 15 tons, this class still contains multiple vehicle types.

(27)

The FHWA vehicle classification contains more classes that resemble the weight-classes but are slightly different. This classification distinguishes between vehicles for passenger transport (classes 1-4) and vehicles for freight transport (classes 5-13) (Randall, 2012). Figure 9 shows these classes. When considering regional transport, most trucks are tractor-trailer combinations which are represented by classes 8-13.

Technical standards

Tractor-trailer combinations are thus the product standard in the system (depicted in figure 10). As stated in the interviews, these combinations can generally be considered as ‘one-size-fits-all trucks’ (sources: appx. A3; A6; A7). This means that they can be deployed on many different trips, ranging from short trips carrying freight of large volumes and low weights, to longer trips carrying heavier freights. This is due to the fact that most tractors can pull up to 20-30 tons and can drive at least 1500 km without refuelling. Besides, refuelling only takes a couple of minutes whereby this generally does not have a large impact on the transport schedule.

Figure 10. Tractor-unit on the left and a tractor-trailer-combi on the right (source: Simon Loos Logistiek BV) When analysing the CO2 emissions of a vehicle, a distinction is often made between a tank-to-wheel

(TTW) analysis and well-to-wheel (WTW) analysis (Curran, Wagner, Graves, Keller & Green, 2014).The TTW analysis only takes into account the use of the vehicle whereby it focuses on the pipeline emissions. The WTW analysis contains two parts; one being the TTW and the other being the analysis of the underlying well-to-tank (WTT) emissions. This WTT part captures the emissions released at the fuel production, and at the transmission and distribution of the fuel, from the point of fuel feedstock extraction to where the fuel is transferred into the vehicle. When aiming at reducing global CO2

-emissions, WTW analyses are the better option because they are more complete than the TTW ones. Considering the mainly fossil fuel-based electricity production in the Netherlands, it must be noted that the electrification of trucks will currently not lead to significant reductions in WTW CO2-emissions

(CBS, 2016b). However, because this electricity production is currently not deemed as a task of the transport sector, but rather of the energy sector, TTW instead of WTW emissions are generally considered for the transport sector (Maverick, 2018). Besides, electricity generated from renewables is expected to increase significantly in the future whereby electrification will ensure serious WTW CO2

reductions (PwC, 2018).

The drivetrain is the main technical standard of the combination and is always situated in the tractor part (Verbeek, van Zyl, van Grinsven & van Essen, 2014). The main component of this drivetrain, in terms of costs, complexity and relevance, is the internal combustion engine (ICE) which has a history of over 100 years (Chan, 2007). A tractor-trailer combination is generally fuelled by diesel and emits

(28)

900 grams of CO2 per kilometre (Verbeek et al., 2014). The transport system, and even more so the

truck OEMs, have a long history with this ICE technology of which they continuously improve the efficiency through many incremental innovations (McKinsey&Company, 2016). Innovations on the technology are thereby also focussed on reducing its emissions but mainly on NOx and PM10. These

innovations are often an important factor in determining the profit streams at OEMs (McKinsey&Company, 2016).

Other fuels

Diesel has been the predominant fuel, but alternative options also exist and are utilized on a small scale in the Dutch system (Verbeek et al., 2014). Biodiesel is an alternative that is compatible with the ICE designed for conventional diesel and has significant lower WTW CO2 emissions than conventional

diesel. Depending on the production method, the CO2 reduction can be up top 90% (produced from

used cooking oil) and the NOx and PM10 emissions can be reduced by 20-60%. (Verbeek et al., 2014).

However, when the biodiesel is produced from palm oil, for example, the CO2 reduction is often much

lower, and the emissions can sometimes even be higher compared to normal diesel (Verbeek et al., 2014).

Natural gas and biogas are alternative fuels that are not compatible with the ICE design for conventional diesel whereby they require another engine technology (Verbeek et al., 2014). When considering natural gas, the NOx and PM10 emissions are approximately similar to diesel and the CO2

emissions are varying from equal to, to 10% lower than for diesel. When bio-CNG (compressed natural gas) or bio-LNG (liquefied natural gas) is used, much larger reductions can be achieved; up to 80% reduction in WTW CO2 emissions and up to 30% NOx and PM10.

Fuel suppliers

Since the drivetrain is mainly built for the combustion of fuels, fuel suppliers should also be included in this regime. Without those suppliers, the drivetrain and therefore the truck could not be used. In the Dutch system, Shell, BP, Esso and Q8 are significant fuel suppliers (AutoleaseWereld, 2018). Their stake is to sell as much fuel as possible. However, some actors in this regime are currently joining the transition towards electrification of vehicles. For example, Shell acquired New Motion, which is a business specialised in charging systems for electric vehicles; BP invests in FreeWire, which is a mobile electric vehicle charging company and Q8 has signed a cooperation with EDF Luminus to in install fast charging stations (EDF Luminus, 2016; Electrek, 2018; fd.nl, 2017). However, although they are joining the transition, this does not inherently mean that they are willing to reduce their fuel production and sales which still is their core business. Moreover, because these companies are very large, have great capitals, and are currently still responsible for our energy supply, they are powerful. Some argue that these actors can use this power to lobby for a policy they favour (Sluiter, 2017).

4.1.2. Users and markets

The regime of users and markets contains the actors that buy the trucks and are thus the customers of the OEMs. Two main stakeholder groups can be distinguished, namely shippers and logistics service providers (LSPs). Shippers are the actors who want their goods transported from location A to location

In transition to electric trucks