Instruments for transition : the impact of future visions of urban mobility in greater Toronto and the Randstad

(1)

U

NIVERSITY

OF

A

MSTERDAM

Instruments for Transition: The

Impact of Future Visions of Urban

Mobility in Greater Toronto and the

Randstad

Florian Langstraat

Thesis ResMSc Urban Studies University of Amsterdam

Florian Langstraat 10454187

florianlangstraat@gmail.com 7 July 2014

Supervisor: Prof. Luca Bertolini

(2)

(3)

Instruments for Transition: The Impact of Future

Visions of Urban Mobility in Greater Toronto and the

Randstad

Florian Langstraat

Thesis ResMSc Urban Studies University of Amsterdam

(4)

Vision without action is merely a dream. Action without vision just passes the time. Vision with action can change the world.

(5)

5

Executive summary

Given the need for a transition to sustainable mobility, a number of transportation scholars have advocated the importance of long-term, regional and normative future visions in urban mobility planning. Yet in the literature, considerable uncertainty about the usefulness of these visions remains. While some authors have heralded these visions as an innovative and promising planning instrument, others are sceptical, and suggest that many end up as little more than inconsequential wish lists.

In light of this debate, this thesis provides a systematic evaluation of the impact of four long-term visions of urban mobility systems: Places to Grow and The Big Move in the Greater Toronto region; and Metropoolregio Amsterdam Duurzaam Bereikbaar and

Masterplan Rotterdam Vooruit in the Randstad. The underlying aim is to identify

factors that either foster or hinder the impact of future visions. The thesis first introduces a causal model for assessing the impact of future visions, based on previous research in planning, but also from other domains. This model is then applied to the four cases to determine what their impact has been, and what factors have determined this impact. Each case is studied through two methods: semi-structured interviews with planners and other stakeholders (43 in total), and analysis of secondary documents.

The results indicate that overall, Places to Grow has had a positive impact on planning in the Greater Toronto region, despite some strategic weaknesses. The impact of The Big Move has been more limited despite the high hopes that were placed on it initially, mainly due to follow-up problems after its completion in 2008. The impact of

Metropoolregio Amsterdam Duurzaam Bereikbaar is the lowest of the four cases. Due

to a lack of any real interest from elected officials at the time, the vision never got a great deal of attention. Masterplan Rotterdam Vooruit is an intermediate case.

Based on the analysis of the four cases, the thesis identifies a number of explanatory factors that either enable or constrain the impact of future visions. In section 7.2, these factors are translated into eight lessons for planning practice: (1) involve stakeholders; (2) make sure there are vision champions; (3) focus on follow-up and implementation; (4) make strategic choices; (5) formulate a clear problem definition; (6) take sufficient time; (7) provide clarity about funding; and (8) align the vision with other visions in the region. The thesis ends with three suggestions for further research.

(6)

6

Self-plagiarism disclaimer

This thesis constitutes the final report for the ResMSc Urban Studies Research Project. Parts of this thesis have been used in the following preceding documents that are part of the same research project:

 Visions of the future in mobility transitions: A conceptual exploration of their

functions. Conceptual paper, Research Apprenticeship, University of

Amsterdam, 27 March 2013.

 Instruments for Transition? The Impact of Future Visions of Urban Mobility in

Canada and the Netherlands. Thesis Proposal, Thesis RMUS, University of

Amsterdam, 27 May 2013.

 Methodological considerations Canadian cases. Thesis RMUS, University of Amsterdam, 15 December 2013.

 Progress report 1: reflection on fieldwork abroad and progress so far. Thesis RMUS, University of Amsterdam, 28 February 2014.

 Progress report 2: Canadian results. Thesis RMUS, University of Amsterdam, 9 April 2014.

 Progress report 3: Dutch results. Thesis RMUS, University of Amsterdam, 11 June 2014.

 Langstraat, F. & L. Bertolini (2014), Assessing the impact of future visions of

mobility: a conceptual exploration. Paper to be presented at the annual

conference of the Association of European Schools of Planning (AESOP), Utrecht, 9-12 July 2014.

Acknowledgements

I am deeply grateful to my supervisor, Professor Luca Bertolini, for his continuous encouragement and highly detailed suggestions and critiques throughout this 18-month research project. This thesis would not have been possible without his extremely valuable advice. The Canadian part of the research project was funded by a scholarship from the Association for Canada Studies in the Netherlands, to which I am greatly indebted. I thank Matti Siemiatycki and Zack Taylor for their assistance with the fieldwork in Canada, and Kim von Schönfeld for her help with the intercoder reliability check. Last but not least, I would like to thank all respondents for kindly answering my many questions about Places to Grow, The Big Move, Metropoolregio Amsterdam Duurzaam Bereikbaar and Masterplan Rotterdam Vooruit.

(7)

7

1. Introduction

1.1. Problem statement: future visions of urban mobility and their impact

It has become widely acknowledged that current patterns of mobility are unsustainable in the long run (Banister, 2005; Schwanen et al, 2011; Givoni & Banister, 2013). Consequently, planners and scholars in the field of urban mobility have increasingly turned their attention to the question of how to achieve a transition towards sustainable urban mobility (Kemp & Rotmans, 2004; Köhler et al, 2009; Grin et al, 2010; Geels et al, 2011; Geerlings et al, 2012; Naess & Vogel, 2012; Switzer et al, 2013). As this focus on sustainability requires a long-term perspective almost by definition, it is hardly surprising that explicit „engagements with the future‟ (Hopkins & Zapata, 2007), have become increasingly common of late, both within the field of urban mobility and outside it (Tomalty, 2009; Van der Helm, 2009).

For transportation planners, there is of course little new about engaging with the future as such – techniques such as forecasting and scenario development have a long and rich history, and are well-established in many planning systems throughout the world. Given the need for a transition to sustainable urban mobility, however, it has been argued that urban planners need to think not just about possible (what can happen?) or plausible futures (what will happen?), but also about desirable futures (what should happen?). Such a normative perspective, it is argued, would allow transportation planners to focus on imagining and facilitating change, rather than merely on accommodating existing mobility patterns. Banister et al (2008: 25-26), for example, suggest that „transport planners need to think more imaginatively about the future,

(9)

9

rather than being content with current trends continuing much as they are at present‟. The sustainability challenge, combined with continuing globalisation and urbanisation, they argue, „[provides] researchers and others (...) with a unique opportunity to challenge existing conventional wisdom, usually based on trend-following analysis, with more radical trend-breaking futures‟ (ibid.: 27). Planners, in other words, need to „think the unthinkable‟ by constructing long-term and normative future visions of mobility systems (Banister & Hickman, 2012).

