Scottish Government

BARRIERS TO MODAL SHIFT

Part 2 - Bridges to mode shift

2.43 There is a very extensive literature on the impacts on travel behaviour of policy measures for travel demand management and mode improvement. Table 2a summarises a range of 'hard' measures which have been shown to be successful, Table 2b summarises the equivalent 'soft' measures and Table 2c summarises the 'complementary' measures (Halcrow 2002, Litman 2002, Rye 2001, DHC 2001). Again, it is through introducing complementary combinations of these factors that success in encouraging mode shift is likely to be greatest.

Table 2a Hard Factors to Encourage Mode Shift

2.44 Improvements to modal alternatives can be effective at increasing patronage, though usually only a minority of the passenger growth comes from former car users ³. In general, light rail or guided bus schemes are more effective at encouraging modal shift from car than conventional bus improvements. Park and Ride schemes are particularly popular with motorists by reducing the need to drive through congested areas. However the design of the schemes, and in particular the location of the "Park" areas, is critical if they are to achieve a net reduction in car kilometres.

2.45 There is now a strong body of evidence showing that long term price and service frequency elasticities are around twice the short term values (Goodwin, 1992), indicating that the benefits of public transport improvements build up over time - and may take five to seven years to achieve their full effect. There are several reasons for this, relating both to 'soft' factors (lack of information, poor image) and structural reasons that cause travel choices to be reappraised (e.g. moving home location, starting a family, retiring).

2.46 In general, restrictions on car use have a stronger effect on behaviour - at least in the short term - than improving alternatives (Jones, 1996). However, as with Park and Ride, careful monitoring is required to assess the overall net effect. The numbers of cars travelling into an area may decrease following an increase in parking charges, but some of these former users might switch to an alternative destination (especially for non-work trips) and in the process travel further by car.

2.47 Research into congestion charging around the world has shown a hierarchy of types of behavioural responses, depending on the level of the charge (Jones, 1992). At quite low levels of charge, drivers who can do so re-route or re-time their trips to avoid the charge. At higher charge levels significant mode shift may occur (depending on modal alternatives), together with some destination switching. At even higher levels, significant trip suppression may result - in cases where there is no reasonable mode or destination alternative.

2.48 Integrated policy approaches where public transport, walking and cycling alternatives are implemented in conjunction with measures to discourage car use in the same area have been shown to be effective both in terms of changing behaviour and in public gaining acceptability for the policies.

2.49 If British towns and cities are compared to their European counterparts such as Munich, Vienna or Zurich, where high quality rail-based public transport services have been provided in conjunction with parking restrictions and limits on car access, higher levels of car use are found in Britain. The lower levels of car usage in the European cities have also been sustained over several decades, despite rising incomes and car ownership (Jones, 2002).

Table 2b Soft Factors to Encourage Mode Shift

2.50 'Soft' improvements to non-car modes largely rely on a combination of better information, fares discounts, marketing and small scale improvements at workplaces and other sites (EU 1999).

2.51 A significant impediment to public transport use is lack of information, or incorrect information. In general, the perceptions of non-users regarding public transport are worse than the reality: they think services are less frequent, slower and more expensive than is actually the case (EU 1999).

2.52 Better pre-trip information can thus remove some barriers to public transport use, through the provision of user-friendly timetables, a telephone enquiry service and information on the internet. The effectiveness of this measure can be enhanced by real-time in-trip information (e.g. real-time information at bus stops and railway stations), and by marketing campaigns that raise awareness of the benefits and advantages of public transport use (Jones 1995). Again, this may also increase use among non car drivers.

2.53 Encouraging car drivers to switch voluntarily to public transport is potentially a long and complex exercise, involving a process of psychological and behavioural change. Figure 3, taken from the on-going EU 'TAPESTRY' research project, identifies up to a seven-stage process of change that may be involved in achieving a permanent change in their travel behaviour (TAPESTRY 2001).

