The development of an option generation tool to identify potential transport policy packages
-
Upload
charlotte-kelly -
Category
Documents
-
view
215 -
download
0
Transcript of The development of an option generation tool to identify potential transport policy packages
ARTICLE IN PRESS
Transport Policy 15 (2008) 361–371
Contents lists available at ScienceDirect
Transport Policy
0967-07
doi:10.1
� Corr
E-m
journal homepage: www.elsevier.com/locate/tranpol
The development of an option generation tool to identify potential transportpolicy packages
Charlotte Kelly, Anthony D. May �, Ann Jopson
Institute for Transport Studies, University of Leeds, Leeds LS2 9JT, UK
a r t i c l e i n f o
Available online 28 February 2009
Keywords:
Option generation
Transport policy packages
Strategy
Web-based knowledgebase
0X/$ - see front matter & 2009 Elsevier Ltd. A
016/j.tranpol.2008.12.008
esponding author.
ail address: [email protected] (A.D. Ma
a b s t r a c t
While techniques for identifying urban transport problems, predicting the effects of possible solutions
and appraising their performance are well developed, there is a significant gap in the decision-making
chain relating to the generation of those possible solutions. A literature review has identified a number
of option generation methods in other sectors, which are of potential application to transport. It
categorises them into ‘inside the box’ methods, which are principally quantitative and draw on a pre-
existing list of solutions, and ‘outside the box’ methods, which are more qualitative but potentially
better able to generate wholly novel solutions. It also distinguishes between applications at the levels of
formulation of an overall strategy and of detailed design of a particular scheme. An ‘inside the box’
method for strategy option generation has been developed building on the capabilities of an existing
web-based knowledgebase, KonSULT. The design principles are described and potential further
developments outlined.
& 2009 Elsevier Ltd. All rights reserved.
1. Introduction
Transport planners face wide ranging challenges in attempting tocreate more sustainable urban transport systems (ECMT, 2002;DGEnv, 2005). Fortunately they have access to an increasingly widerange of policy instruments. No longer is the focus simply on addingto the transport infrastructure and managing it more effectively;possible solutions can be found also in the fields of land useplanning, awareness raising, information provision and pricing.While information on the performance of some of the more recentlydeveloped policy instruments is limited, some guidance is availablefrom a number of sources. The principal ones are the VTPI TDMencyclopaedia (www.vtpi.org/tdm) and the KonSULT knowledgebase(KonSULT, 2008; www.konsult.leeds.ac.uk). Other sources such asELTIS provide case studies of successful policy interventions(www.eltis.org). However, very little guidance is available on howto select potentially suitable policy instruments in the first instance,a challenge that has become more significant as the number ofpossible policy instruments has expanded.
We refer to this process as option generation. In the logicalstructure recommended by PROSPECTS, shown in Fig. 1 (KonSULT,2008) or in greater detail elsewhere (WEBTAG, 2006), plannersneed first to specify the objectives that they seek to achieve, andthe problems that they face, now or in the future, and that are
ll rights reserved.
y).
evidence of the objectives not being met. This is the point atwhich they can begin to identify possible solutions. These can beoverall strategies, a specific policy instrument on its own, such asroad pricing, or the detailed design of a given policy instrument,such as the specific location of a road pricing cordon. The term‘strategy’ can be thought of as a broad approach to policy, such asreducing the level of car use (May et al., 2005a), or a specificpackage of policy instruments, such as a combination of roadpricing, bus service improvements and fare changes (May et al.,2005b). Option generation is needed at all of these levels: thestrategy, packages of policy instruments, specific policy instru-ments, and detailed scheme design. The methodology described inthe paper relates to the second and third of these: packages andspecific instruments.
What evidence there is suggests that option generation israrely regarded as a key stage in the strategy or schemeformulation process. A study by Atkins (2007) for the UKDepartment for Transport of its Local Transport Plan processsuggests that local authorities, in England at least, tend not toinnovate, but rather to pursue schemes that have been underconsideration for a long period, and to focus on infrastructureprojects and management-based improvements to the infrastruc-ture, rather than considering enhancements to public transport orways of managing demand. The final report on the studyconfirmed that policy instruments involving public transportimprovements, demand management and environmental en-hancement were given less emphasis in both the first and secondround of the Local Transport Plans. The UK Eddington Report
ARTICLE IN PRESS
Objectives/Indicators (7)
Assess problems (8)
Possible instruments (9)
Predict impacts (12)
Compare solutions (13)
Implement (15)
Evaluate performance (15)
Monitor (15)
Barriers (10)
Possible strategies (11)
Optimisation (14)
Appraisal (13)
Scenarios (11)
Fig. 1. The logical structure (source: May et al., 2005a).