In planning practice, such future visions have indeed become increasingly prevalent in many countries (Banister et al, 2008; Van der Helm, 2009; Banister & Hickman, 2012). There is however still considerable uncertainty regarding the usefulness of such visions. Many authors are sceptical, and suggest that these visions are often well-intended but ultimately inconsequential wish lists (Myers & Kitsuse, 2000; Berkhout et al, 2004). Geels and Schot (2010: 84), for example, argue that „many of these [visioning] exercises have become rituals, where actors express good intentions as a form of public impression management‟. A related problem is that the term „visioning‟ is not always clearly defined, so that the term „can be applied loosely and conveniently to legitimise plans as creative, innovative and consensual‟ (Gaffikin & Sterrett, 2006: 163; see also Shipley, 2002). Still, it is also clear that challenges such as climate change, peak oil and the ever-increasing demand for transport are such that traditional trend-following approaches will simply no longer suffice (Banister et al, 2008; Givoni & Banister, 2013). The question therefore is how to make sure that future visions „move beyond rhetoric‟ (Gaffikin & Sterrett, 2006: 176), and become effective planning instruments.

(10)

10

1.2. Research goals: towards effective visioning

Given this debate, it seems an important task for planning research to evaluate the impact of existing future visions of urban mobility, so that lessons can be drawn that may help create future visions more effectively. Yet in the field of urban and transport planning, there seems to be a dearth of such systematic ex-post assessments. Somewhat surprisingly perhaps, very little has been done by way of analytical empirical research to explore how different stakeholders engage with visions of the future once they have been created, and why their actions are influenced by some of these visions, but not others.

It is this research gap that this thesis seeks to address. This thesis aims to systematically evaluate the impact of existing long-term visions of urban mobility systems, by way of a cross-national research design – a comparison of four visions, two in Canada and two in the Netherlands (see Chapter 3 for details). The goal of this evaluation is to identify factors that either foster or hinder the impact of these future visions.

This evaluative research project proceeds in a number of theoretical and empirical steps. The theoretical part of this thesis introduces a causal model for assessing the impact of future visions of urban mobility systems. It does so by drawing on existing research not only from the field of planning, but also from other literatures in which future visions play a prominent role, such as transition studies, technology development studies and backcasting studies. In the empirical part of the thesis, this model is then applied to the four cases.

These empirical ex-post evaluations, which form the heart of the current thesis, are intended to contribute to both planning theory and planning practice. The main

(11)

11

theoretical contribution lies in addressing the research gap identified above. Most of the literature on future visions in planning to date has only concerned itself with debating their impact, not so much with systematic ex-post evaluations aiming to identify the factors that can explain this impact.

Apart from this theoretical innovation the thesis also aims to make a contribution to planning practice. Visioning exercises are obviously expensive and time-consuming projects, so from a practical standpoint, it is crucial that the resulting future visions do not „devolve into inconsequential and expensive wish lists for the future‟ (Myers & Kitsuse, 2000: 228). By comparing historical cases of recently created future visions in Canada and the Netherlands, the thesis aims to draw lessons from these cases for practitioners in these two countries – and potentially also elsewhere – that might help to create future visions more efficiently and effectively.

1.3. Research questions

The main research question reads as follows:

What factors explain the impact of future visions of urban mobility?

The main research question can be divided into the following seven sub-questions. For clarity, the chapter in which each of these sub-questions is addressed is shown in parentheses.

(12)

12

1. How can the concept of „future visions of urban mobility‟ be defined? (Ch. 2.1) 2. What functions can future visions fulfil in a transition towards sustainable urban

mobility? (Ch. 2.2)

3. How can the different functions of future visions be operationalised, so that their

impact may be assessed? (Ch. 2.3 and 3)

4. What has the impact of these future visions been in the medium term (5 to 10 years)? (Ch. 5 and 6)

5. What factors have determined this impact? (Ch. 5 and 6)

6. How might the insights from historical case studies be used to create future visions more efficiently and effectively? (Ch. 7)

1.4. Structure of the thesis

The rest of this thesis is organised as follows. The next chapter provides the theoretical foundation for the research project. Based on existing literature, it provides a definition of the concept of „future visions‟, reviews their main functions, and introduces a causal model for future visions and their impact, which will be the basis for the empirical evaluations in the following chapters. Chapter 3 discusses the methodology used for the study, including case selection, empirical research design, data collection and data analysis. The research design is split up into two phases: first, an inductive phase to check the validity of the causal framework developed in Chapter 2; followed by a deductive phase applying the causal framework to the four cases.

Chapter 4 briefly introduces the four cases. The results of the empirical evaluation of these four cases are discussed in Chapters 5 and 6. Chapter 5 discusses the results of

(13)

13

the first inductive research phase, Chapter 6 focuses on the second deductive phase. By way of conclusion, Chapter 7 reflects on the causal model used in this thesis, identifies a number of lessons for planning practice based on the results of the study, and provides suggestions for further research on future visions of urban mobility and their impact.

(14)

14

2. Theoretical framework

This chapter provides the theoretical foundation for the empirical research that will be used to answer the main research question. The discussion is structured around the first three sub-questions defined in section 1.3. Section 2.1 first provides a definition of the term „future vision‟, contrasting it with a number of closely related but distinct concepts (sub-question 1). Section 2.2 reviews the different functions of future visions as they appear in the existing literature (sub-question 2). Five different strands of literature are used to answer this question: (1) the multi-level perspective (MLP) on transitions; (2) transition management studies; (3) technology development studies, particularly the German tradition around the Leitbild concept; (4) backcasting studies; and (5) the broader urban planning literature, some of which has also considered the functions of future visions. Finally, section 2.3 introduces the causal model that will serve as the basis for the empirical evaluations (sub-question 3). The causal model is based on a reworking of similar research from outside the field of urban mobility planning – in particular, existing evaluative research by Quist (2007; see also Quist et al, 2011).

