1、Travel Transitions How Transport Planners and Policy Makers Can Respond to Shifting Mobility TrendsTravel Transitions How Transport Planners and Policy Makers Can Respond to Shifting Mobility TrendsResearch Report 2021 The International Transport Forum The International Transport Forum is an intergo

2、vernmental organisation with 63 member countries.It acts as a think tank for transport policy and organises the Annual Summit of transport ministers.ITF is the only global body that covers all transport modes.The ITF is politically autonomous and administratively integrated with the OECD.The ITF wor

3、ks for transport policies that improve peoples lives.Our mission is to foster a deeper understanding of the role of transport in economic growth,environmental sustainability and social inclusion and to raise the public profile of transport policy.The ITF organises global dialogue for better transpor

4、t.We act as a platform for discussion and pre-negotiation of policy issues across all transport modes.We analyse trends,share knowledge and promote exchange among transport decision-makers and civil society.The ITFs Annual Summit is the worlds largest gathering of transport ministers and the leading

5、 global platform for dialogue on transport policy.The Members of the Forum are:Albania,Armenia,Argentina,Australia,Austria,Azerbaijan,Belarus,Belgium,Bosnia and Herzegovina,Bulgaria,Canada,Chile,China(Peoples Republic of),Colombia,Croatia,Czech Republic,Denmark,Estonia,Finland,France,Georgia,Germany

6、,Greece,Hungary,Iceland,India,Ireland,Israel,Italy,Japan,Kazakhstan,Korea,Latvia,Liechtenstein,Lithuania,Luxembourg,Malta,Mexico,Republic of Moldova,Mongolia,Montenegro,Morocco,the Netherlands,New Zealand,North Macedonia,Norway,Poland,Portugal,Romania,Russian Federation,Serbia,Slovak Republic,Sloven

7、ia,Spain,Sweden,Switzerland,Tunisia,Turkey,Ukraine,the United Arab Emirates,the United Kingdom,the United States and Uzbekistan.International Transport Forum 2 rue Andr Pascal F-75775 Paris Cedex 16 contactitf-oecd.org www.itf-oecd.org ITF Research Reports ITF Research Reports are in-depth studies o

8、f transport policy issues of concern to ITF member countries.They present the findings of dedicated ITF working groups,which bring together international experts over a period of usually one to two years,and are vetted by the ITF Transport Research Committee.Any findings,interpretations and conclusi

9、ons expressed herein are those of the authors and do not necessarily reflect the views of the International Transport Forum or the OECD.Neither the OECD,ITF nor the authors guarantee the accuracy of any data or other information contained in this publication and accept no responsibility whatsoever f

10、or any consequence of their use.This document and any map included herein are without prejudice to the status of or sovereignty over any territory,to the delimitation of international frontiers and boundaries and to the name of any territory,city or area.Cite this work as:ITF(2021),Travel Transition

11、s:How Transport Planners and Policy Makers Can Respond to Shifting Mobility Trends,ITF Research Reports,OECD Publishing,Paris.Acknowledgements This report sets out the findings of a Working Group facilitated by the International Transport Forum(ITF)and chaired by Kiron Chatterjee of the Centre for T

12、ransport and Society at the University of the West of England,Bristol,United Kingdom.Substantive directions and inputs were provided by Philippe Crist,Advisor for Innovation and Foresight at the ITF.The Working Group was facilitated by Asuka Ito,Policy Analyst,and Ombline de Saint Lon Langls,Researc

13、h Officer,both at the ITF.The report was reviewed by Stephen Perkins,Head of Research and Policy Analysis at the ITF and editing support was provided by Gemma Nellies,independent editor,and Hilary Gaboriau,Content Production Assistant at the ITF.The principal authors and section co-ordinators were:K

14、iron Chatterjee(University of the West of England,Bristol,United Kingdom):“Introduction”and“Trend breaks and travel transitions”Javier Burrieza Galn(Nommon Solutions and Technologies,Spain):“Looking back to look ahead”Glenn Lyons(University of the West of England,Bristol,United Kingdom):“Handling un

15、certainty in assessing travel transitions”Karolina Isaksson(National Road and Transport Research Institute,Sweden)and Greg Marsden(University of Leeds,United Kingdom):“Governance challenges and opportunities”.Other contributors to the report include:Alexandra Millonig(Austrian Institute of Technolog

16、y),Marcin witaa(Road and Bridge Research Institute,Poland),Mehmet Yazc(Ministry of Transport and Infrastructure,Turkey),and Peter Jorritsma(KiM,Netherlands Institute for Transport Policy Analysis).Working Group participants were:Austria,Chile,France,Germany,Ireland,Latvia,Mexico,the Netherlands,New

17、Zealand,Poland,Portugal,Spain,Sweden,Turkey,the United Kingdom,the United States and the European Commission.The full list of participants at the Urban Travel Transition and New Mobility Behaviors in Light of Covid-19 Working Group Meeting held on 7-8 December 2020 appears in Annex A.TABLE OF CONTEN

18、TS 4 TRAVEL TRANSITIONS OECD/ITF 2021 Table of contents Executive summaryExecutive summary .7 7 IntroductionIntroduction .1010 Unexpected trend breaks.11 Realisation that existing practices are inadequate for future transport planning.12 Readers guide to this report.14 Trend breaks and travel transi

19、tionsTrend breaks and travel transitions .1515 Travel behaviour,habits and change.15 Drivers of change in travel behaviour.21 The impact of the Covid-19 pandemic.28 Lessons from past travel transitions.31 Indicators of change.32 New data opportunities for monitoring travel trends.34 Recommendations

20、on scanning for emerging trends.35 Looking back to look aheadLooking back to look ahead .3737 Multiple reasons to look ahead.37 Established approaches to travel demand futures.42 The uncertain nature of travel demand futures.47 A certainty:Looking ahead always influences future pathways.51 Handling

21、uncertainty Handling uncertainty in assessing travel transitionsin assessing travel transitions .5252 Transition and deepening uncertainty.52 The need for a transition in transport planning and appraisal.57 Ways of exploring the future to inform strategic planning.61 State of practice in handling un

22、certainty.65 Looking to the future.68 Governance challenges and opportunitiesGovernance challenges and opportunities .7070 Key challenges for contemporary transport governance.70 Making transport governance more transformative.73 Transformative capacity in practice.76 Transport governance for a chan

23、ging world.78 NotesNotes .8080 ReferencesReferences .8181 Annex A.Annex A.Working Group participants listWorking Group participants list .9595 TABLE OF CONTENTS TRAVEL TRANSITIONS OECD/ITF 2021 5 Figures Figure 1.Growth in passengerFigure 1.Growth in passenger-kilometres by private car and light van

24、s in six advanced kilometres by private car and light vans in six advanced economies,1990economies,1990-2009(indexed to 1990=100)2009(indexed to 1990=100).1111 Figure 2.Hierarchical structure of decisionsFigure 2.Hierarchical structure of decisions .1616 Figure 3.Conceptual framework for influences

25、on travel patternsFigure 3.Conceptual framework for influences on travel patterns .2121 Fi Figure 4.Number of motor vehicles per 1gure 4.Number of motor vehicles per 1 000000 inhabitants vinhabitants versuersus GDP per capita,2014s GDP per capita,2014 .2323 Figure 5.Shared mobility service options b

26、y trip lengthFigure 5.Shared mobility service options by trip length .2626 Figure 6.Daily trips between Madrid and Barcelona,14Figure 6.Daily trips between Madrid and Barcelona,14 February 2020 to 14February 2020 to 14 February 2021February 2021 .3535 Figure 7.The futures coneFigure 7.The futures co

27、ne .3939 Figure 8.Predict and provide cyclical processFigure 8.Predict and provide cyclical process .4444 Figure 9.Sources of uncertaintyFigure 9.Sources of uncertainty .5050 Figure 10.Time series data for selected OECD coFigure 10.Time series data for selected OECD countries depicting potential sig

28、ns of regime untries depicting potential signs of regime transition,1995transition, .5454 Figure 11.Hype cycle depicting a path for emergent technologiesFigure 11.Hype cycle depicting a path for emergent technologies .5555 Figure 12.Plausible uncertainty trajectoriesFigure 12.Plausible

29、uncertainty trajectories .5656 Figure 13.Illustrative hedgehog diagram of road traffic forecastingFigure 13.Illustrative hedgehog diagram of road traffic forecasting .5757 Figure 14Figure 14.The adaptability of the Triple Access System and its use during the.The adaptability of the Triple Access Sys

30、tem and its use during the CovidCovid-19 pandemic19 pandemic .5959 Figure 15.Contrasting a predict and provide approach with a decide and provide approach to Figure 15.Contrasting a predict and provide approach with a decide and provide approach to strategicstrategic transport planningtransport plan

31、ning .6060 Figure 16.StressFigure 16.Stress-testing policy options against uncertaintytesting policy options against uncertainty .6565 Figure 17.The transformative capacity frameworFigure 17.The transformative capacity framework k .7676 Tables Table 1.Travel behaviour studies during the CovidTable 1

32、.Travel behaviour studies during the Covid-19 pandemic19 pandemic .2929 Table 2.LEVITATE goals and indicatorsTable 2.LEVITATE goals and indicators .3333 Table 3.An overview of metaTable 3.An overview of meta-studies on projectstudies on project-based travel demand forecasting based travel demand for

33、ecasting performanceperformance .4646 Table 4.Walker uncertainty levelsTable 4.Walker uncertainty levels .4848 Table 5.Covey circlesTable 5.Covey circles .4949 TABLE OF CONTENTS 6 TRAVEL TRANSITIONS OECD/ITF 2021 Boxes Box 1.Monitoring travel demand in Spain through mobile network data during the Bo

34、x 1.Monitoring travel demand in Spain through mobile network data during the CovidCovid-19 pandemic19 pandemic .3535 Box 2.The evolution of Finnish national transport futures stuBox 2.The evolution of Finnish national transport futures studiesdies .4242 Box 3.A sixBox 3.A six-stage approach to apply