C. Kelly et al. / Transport Policy 15 (2008) 361–371362
(Eddington, 2006) outlines the need succinctly: ‘‘Unless a widerange of appropriate options is considered, there is a risk that thebest options are overlooked and money could be wasted. A goodoption generation process is crucial to ensure that the transportinterventions that offer the highest returns can be found. The fullrange of options should look across all modes and include makingbetter use of the existing transport system, including betterpricing; investing in assets that increase capacity y; investmentin fixed infrastructure; and combinations of these options.’’
This is one of the challenges that has been addressed in a4-year UK research programme, DISTILLATE (Design and Imple-mentation Support Tools for Integrated Local Land use, Transportand the Environment), which is described more fully in acompanion paper (May et al., 2009). The programme involvedseven integrated projects: one which reviews the barriers that localauthorities face in seeking sustainability, one on the use ofindicators (Marsden et al., 2006), the one reported here on optiongeneration, one on overcoming financial barriers (Brannigan andPaulley, 2009), one on enhancements to predictive models(Shepherd et al., 2009), one on enhanced appraisal (Jopson et al.,2007) and one on institutional structures and processes. At an earlystage in DISTILLATE the 16 local authority partners were asked toassess the barriers that they faced (Hull and Tricker, 2005). Halfconsidered the generation of strategy options problematic, whileonly a quarter found generation of scheme options difficult. At astrategic level, national guidance and professional judgmentdominated the sources of information, followed by stakeholderinput. Under a quarter identified option generation tools ascurrently making an important contribution. Dissatisfaction wasgreatest with the lack of available tools, followed by that with
national guidance. Over half of those who responded thought itimportant to have access to option generation tools.
This survey confirmed the importance of a focus on optiongeneration and, in particular, on tools for strategy optiongeneration. The project pursued these initially through a literaturereview in transport and in other sectors, the principal findings ofwhich are outlined in Section 2. That review identified two broadtypes of approach: ‘inside the box’ methods, which are typicallyquantitative in nature and draw on pre-existing concepts andcomponents, and ‘outside the box’ methods, which are typicallyqualitative and are able to identify radically different approaches(Jones and Lucas, 2006). It was agreed to develop four specificoption generation methods for further testing: two ‘inside the box’and two ‘outside the box’ methods with one of each operating atthe strategic and one of each at the scheme level.
This paper focuses on the development of an ‘inside the box’method for generating individual policy instruments and packages,and can therefore be considered as an example of ‘inside the box’strategy option generation. It is based on an existing knowledge-base, KonSULT (Knowledgebase on Sustainable Urban Land use andTransport). Further information on the three other methods can beobtained from the programme website (www.distillate.ac.uk).Section 3 describes KonSULT. Section 4 explains the enhancementsthat have been made to KonSULT to allow it to operate moreeffectively as an option generator for the sets of policy instrumentsto be potentially included in a strategy. Section 5 describes theapproach taken to developing packages of policy instruments. Thesetwo sections include a worked example of the whole process.Section 6 describes the approach adopted to testing the new tool inpractice and proposals for further development.
2. Literature review of option generation methods
2.1. The range of methods
A comprehensive review of option generation techniques(presented in Jones and Lucas, 2006) was conducted to identifynew techniques and approaches to assist transport planners ingenerating alternative options at the strategy and scheme level.This section looks in turn at some of the different approaches thatwere identified as potential option generation techniques thatare either currently widely applied in the transport field (e.g.databases such as KonSULT) or could be transferred from othersectors (e.g. mind mapping).