2.1. Defining ‘future visions’

The objects of this research project are visions of the future, or „future visions‟ (these two terms are used interchangeably throughout this thesis). At first sight, it might seem obvious what a future vision of urban mobility would be, and it might not be particularly hard to think of examples. Defining the term is however much less

(15)

15

straightforward – the term „future vision‟ appears to be a typical „fuzzy concept‟ (Markusen, 2003) on closer inspection. Indeed, this problem has long been acknowledged in the planning literature. Shipley and Newkirk (1999: 573; see also Gaffikin & Sterrett, 2006), for example, contend that „(...) the terms vision and visioning are neither simple nor at all well understood‟. Their concern is that „these words [vision and visioning] have simply been misappropriated in suspicious ways.‟ (ibid.: 589). It is therefore important to be clear about what exactly the object of research is.

This section provides an operational definition of the term „future vision‟ in two steps. It first proposes six criteria by which it may be determined whether or not something is in fact a future vision. For the sake of clarity, the term „future vision‟ is then contrasted to other closely related but distinct terms that are often used in the planning literature. Two caveats should be kept in mind: this section‟s attempt at defining the term „future vision‟ is not aiming to be comprehensively descriptive – that is, it does not attempt to describe in full all meanings that various practitioners themselves ascribe to the terms „vision‟ and „visioning‟. The operational definition is not meant to be in any way prescriptive either – this section aims to address the question of what future visions are, not what they ought to be.

Six definitional criteria for the term „future vision‟ can be proposed to answer Shipley and Newkirk‟s (1999) question: what do these terms really mean? First, in defining „future visions‟, it is helpful to distinguish between „hard‟ definitional criteria and „soft‟ definitional criteria. Hard criteria are criteria that any potential future vision must by definition fulfil in order to be called one. Soft criteria are indicators the presence of which strongly hint at the object under consideration being a future vision,

(16)

16

though it is conceivable, in theory at least, that it might not fulfil these criteria and still rightfully be called a future vision. The three hard criteria are:

1. A future vision is inherently normative in nature. This is the essence of Banister and Hickman‟s (2012) „thinking the unthinkable‟. Visionary planning, Gaffikin and Sterrett (2006: 163) argue, is concerned not so much with the „possible‟ or with the „probable‟, but with the „desirable‟, that is, with „how to create a widely agreed view of where the city or community wants to be in two decades or so hence‟ (ibid., emphasis in original).

2. Future visions are long-term constructs. This sets visioning apart from other ways of engaging with the future in planning, such as Delphi surveys or megatrend forecasting (Banister et al, 2008). In a sense, this criterion follows logically from the first: thinking normatively about the short-term future makes little practical sense.

3. Third, and perhaps very obviously, future visions are planning tools. Normative depictions of the long-term future aimed merely at the general public, either in fictional or non-fictional form, do not qualify in this definition.

The following three soft criteria can be added to this list:

1. Future visions often contain some form of imagery. This may take the form of maps, computer-generated pictures or videos, or any other visual medium. It might be conceivable that the contents of a future vision might be transferred only via the use of words, but this seems unlikely in practice.

2. Future visions are generally strongly narrative-based. This corresponds to what Banister et al (2008: 30) call „qualitative‟ rather than „quantitative‟ plans. The term „qualitative‟ does not mean that a future vision cannot be supported by

(17)

17

models and statistics, but these support the narrative that describes the desired future state, not the other way around.

3. Finally, future visions are participatory. Quist (2007: 33) argues that visions of the future are always „shared multi-actor constructions‟, and it is indeed difficult to see how a future vision, which aims to guide stakeholder behaviour, might be created without at least some form of involvement from those stakeholders. Future visions might also be created with the help of the community at large (Grant, 2007), though this is not necessary, and hence, not a hard definitional criterion.

Taken together, these six criteria make clear what is distinct about future visions as opposed to other related concepts in the field of planning. First, a future vision is different from a planning concept. Future visions have long-term time spans but, unlike planning concepts, they do have an explicit time span; in other words, future visions are one-off projects, planning concepts are not. A future vision is also distinct from a

strategic plan: strategic plans, like future visions, aim to guide stakeholder behaviour

rather than to provide a blueprint (Faludi, 2000), but they are not necessarily long-term, and not necessarily normative (strategic planning in general can, however include a visioning element alongside plans for short-term actions; see Albrechts, 2004). A future vision is also different from a scenario. As discussed, scenarios, unlike visioning exercises, are not normative in nature. Rather than being concerned with desirable futures, scenarios are concerned with sketching a range of possible or plausible futures based on certain assumptions (Annema & De Jong, 2011; Banister & Hickman, 2012), not instruments for „thinking the unthinkable‟.

(18)

18

Finally, though the two terms are in fact very closely related, „future vision‟ does not completely overlap with backcasting experiment either. Backcasting experiments are usually considered to be a longer process, of which visioning forms a part (Geurs & Van Wee, 2004). Quist et al‟s (2011) definition of backcasting experiments distinguishes between five steps in backcasting, of which the development of a future vision is the second step. This future vision is then followed by a „backcasting analysis‟, which sketches pathways from the desired future state back to the present. Strictly speaking therefore, a backcasting experiment is more than a future vision alone, though it has become quite common to design future visions using backcasting methods (Banister & Hickman, 2012).

In sum, then, this thesis takes „future visions‟ to mean the following.

A future vision is a planning tool that is normative and long-term in nature, uses visual imagery, and is narrative-based and participatory.

2.2. Reviewing the functions of future visions

We now turn to the functions of these visions in urban and transport planning – in short, what are they good for? The rest of this section reviews five strands of literature to answer this question: (1) the multi-level perspective (MLP) on transitions; (2) transition management studies; (3) technology development studies, particularly the German tradition around the Leitbild concept; (4) recent work in backcasting studies; and (5) the broader urban planning literature, some of which has also considered the potential functions of future visions. These five strands of literature were chosen as each either

(19)

19

has already been applied, or has the potential to be applied to the field of urban transportation planning. Other perspectives, such as Strategic Niche Management (SNM), also discuss the role of future visions, but are more concerned with the introduction and diffusion of new technologies as such (Caniëls & Romijn, 2008), and hence less relevant for this thesis.