35、ing decide and providestage approach to applying decide and provide .6262 Box 4.Question posed to International Transport Forum country rBox 4.Question posed to International Transport Forum country representativesepresentatives .6767 EXECUTIVE SUMMARY TRAVEL TRANSITIONS OECD/ITF 2021 7 Executive su

36、mmary What we did This report considers why breaks in past travel trends and the emergence of new urban mobility behaviours were not foreseen.It assesses what has been learned about the causes of previous trend breaks and highlights how the future evolution of travel demand can be better anticipated

37、.It reviews how transport planners and managers use projections of future movement of people and goods to guide decision making and identifies the limitations of established approaches for forecasting travel demand.The report considers the implications of deep uncertainty for strategic transport pla

38、nning and suggests new ways of thinking and planning for more resilient investment decisions.It also considers the governance challenges to bring about change and the associated institutional implications.What we found Travel trends in industrialised,urbanised countries at the start of the 21st cent

39、ury have confounded expectations.Scientific studies have used theories of behaviour change and longitudinal data to analyse these travel trends.These studies have shed some light on travel transitions but have rarely fully answered why these transitions have occurred.It is now clear that long-term t

40、ravel trends are primarily influenced by changes to broader society and lifestyles rather than by internal changes to the transport system,although the interaction between these factors is important.The Covid-19 pandemic may cause further long-term changes to travel behaviour be that due to people b

41、ehaving differently during the pandemic and continuing these new behavioural patterns afterwards,or due to broader changes in society,which arise as a result of the pandemic(e.g.changes to working practices).Public authorities have been slow to identify breaks in travel trends and to put effort into

42、 understanding these changes.Although efforts to track changes in travel patterns in response to transport investments have increased in the last 20 years,a more proactive approach should be taken to anticipate future travel transitions.Such analysis will need to look at changes in socio-economic sy

43、stems,advances in transport and telecommunications,and changes in travel preferences and patterns.The“predict and provide”paradigm,which has dominated transport planning since the mid-20th century,typically utilises one or more forecasting models to predict future demand.However,evidence on the perf

44、ormance of forecasting methods suggests two major limitations:a failure to capture the uncertain nature of travel demand and unsuitability for supporting new decision-making methods in transport planning.The deterministic travel demand forecasts widely applied under this paradigm were not intended t

45、o shape or limit mobility.However,the use of such forecasting methods often seems to have led to a cyclical reinforcement of undesirable trends.EXECUTIVE SUMMARY 8 TRAVEL TRANSITIONS OECD/ITF 2021 The confluence of the motor age and the digital era has created a state of flux in transport and societ

46、y generally,further exacerbated by the Covid-19 pandemic.This instability and change challenges the forecast-led paradigms effectiveness in many contexts and suggests the need for a transition in transport planning and appraisal from“predict and provide”towards a“decide and provide”approach.The deci

47、de-and-provide paradigm gives primacy to access,i.e.a combination of physical mobility,spatial proximity and digital connectivity.It is also vision-led rather than forecast-led,i.e.it sets out a preferred future and charts a course towards it that allows for uncertainty.This approach has been applie

48、d in the United Kingdom,where scenario-based representations of uncertainty have been used to test the resilience of policy interventions.Signs point to a change in thinking and practice in several other countries as well.However,the formal and informal frameworks that condition the development and

49、implementation of new approaches can hold back progress.The challenge is to develop and integrate target-oriented transformative governance processes,even in fragmented institutional contexts characterised by established norms and power relations.What we recommend Scan for emerging travel trends usi

50、ng a combination of traditional and new data sources Scan for emerging travel trends using a combination of traditional and new data sources Big data sources(such as mobile phone records)can be used to rapidly detect changes in travel patterns at a high degree of spatial and temporal resolution.Howe

51、ver,traditional travel surveys also remain essential for monitoring longer-term trends and relating these to socio-demographic and lifestyle factors.Detailed studies should be undertaken of outlier groups and areas exhibiting new mobility behaviours.Measure the performance of the transport system wi

52、th indicators that reflect how mobility contributes to Measure the performance of the transport system with indicators that reflect how mobility contributes to societal objectsocietal objectivesives Travel is not an end in itself and indicators of travel activity need to be helpful in assessing how

53、mobility contributes to achieving societal goals.In this light,transport planners should reflect on whether the indicators for the travel activity they track are the most useful ones.Instead of focusing on total or average travel across a population,they should also look at the spread in values acro

54、ss the population to understand equity impacts.Take a proTake a proactive approach to anticipating travel transactive approach to anticipating travel transitions by scanning developments inside and outsideitions by scanning developments inside and outside thethe transporttransport sectorsector Trans

55、port planners should proactively scan developments inside and outside the transport sector in order to assess the implications of such developments for travel preferences and behaviours.This requires collaboration with scientists and researchers in other fields.Account for uncertainty when making pr

56、edictions and be explicit about the different sources of uncertainty Account for uncertainty when making predictions and be explicit about the different sources of uncertainty The level of uncertainty regarding future travel demand can vary from complete certainty to complete(or“deep”)uncertainty de

57、pending on the transport system and time horizon under consideration.The uncertainty in predictions stems from uncertainty in values of explanatory variables,assumed relationships and underlying processes.Uncertainty in transport often arises from the diverse stakeholders involved and their autonomy

58、 in decision making.EXECUTIVE SUMMARY TRAVEL TRANSITIONS OECD/ITF 2021 9 Shift from a Shift from a“predict and providepredict and provide”approach towards a approach towards a“decide and providedecide and provide”approach in the face of approach in the face of deepdeep uncertaintyuncertainty Determi

59、ning what to do about deep uncertainty when making planning and investment decisions requires a new approach to strategic transport planning,as well as a new way of thinking about future trends.This shift entails moving from a purely predict and provide approach towards decide and provide an approac

60、h that is vision-led rather than forecast-led and in which a preferred future is identified and a pathway towards that future is conceived that can negotiate uncertainty ahead.Decide and provide approaches should be part of the suite of methods used to assess the resilience of planning and investmen

61、t decisions.Inherent to such an approach should be a move away from focusing only on mobility and a move towards recognising that the main purpose of transport to enable access to opportunities can also be achieved by spatial proximity and digital connectivity.Change the mindset and enhance the skil

62、lset of the Change the mindset and enhance the skillset of the transporttransport-planningplanning workforceworkforce Transport planners will need a different mindset and new skills to enable the sector to move to vision-based transport planning.For example,new notions and norms of analytical robust

63、ness are required in strategic transport planning which focus more on plausibility than precision it is better to be approximately right than precisely wrong.Foster a strengthening of international knowledge sharing and coFoster a strengthening of international knowledge sharing and co-operation via

64、 a“learning by doing”operation via a“learning by doing”approachapproach Learning from successful practical examples and building upon them will create the necessary confidence that new approaches to dealing with uncertainty are fit for purpose.Fostering international knowledge sharing and co-operati

65、on in a learning-by-doing context will help the transition to more effective planning approaches to gain momentum.Adapt transport governance to better accoAdapt transport governance to better account for uncertainty in planningunt for uncertainty in planning The current situation of uncertainty and

66、flux with respect to future travel demand is an opportunity to acknowledge the limitations of existing processes and transform transport governance.However,developing transformative capacity requires adaptation and adjustment to specific contexts,actor constellations and situations.Sweden offers an

67、example of how this is possible.The Swedish Energy Agency led a commission to decarbonise the transport sector,demonstrating how target-oriented transformative governance processes can be developed and integrated,even in fragmented institutional contexts characterised by established norms and power

68、relations.INTRODUCTION 10 TRAVEL TRANSITIONS OECD/ITF 2021 Introduction Travel trends in industrialised,urbanised countries around the world have confounded expectations at the start of the 21st century.Growth in motorised travel has slowed down and car travel per capita has decreased in some countr

69、ies.Public transport use and cycling have increased in many cities.Many different factors appear to be at play in influencing peoples desires and needs for mobility changes to socio-demographics,economies and lifestyles are occurring at the same time as fast-paced technological change.On top of this

70、,the Covid-19 pandemic has arrived,without warning,to accentuate the sense of uncertainty about what the future holds for personal travel.Will the changes Covid-19 has brought about permanently alter the way people live now that societies have adapted to perform many essential activities without nee

71、ding to travel,or will it make people appreciate travel more than ever and seek out greater travel opportunities?This report is concerned with“travel transitions”,those changes that induce a break from existing and habitual patterns of behaviour and lead from one state or condition to another.The tr

72、avel behaviour of urban residents is the focus of this paper,as it is urban residents whose travel behaviour has changed the most in the last 20 years in industrialised countries and urban areas that face many of the most serious challenges that lie ahead.However,trends are often reported at a broad

73、er level than just urban areas,so the report considers travel at a variety of levels,from local to national.Although growth in travel has decreased in industrialised countries,the need to anticipate future demand for travel and make plans to accommodate or influence these developments has never been

74、 more important.A good understanding of travel activity is required to provide transport infrastructure and services that serve the needs of all in society and to support economies.It is important to understand and account for travel transitions in infrastructure provision,transport and land-use pla

75、nning and the regulation of mobility services.For those world regions still experiencing rising motorised travel,the travel transitions in industrialised countries may indicate a need to re-evaluate perceived wisdom regarding appropriate transport policy design and deployment.Transport has a key rol

76、e to play in tackling existential threats to society such as climate change,air pollution and inactive lifestyles.Hence,there is a need for public authorities to be proactive in shaping the future of transport and demand for travel in order to meet societal imperatives.Prior to the pandemic,transpor

77、t authorities in industrialised countries could be seen as largely reactive to external developments and struggling to understand the implications of changes taking place both externally and within the transport sector(such as the introduction of new mobility services).This report explains why it is

78、 no longer reasonable to make confident predictions about future travel activity given the diverse set of factors that influence mobility and the complex inter-relationships between them.The report argues,based on the experience of the last two decades,that the analytical and decision-making methods

79、 of the past are no longer fit-for-purpose and a change is required,not only in methodology but also in the ways of thinking about the future.There is a need to directly engage with uncertainty in a way that recognises that there are multiple possible futures and to acknowledge that actions taken to