Hull and Tricker (2005) found that transport planners narrowdown the potential options available to them based on theirexperience of ‘what works’. Jones and Lucas (2006) concluded thatthis approach works under many scenarios, but ‘‘where asignificant step change in policy making and travel behaviour isrequired more innovative and radical solutions may be necessary’’.These two approaches (applying ‘what works’ and the moreinnovative options) can be classified respectively, as ‘inside thebox’ and ‘outside the box’ techniques. Table 1 provides a summaryof these two categories. ‘Inside the box’ option generationtechniques make use of existing knowledge and either apply itdirectly to new contexts (e.g. library-based approaches) or use thefoundations of this knowledge to create alternative options (e.g.morphological box). ‘Outside the box’ methods tend to involvemore qualitative techniques, allowing decision makers to generatenew forms of options that have previously not been considered.These two categories of methods do not work in isolation; anysolutions generated by the ‘outside the box’ techniques willbecome part of the ‘inside the box’ solutions once they have beenapplied and evidence on effects collated. Some examples ofmethods identified by the review are provided in Table 1.
ARTICLE IN PRESS
Table 1Methods for outline option generation.
Categories Description Approach Examples
‘Inside the box’ Pre-existing solutions Library-based approaches KonSULT (2008), VTPI (2008)
Existing parameters, new permutations Morphological box approach Road user charging (Kocak et al. (2005),
Zwicky (1947)
Constrained options Constrained by specific parameters Priority evaluator (Hoinville, 1971)
‘Outside the box’ Generating new forms of options including
some new parameters
Structured approaches Mind/cognitive mapping, six thinking hats
(De Bono, 1985)
Unstructured approaches: verbal stimuli Brainstorming
Unstructured approaches: visual stimuli Eureka (Townsend and Faviour, 1991)
Source: Adapted from Jones and Lucas (2006).
C. Kelly et al. / Transport Policy 15 (2008) 361–371 363
2.2. Inside the box methods
‘Inside the box’ methods are commonly used in transportplanning utilising both professional knowledge on ‘what works’and an evidence base built up from the application of certainpolicy instruments and the results they achieve. This category canbe split into three main approaches: library-based approaches,morphological box approach and constrained options approach.
Library-based approaches have been used for many years bydecision makers seeking to solve specific problems or meet agreedpolicy objectives by referring to databases of evidence andresearch on policy instruments. For example KonSULT providesevidence on the likely effects on specified objectives of a range ofpolicy instruments. Decision makers then use this information todetermine the suitability of specific policy instruments for theirpurpose. In recent years, library-based approaches have utilisedthe internet to disseminate information. A number of library-based approaches in the transport industry, including KonSULT(2008) and the VTPI transport demand management encyclopae-dia (VTPI, 2008), have developed internet-based search facilitiesfor assessing policy instruments. The key advantage of theselibrary-based approaches is the wide selection of evidence that isbrought together, enabling a large number of options to beconsidered by transport planners. The key disadvantage is thatthere may be solutions not in the database that could be moresuitable for the specific objectives or locations that are beingconsidered.
Morphological box analysis was first developed by Fritz Zwicky(1947) as a structure for systematically identifying the completeset of potential permutations for designing a particular aero-nautical system. Current applications of morphological analysisutilise this basic methodology to provide a structure to illustratethe range of choices available for a particular instrument orstrategy. Once all the options are identified and input into thematrix (morphological box), and any infeasible options elimi-nated, it is then easier to consider all the potential combinationsthat could be used to design a particular policy and also identifywhere gaps exist in the current options available. The mainchallenge with the morphological box approach is to narrowdown the very large number of potential combinations, byrejecting infeasible sub-sets of combinations. This methodologyhas been recently been transferred to the transport planningsector through the work presented in Kocak et al. (2005). Theydeveloped a web-based tool to assist local authorities in the UK toinvestigate the suitability, and then the possible design, of a roaduser charging scheme for their local area. This approach relies ondecision makers knowing all the possible permutations (policyparameters), making use of the library-based approach togenerate this information. Its advantage is that it allows optionsthat are infeasible for specific criteria to be removed from the
decision, and at the same time potentially creates new permuta-tions that have not previously been considered by decisionmakers.