2.2.1. Functions of future visions in the Multi-Level Perspective (MLP)

The multi-level perspective on transitions is one of two main approaches to analysing socio-technical changes or „transitions‟. The other strand of transition studies, transition management, is discussed below. As the name itself indicates, the main argument of the Multi-Level Perspective is that socio-technical transitions occur at multiple levels: niche, regime and landscape (Geels et al, 2011). Niches are the domain of radical innovation, in which novelties emerge. The regime is „the locale of established practices and associated rules‟ that govern the socio-technical system (ibid.: 52). The landscape consists of the exogenous context that is outside the sphere of influence of actors at the niche and regime levels. Transportation planners have recently turned to the MLP, as it is felt that the MLP might be a helpful tool for identifying pathways towards a more sustainable form of mobility. Examples of existing studies that have applied the MLP in this way are Elzen et al (2002), Nykvist and Whitmarsh (2008), Köhler et al (2009), Naess and Vogel (2012) and Switzer et al (2013).

The MLP is largely silent on the question of what functions future visions may fulfil in bringing about a future transition to sustainability. Yet, while none of the MLP studies cited above list a comprehensive list of functions future visions might fulfil, the MLP framework itself may potentially still be a useful framework to shed light on the role of future visions in a transition process towards sustainability. From an MLP

(20)

20

perspective, an important question is whether the primary function of future visions lies at the niche level, at the regime level, or both (Quist et al, 2011) – should future visions be conceptualised as niches, or might they be used for regime transition, too?

Naess and Vogel (2012: 45) suggest a combination of both – they recommend that future visions be created „showing how environmentally sound “niche solutions” (...) for urban development could become dominating, and what kind of resulting urban structure this would entail‟, thus challenging the dominant regime from below. Similarly, Späth and Hochracher (2010: 457) note that „although the primary function [of regional future visions] is to coordinate and orient change at the level of regional niches (...), [they] often serve as an argument for the feasibility of certain changes at the meso-level (...), [and] hence also facilitate transitions at the regime level‟. Future visions thus fulfil different functions at different levels, in MLP terms. Apart from these two contributions however, studies linking the MLP to future visions and their functions remain few and far between.

2.2.2. Functions of future visions in Transition Management (TM)

The second main strand of transitions research is what is commonly called the Transition Management approach (e.g. Rotmans et al, 2001; Kemp & Rotmans, 2004; Kemp et al, 2007). The TM approach often draws on the MLP, but is less descriptive and less historical. Instead, the main focus is on governing transitions. The TM approach has been particularly influential in the context of transitions to sustainable development (Kemp et al, 2007). It has been applied extensively in the Netherlands to help create transitions towards sustainability in a variety of sectors, including energy, agriculture and health care (KSI, 2013).

(21)

21

Visions of the future are explicitly named in the TM tradition as important ingredients for bringing about a transition. In TM, future visions are conceptualised mainly as instruments that precede transition experiments in protected niche environments (Grin et al, 2010). First, a vision is defined, and from that starting point, „the objective for transition management is to steer bottom-up, niche-to-regime processes of transformation‟ towards that vision (Berkhout et al, 2004: 56). There is, however, still some debate regarding the precise status of future visions in transition management. While it is commonly argued that visions are the starting point for the transition management process, historical examples of transitions suggest that widely shared, overarching visions are neither sufficient nor necessary to achieve a transition (Berkhout et al, 2004). Geels and Schot (2010: 84) are also critical of the strong emphasis on visioning practices in TM, suggesting that many of them have become „rituals, where actors express good intentions as a form of public impression management‟.

A number of studies in transition management have attempted to list the different functions of future visions. Berkhout et al (2004) see two valuable functions that visions of the future might fulfil: facilitating deliberation and mutual learning, and creating debate, plurality and dissent. Smith et al (2005) state that visions about future system innovations have five important functions: mapping a possibility space, a heuristic, a stable frame for target-setting and monitoring, a metaphor for building actor-networks, and a narrative for focusing capital and other resources. Wiek et al (2006) also distinguish five functions of future visions: representing future system knowledge, contributing to decision-making, integrating different types of data and different thematic fields, allowing for competence-building, and facilitating transdisciplinary participation. Table 2.1 provides an overview.

(22)

22

Authors Functions of future visions (Transition Management tradition) Berkhout et al (2004) (1) facilitates deliberation and mutual learning

(2) creates debate, plurality and dissent

Smith et al (2005) (1) maps a possibility space (2) a heuristic

(3) a stable frame for target-setting and monitoring progress (4) a metaphor for building actor-networks

(5) a narrative for focusing capital

Wiek et al (2006) (1) represents future system knowledge (2) contributes to decision-making

(3) integrates different types of data and different thematic fields (4) allows for competence building

(5) facilitates transdisciplinary participation Table 2.1. Functions of future visions in the Transition Management literature.

2.2.3. Future visions (Leitbilder) in German technology development studies

The third strand of literature is the German tradition of technology development studies, which has advanced the concept of Leitbilder, or „guiding images‟ (Dierkes et al, 1996; Kuusi & Meyer, 2002; Späth, 2008; Potschin et al, 2010). The Leitbild concept was first used by Dierkes et al (1996) in a study of technological development trajectories (Quist, 2007). An effective Leitbild is a vision of the future that „creates a shared overall goal, offers orientation toward one long-term overall goal, and provides a basis for different professions and disciplines to work in the same direction‟, thus functioning as an „interpersonal stabilizer‟ (Kuusi & Meyer, 2002: 625). According to Dierkes et al (1996), the function of a Leitbild is twofold: it provides guidance (Leit), and it provides an image (Bild), which serves as a focal point that mobilises actors. Each of these two functions can be split into three sub-functions (Quist, 2007; see table 2.2). The Leitbild concept has recently been applied to fields such as regional energy policy (Späth, 2008) and spatial planning (Potschin et al, 2010).

(23)

23

Authors Functions of future visions (Leitbilder tradition) Dierkes et al

(1996)

(cited in Quist,

2007)

(1) Guidance

1a. Collective projection: creates an overall goal

1b. Synchronous pre-adaptation: aligns interaction and learning processes among actors 1c. Functional equivalent: makes space for developing new rules, replacing existing systems

(2) Image

2a. Cognitive activator: generates knowledge for realising the vision 2b. Individual mobiliser: mobilises individual actors

2c. Interpersonal stabiliser: binds heterogeneous actors

Kuusi & Meyer (2002)

Same as above

Späth (2008) (1) synchronises expectations (2) aligns actors and resources

(3) reduces uncertainty about the behaviour of other actors

Potschin et al (2010)

(1) provides a set of guidelines that shape actions

(2) provides a framework within which the impact of particular developments can be judged and socially negotiated

Table 2.2. Different functions of future visions in the Leitbilder literature.