80、day will influence the future and can help chart a course towards desirable outcomes.The INTRODUCTION TRAVEL TRANSITIONS OECD/ITF 2021 11 report also considers the governance challenges in bringing about the necessary changes in perspectives and approaches to transport planning and what these mean f

81、or institutional arrangements.The remainder of this introduction provides further background to the report and explains why a change in perspective is needed by public authorities in order to anticipate future transport developments and make strategies and plans which are resilient and address the m

82、ajor challenges that they face.Unexpected trend breaks Growth in travel has not developed as expected in urban,industrialised economies in the last two decades.This was observed as long ago as 2011,when an ITF report highlighted the lack of growth in car travel in six advanced economies(Australia,Fr

83、ance,Germany,Japan,the United Kingdom and the United States)in the first ten years of the 21st century(see Figure 1),in contrast to the strong growth that had been seen in preceding decades(ITF,2011).Figure 1.Figure 1.Growth in pGrowth in passeassengernger-kilometres by private car and light kilomet

84、res by private car and light vanvans s in six advanced economiesin six advanced economies,19199 90 0-2009 2009(indexed to indexed to 1990=100)1990=100)Source:ITF(2011).Following these observations,there has been widespread discussion of whether“peak car usage”has been reached and what the future hol

85、ds for car travel and the use of alternative methods of transport(Goodwin,2012).It has proven difficult to explain why car use has stopped growing in these and other countries experiencing similar trends.It is clear that the previously observed strong link between economic growth and travel growth h

86、as weakened(Millard-Ball and Schipper,2011).Suggestions have been made that a saturation point has been reached where citizens cannot benefit from more travel(Metz,2013).This is particularly argued for cities,as illustrated by Newman and Kenworthy(2011),who showed that growth in car use in cities in

87、 Australia,Canada,Europe and the United States has been slowing down in each decade since 1960.In summarising the search for explanations,Goodwin and van Dender(2013)concluded that:New INTRODUCTION 12 TRAVEL TRANSITIONS OECD/ITF 2021 econometric work suggests that an aggregate model focusing on gros

88、s domestic product(GDP)effects and fuel prices is too crude to capture the diversity and various dynamics underlying aggregate car travel demand and how it changes”.Researchers have therefore focused their attention on different groups of the population.For example,in the United Kingdom there have b

89、een:decreases in car driving among men,but little change in driving among women;decreases in car driving by young adults,but increases in driving by older adults;and decreases in driving in cities with little change in driving in smaller urban areas and rural areas(DfT,2015).BMWs Institute for Mobil

90、ity Research commissioned an international comparative study in 2010 which assessed how the Generation Y cohort(defined as being born between the early 1980s and early 1990s)differed from previous cohorts in their travel behaviour(IFMO,2013).It compared trends in France,Germany,Great Britain,Japan,N

91、orway and the United States.It found a similar trend of decreased distance travelled by car in all six countries,although noted differences between countries in trends such as public transport use(for example,increasing strongly in Germany,but unchanged in the United States).The study found that one

92、-half of the decrease in car ownership in Germany could be explained by socio-economic changes(such as decreasing real incomes,an increase in the proportion of urbanites and a higher average age for starting a family),but the other half could not be explained by those factors.For Great Britain,only

93、one-third of the decrease of car ownership could be explained by socio-economic factors.An in-depth study commissioned by the UK Department for Transport(DfT)sought to understand the reasons why young people in Great Britain are driving less than previous generations(Chatterjee et al.,2018).It asses

94、sed the role of 22 possible contributary factors and concluded that there are multiple inter-connected reasons for the changes in transport use.They include changes in the socio-economic,living and family situations of young people,as well as changes in transport costs and the relative importance at

95、tached to driving in the digital age.An important point to note is that travel patterns and trends can vary enormously,even within a country or city.A comparison of the travel trends in six city regions(Atlanta,Brisbane,London,Melbourne,New York and the combined English metropolitan areas of Greater

96、 Manchester,Merseyside,West Midlands and Tyne and Wear)found“markedly different trends in young adult travel behaviour that are unlikely to be explained by economic differences alone”and suggested that“changes to the transport systems in these cities are likely to be playing an under-recognised role

97、 in shaping travel behaviour”(Delbosc et al.,2019).In the Discussion Paper prepared for this working group,Chatterjee identified ten different travel transitions which have occurred in the last 20 years in Great Britain(Chatterjee,2020).Most of these transitions are connected to the broader trend br

98、eak in car travel which has seen per capita car use decrease since 2002.Nearly all of the travel transitions can be traced back to the 1990s,or earlier,and therefore have been in existence for 20 years or more.However,it has taken a long time for the transitions to be recognised as the subject of si

99、gnificant studies,and when studies have been carried out,they have not been able to fully ascertain why transitions have occurred.Realisation that existing practices are inadequate for future transport planning There have been large discrepancies between forecasts and actual outcomes of travel deman

100、d in industrialised countries in the first two decades of the 21st century.For example,it was forecast in 2000 INTRODUCTION TRAVEL TRANSITIONS OECD/ITF 2021 13 that traffic would grow by 22%by 2010 in Great Britain,but it only grew by 8%up to 2007(immediately before the great recession)and was only

101、5%higher in 2010 than in 2000(DfT,2019a).The DfT has explained its tendency to overestimate future traffic growth on roads as“substantially attributable to over-forecasts in key inputs to the model rather than modelling error”(Marsden et al.,2018:15).Up to 2015,the national forecasts for Great Brita

102、in included a central,most likely forecast along with low and high forecasts.In the 2015 and 2018 forecasts,the approach changed,with forecasts given for a set of equally plausible scenarios,without any one of them identified as most likely.The 2018 national forecasts for Great Britain have a range

103、of predictions of future traffic growth(between 17%and 51%for total traffic by 2050)for seven different scenarios(DfT,2018a).Despite all seven scenarios being described as plausible,one scenario is identified as a reference scenario(entailing a 35%growth in road traffic to 2050).Lyons and Marsden(20

104、19)provide a critique of DfTs treatment of uncertainty in national forecasting and are positive regarding the increasing recognition of uncertainty,especially the abolishment of a central forecast with sensitivity tests around it.They suggest this has“opened out”the degree of consideration of uncert

105、ainty in forecasts.They note,however,that this is then“closed down”when it comes to using forecasts for scheme appraisals and policy decisions where“DfT guidance requires the modelling of a core scenario that is based on central projection data from the National Trip End Model(NTEM)that forms part o

106、f the National Transport Model(NTM)”.Sensitivity testing is carried out around this with high and low growth scenarios,but it is noted that“the choice of which scenario is used to bound assessment of uncertainty,among a set of(equally)plausible options,matters hugely to what levels of demand are con

107、sidered in closing down and which scenarios are included or excluded as a result of that”.Even where aggregate forecasts of demand have been reasonably accurate,there have been notable divergences for underlying components of travel demand.Furthermore,some unexpected transitions in travel patterns h

108、ave only been noticed some years after they occurred.At the same time,bold assertions are often made of an imminent transformation in how people will travel usually by stakeholders with a vested interest and this can have a disproportionate impact on transport policy decisions.The mismatch between a

109、ggregate forecasts at the national scale,even when they are accurate,and the need to account for disruptive and outlying changes in travel behaviour at the local and regional scale causes tensions going forward.At the national level,there has generally been a focus on projecting future travel demand

110、 in order to adapt to it(e.g.by providing sufficient infrastructure).Local and regional authorities,on the other hand,have generally shown more appetite to positively influence what lies ahead,rather than just waiting for and adapting to change.There is a further mismatch between what can be gathere

111、d from aggregate forecasts and insights into outlying behaviours among segments of the population that may spread to other sectors and indicate the emergence of a“new normal”.Against all of this background,the emerging short-,mid-and long-term impacts of the Covid-19 pandemic further add to the chal

112、lenge of forward planning and policy in transport.Given doubts about the capabilities for accurate quantitative forecasting of travel demand,there is growing interest in alternative approaches that explicitly grapple with both uncertainty and disaggregate but important changes in behaviour at differ

113、ent spatial scales and among certain sectors of the population.Transport planning as a“wicked problem”The digital age has dramatically changed modern life over the last 25 years(Lyons et al.,2018)and it is increasingly apparent that it is impossible to predict future developments with any confidence

114、.It has been suggested that this is a period of deep uncertainty for developments that affect travel behaviour,with INTRODUCTION 14 TRAVEL TRANSITIONS OECD/ITF 2021 divergent opinions among experts and stakeholders and a lack of empirical evidence(Lyons and Davidson,2016).Lyons and Marsden(2019)have

115、 suggested that the extent of uncertainty in the transport sector means that decision makers are dealing with a“wicked problem”.Kolko(2012)states that a wicked problem is a social or cultural problem that is difficult or impossible to solve for four reasons:incomplete or contradictory knowledge numb

116、er of people and opinions involved large economic burden interconnected nature of these problems with other problems.The uncertainty over developments in transport and future demand for travel,the number of parties with an interest,the significance of transport for modern economies and the role of t

117、ransport in tackling wider societal problems confirms that transport planning in this period of deep uncertainty can be seen as a wicked problem.Readers guide to this report What the previous discussion has indicated is the need to take stock of the approaches used to forecast travel demand and plan

118、 transport systems.The next section,“Trend breaks and travel transitions”,discusses what is known about the causes of travel transitions and new mobility behaviours and asks whether it is possible to learn enough about drivers of travel behaviour change processes to be able to anticipate future tren

119、ds.It makes suggestions for how to more quickly identify counter-trends.The following section,“Looking back to look ahead”,provides a critical assessment of current capabilities in transport planning and considers data,theory,models and assessment frameworks and their varied applications.It covers l

120、ong-established approaches which continue to be used today,as well as examples of novel approaches currently in practice.The section“Handling uncertainty in assessing travel transitions”considers the challenge of how to deal with uncertainty,especially deep uncertainty,in strategic transport plannin

121、g.It considers how policy makers and other transport sector stakeholders can,and are,making sense of this feeling of deep uncertainty and responding to it.It introduces ways to embrace this uncertainty such that planning and investment decisions can be taken that are more resilient in the face of it