The final group of ‘inside the box’ methods is constrainedoption solutions. These methods are similar to the morphologicalbox approach as they present a range of possible permutations ofoptions to the decision maker; however the difference is thatthese options are constrained, in most cases by imposing a budgeton the options that can be selected. The priority evaluator is anexample of this type of methodology. Hoinville (1971) developedthe priority evaluator method to gauge community preferencesfor improvements to a local area. Participants were presentedwith a range of attributes in their local community that could beimproved and a range of levels to which each option could beimproved (including not improving a specific attribute). Theywere then presented with a budget and told to allocate thisbudget among the given attributes, indicating what they wouldprefer to improve. This type of methodology has also been appliedby the ARTISTS project (ARTISTS, 2006). In this project the publicwere asked how they would like to reallocate highway spacealong a selection of mixed-use urban streets and instead of abudgetary constraint a spatial constraint was imposed on thedecision. The key advantage of constrained option solutions is thatthey introduce an element of feasibility.
2.3. Outside the box methods
Jones and Lucas (2006) identified three main types of ‘outsidethe box’ methods that are commonly used: structured approachesand visual and verbal unstructured approaches.
One of the main examples of a structured approach to ‘outsidethe box’ option generation is mind mapping. It is structuredbecause it is based around a central theme and the aim is tocollate all possible links to this central theme. Under thisapproach, potential solutions that had not previously beenconsidered can be generated from new previously unconsideredlinks to this central theme. Another example of the structuredapproach is the de Bono (1985) ‘six thinking hats’ technique.Under this approach, participants are led through a problem tofind potential solutions by taking a different perspective on theproblem each time they put on one of the six hypotheticalcoloured hats. Each of the colours stands for a differentperspective. For example when putting on the black hat theparticipant focuses on the negative analytical approach and whenputting on the green hat she focuses only on new ideas. Thesestructured approaches allow different and potentially newsolutions to be generated for a specific problem. The negativeside is that they may generate options that may be unfeasible,which is the key advantage of the ‘inside the box’ methods.
ARTICLE IN PRESS
C. Kelly et al. / Transport Policy 15 (2008) 361–371364
The other methods identified by the review use verbal or visualstimuli to generate options. An example of verbal stimuli isbrainstorming, which involves suggestions, discussion on thosesuggestions and then agreement. Other examples are lateral andcreative thinking techniques, which through the use of discussionallow individuals and groups to generate options to specificsolutions that may not have been considered. This process isdescribed by de Bono as a ‘‘way of thinking that seeks a solution toan intractable problem through unorthodox methods or elementsthat would normally be ignored by logical thinking’’ (de BonoHomepage, 2006). Jones and Lucas (2006) identified Eurekaas an example of using visual stimuli to generate options. Thistechnique ‘‘presents participants with unrelated objects or ideas,and asks people to find a significance or association in theirjuxtaposition (Townsend and Faviour, 1991)’’. The results are thenrelated back to the problem in question. The advantage of theseunstructured approaches is that the participants using the optiongeneration methodology can think about the problem without thebias of considering what the ‘experts’ think works or knowingwhat policies have worked well in similar scenarios, so potentiallygenerating feasible new options.
3. The KonSULT-based option generation tool: background
As indicated in Section 1, this paper describes an ‘inside thebox’ tool being developed for strategic option generation. Itcombines the principles of a library tool, KonSULT, described inthis section and further developed in Section 4, and a morpho-logical box analysis, which is described in Section 5.
3.1. Development
KonSULT is a web-based knowledgebase designed to provideup to date evidence on the performance of a wide range oftransport and land use policy instruments. It is aimed atprofessional transport planners at all levels of responsibility, butalso at decision makers, interest groups and members of thepublic concerned about transport problems. A description of itsdevelopment can be found in Jopson et al. (2004) and May andTaylor (2002).
A list of 60 policy instruments for potential inclusion wasdeveloped on the basis of a taxonomy from earlier work by Mayand Still (2000) and extended by Matthews and May (2001). Theseinstruments are grouped into six categories: land use interven-tions, behavioural and attitudinal measures, infrastructure pro-jects, management and operational measures, informationprovision, and pricing. To date the knowledgebase has beenpopulated with 42 of the 60 instruments. Further case studies forthese, and additional policy instruments, will be added asinformation becomes available (www.konsult.leeds.ac.uk).