Figure 3 Process of Travel Behaviour Change

2.54 First, there needs to be an awareness that there is a problem that has to be addressed (e.g. traffic congestion, air pollution). Next the driver has to accept personal responsibility - both for the problem and to recognise that changing their own behaviour would make a difference and therefore would be worthwhile. Drivers then need to be aware of alternative modes, identify whether there are any viable options - given their particular circumstances - and decide to make a change. Given that an alternative has been identified, the driver then experiments with this alternative and - only if the experience is positive - does a permanent change in behaviour take place.

2.55 Studies in the UK and elsewhere (e.g. Ciaburro et al 1994) have found that drivers typically report that about 10% of trips they currently make by car could easily be made in some other way, or not at all. More in-depth investigations have found that the true figure is nearer to 20% of trips currently made by car are not car dependent (RAC, 1995), e.g. as noted previously for school travel, around half of parents who drive their children to school say that they would rather not have to undertake this task.

2.56 Recent research for the Department for Transport (DfT 2002) reviewed the effectiveness of a wide range of 'personalised journey planning techniques', from computer-based journey planners, through dialogue marketing approaches, to teachers' packs, mobility centres and travel awareness campaigns. This identified that one of the most promising and successful approaches was the 'Individualised Marketing' approach, developed by Socialdata in Munich. Here all households in an area were directly approached, and asked if they would like information on travel alternatives to the car; customised information packs were then sent out based on requests received.

2.57 A separate study (UITP 1998), reports on the success of 32 small scale pilot projects carried out in ten European countries, with the aim of increasing public transport use. Across the 32 sites, the average increase in public transport trips was 18%.

2.58 Recent experiments in Perth, Australia (Rose 1997) have used the same approach, but aiming to increase walking and cycling as well as public transport use. As a result of households being sent customised information packs and incentive packages, reductions in car trips of about 14% and in trip distances of around 17% have been achieved in a project involving all households in South Perth. These changes seem to have been sustained over several years. The city authorities have now decided to extend the approach to the whole city, on a rolling programme of area initiatives.

2.59 A recent small scale pilot study in Gloucester using Individualised marketing achieved a 9% reduction in car use (SUSTRANS, 2002). The Department for Transport is currently piloting fourteen studies using this kind of approach, and Transport for London is setting up four similar pilot projects in London.

2.60 Other types of soft measures are to be found in a range of Travel Plans, that are being introduced at workplaces, schools, hospital sites and major leisure sites (e.g. football grounds). These measures are usually targeted at particular modes. For example cycling is promoted with secure bicycle parking, with lockers and showers, plus a business mileage allowance (tax free). The emphasis is on practical measures to facilitate alternatives to car commuting such as the development of a database for car pooling, and a guaranteed ride home. These are being assessed in another current EU research project, MOST ⁴. Initial results from this research and other UK based research has been encouraging (e.g. Rye 2001, DHC 2001 Halcrow 2002).

2.61 International experience on the results of the implementation of hard and soft measures is growing. Success in encouraging mode shift is found to be context specific and heavily dependent on combinations of measures for their effectiveness (Halcrow 2002). Survey work on attitudes to potential policies suggests that more reliable and cheaper public transport, shorter journey times and integrated ticketing would be the most effective measures in promoting mode shift (Steg 1997).

2.62 However, over half of car drivers in Scotland already make regular or occasional use of bus or train, and nearly a third of drivers would like to use their cars less (NFO 2001). This, therefore, appears to suggest that there is a broad willingness to use public transport where it is perceived to be a practical alternative.

2.63 There are measures that can be taken to encourage a reduction in car travel, other than by improving modal alternatives (or awareness of them), or directly restraining car use. These 'complementary measures' include initiatives that can be taken by organisations in other sectors that will assist in achieving a mode shift from car. Table 2c lists some examples.