122、.The final section,“Governance challenges and opportunities”,considers challenges and opportunities to govern a transformation towards a regime of transport planning with a higher capacity to face transformation and uncertainty.TREND BREAKS AND TRAVEL TRANSITIONS TRAVEL TRANSITIONS OECD/ITF 2021 15

123、Trend breaks and travel transitions The unexpected discontinuity of past travel trends and the emergence of new urban mobility behaviours prompt reflection on how transport planning is undertaken.It is important to consider why these trend breaks were not foreseen.This requires consideration of the

124、processes which can lead to counter-trends and transitions in travel behaviour and the contributions made to these processes by different drivers of change.This section looks at the role of different types of drivers of change in travel behaviour,what is known about their influence and what gaps in

125、knowledge exist.Recent studies have provided welcome illumination on travel transitions,but have usually needed to acknowledge they have not been able to fully establish why transitions have occurred.The experience from the Covid-19 pandemic is used to consider how effectively it has been possible t

126、o monitor and understand the rapid changes in travel behaviour that have occurred during the pandemic and the lessons from this for future travel monitoring and detection of trend breaks.The final part of the section looks at promising avenues for monitoring travel and detecting trend breaks and dis

127、cusses how these might be used to support transport authorities in their anticipatory work.It is suggested that pro-active monitoring of developments external to transport and within transport can enable quicker identification of counter-trends and support adaptive responses to emerging phenomena.Su

128、itable data and research methods and a long-term,process-based perspective are needed to examine hypotheses for how changes to the socio-economic system and to transport and telecommunications are influencing travel.Population-representative data sets will continue to be crucial for assessing the pr

129、evalence of travel transitions and new mobility behaviours in the general population,but focused studies of outlier groups and areas will be invaluable for examining the plausibility of different future trends for wider society.Travel behaviour,habits and change How can changes in travel patterns ov

130、er time be explained?Travel patterns in a neighbourhood,city,region or country are comprised of thousands or millions of individual decisions,a large proportion of which at any time are habitual decisions.Over time,individuals change their travel behaviour,sometimes because their travel needs change

131、(e.g.when their workplace changes),sometimes because their transport options change(e.g.when public transport services are modified)and sometimes for reasons that are hard to grasp(e.g.when wanting to do something different).As time passes,the population itself changes as people move into or out of

132、an area.Aggregate change in travel patterns is the net effect of all these underlying changes.If there is stability at the aggregate level it is because underlying changes are balancing each other out.For example,observations one year apart for a representative sample of 19 545 English households sh

133、owed the same percentage of households increased the number of cars owned(9%)as decreased the number of cars owned(9%)(Clark,Chatterjee and Melia,2016).Aggregate change(either a continuation of a past trend or a reversal)arises when there is an imbalance in the underlying changes.This has been refer

134、red to TREND BREAKS AND TRAVEL TRANSITIONS 16 TRAVEL TRANSITIONS OECD/ITF 2021 as“asymmetric churn”because the change in one direction is different in size to the change in the opposite direction.For example,a panel study undertaken in the United Kingdom during the Covid-19 pandemic showed 15%of the

135、 panel increased the frequency they drove a car between June/July 2020(when restrictions had been eased after a first national lockdown)and November/December 2020(when a second lockdown was in place),while 22%decreased the frequency of car driving and 63%reported unchanged car driving frequency betw

136、een these periods(Marshall,Bizgan and Gottfried,2021).While the second lockdown appeared to have had contrasting effects on different members of the population,there was a net decrease in car driving.Individual-based theories Various theories have been used to understand travel behaviour,some focusi

137、ng on the behaviours of individuals and others on the behaviours prevalent in society more generally.Perhaps the most commonly employed basis for understanding travel behaviour is rational choice theory,which assumes travel entails disutility(notably time and cost)and that travellers seek to minimis

138、e disutility to reach destinations.Rational choice theory helps illustrate how supply-side factors related to transport provision(such as travel times and costs)influence travel choices.Figure Figure 2.Hierarchical structure of decisions2.Hierarchical structure of decisions Source:Van Acker,Mokhtari

139、an and Witlox(2011).More advanced theories focus on the context for people making journeys.The activity-based approach to travel behaviour considers how people organise their travel in the context of the activities they wish to pursue and constraints of time and place(McNally and Rindt,2007).This ap

140、proach helps illustrate how household needs and organisational factors influence travel choices.Socio-psychological theories move away from the assumption that people are rational decision makers(seeking to minimise travel disutility)and consider how subjective factors,such as attitudes and social n

141、orms,influence travel behaviour.They TREND BREAKS AND TRAVEL TRANSITIONS TRAVEL TRANSITIONS OECD/ITF 2021 17 help describe how people perceive the options available to them and are motivated to perform particular behaviours.Combining the above theoretical perspectives,van Acker,Mokhtarian and Witlox

142、(2011)proposed a conceptual framework(see Figure 2)which suggests travel behaviour decisions are part of an extended choice hierarchy with lifestyle choices at the top level of the hierarchy,representing the long-term view of“what life should be like”and manifested in decisions on family,employment

143、and leisure.Medium-term decisions,such as residential location and car ownership and short-term decisions such as choice of destination and mode of transport are made in line with lifestyle choices and attitudes.The framework also recognises that lifestyle preferences and travel decisions are influe

144、nced by the wider social-economic and demographic context.Turning to theories of behavioural change,habit theory hypothesises that behaviour when first initiated is the product of rational decision making,but becomes automatic when repeated in a stable context.In particular,the habit-discontinuity h

145、ypothesis posits that habits may become weakened when they are interrupted by a contextual change(Verplanken et al.,2008).This has led to interest about the nature of events which bring about a reconsideration of behaviour.Events can be at the micro-level(relating to the life of an individual and th

146、eir immediate social network),or at the macro-level(relating to the wider social system,including the transport system)(Chatterjee and Scheiner,2015).The life course perspective is a multidisciplinary paradigm for the study of peoples lives,structural contexts and social change.It is helpful in cons

147、idering why travel behaviour changes during the course of peoples lives.There are four primary analytic themes of the life course perspective(Elder,1998):1.Historical time and place the life course of individuals is embedded in and shaped by the times and places they experience over their lifetime.T

148、his signals the importance of cohort effects,where distinctive formative experiences are shared at the same point in the life course by birth cohorts.2.Timing of lives the impact of life transitions or events is contingent on when they occur in a persons life.3.Linked lives lives are lived interdepe

149、ndently,and social and historical influences are expressed through this network of shared relationships.The family has been the prime focus of life course research in this respect,but social relationships can be considered in a wider sense.4.Human agency individuals make their own decisions and cons

150、truct their own life course through the choices and actions they take,within the opportunities and constraints of their history and social circumstances.The life course perspective has been applied to travel behaviour through the concept of mobility biographies.Lanzendorf(2003)proposed that mobility

151、 biographies consist of lifestyle,accessibility and mobility domains,and that these three domains are interlinked,with events in one domain affecting the others.The emphasis in mobility biographies research has been to study how events in the course of life influence a change in travel habits.Studie

152、s have highlighted the importance of changes in household composition,driving licence availability,residential and workplace location(Chatterjee and Scheiner,2015).While researchers have given much attention to behavioural change induced by disruptive events(whether micro or macro events),there are

153、theories that suggest that change is a more gradual process and occurs in stages.The transtheoretical model of change(Prochaska and Di Clemente,1983)assumes that individuals progress through stages of change over time,ranging from not contemplating change through to contemplating and preparing for c

154、hange,enacting change and maintaining changes.At each TREND BREAKS AND TRAVEL TRANSITIONS 18 TRAVEL TRANSITIONS OECD/ITF 2021 stage,there is a“decisional balance”of pros and cons for change.Learning theories,such as social learning theory(Bandura,1977),are helpful in highlighting that behavioural ch

155、ange occurs through a combination of trial and error and in observing what others do.It is suggested that“learning is more likely to happen when there is a change in the situational context(or behavioural goal),when deliberation is prompted by information or when the situation is uncertain”(Sunitiyo

156、so,Avineri and Chatterjee,2013:259).Systems theories The theories highlighted above emphasise individual agency in travel decision making,but there are sociological theories that argue that individual behaviour is determined by societal structures.Instead of placing attention on individual behaviour

157、,social practice theory focuses on collective practices and looks at the elements that are needed to maintain these practices(Shove,Pantzar and Watson,2012).Elements are categorised under the headings of materials,competencies and meanings and can be interpreted for transport as access(physical acce

158、ss to transport services),ability(know how to use a transport mode)and ambition(willingness to use a mode of transport)(Millionig,2021).While efforts are being made to reduce physical and financial limitations restricting access to car alternatives,there are still few solutions for overcoming compet

159、ence deficits and emotional barriers.Although access is an essential prerequisite for behaviour change,simply ensuring access is,of itself,insufficient to drive behavioural change.More comprehensive efforts are required to address barriers related to ability and ambition in order to achieve greater

160、levels of behavioural change.This explains,from a systems perspective,why travel behaviour change is a slow and gradual process at the aggregate level,even if change is faster and more substantial among certain individuals and cohorts.One application of social practice investigated the elements that

161、 help to maintain car-based commuting practices in the United Kingdom and what might destabilise these practices(Cass and Faulconbridge,2016).It concluded that structural barriers to bus and cycling use need to be addressed,(such as the frequency of bus services,the availability of cycling equipment

162、,knowing how to navigate bus timetables or ride a bike safely and appreciation of exercise gained through cycling or relaxation when using the bus),as well as the re-organisation of linked social practices(such as working hours)that hinder bus-based and cycling-based commuting practices.Everett Roge

163、rs proposed the diffusion of innovations theory to explain how new ideas and technologies spread in a population and is helpful in considering how large-scale system change occurs.The theory states that the diffusion of an innovation(or a new practice)depends on its relative advantage over previous

164、practices,its compatibility with the needs,habits or values of those who will potentially adopt the practice,its complexity(or ease of use)and its potential for trialling(Rogers,2003).Some people are more likely to adopt a new practice than others with five groups identified:innovators,early adopter