3.2. Structure
A consistent approach has been undertaken for describing thestructure of each instrument. This covers:
�
a taxonomy and description that defines the instrument, itsaims and technological requirements; � a first principles assessment that looks at why that instrumentshould be introduced, considers its anticipated demand andsupply impacts, assesses the resulting positive or negativecontributions to key policy objectives and problems,and identifies likely winners or losers and barriers toimplementation;
�
evidence on performance, illustrated with a series of casestudies describing specific interventions, and empirical evi-dence on their impacts on the same set of objectives andproblems examined within the first principles assessment; � a summary of the contribution that the instrument is expectedto make and the contexts in which it is likely to be mosteffective and
� an identified set of complementary instruments that wouldwork well with the selected instrument by helping to over-come barriers or enhance its positive impacts.
To illustrate the impacts of each instrument and ensure aconsistent methodology the same lists of demand and supplyimpacts, objectives, problems, barriers and contexts are used toassess each instrument. The contribution of each policy instru-ment to each of these is assessed on a common 11-point scale.This is illustrated in Fig. 2, in which the instrument concernedscores four out of a possible five in terms of its contribution toliveable streets and equity, three for efficiency, safety and theeconomy, and two for environment. It has an uncertain impact onfinance, but has no negative impacts on any of the objectives. Thescores assigned are judgmental rather than evaluative, and takeaccount of qualitative and quantified performance, public percep-tions and professional assessments. They are thus designed not toreplace a formal cost–benefit analysis but to suggest policyinstruments that might be developed for subsequent moredetailed appraisal.
3.3. Current search facilities
Within KonSULT, users have the option to search for a specificinstrument using a series of drop-down menus, a keyword search,hyperlinks from the complementary instruments section or afilter search. This paper describes the development of the pre-existing (library-based) filter search within KonSULT to enable amore sophisticated level of option generation from the informa-tion already available within KonSULT. The previous filter isillustrated in Fig. 3, which shows that users were able to specifytheir role and area type and select the policy objectives, problemsand strategies that they wished to investigate. The filter then usedthe scores for each policy instrument for these selected attributeswithin KonSULT to produce a ranked list of policy instrumentsagainst the selected criteria. This ranking was offered not as aprescribed solution but as a list of options for further examination.
3.4. Weaknesses of the filter search
The filter search was developed experimentally at an earlystage in the preparation of KonSULT, and used a simple additiverule for the scores for all the attributes that the user specified. Atthe outset of work on DISTILLATE, it was identified as a potentiallyvaluable option generation tool. However, a number of weak-nesses were evident:
(a)
only 10 policy instruments were included in the filter and itsdesign was too inflexible to permit the inclusion of additionalpolicy instruments;(b)
the implications for different user types were unclear; (c) the listing of area types was somewhat inconsistent, and itwas unclear how instruments could be selected for differentareas in ways that achieved some spatial consistency;
(d)
the ability to select both objectives and problems led to adegree of double counting;(e)
there was no facility for identifying the relative importance ofdifferent objectives (or problems);ARTICLE IN PRESS
Fig. 2. An example of the instrument contribution scores.
C. Kelly et al. / Transport Policy 15 (2008) 361–371 365
(f)
there had been a growing emphasis on the use of indicatorsand targets (Marsden et al., 2006), yet users were unable toselect or prioritise indicators;(g)
the additive nature of the selection rules led to inconsistenciesin the shortlisted instruments and their performance scores;and(h)
there was no function for saving the output for furtheranalysis.In addition the method was only suitable for identifyingindividual instruments. It could not suggest a package of policyinstruments that might form an effective integrated strategy.Sections 4 and 5 of this paper now describe the development ofthe strategy option generation tool in response to these identifiedweaknesses.
4. The KonSULT-based option generation tool: development
4.1. The approach adopted
The development of KonSULT has proceeded in two stages. Thefirst stage is designed to address weaknesses (a)–(h) above, and isdescribed in this section. The second stage tackles the issue of
generating packages of instruments, and is outlined in Section 5. Afull description of the development is provided in Kelly et al.(2008).
Work on the first stage involved enhancing the individual fieldswithin the filter search, enhancing the logic for scoring inindividual fields and aggregating those scores, and determiningthe form of output. A deterministic, transparent approach hasbeen adopted that retains as much flexibility as possible for theuser, while enabling both user and researcher to identify, in thetesting stage, any attributes that cause problems. The key focus ofthe improvements described in this section was to use the existingprofessional knowledge on policy instruments already in KonSULTto enhance the tool.