2.2.4. Functions of future visions in backcasting studies

In their recent study, Quist et al (2011) discuss the impact of three participatory backcasting experiments (see also Quist, 2007). Though backcasting experiments are, as we have seen, strictly speaking bigger than future visions, Quist et al‟s conceptualisation of the impact of backcasting experiments is nevertheless useful here. According to Quist et al, backcasting experiments fulfil three main functions that may generate impact (ibid.: 887). These three building blocks are (1) network formation, that is, alignment of activities, actors and resources; (2) orientation (where to go) and guidance (what to do) – Quist et al adopt this from the Leitbild concept; and (3) institutionalisation, that is, the extent to which the backcasting experiment „reflects changes in institutions, practices and rules‟ (ibid.: 887). Table 2.3 provides a summary.

Authors Functions of future visions (backcasting studies) Quist et al (2011) (1) network formation

(2) orientation and guidance (3) institutionalisation

(24)

24

2.2.5. Functions of future visions in planning

Despite the increasing popularity of visioning in urban planning, there is still, as Gaffikin and Sterrett (2006: 162) point out, a „paucity of academic literature in which this planning approach has been theorized and evaluated‟. Much of the planning literature has viewed the spread of visioning approaches with some scepticism, arguing either that visioning is not in fact original or new (Shipley, 2002), or that it favours rhetoric over substance and often fails to deliver on its ambitious promises (Gaffikin & Sterrett, 2006).

Some, however, are more positive, and argue that the development of future visions does have significant added value. For example, a future vision can „constitute an effective platform for collaborative planning‟ (Ratcliffe & Krawczyk, 2011: 646; see also Grant, 2007). Detailed discussions of the different functions of future visions remain rare in this type of literature – though table 2.4 lists two exceptions.

Authors Functions of future visions (planning tradition) Gaffikin & Sterrett

(2006)

(1) „imagineers‟ futures beyond traditional land-use planning (2) connects people better to the politics of place

(3) contributes to co-ordinated thinking (4) facilitates search for consensus

Ratcliffe & Krawczyk (2011)

(1) provides an integrated approach to urban stewardship (2) constitutes an effective platform for collaborative planning (3) mobilisation

Table 2.4. Different functions of future visions in the urban planning literature.

2.2.6. Discussion

What conclusions can be drawn from the above literature review? At first sight, the five perspectives discussed in section 3 do not seem to have a lot in common – the MLP is largely silent on the different functions of future visions, while the TM approach features some very explicit and lengthy discussions of the issue; the planning perspective generally takes a very sceptical view, whereas the German literature around

(25)

25

the Leitbild concept seems far more receptive to the idea that a well-articulated future vision may guide change. Three common elements nonetheless emerge on closer examination. One could say that together, they form the core functions of future visions. These three common elements are:

 First, a future vision can provide guidance, by creating a common reference point that actors can work towards. This is what Gaffkin and Sterrett (2006) call „co-ordinated thinking‟. With the exceptions of Späth (2008) and Berkhout et al (2004), all studies in tables 2.1 and 2.2 include this element in their definitions in some way. For Dierkes et al (1996), a vision „provides an overall goal‟; for Quist et al (2011), it provides „orientation and guidance‟; in Potschin et al (2010), it „provides a set of guidelines that shape actions‟; it is a „frame for target-setting and monitoring progress‟ (Smith et al, 2005); and it „contributes to decision-making‟ (Wiek et al, 2006).

 Secondly, a future vision can bind together a group of diverse actors, each with their own agenda. In other words, it is an inter-actor „stabiliser‟ (Dierkes et al, 1996) or „platform‟ (Ratcliffe & Krawczyk, 2011) that aligns the interests (Smith et al, 2005), expectations and resources (Späth, 2008) of different actors. The vision facilitates network formation (Quist et al, 2011), and mutual learning (Berkhout et al, 2004) in the sense of learning about one another – which can result either in the creation of consensus and a strong actor-network (Smith et al, 2005; Gaffikin & Sterrett, 2006), or it generates debate, dissent and controversy (Berkhout et al, 2004). A future vision can therefore bind different actors together through agreement as well as through contestation.

(26)

26

 Thirdly, a future vision can generate new knowledge. This is another form of mutual learning (Berkhout et al, 2004) – that is, learning from one another. It is what Wiek et al (2006) call „learning-by-planning‟, similar to Dierkes et al‟s (1996) „cognitive activator‟. Collectively designing a vision of the future that integrates insights from different groups of stakeholders and different areas of expertise ultimately results in new future system knowledge (Wiek et al, 2006; Gaffikin & Sterrett, 2006).

Figure 2.1 summarises the above observations.

Figure 2.1. A conceptual overview of possible functions of future visions.

Figure 2.1 present an ideal type of future visions – these are functions that visions of the future may fulfil. As mentioned in the introduction, many authors have pointed out that a lot of future visions often end up falling short of expectations, failing to move from mere consensual rhetoric to real-world impact. In-depth empirical analysis of existing future visions, their impact on the subsequent actions of stakeholders, and their role in bringing about a transition process therefore seems like a highly pertinent research objective.

(27)

27

2.3. A causal model for future visions and their impact

Having identified three main functions that future visions may fulfil, a causal model can now be developed. Our main interest, after all, is to assess the impact of these future visions. The „impact‟ of future visions can be defined as the extent to which the vision has fulfilled each of the three functions identified in figure 2.1. In other words, a vision with high impact is a vision that has successfully provided guidance, bound actors together and generated new knowledge; a vision with low impact is a vision that has not done so.

We now work our way backwards from the effect we seek to explain – the impact of future visions – to the underlying explanatory factors. This approach is a typical example of what Goertz and Mahoney (2012: 41) call a „causes-of-effects‟ type of research design. The starting point, or „effect‟, is the impact (Y) of future visions of urban mobility, that is, the extent to which they provide guidance (Y1), bind actors

together (Y2) and generate new knowledge (Y3). The explanatory variables (the Xs) are

the factors that determine this impact.