165、s,early majority,late majority and laggards.While diffusion of innovations theory has often successfully been applied to profile users of new transport technologies,it is less well suited to understand how broader travel behaviours will evolve over time.A more expansive theory of system change is th

166、e multi-level perspective(MLP),which adopts a socio-technical approach to the study of transitions and assumes“that transitions are non-linear processes that result from the interplay of multiple developments at three analytical levels:niches(the locus for radical innovations),socio-technical regime

167、s(the locus of established practices and associated rules),and an exogenous socio-technical landscape”(Geels and Kemp,2012:53).Niche actors work on radical innovations that deviate from existing regimes and hope that their promising novelties are eventually used in the regime or even replace it.Howe

168、ver,the existing regime is stabilised by many lock-in mechanisms.For example,a car-based transport system is stabilised by“sunk investments(in road infrastructures,TREND BREAKS AND TRAVEL TRANSITIONS TRAVEL TRANSITIONS OECD/ITF 2021 19 plants,skills),user patterns and lifestyles oriented around the

169、car,favourable regulations,cultural values(such as speed,freedom,individuality,identity),resistance from vested interests(industry,car lobby,road-building lobby)”(Geels and Kemp,2012:58).These lock-in mechanisms are persistent and pervasive and amount to what Mattioli et al.call a“system of provisio

170、n”for car-based transport that generates significant friction to change by niche or other actors(Mattioli et al.,2021).When it comes to considering timescales of change,Tilley(2017)has proposed a dynamic framework for understanding the multi-level forces stimulating changes in travel behaviours.This

171、 framework identifies three types of multi-level forces that influence change in travel behaviour over time:period effects shorter-term effects that apply to whole populations,such as macroeconomic processes of growth and recession mid-structural effects structural changes operating at a moderate ra

172、te of change,such as post-war planning and the resulting processes of suburbanisation and counter-urbanisation deep structure effects cultural changes occurring at an almost imperceptible rate of change and which contribute to the development of socially constructed norms regarding mobility,which in

173、 turn influence travel patterns.This is a helpful basis from which to think about the rate of change of travel patterns and whether a transition might represent a swift adjustment to a time-limited event or a long-term,gradual evolution of behaviour in response to structural effects.Empirical studie

174、s of travel trends Having introduced the above theories,it is now shown how empirical analysis can apply these theories to help analyse and interpret travel trends.When looking at longer-term change in travel trends,the most easily obtained data is time-series data of aggregate travel,such as annual

175、 observations of vehicle-kilometres travelled.Econometric analysis is often conducted with such data to explore how socio-economic conditions influence travel trends over time with the underlying assumption of rationality,such that travel increases with more income and decreases with higher transpor

176、t costs.A good example of an econometric analysis of time-series data is that of Bastian,Brjesson and Eliasson(2016),who estimated multiple regression models of log vehicle-kilometres travelled per capita against log gross domestic product(GDP)per capita and log gasoline price for six countries(Aust

177、ralia,France,Germany,Sweden,the United Kingdom and the United States)based on 1980-2014 time-series data.They found the models explained the observed trends very well through the full period.They also found that GDP per capita elasticities have decreased over time,whereas gasoline price elasticities

178、 have increased,which indicates saturation of car ownership and use among higher-income groups and increased sensitivity to fuel prices when they are at high levels.The authors“conclude that economic variables are sufficient to explain the aggregate trends in car use”but“do not rule out the existenc

179、e of alternative explanations”.While the regression models estimated by Bastian,Brjesson and Eliasson accurately reproduced the time series of observations from which they were estimated,this does not guarantee their accuracy for future predictions.Even if trend breaks have been driven by macroecono

180、mic factors,the adaptations that arise(e.g.investment in car alternatives or take-up of alternative lifestyles)may exert longer-term influence,as implied by the decreasing GDP elasticities.It is,therefore,still important to look at other influences alongside economic factors and to study different s

181、ocio-economic groups.While analysis of this kind can assess the extent to which population-wide aggregate travel trends are explained by socio-economic TREND BREAKS AND TRAVEL TRANSITIONS 20 TRAVEL TRANSITIONS OECD/ITF 2021 variables,it is not able to explain why differing trends might be taking pla

182、ce within the population.For example,the analysis could not explain the significant reduction in car travel by young adults.Where disaggregate,longitudinal data is available from repeated cross-sectional surveys(such as annually conducted national travel surveys),more sophisticated analysis can be p

183、erformed and a greater understanding gained of reasons for change in travel behaviour.In 2013,IFMO reported on the use of a trend decomposition technique to analyse national travel survey data for France,Germany,Great Britain,Japan,Norway and the United States.This data was used to look at changes i

184、n travel behaviour by age group and assess the impact of population aging on the levelling off or reduction in car distance travelled observed in these countries.For example,in Great Britain between 1996 and 2005 increased car ownership and use among older people(aged 60 or above)contributed to grea

185、ter overall distance travelled by car,however,decreased total travel and car mode share of young people(aged 20-39)counteracted this.However,this study could not explain why total travel and car mode share had reduced for young people.Cohort analysis is a powerful approach to understanding trends.Be

186、fore analysis,a data set is broken down into related groups,where these groups or cohorts share common characteristics or experiences within a defined time span.Cohort analysis is often applied to distinguish between three types of time-related variation:age effects variations associated with age th

187、at remain more or less stable over time period effects variations over time that affect everyone simultaneously,irrespective of their age cohort effects changes across groups of individuals who experience an initial event together,such as their birth year.When seeking to understand changes over time

188、,cohort analysis is potentially highly illuminating at identifying whether changes can be attributed to explanatory factors,such as socio-demographics or transport provision,or whether unexplained change applies to the whole population or particular cohorts.McDonald(2015)used United States National

189、Household Travel Survey data for 1995,2001 and 2009 to compare daily car mileage of the Generation X cohort(those born in the late 1960s to the late 1970s)and the millennial cohort(those born in the last two decades of the 20th century).McDonald noted that car trips and mileage decreased between 199

190、5 and 2001 for 19-30 year-olds,suggesting“a long-term decrease in automobility that started in the late 1990s with younger members of Gen X and has continued with the millennial generation”.McDonald went on to note that there was only a very modest increase in the use of public transport,walking and

191、 cycling over the period in which car use decreased.McDonald(ibid.)used multiple regression modelling to quantify socio-demographic,age-specific,period-specific and cohort-specific effects on car trips and mileage of 19-42 year-olds.A decomposition of the different contributors to changes in car mil

192、eage showed that lifestyle-related socio-demographic changes accounted for 10-25%of the reduction in car mileage from 1995 to 2009.Changes over time specific to millennials and younger members of Generation X accounted for 35-50%of the reduction and general dampening of car mileage travel that appli

193、ed across all age groups accounted for the remaining 40%reduction.McDonald interprets the 35-50%reduction specific to younger Generation X and millennials as“Millennial-specific factors such as changing attitudes and use of virtual mobility(online shopping,social media)”.It is not possible to conclu

194、de further from this study to what extent“millennial-specific factors”comprise structural factors untested in the analysis or changes to values,attitudes or social practices.Panel data which tracks the same individuals over time can be used to analyse the dynamics of behaviour and understand process

195、es of change.For example,analysis of panel data from the UK Household TREND BREAKS AND TRAVEL TRANSITIONS TRAVEL TRANSITIONS OECD/ITF 2021 21 Longitudinal Study has revealed the circumstances in which households are more likely to change car ownership(Clark,Chatterjee and Melia,2016).Events associat

196、ed with a transition to adulthood(e.g.acquiring a driving licence,entering employment,partnership formation or having a child)increase the likelihood of becoming a car-owning household.If these events are postponed or foregone altogether,then reduced car ownership is to be expected.Additionally,the

197、panel data showed that the volatility of young peoples living and socio-economic circumstances leads to instability in car ownership and this is likely to have increased in recent years given more unstable employment.It is clear that to understand travel behaviour change it is necessary to gather lo

198、ngitudinal data.Time-series data at an aggregate level will support analysis of overall dynamics,while periodically collected observations of individual behaviour will enable understanding of the underlying components of change.Meanwhile,panel data can shed light on the process of behavioural change

199、 experienced by individuals,which can help with the interpretation of aggregate trends.Drivers of change in travel behaviour Travel can be predictable at an aggregate level when the drivers of travel demand are stable but rarely,if ever,is this the case.Drivers of change can apply to the whole popul

200、ation or certain groups within the population,but either way can bring about fundamental change.Taking a high-level systems perspective that draws upon the theories previously introduced,travel patterns can be considered to be a function of three domains:the wider socio-economic system;peoples activ

201、ity or travel preferences;and the transport and telecommunications options available(see Figure 3).Changes in each of these three domains,and in the relationships between them,will affect travel patterns.This makes it difficult to anticipate future travel patterns.Typically,forecasts of future trave

202、l patterns are based on simple assumptions about future changes to the socio-economic system and the transport system with relationships assumed to remain constant over time.Figure 3.Conceptual framework for influences on travel patternsFigure 3.Conceptual framework for influences on travel patterns

203、 Source:Chatterjee(2020).TREND BREAKS AND TRAVEL TRANSITIONS 22 TRAVEL TRANSITIONS OECD/ITF 2021 The conceptual framework above highlights that to understand what is causing changing travel patterns,it is necessary to look beyond the transport system and changing population characteristics.Considera

204、tion needs to be given to broad changes in the socio-economic system and how they influence the preferences of the population and the provision of transport and telecommunications.It is also necessary to regularly challenge the understanding of how activity and travel preferences are resolved subjec

205、t to the available transport and telecommunications systems.This section will look at the role of demand-side drivers of change(such as economic change and cultural shifts),supply-side drivers of change(such as new technologies within transport and telecommunications)and,finally,new behaviours and p

206、ractices.These drivers of change will be evaluated based on experiences in the ITF member countries that participated in this working group.Social and economic change The socio-economic system is complex in nature,but traditionally only certain key characteristics are considered when predicting futu

207、re travel demand.This raises questions:How fully do the socio-economic factors considered in transport models explain travel behaviours and how travel behaviours change over time?How important are the socio-economic factors that are not considered in transport models?In the Discussion Paper for this