Table 2c Complementary Factors to Encourage Mode Shift

2.64 In the business area there is scope for reducing the need to travel, by encouraging tele-working and tele-conferencing. There are mixed views, however, as to whether overall car travel declines as a consequence of adopting such practices. There are concerns that car use for non-work travel might increase to compensate (under the constant travel time budget hypothesis), or that commuters may move to more rural locations further from their workplace, if they only have to travel in on a limited number of days per week - perhaps maintaining the same weekly commuting mileage/time as previously (Lehto 2002, Zumkeller 2001).

2.65 There are initiatives that other sectors can take which will also help to reduce car use. The most obvious example is in the health sector, where doctors may encourage walking and cycling as a form of physical exercise, in order to reduce the incidence of heart disease. One such initiative is currently being implemented in Gavle, Sweden, as part of the TAPESTRY project (see www.eu-tapestry.org).

2.66 It is increasingly becoming recognised that changes in habitual travel behaviour are most likely to be achieved when habits are broken as a result of major life events. These may be associated with changes in occupational status (getting a job, retiring), in life cycle stage (e.g. having a baby), or in physical fitness. Another occasion when travel choices are reappraised in when people move house. In a recent German research project called Mobiplan, web-based software was developed to enable prospective moves to assess the likely travel implications of moving to particular sites. Details can be found at: www.rwth-aachen.de/mobiplan.

2.67 The software provides information on the availability of a range of services from a chosen location (e.g. schools, shops, libraries). In addition, by inputting details of work location and the activities that would be undertaken in a typical day or week, the software indicates how these activities could be reached using a variety of transport modes (car, public transport, walking and cycling). In each case it estimates total travel times and costs, and environmental impacts in terms of air pollutants and CO ₂ production.

Part 3 - Implications for transport modelling

2.68 As discussed above, rather than simply predicting the demand for travel, transport planners now need to understand travel behaviour (Jones 2002). This involves analysis of what motivates people to travel and how they will respond to changes. Relevant costs in transport planning can therefore be as diverse as: the differences in land prices between a rural greenfield and city centre brownfield development, or the costs of travellers purchasing and using a mobile phone to keep in touch when travelling. Benefits might include issues as diverse as improved social cohesion, better air quality, and local economic growth.

2.69 Faced with no established analytical framework within which to consider these diverse and complex issues, current integrated transport planning (SE 2001) draws heavily from inherited modelling approaches originally developed to support "predict and provide". Analysis of wider costs and benefits continues to be much less rigorous than for transport costs pending the development and application of new modelling techniques (Simmonds et al 2001).

2.70 Surveys of travel behaviour and barriers to mode shift need to fit within a robust behavioural analysis framework if they are to be used in practical transport planning. Behavioural analysis can include: activity analysis including trip chaining, longitudinal analysis looking at trends in travel behaviour, analysis of cultural factors including structural or conscious dependence on a mode, accessibility constraints with travel cost thresholds by trip purpose taking account of lifestyle factors, and analysis of the cost and value of travel (Hensher 2001).

2.71 Despite the large volume of research on travel behaviour, there has been relatively little effort to implement the findings of this in practice. Current state of the art evaluation procedures lag significantly behind the understanding of travel behaviour (Jones 2002). One of the reasons why travel behaviour research has been largely ignored in modelling is that it is not easy or perhaps practical to include a range of important behavioural relationships within the current four stage demand models which are widely deployed.

2.72 One exception has been the move in some cases from trip-based to (simple) round-trip tour based modelling. It is intuitively obvious that mode choice decisions, between travelling by car or by public transport, are not made on a (one-way) trip by trip basis, but looking at the whole sequence of travel, from home until returning home: it is the 'weakest link' on this chain that determines whether a non-car option is feasible.

2.73 Yet, only the most recent models in Scotland have been based on tour data. The advent of modelling exercises designed to look at the impacts of congestion charging (where total costs incurred depend both on the numbers of affected trips in a tour and the times of day at which boundaries are crossed), have required modellers to develop tour-based models. These were first used in London in the 1980s, and are now being employed in the current study for City of Edinburgh Council to model congestion charging options in the city (MVA 1999).