As mentioned earlier, there is very little evaluative research on the impact of future visions of urban mobility to build on. For the explanatory factors (the Xs), we therefore have to rely on existing studies from other disciplines. The basis for the causal framework is provided by Quist et al (2011; see also Quist, 2007, for a more extensive discussion of the same research project), who evaluate the impact of three future visions five to ten years after their completion – on the topics of sustainable meat production, sustainable food consumption, and sustainable land-use in rural areas. Quist et al (2011) first distinguish between internal factors (XINT) and external factors (XEXT) that either

(28)

28

creation process. External factors are variables related to the socio-technical system and the context that surrounds it (figure 2.2).

Figure 2.2. A simplified causal framework for future visions and their impact. Source: adapted from Quist et al (2011).

Crucially, this causal framework is concerned with the impact of future visions 5 to 10 years after their completion. This may seem paradoxical, since section 2.1 defined future visions as planning tools with long-term planning horizons (typically 25 to 40 years). Our argument here however is that the impact of these visions, which themselves are long-term, is actually best observed in the medium term. Put another way, we would argue that to take the same 25 to 40 year time horizon in order to evaluate the impact of future visions is in fact to misconstrue their main function. The value of these visions is to be found not so much directly on the ground, but in the minds of planners and other relevant stakeholders: they may be instruments for change by creating a common reference point, by bringing actors together and by generating new knowledge. These are typically processes we should observe within the first 5 to 10 years. This crucial medium term window is therefore the focus of the model.

This simplified causal model is now extended by adding a number of explanatory variables. These explanatory variables are drawn from the transition studies and planning literatures, which were reviewed in the previous section. X1, for example, is

(29)

29

supported by Van de Kerkhof and Wieczorek‟s (2005) work on stakeholder learning processes; Ratcliff and Krawczyk‟s (2011) discussion of leadership underpins X2, while Gaffikin and Sterrett‟s (2006) in-depth case studies provide support for X6. The most important source however is Quist (2007) / Quist et al (2011), as it is, by some distance, the most rigorous study to date in terms of explicit hypothesis formulation. Table 2.5 lists these seven explanatory variables.

X variable Definition Sources

Internal variables

X1

Stakeholder involvement

Involvement of relevant stakeholders in the region

Quist (2007) Quist et al (2011)

Van de Kerkhof & Wieczorek (2005)

X2

Presence of vision champions

Leadership that harnesses new ideas in the vision and mobilises civic energy to follow through on them

Van de Kerkhof & Wieczorek (2005)

Gaffikin & Sterrett (2006) Ratcliff & Crawczyk (2011)

X3 Focus

Prioritised goal of the vision

(follow-up and implementation or academic achievements)

X4

Number of visions

Number of visions Quist (2007)

Quist et al (2011) External variables

X6

Power balance between actors

Stability of power relations in the region Quist (2007) Quist et al (2011)

Gaffikin & Sterrett (2006)

X7 Higher government funding

Presence and nature of higher government funding

Quist et al (2011)

X8

Coordination with other visions

Coordination with other emerging or existing visions

Quist et al (2011)

Table 2.5. Hypotheses for internal causal variables (X1-4) and external causal variables (X5-7).

Figure 2.3 extends the causal framework to include the seven explanatory variables. Both the X and Y variables are operationalised further in Chapter 3.

(30)

30 Figure 2.3. Causal model for assessing the impact of future visions. Source: adapted from Quist et al (2011).

(31)

31

3. Methodology

3.1. Case selection

In terms of empirical research, this thesis consists of four in-depth case studies of future visions of urban mobility systems and their subsequent impact, according to the causal model introduced in Chapter 2. These future visions will be drawn from two urban regions. The empirical part of the thesis can thus be described as an „embedded multiple-case design‟ (Yin, 2009: 46), as there are two levels of case selection. The first is the level of urban regions, for which we have two cases. The second level is the level of the actual future visions, for which we have two further cases in both regions.

Level 1: selecting two urban regions

The following four criteria were used to select the cases for the first level of the embedded multiple case study, the urban regions:

1. The urban regions must be sufficiently large (population over 1 million); 2. The urban regions must lie within a developed Western democracy; 3. The urban regions must have a history of long-term mobility planning; 4. The urban regions must face significant mobility challenges, in terms of

capacity, sustainability or otherwise.

The Greater Toronto Area in Canada and the Randstad in the Netherlands fulfil each of these four criteria to a particularly high degree. The first two points are self-explanatory.

(32)

32

Regarding the third criterion, both are densely populated urban regions where the creation of a high-quality mobility system has been a policy priority for many decades (see Filion & McSpurren, 2007, for the Greater Toronto area, and Alpkokin, 2012, for the Randstad).

With regard to the final criterion, the mobility systems of both regions are said to be under pressure from sharp increases in demand as a result of long-term population growth. We can expect therefore that the need to „think the unthinkable‟ (Banister & Hickman, 2012) is especially pressing in these two regions. Most experts have described transportation in the Greater Toronto region in negative terms: it is a „bottleneck‟ for the region‟s competitiveness (Keil & Young, 2008), and most policy initiatives aimed at creating a sustainable, high-quality urban mobility system over the past decades have largely flattered to deceive (Filion & McSpurren, 2007). The Randstad also faces significant mobility issues. Its problems may not be quite as severe – the Randstad‟s transportation network has been described as still „quite good‟ compared to neighbouring countries (Hilbers & Wilmink, 2002) – but it, too, faces real challenges: there is limited capacity on the existing network to accommodate future growth (ibid.), the fine-tuning of the demand for and supply of accessibility still leaves a lot to be desired (Bertolini & Le Clerq, 2003), and there is a lack of coherence in regional public transport (OECD, 2007).

Both the Greater Toronto Area and the Randstad are thus particularly appropriate cases according to the four selection criteria. There are, of course, also differences between the two regions. There are two differences between the two regions that are especially relevant for our research design. The first is the difference between monocentricity (in the Greater Toronto Area) and polycentricity (in the Randstad). This could mean, for instance, that the power balance between actors is different (X5) and

(33)

33

that government funding is distributed differently (X6). The second, more general difference is the difference in „planning cultures‟ (Sanyal, 2005). The Dutch planning culture is famous for being one of the strongest and densest in the world (Faludi, 2005; Roodbol-Mekkes & Van der Valk, 2012), whereas the Canadian planning culture shows more similarities to the planning culture in the United States (Rothblatt, 1994). It is expected that this difference has an influence both on the internal explanatory variables and on the external explanatory variables. Hence, for both these differences, no new X has been created – rather, these two differences between the cases should be seen as factors underlying the existing Xs.