208、 working group(Chatterjee,2020),these questions were considered with respect to national road traffic forecasting in Great Britain.The UK Department for Transport(DfT)evaluated the performance of its road traffic forecasts carried out between 2009 and 2015,with a view to understanding why its foreca

209、sts had overestimated traffic growth(DfT,2018a:18-25).It wanted to know to what extent differences between forecasts and outcomes between 2010 and 2017 could be explained by the changing relationships between travel and its key drivers(including the emergence of new drivers)or by input over-and unde

210、r-forecasting for key drivers such as GDP,population and fuel costs.The forecasting methodology was found to provide aggregate results close to outcomes when adjustments were made for actual values for GDP,population,etc.This is similar to the conclusion of Bastian,Brjesson and Eliasson(2016)who,as

211、mentioned earlier in this section,found that national trends in vehicle-kilometres between 1980 and 2014 in six highly industrialised countries were well explained by changes in GDP and gasoline prices.The DfT did note,however,that forecasts overestimated traffic growth in London and underestimated

212、traffic growth on longer distance inter-urban roads.It concluded that the latest version of its model is fit-for-purpose at an aggregate level,but“has difficulties replicating travel patterns at local levels where travel behaviour is substantially different from the national picture”(DfT,2018a:25).A

213、s well as doubts as to whether national-scale forecasting can be informative for anticipating travel trends at sub-national scales,the experience from the last 20 years in Great Britain suggests there are also doubts about how well forecasts are able to anticipate travel trends for different socio-d

214、emographic groups.An in-depth study that sought to understand the large reduction in driver licence rates and car travel by young people in Great Britain since the 1990s,concluded that changes in demographics,socio-economics and living circumstances only provide a partial explanation,with changes in

215、 travel attitudes and substitution of travel by online communication also likely to be important factors(Chatterjee et al.,2018).The DfT has been undertaking work to improve its forecasting system to account for the latest evidence on travel behaviour and its determinants.Driving licence rates,car o

216、wnership,trip rates and usage of different transport modes are predicted in the National Transport Model based on exogenous demographic and socio-economic projections for population,employment,housing supply,income and transport costs.For the latest 2018 forecasts,the National Transport Model was up

217、dated to account for TREND BREAKS AND TRAVEL TRANSITIONS TRAVEL TRANSITIONS OECD/ITF 2021 23 recent travel trends.One of the most significant developments was updating the trip rate models based on a larger number of socio-demographic variables found to influence trip rates(AECOM/Imperial,2017).Howe

218、ver,it was found that these variables could not provide much explanation for observed decreases in trip rates in the last two decades.This further highlights that socio-economic factors traditionally included in transport models have not been especially useful in explaining breaks in past travel tre

219、nds in the last 20 years.This is further supported by an analysis by the DfT of National Travel Survey data for England for 1995-2012,which looks at demographic and socio-economic determinants of having a driving licence,car access and car mileage over time(DfT,2018b).This analysis found age,gender,

220、household composition,employment status,job type,education,personal income,type of residential area and access to public transport to all be influential factors.However,it found that the positive association between personal income and licence holding,and car mileage weakened over the period.It also

221、 found that household composition made less difference over time,while employment status became more important.After accounting for demographic and socio-economic factors,it found increases in driving licence holding across the population over the period(linked to people with driving licences growin

222、g older and eventually replacing the generation above)and decreases in car use over the period.However,it also found more recent cohorts had a lower probability of licence holding and car access than previous cohorts(after accounting for other factors)and had lower car mileage.These results show tha

223、t there are travel trends over time,which vary between generations,which are unexplained by the demographic or socio-economic characteristics of the population.Figure Figure 4.4.Number oNumber of motor vehicles per 1f motor vehicles per 1 000000 inhabitants vinhabitants versuersus GDP per capita,s G

224、DP per capita,20142014 Source:Our World Data(2014).The observations above suggest that it may be possible to make reasonable forecasts of short-term aggregate(national-level)travel trends based on demographic and economic projections,where these projections are reliable.However,these are unlikely to

225、 be informative for sub-national travel trends or for the travel trends of specific socio-demographic groups which require their own,bespoke analysis,and are TREND BREAKS AND TRAVEL TRANSITIONS 24 TRAVEL TRANSITIONS OECD/ITF 2021 likely to be subject to influence by social and economic change which

226、is not easily captured in traditional forecasting methodologies.In highly motorised countries,the widely observed phenomenon that the relationship between income and car use has weakened lends support to the argument that saturation has occurred of car ownership and use among higher-income groups.St

227、udies show that employment status and job type remain important discriminators of car use,which suggests that attention needs to be given to changes to labour-force composition when looking at future travel demand.The locations where people live have been shown to be of increasing significance,with

228、increasingly low levels of car use in large cities,but continued high car use in less populated areas,which suggests it is vital to pay close attention to the future spatial distribution of the population.Furthermore,it is clearly necessary to pay attention to attitudinal and lifestyle factors that

229、are reshaping travel behaviours among different groups in society.When looking across a range of countries with different levels of industrialisation it can be seen that there is a positive association between road vehicle ownership and GDP per capita(see Figure 4).This would suggest that economic g

230、rowth measured in this way will continue to play an important role for travel trends in industrialising countries.However,it should be recognised that there are large variations in vehicle ownership rates between countries with high GDP per capita with,for example,Australia,Italy and the United Stat

231、es having much higher rates than Denmark,Korea and Saudi Arabia.Technological change It was clear,even before the Covid-19 pandemic,that the digital age has been having fundamental effects on peoples lifestyles and everyday lives.There has been much speculation on whether information and communicati

232、on technology(ICT)is substituting,stimulating,supplementing or redistributing travel(Lyons,2015).Research has been inconclusive,however,and it is suggested that instead of directly asking this question there is a need to recognise that changes are taking place gradually to our lifestyles and there i

233、s a need to focus attention on“how mobile ICTs are transforming many aspects of our daily lives and especially how they are helping to reshape the temporal and spatial organization of everyday activities”(Aguilra,Guillot and Rallet,2012:667).Studies that have investigated the relationship between IC

234、T use and travel have generally found that those who use ICTs more also travel more(e.g.Kroesen and Handy,2015),leading to the conclusion that the digitisation of society is not a contributor to reduced travel.However,it is questionable how useful cross-sectional data used in such studies is for hig

235、hlighting the effect of digitisation over time,as it cannot control for other factors(such as socio-economic status)which explain both higher levels of ICT use and travel.Longitudinal data tracking of individuals ICT use and travel over time would be more helpful in assessing this relationship.The p

236、otential impacts of ICTs in reducing the number of trips for different activities are considered below.Teleworking Taking the example of the United Kingdom,working from home was becoming more common prior to the Covid-19 pandemic,both occasionally and on a regular basis,and there had been a decline

237、in the number of commute trips made by workers in England from 7.1 journeys per worker per week in 1988-92 to 5.7 in 2013/14,with a similar trend observed in the United States(DfT,2017).It has also been shown that over time a greater proportion of the English working population has either spatially

238、variable working patterns or are infrequent commuters(Crawford,2020).Analysis of the Swiss Mobility and Transport Microcensus suggests teleworking increases travel due to teleworkers living further away from the workplace than other workers and replacing commuting trips TREND BREAKS AND TRAVEL TRANS

239、ITIONS TRAVEL TRANSITIONS OECD/ITF 2021 25 with journeys for other purposes(Ravalet and Rrat,2019).However,an analysis of data from the Netherlands for 2000-16 concludes that different types of flexible working,including teleworking,have contributed to a 2%reduction in car-kilometres on working days

240、 and a 7%reduction in traffic during peak hours on all roads(van der Loop,Haaijer and Willigers,2019).In the future,job profiles are expected to change fundamentally in many sectors.The main drivers of a new world of work and mobility are advancing digitalisation and connectivity,artificial intellig

241、ence and autonomous driving.Flexible working hours,desk sharing and teleworking will,in all likelihood,play a greater role in this future working world than they do today(Wipperman,2018).During the Covid-19 pandemic,the acceptance of teleworking among employees and employers has increased and much e

242、xperience has been gained.What is uncertain now is whether sustained increases in teleworking will lead to corresponding decreases in traffic volumes,moves away from urban areas to locations where greater travel distances are required to meet daily needs,and whether workers will compensate for decre

243、ased time spent commuting with more leisure travel.Distance learning Distance learning refers to the replacement or partial replacement of school attendance by the provision of digital learning content.While this form of education had been quite rare before the Covid-19 pandemic,it dramatically incr

244、eased during 2020.However,the drawbacks of distance learning for young people also became quickly apparent(see,for example,OECD,2020).These include more limited access to learning resources for socio-economically disadvantaged students,additional care obligations for working parents,potential isolat

245、ion of students,insufficient development or decline in social skills and less physical activity for home learners.Thus,even if distance learning concepts are maintained to some extent in the future,it is unlikely that this will contribute to a significant reduction in traffic volume.E-commerce Prior

246、 to the pandemic,online shopping was growing at around 10-12%per year in the United Kingdom and represented almost 17%of total retail sales(Marsden et al.,2018).At the same time,there was a long-term trend of decreasing personal trips and distance travelled for shopping.In 2019,4.4 million Austrians

247、 used online shopping,which is approximately half of the total population,according to the Austrian Trade Association(2019).Around 10%of consumer spending flows into internet retail.However,domestic retailers only benefit from the growing market to a limited extent,as more than half of people order

248、items from abroad,thus exacerbating the outflow of purchasing power and potentially increasing global transport.In the second-quarter of 2020,at the height of official Covid-19 movement restrictions in Central Europe,turnover in German mail order and online trade increased by 29%compared to the same

249、 period the previous year,and turnover in the e-food sector rose by 90%(KPMG,2020).With regard to transport impacts,several studies conclude that no significant carbon emissions savings can be expected once it is understood that reduced personal car travel is offset by increased goods vehicle moveme

250、nts,missed deliveries and collections of returned items(Lengauer et al.,2015).There is,however,some transport saving potential regarding online shopping when orders do not result in the physical delivery of goods.Examples include 3D printing(although material for printing has to be delivered)and dow