2.74 Activity based transport modelling would provide a framework for more robust consideration of travel behaviour, but is currently regarded as a distant goal (Simmonds 2001). Therefore, whilst travel behaviour research increasingly works with a general presumption that activity based models will become practical in the future, at present there remains a significant gap between theory and practice.

2.75 In some countries, limited activity based modelling is now underway, particularly in The Netherlands, Japan and the USA (Arentze et al 2001). More general activity-based analysis is now being applied where this supports policy goals (e.g. maximising time spent on particular activities) (Jones 2002). However, in Scotland there are no national policy goals related to travel time or time budgets, so travel behaviour research needs to be fitted within an analytical framework designed to support transport policy objectives for economy, accessibility, integration, environment and safety (SE 1998).

2.76 Recent research in England (Halcrow 2002) classified influences on travel demand according to whether or not they were included in current transport models for multi-modal studies. The changes most likely to impact significantly on car travel demand, which are not generally included in transport models were considered to be:

• Electronic communications and travel - e.g. tele-working, and videoconferencing.
• Focused initiatives to change travel behaviour - Individualised marketing campaigns, targeted public transport information, workplace travel plans, and school/higher education travel plans.
• Public transport improvements - Particularly bus quality partnerships and fares and ticketing initiatives.

2.77 Most of the current modelling activity concentrates on transport economic efficiency and presents the results as present value benefits within a multi-criteria analysis framework. However the results of such cost-benefit evaluation have often led to recommendations that are at odds with what the public and politicians regard as preferred solutions (Jones 2002). This is partly because the aspects of transport that particularly interest the public and politicians are the links between transport and the wider economy, society and the environment. Yet these links are rarely considered robustly in appraisal (Mackie 2000). SACTRA (1999) highlighted the wide range of spurious claims about the impacts of transport on the wider economy, but the reality remains that these spurious claims probably still have more influence over transport decisions than established transport modelling (Shaw 2002).

2.78 When looking at links between transport and non transport factors, it has been identified that accessibility models can include a very wide range of non-transport factors (DHC 2000, DfT 2002). They can therefore be used to reflect a wider range of impacts than is usually practical with demand models. By considering the impacts on people rather than on vehicles, they also offer advantages when looking at barriers to mode shift as shown in Figure 4.

Figure 4 A Possible Analytical Framework

2.79 Although there are many uncertainties in aspects of travel behaviour, the accessibility analysis allows a highly disaggregate approach to be taken to the important variables, including types of people and types of trips. With this approach the best available research on soft factors and complementary factors can be included in the accessibility analysis alongside the hard factors.

2.80 Another approach which is gaining favour based on demand modelling, is to gear the analysis to assess the sensitivity of travel patterns to specific interventions (Halden 1996). In such analysis, the design of policy models need not be constrained to transport factors, but can include whatever relationships are needed to assess the range of policy scenarios being investigated (Walker 2001). However, the accuracy of the predictions from strategic policy models will only be as good as the algorithms which define the behavioural responses to diverse interventions. This relies upon effective translation of findings from behavioural modelling into policy modelling.

2.81 There is already a very extensive body of research which can be used to value particular variables within transport models relating to interchange, information and other factors (e.g. Cook 1999). The focus of much of this is on discrete elements or components of a journey. However in travel behaviour theory (Mokhtarian 2001) it is the travel choices which are discrete, and the different components of a journey such as cost, time, effort, route, interchanges, etc. are viewed as a package by travellers.

2.82 Despite the development of many complex models for the recent national transport corridors study in Scotland (MVA 2002), including network models, four stage models, and land use transport interaction models, it was still necessary to undertake relatively simple logit choice analysis to estimate park and ride demand. A similar approach is likely to be needed for other "barrier modelling", interfacing such analysis with policy models, network models and accessibility models as required to ensure appropriate decision support.