It should be noted that the names „Greater Toronto Area‟ and „Randstad‟ are indicative, as the cases might not cover these exact areas. In Canada, the city of Hamilton is often included in spatial plans to form the Greater Toronto and Hamilton Area (GTHA); the Niagara peninsula may also be included, resulting in an area known as the Golden Horseshoe or, if its hinterland is also included, the Greater Golden Horseshoe (GGH). In the Netherlands, planning policy commonly divides the Randstad into a Noordvleugel, which includes Amsterdam and Utrecht, and a Zuidvleugel, which includes Rotterdam and The Hague. The research design, in other words, is geographically flexible; our research question remains the same regardless of the exact scale.

Level 2: selecting two future visions for each region

The second level of case selection is the selection of future visions within both urban regions. Three criteria were used to select the cases within the Greater Toronto area and the Randstad area. First, the case must conform to all aspects of the definition of a „future vision‟ developed in section 2.1: it must be a planning tool that is normative and

(34)

34

long-term in nature, uses visual imagery, and is narrative-based and participatory. The cases selected should thus be „typical cases‟ (Gerring, 2007) for the phenomenon under study, future visions of urban mobility systems. Second, the future vision must be between 5 and 10 years old. As discussed in section 2.3, the causal model focuses on this time window, since it is in this medium term that a future vision will generate most of its impact. Apart from this theoretical argument, this time frame is also most reasonable and workable from a methodological point of view: less than 5 years would make it difficult to assess medium or long-term „impact‟, while more than 10 years would likely result in data availability problems. Third, there must be sufficient information available about each case. Note that since it is very difficult to know in advance how the cases will score on either the X or the Y variables, it was not possible to explicitly select „diverse cases‟. However, the working assumption was that selecting four cases from two different countries would still guarantee a sufficient degree of diversity.

A brief scan of reports by various media outlets as well as publications by relevant governmental agencies was carried out to identify potential cases in both urban regions. After ruling out potential cases that did fully not meet the definitional criteria and cases that were either too old or too recent, two future visions remained as fully suitable cases in both urban regions:

(35)

35

Greater Toronto Area

1. Places to Grow, published in 2006, 25-year time horizon 2. The Big Move, published in 2008, 25-year time horizon

Randstad

3. Metropoolregio Amsterdam Duurzaam Bereikbaar, published in 2008,

32-year time horizon

4. Masterplan Rotterdam Vooruit, published in 2009, 31-year time horizon

These four cases are described in more detail in Chapter 4.

3.2. Empirical research design: two phases

Now that the four cases have been selected, the question is how the causal framework outlined in Chapter 2 might be applied to these cases. It should be kept in mind that the basis for this causal framework comes from outside the field of transportation. The research design for this study should be understood against the backdrop of this state of the literature. On the one hand, we do have a relatively specific causal model with X and Y variables that can be applied to our cases. On the other hand, since we are borrowing from research outside the area of transportation, there is also a need to test, as a first step, whether the causal framework itself is accurate and comprehensive enough when applied to visions of urban mobility.

With this in mind, the most logical option is to adopt a mixed research design, based on two distinct methodological approaches. The first part of the empirical research design is what could be called the inductive phase. This first phase does not depart from the causal framework, but instead aims to check the applicability and

(36)

36

validity of the causal framework itself. This implies a loosely structured form of analysis, which aims to explain the Y („impact‟), but does not depart from a preconceived set of possible explanations. The results of this inductive analysis can then be compared to the causal model: if the inductively derived explanations overlap with the variables in the causal model, it can be concluded that the model is accurate and comprehensive; if, on the other hand, the inductively derived explanations and the model do not match, the model needs to be adjusted.

The second step in the empirical research design is what could be called the

deductive phase. Having checked the validity of the theoretically derived causal model

in the first phase (and having made adjustments to the causal model where necessary), this second part of the research then applies the refined causal framework to the cases. Each of the variables in the causal framework can be operationalised into measurable statements, which can be scored based on deductive data analysis.

Figure 3.1 provides an overview of the two steps that make up the empirical research design. Section 3.3 discusses the two data sources that serve as input for this research design: semi-structured interviews and secondary documents. The final two sections in this chapter then outline the analytical steps in phase I and phase II of the research design, respectively.

(37)

37

3.3. Data collection

As figure 3.1 shows, two sources of data were used for the fieldwork in both countries: interviews with key actors and secondary documents. The data collection procedure that was followed for these two data sources is described below.

Interviews

The first and most central data source is semi-structured interviews with key actors. Four types of interviewees were identified in advance: (1) creators of the vision; (2) other governmental parties in the region, e.g. municipal planners; (3) non-governmental stakeholders; and (4) external experts. It was decided not to focus on the parties behind the creation of the vision in question alone, but to also include other stakeholders and external experts. This made it possible to secure a much richer diversity of perspectives amongst the interviewees.

In total, 43 interviews were held; 24 in Canada and 19 in the Netherlands. Table 3.1 provides an overview. Note that a number of interviewees in Canada were respondents for both Places to Grow and The Big Move. The 19 interviews in the Netherlands all focused on only one case. Of the 43 interviews, one was conducted with two respondents, and two were conducted with three respondents. These interviews are shown as one interview in table 3.1. The interviews in Canada were conducted in the fall of 2013, the interviews in the Netherlands were conducted in the spring of 2014. Only two potential respondents, both in Canada, declined to participate (not included in table 3.1).

(38)

38 Places to Grow The Big Move Both Canadian cases Metropoolregio Amsterdam Duurzaam Bereikbaar Masterplan Rotterdam Vooruit 1. Creators of the vision 1 1 - 7 8 2. Other governments 2 2 7 2 - 3. Non-governmental stakeholders - 4 1 - 1 4. External experts 3 2 1 - 1 Total 6 9 9 9 10 24 19

Table 3.1. Overview of semi-structured interviews conducted for the four cases.

The above table shows some differences between the types of interviews for the Canadian cases and the interviews for the Dutch cases. However, two things should be kept in mind. First, the Canadian visions were created by one government body (the Ontario Ministry of Infrastructure for Places to Grow, and the regional transportation authority Metrolinx for The Big Move), whereas the Dutch visions were created by a number of government parties working together. As a result, the two Dutch cases have more interviews in the category 1, but fewer in category 2. Secondly, table 3.1 only shows the current position of the respondents, but this position may be different from their position at the time the vision was created. In Canada for example, two municipal planners and two external experts were involved in the creation of Places to Grow and The Big Move at the time, but were classified as category 2 and 4 here based on their current positions.