251、nloads of e-books,music,movies,games or software.New transport options There are said to be three transport revolutions occurring concurrently:electrification of the vehicle fleet,automation of driving and adoption of shared mobility(Marsden et al.,2018).The autonomous vehicle has been heralded as a

252、 transformative technology which will drive significant social change.By reframing TREND BREAKS AND TRAVEL TRANSITIONS 26 TRAVEL TRANSITIONS OECD/ITF 2021 the terms of accessibility through changes in attitudes to the time people are willing to spend on travel,speed of travel and transportation cost

253、s the autonomous vehicle is likely to change lifestyles in many ways,with important consequences for various aspects of daily and long-distance mobility practices(especially frequency,average distances and modal choice).Anticipating these changes and preparing for them(Bali,Capano and Ramesh,2019)is

254、 of critical importance in regulating the deployment of autonomous vehicles in order to limit the negative environmental and social impacts and to promote the benefits of this kind of transport(Harb et al.,2021;Narayanan,Chaniotakis and Antoniou,2020).The medium-and long-term decisions concern innov

255、ation policies,but also planning policies,such as infrastructure investments and land-use policies.Anticipating these changes is nonetheless a considerable scientific challenge since,on the one hand,the technology is not mature and,on the other hand,lifestyles are the result of complex,interdependen

256、t decisions involving many factors and different temporalities.Shared mobility services offer a range of alternatives to accessing mobility through personal ownership of vehicles or using public transport and,currently,are perhaps the most notable development which could influence future travel beha

257、viour.Figure 5 shows different shared mobility services which could meet the needs for travel over different distances.Figure 5.Shared mobility service options by trip length Figure 5.Shared mobility service options by trip length Source:NHTS(2018).In France,80 million trips were made by shared mobi

258、lity services in 2020 and station-based bike services accounted for 73%of these trips(Fluctuo,2021).Turkish carsharing service provider,MOOV by Garenta,announced that it has had 2 million rentals,with more than 100 000 active users within two years and that 69%of the trips made during the Covid-19 p

259、andemic were for commuting,implying they were being used as a substitute for public transport(Dunya,2020).Looking at the United Kingdom,it has been noted there have been increasing numbers of shared mobility users since 1998 when services emerged,but users remain concentrated in London and among you

260、nger,higher-income residents.The number of carshare members has increased from 32 000 in 2007 to nearly 250 000 in 2017 in the United Kingdom,while the equivalent growth in Germany has been from 100 000 TREND BREAKS AND TRAVEL TRANSITIONS TRAVEL TRANSITIONS OECD/ITF 2021 27 to 1.7 million(Marsden et

261、 al.,2018,p30).However,it is generally not possible to quantify the share of the travel market carved out by these services since use of shared mobility services is not distinguished from personal vehicles in most travel surveys.A study on carsharing in the Netherlands estimated that 1%of the popula

262、tion has used carsharing and that carsharing accounts for 0.02%of the total car journeys in the Netherlands(KiM,2015).New behaviours and practices The different societal and technological trends described previously have the potential to give rise to new mobility behaviours.The European Transport an

263、d Mobility Forum identified,among others,the following trends that are expected to create new patterns of behaviour(Mobility4EU,2016):Digitalisation and personalisation the wealth of data collected about mobility service customers is giving rise to increasingly customised products and services.This

264、will lead to more individualised and flexible behaviour patterns.Climate change and resource efficiency in view of the impending consequences of climate change,measures to reduce CO2 emissions will become increasingly important,which will also lead to stricter regulations and stronger pricing mechan

265、isms.More sustainable behaviours need to be enforced,which can be driven,to some extent,through the use of incentives.Sharing instead of owning new concepts of resource efficiency also lead to new ownership models,such as the sharing economy and collaborative consumption,which allow for a more flexi

266、ble use of different mobility offers.This is further supported by digital services.The trends described above are already evident in the behaviour patterns of younger,urban travellers.Many of these behaviour patterns are closely related to the use of digital information in the context of mobility.Wi

267、thin the framework of a study conducted in Austria,for example,six different mobility types were identified,which differ in their behaviour patterns and information habits(Markvica et al.,2020).Three of these mobility types are particularly prevalent in the younger age groups.One such type is called

268、“highly-informed sustainability”(17%).People belonging to this group proactively search for or receive transport-or mobility-related information.They deal with current issues and topics in detail and are characterised by their effort to organise their own daily routine in a sustainable and environme

269、ntally friendly way with a fundamental interest in new things,such as new developments and innovations in the field of mobility.The type“Spontaneous on the go”(6%)is not yet common,but will become more significant in the future.This group is characterised by the speed and spontaneity with which they

270、 expect,absorb and process information.Due to a mobile,flexible and non-routine lifestyle,this group is dependent on a lot of information on a daily basis.This type is very mobile and uses many different means of transport for their travel needs.They do not stick to one means of transport and are,in

271、 principle,open to all transport options.In contrast,the“Efficiency-oriented”type(16%),is only interested in the information that it needs to fulfil a certain goal.Characteristic of this type is a commitment to one means of transport and more resistance to changes in behaviour than the two previous

272、types.Owning a car is important;they see the car as something very personal.However,they also use sharing services,often as a supplement to their own car.All three of these mobility types use digital and mobile information to organise their mobility and will benefit from increasing digitalisation an

273、d personalisation.The remaining three types,“Interested-Conservative”(35%),“Low Demand”(16%)and“Digital Illiterates”(10%)will decline over time as younger TREND BREAKS AND TRAVEL TRANSITIONS 28 TRAVEL TRANSITIONS OECD/ITF 2021 generations come of age in the population.However,these less digitally re

274、liant behavioural types will be prevalent for longer outside of cities,which is why there may be even more divergence between urban and rural mobility patterns.The impact of the Covid-19 pandemic The Covid-19 pandemic had an extraordinary impact on the way people live,work and travel.There is a weal

275、th of information available on how the pandemic has affected the transport sector(see,for example,ITFs Covid-19 insights:The Compendium ITF,2021).Of particular interest for this report are the travel behaviour changes that have taken place and how these have been measured using existing and new moni

276、toring approaches.Given that the pandemic is still placing restrictions on the activities of citizens in most parts of the world,now is not the right time to judge whether there are enduring shifts in travel patterns compared to pre-2020.However,comments can be made on what monitoring approaches wil

277、l be valuable in the future to assess ongoing trends in travel behaviour.Established approaches for collecting travel data,such as road traffic counting systems,public transport passenger counting systems and national travel surveys have generally continued during the pandemic,even if they have had

278、to adapt in response to health protection measures.These surveys are contributing to understanding the impacts of the pandemic and will play a particularly important role in assessing longer-term implications.As an example,the Union of Municipalities of Turkey collated information for 30 cities in T

279、urkey and found decreases in public transport use of 34%to 87%during the first wave of the Covid-19 pandemic in Turkey,between March and June 2020,when social restrictions were in place(UoM,2021).There is often a considerable time lag between data being collected and results being published,due to t

280、he time needed for processing and analysing the data.For example,in England,results from the National Travel Survey1 are published in July for data collected during the previous calendar year,hence results for 2020 are not currently available at the time this report is being written.Global informati

281、on technology companies have helped to provide up-to-date data on travel activity during the pandemic.Apple has published daily“Mobility Trends Reports”since 13 January 2020 on the levels of driving,walking and public transport in different countries,regions and cities,based on the number of request

282、s for navigation by transport mode(Apple,2021).Similarly,Google has published daily“Covid-19 Community Mobility Reports”since 15 February 2020 on visits to different types of destinations(e.g.retail and recreation,public transport hubs,etc.)based on anonymised data collected from Google apps on mobi

283、le devices(Google,2021).These two data sources provide an indication of changes in aggregate travel activity over time,but they do not provide information about the characteristics of those making trips and only limited categorisation of the type of trips made.Hence,it is not possible to analyse tre

284、nds for different population groups and types of travel.Given the limitations during the pandemic of obtaining sources of data to monitor travel behaviour,countries have been interested in collecting data to enable rapid reporting of travel behaviour and to provide illumination on specific issues of

285、 interest or concern during the pandemic.Table 1 provides details of three major travel behaviour studies commissioned and organised at short notice to obtain an in-depth understanding of travel behaviour impacts of the pandemic.TREND BREAKS AND TRAVEL TRANSITIONS TRAVEL TRANSITIONS OECD/ITF 2021 29

286、 Table 1.Travel behaviour studies during Table 1.Travel behaviour studies during the the CovidCovid-1919 pandemicpandemic Country/studyCountry/study Data collectedData collected Impact of pandemic on travel Impact of pandemic on travel behaviourbehaviour LongerLonger-teterm rm implications implicati

287、ons Netherlands Institute for Transport Policy Analysis(KiM).Special survey of Netherlands Mobility Panel(MPN)participants(de Haas,Hamersma and Faber,2020).2 000 panel members asked to keep a 3-day travel diary in March-April 2020,and complete a questionnaire.The questionnaire was designed to identi

288、fy the causes of any changes in travel behaviour,related perceptions and experiences,and expectations for the future.44%of workers started to work from home or worked more from home.Number of trips decreased by 55%and the total distance travelled decreased 68%.Increased modal share for walking and d

289、ecreased modal share for all other types of transport.27%of home workers expect to work from home more after the pandemic than before.20%think they will walk and cycle more and around 20%say they will fly less.Switzerland Institute for Transport Planning and Systems(IVT)at ETH Zurich and the Faculty

290、 of Business and Economics(WWZ)at the University of Basel.MOBIS-Covid19 sample recruited from participants of the 2019 MOBIS(MObility Behaviour in Switzerland)study(Molloy et al.,2021).App-based(GPS)tracking of travel behaviour of 1 439 Swiss residents supplemented with occasional online questionnai

291、res to collect personal details.Data weighted to be representative of 22 000 respondents to MOBIS introductory survey.Large reductions initially in distance travelled by all modes,except cycling which increased.Cycling distance remained 100%higher in August 2020 compared to the baseline,while walkin