Depending on the type of interview, the interviewees were asked about the creation of the vision, and/or the ways in which they and their organisations have internalised and acted upon the vision after its completion. Because the research design starts with an open-ended, inductive phase, the interviews did not depart from the theoretically derived Xs and Ys. Instead, a form of narrative interviewing was used. During the

(39)

39

conversation, the interviewees were invited to „tell the story‟ of the life of the future vision (Riessman, 2001; Bryman, 2008; see Giezen, 2012, for a recent example of narrative interviewing in planning research). For this type of interviewing, „preliminary work‟, such as introducing oneself to the respondent and explaining the purpose of the interview, is particularly important for „evoking the respondent‟s story‟ (Holstein & Gubrium, 1995: 40). The interviewer should keep guidance to a minimum, ask short prompting questions and ask for clarification if necessary (Kvale & Brinkmann, 2009). Overall, this strategy was successful, as most of the interviewees felt free to talk about their experiences of creating and/or working with the vision, their views on the strengths and weaknesses of the vision, and on lessons they felt they had learned.

For the interviews with the vision creators (category 1 in table 3.1), a potential issue is that they may exaggerate the impact of their vision, because they regard it as a personal achievement that they are proud of. To some extent, this problem is unavoidable, as the there is no alternative for the first-hand knowledge that these interviewees have. It was therefore of great importance to phrase the interview questions particularly carefully to avoid any suggestion that the interviewee was being „judged‟. In hindsight, there is no indication that the creators of the visions are more positive about „their‟ vision compared to the other types of interviewees, which suggests that this issue did not play a significant role, and the interviews can be seen as reliable.

Secondary documents

In addition to the interviews, an overview of relevant secondary documents was created for each case. For the most part, these secondary documents were found through searches of the websites of the responsible organisations or through general internet

(40)

40

searches. The full list of secondary documents can be found in Appendix A. Generally, the following types of documents were retrieved:

1. For the vision creators: the visions themselves, including supporting documentation (e.g. discussion papers, drafts and progress reports);

2. For other governments: Official Plans, Transportation Master Plans, and other strategic plans published or amended since the creation of the vision;

3. For non-governmental stakeholders: reviews and submitted opinions; 4. For external experts: commentaries and reports.

To make sure that no important information would be overlooked, the interviewees were asked if there were any additional secondary documents published by their organisation they thought would be relevant to include. No important publicly available documents seem to have been overlooked, though some stakeholders did provide some additional internal documentation. Because of this extra check, it seems safe to assume that the list of secondary documents is comprehensive.

Given the sheer number of secondary documents found in this way, it was necessary to distinguish between key secondary documents, which are directly related to the research question; and supporting secondary documents, which are only partially related to the research question. All municipal transportation plans retrieved, for instance, comprise thousands of pages of documentation, but only certain parts of these relate directly to the visions that are the object of this research project. The documents identified as key secondary documents are listed in Appendix B (numbers correspond to the full list in Appendix A).

(41)

41

3.4. Data analysis phase I (inductive analysis)

As discussed in section 3.2, the first phase of the research design is an inductive form of analysis (figure 3.2). The goal of phase I is to check the validity of the causal framework, which was borrowed from a different domain, when it is applied to the field of urban mobility, and to make adjustments to the model where needed. The adjusted causal model can then serve as the basis for the deductive phase II in the research design.

Figure 3.2. Research design phase I: inductive analysis.

In practice, the „open coding‟ involved the following steps. All interviews were transcribed and, together with the key secondary documents, entered into the qualitative analysis program Atlas.ti 6.2. For each case, all transcripts and documents were coded to look for reoccurring themes. This was done in three rounds. First, the data was searched for themes relating to impact (these were marked with codes starting with „Y_‟). The data was then searched for topics that could explain this impact (codes starting with „X_‟). Finally, other potentially relevant themes that did not completely fit either of the first two criteria were then coded (codes being just the theme itself). A list of all codes and the number of quotations attached to each code can be found in

(42)

42

Appendix C. Overlapping topics were then merged into single codes to draw out the most important themes emerging from the data. The resulting codes were then sorted by the number of quotations attached to each code1. The top 6 or 7 codes were then identified as „key factors‟ for each of the four cases.

Based on these key factors resulting from the open coding, thick descriptions were created for each of the four cases, centred around 6 or 7 key „lessons learned‟. Sections 5.1 to 5.4 discuss each of these factors in detail on a case-by-case basis. Section 5.5 then makes adjustments to the causal model as introduced in Chapter 2 based on these factors.

3.5. Data analysis phase II (deductive analysis)

Phase II of the empirical research then applied the causal model, having been checked and adjusted in the first phase, to the four cases (figure 3.3). Now that adjustments were made to make sure the causal model can be applied in the context of urban mobility visions, the remaining task in this deductive phase is simply to operationalise and score the X and Y variables to test if a relationship between X and Y can be observed in our four cases, as the causal model hypothesises. This operationalisation procedure is described below.

1

Note that selecting the topics that were mentioned first by the interviewees, another possible way of identifying important codes, would not be appropriate here, since most interviews were structured chronologically. There is therefore no relationship between which topics are mentioned first and which are the most important.

(43)

43 Figure 3.3. Research design phase II: deductive analysis.

Operationalisation of Y variables (impact)

In the causal model introduced in Chapter 2, „impact‟ was defined in terms of three variables: guiding (Y1), binding together (Y2) and learning-by-planning (Y3; see figure 3.4). These three Y variables can be operationalised in two steps. First, by going back to the literature, these three abstract concepts can each be divided into three more measurable statements. These are shown in table 3.2.

Instruments for transition : the impact of future visions of urban mobility in greater Toronto and the Randstad

U

NIVERSITY

OF

A

MSTERDAM

Instruments for Transition: The

Impact of Future Visions of Urban

Mobility in Greater Toronto and the

Randstad

Florian Langstraat

Instruments for Transition: The Impact of Future

Visions of Urban Mobility in Greater Toronto and the

Randstad

Florian Langstraat

Executive summary

Self-plagiarism disclaimer

Acknowledgements

Table of contents

1. Introduction

2. Theoretical framework

3. Methodology