292、g and use of cars returned to pre-pandemic levels and public transport was still 50%lower.The proportion of active(mobile)days decreased from around 90%pre-pandemic to 70%at start of first wave and 80%in August 20.Not directly investigated,but the tracking study continues as of May 2021.UK Departmen

293、t for Transport(DfT).New survey“All change?”(Marshall,Bizgan and Gottfried,2021).Online longitudinal survey repeated three times(May-June 2020,June-July 2020,November-December 2020)with one further wave planned for Spring 2021.4 059 adults aged 16-75 took part in wave 1.Waves 1 and 2 had 2 782 parti

294、cipants and waves 2 and 3 had 2 847 participants.Data weighted to be representative of United Kingdom.Main topics covered:frequency of travel by mode,purposes of travel undertaken,future expectations of travel.Travelled by car as a driver once a week or more often during previous four weeks:63%in Ja

295、nuary-March(pre-pandemic),47%in May-June 2020,60%in November-December 2020.Used a bus once a week or more often during previous four weeks:30%in January-March 2020(pre-pandemic),6%in May-June 2020,14%in November-December 2020.Walking and cycling at similar rates during the pandemic as pre-pandemic.F

296、requency of travel changed more during pandemic than mode switching.Outlook for longer-term behaviours to be considered with last wave of data.The value of longer-term ongoing studies is shown with these examples.The Netherlands Institute for Transport Policy Analysis(KiM)runs the Netherlands Mobili

297、ty Panel(MPN)2 which collects travel behaviour data on an annual basis for a fixed group of individuals from around 2 000 households.The special survey in March-April 2020 enabled a direct comparison to be made between travel behaviour during the first wave of the pandemic to pre-pandemic baseline t

298、ravel behaviour already reported in the main survey(de TREND BREAKS AND TRAVEL TRANSITIONS 30 TRAVEL TRANSITIONS OECD/ITF 2021 Haas,Hamersma and Faber,2020).Like most countries,the United Kingdom does not have a mobility panel,but the DfT saw the value in commissioning a longitudinal panel to specif

299、ically capture travel behaviour data during the pandemic(Marshall,Bizgan and Gottfried,2021).Because there was no equivalent study of travel behaviour prior to the pandemic,participants were instead asked to think back to the January-March 2020 period and,retrospectively,report their pre-pandemic tr

300、avel behaviour with this serving as a baseline.The MOBIS-Covid19 study in Switzerland was made possible by the 2019 MOBIS study.Although the original MOBIS study had finished collecting data via the tracking app in November 2019,300 participants were still using the app March 2020 and around 1 600 p

301、articipants responded positively to the invitation to reactivate their app(Molloy et al.,2021).As with the Netherlands Mobility Panel,the availability of pre-pandemic data for the study participants enabled accurate comparisons to be made of travel behaviour.Also,in this case,the tracking app has en

302、abled data to be collected continuously throughout the pandemic and this data collection can potentially be continued afterwards(from participants willing to continue and some refreshment of the participant sample).Around the world,researchers have also taken the initiative to carry out smaller-scal

303、e studies.One notable example,which responded very quickly to the developing situation,was an initiative of researchers based in Istanbul(Turkey),Leeds(United Kingdom)and Sydney(Australia)who organised a panel survey of residents in Istanbul and collected data in January-February 2020,February-March

304、 2020 and March-April 2020(Shakibaei et al.,2020).A snowball technique was used to recruit participants given the lack of immediate access to a market research firm or online platform.Analysis of data from 144 people participating at each of the three waves showed minimal change in travel activity b

305、etween the first two waves,but large changes in travel to work and travel for social,recreational or leisure activities between the second and third waves when social restrictions were introduced.Although such smaller-scale studies cannot provide robust estimates of population-wide travel behaviour

306、trends,they can identify interesting phenomena worthy of more detailed investigation.Public authorities have needed to respond to the pandemic and make alterations to the transport system to accommodate modified demand for different modes of transport and to protect health.Many of the measures intro

307、duced were expected to be temporary,but in some cases,there has been an eye to the future with the possibility of the measures being maintained in the longer term if they are beneficial.In either case,it is helpful to monitor the impacts and see what changes in travel behaviour occur over time.In th

308、e United Kingdom,the DfT introduced the Active Travel Fund to make cycling and walking safer and to facilitate more trips via active transport modes at a time when public transport capacity is reduced due to social distancing,pavements may not be wide enough to provide sufficient space for pedestria

309、ns and roads are at risk of congestion if significant numbers of people switch their commuting mode from public transport to cars(DfT,2021a).It initially invited applications from local authorities in June 2020 for the installation of temporary projects.It then invited applications in November 2020

310、for the creation of longer-term projects.The second phase schemes have been rolled out without time to organise monitoring activities for the first phase.In July 2020,the DfT legislated to enable 12-month e-scooter trials to take place in England and invited local authorities to come forward with pr

311、oposals.Trials are taking place in 32 areas as of 19 April 2021(DfT,2021b).It has commissioned a national monitoring and evaluation study to assess the outcomes of these trials on travel behaviour.It seems likely from the account above that the disruptive event of the Covid-19 pandemic will bring ab

312、out long-term changes to travel behaviour due to people doing things differently during the pandemic and continuing these behavioural changes afterwards,or due to broader changes to society which arise as a TREND BREAKS AND TRAVEL TRANSITIONS TRAVEL TRANSITIONS OECD/ITF 2021 31 result of the pandemi

313、c(e.g.changes to working practices).For all these reasons,it is important to monitor changes and seek to understand them.Lessons from past travel transitions As mentioned in the Introduction,the Discussion Paper for this working group(Chatterjee,2020)reviewed studies of ten different travel transiti

314、ons,which have occurred in Great Britain since the 1990s and drew a number of lessons from the findings of these studies.These are now summarised.Nearly all the transitions could be traced back to the 1990s or earlier,but it took a long time for the transitions to become the subject of significant i

315、nvestigations.It takes years(usually at least two)for a break in a trend to be identified as a persistent change,rather than a short-term“blip”,but a delay of ten years or more for studies to be carried out is noteworthy.It can be speculated that the delays in instigating studies occurred because th

316、e transitions were not expected,hence it took longer for them to be recognised as genuine breaks in trends.For most of the studies,no prior hypotheses were put forward to explain the transition and an array of potential contributing factors were examined.In some cases,there was strong belief in the

317、importance of a particular factor,but evidence only supported this playing a modest role.It was generally not possible to use statistical analysis to quantify the relative contribution of different factors in explaining transitions.Suitable time-series data(in particular,repeated cross-sectional dat

318、a)was either not available or,where available,it did not include factors thought to have played a role.Instead,qualitative judgement has been needed in order to explain transitions.This is exemplified by the assessment of 22 putative factors contributing to reduced car driving by young people in Cha

319、tterjee et al.(2018).This involved looking at the trend direction over time for each potentially influential factor and using the most up-to-date knowledge of the relationship between that factor and the travel indicator of interest in order to make a judgement whether the factor had contributed to

320、the trend.The studies provided welcome illumination on the travel transitions by allowing better characterisations of the nature of the transitions and identifying which population groups were the main contributors to these changes(for example,see a study on a decline in bus use by Le Vine and White

321、,2020).Such studies have usually needed to acknowledge that they have not been able to fully answer why the transition has occurred(for example,studies on the income-car travel relationship have not been able to explain why those with high incomes are using cars less than before).A common issue is t

322、hat studies have not had longitudinal data on individuals belonging to specific groups of interest and which can help to explain their travel behaviour histories.This is a key evidence gap.These groups may represent trend-setters,from whom it can be learned whether the transition might be expected t

323、o transfer across to other groups.For example,the profile of early adopters of shared mobility services is distinctive and there are doubts whether other groups in the population will follow them as users,but this could be investigated further by finding out more about the motivations of early adopt

324、ers of these services.The question is often asked in these studies whether,from what has been learned about the transitions,it will be possible for forecasting models to account for them.There have been mixed conclusions on this.In some cases,there is inadequate data available on the travel behaviou

325、r of interest(e.g.shared mobility use),which means the behaviour cannot be well represented in models.For some transitions,travel behaviour and travel behaviour change vary significantly across the population.However,models are not set up to include sufficient segmentation of the population.For some

326、,transitions changes in travel TREND BREAKS AND TRAVEL TRANSITIONS 32 TRAVEL TRANSITIONS OECD/ITF 2021 behaviour have not been explained by variables that can be included in models,but have simply shifted over time(e.g.lower trip rates)any explanation lies beyond the specification of models and diff

327、erent approaches will be required to consider their future significance.Indicators of change Forecasts generally assume that the indicators of travel activity that are important to measure and track over time have already been identified.Examples include the number of cars per capita,trip rates and

328、vehicle-kilometres travelled.However,it must be considered whether the most important and useful indicators are in fact being measured and consider whether it is enough to obtain overall or average measures for these indicators or whether it is necessary to know the spread in values across the popul

329、ation.For example,it is useful to know the trend in the average daily trip rate,but it is also important to know what proportion of the population is mobile on any day and what proportion is immobile.Immobility should also be considered for different groups.An increase in immobility among workers mi

330、ght reflect voluntary working from home,while an increase among retired people might represent difficulty accessing transport.While this working group has a primary focus on the daily mobility of urban citizens,it is important to recognise the role of long-distance travel(whether conducted by car,bu

331、s,coach,rail or air).Although urban residents make long-distance trips less often than short-distance urban trips,these longer trips contribute substantially to total personal travel mileage and emissions.Furthermore,there are large differences in the amount of long-distance travel undertaken by dif

332、ferent groups within the population.Monitoring of personal travel activity should therefore also capture long-distance travel.Taking a broader view,travel activity is not an end in itself but is part of the broader socio-economic system and travel needs to play its role in contributing to societal g

333、oals.It is therefore important to obtain indicators of travel activity that are useful in assessing the achievement of wider societal goals.In the Horizon 2020 LEVITATE project,quantified policy goals were defined to identify desirable urban visions as a starting point for a backcasting approach to identify policies and measures which enable a path to be taken towards the vision.Indicators were de