A database has been produced that allows new policyinstruments to be added to the system with minimal effort. Allof the 42 policy instruments currently included in KonSULT arewithin the option generation tool.
4.2. The user type field
The reason for having a user type field is to identify only thosepolicy instruments that the user has the ability to directlyinfluence and implement. For example in the UK, road tax is a
ARTICLE IN PRESS
Fig. 3. Original KonSULT filter search screen.
C. Kelly et al. / Transport Policy 15 (2008) 361–371366
policy tool that only the national government can modify,whereas parking charges are the responsibility of the localauthority. Using this information and the three key types ofgovernance within the UK, the user types have been reviewed andre-specified for the new tool as:
�
decision makers from national organisations (e.g., government,regulatory bodies, national infrastructure operators); � decision makers from regional organisations (e.g., regionalassemblies, regional development agencies) and
� local authority decision makers (e.g., city councils, local landuse and transport planners).
A binary (1, 0) code against each policy instrument in the databaseindicates whether it is available to each of the specified user types(1 if suitable).
4.3. The area type field
A wide range of literature has looked at how effective differentpolicy instruments are when applied in different spatial locations.For example, Sumalee et al. (2002) considered how to determineefficient locations for road pricing cordons, while a review oflight-rail schemes by Babalik-Sutcliffe (2002) identified that high-density development and the existence of radial corridors arecritical factors for urban rail success. Using such case studies andprofessional knowledge the potential effectiveness of a giveninstrument in each specified spatial context has been determinedusing a 0–5 rating scale (not suitable–very suitable). For examplethe policy instrument park and ride scores 5 for applicability alonga transport corridor, but only 1 in a district centre. For the tool this
score is normalised to a 0–1 scale. Users are now able to select onearea type, either by settlement type (large town or city 4100k;small town or city o100k), or city area (town or city centre, innersuburb, outer suburb, district centre, corridor) or to specify ‘anyarea type’. If ‘any area type’ is selected a normalised default scoreof 1 is used.
4.4. Treatment of objectives, problems and indicators
The treatment of objectives, problems and indicators is themost complex aspect of the filter search. Users are able togenerate policy instruments according to the objectives, problemsor indicators that they are most concerned about, by selectingfrom the appropriate list. Whilst the lists of objectives andproblems have stayed the same as in the original filter, a new setof indicators has been added. This indicator set is based on workin one of the other DISTILLATE projects (Marsden et al., 2006) andis shown in Fig. 4. Users are able to select only from one of theselists to avoid double counting; so their first step is to decidewhether they wish to focus on objectives, problems or indicators.Users are able to select several options from their chosen list, andare then given the opportunity to weight their selected objectives(or problems or indicators) in terms of their relative importanceto them. The weighting system uses a score of 1–5 (5—highimportance, 1—low importance), with weightings then beingnormalised to give weights summing to 1. The default is that allselected objectives (or problems or indicators) have equal weightssumming to 1.
For example the user in Fig. 4 has chosen to focus on objectivesand has selected the categories of efficiency, liveable streets,protection of the environment, equity and social inclusion and
ARTICLE IN PRESS
Fig. 4. KonSULT objectives, problems and indicator filter.
C. Kelly et al. / Transport Policy 15 (2008) 361–371 367
safety. Efficiency has been weighted as being of high importance(5) in this preference list and safety has been weighted as 3, whileall the others have been weighted as 1 (low importance).
The tool then draws on the score from within KonSULT, whichassesses the contribution of each instrument to the selectedobjectives (or problems or indicators) on the 11-point �5 to +5scale (as described in Section 3), which is then normalised to �1to +1. A negative score implies that the policy instrument wouldhave a negative impact on the objective (or problem or indicator)being considered. Where several objectives (or problems orindicators) are selected, the scores for a given instrument arefactored by the weights determined by the user, summed overthe selected objectives (or problems or indicators) and thennormalised.
4.5. Treatment of strategy
The term strategy is used here in the first sense identified inSection 1; that is a broad approach to policy. While the types ofstrategy have not changed from the original filter the functionalityof being able to weight them has been included. The weightscoring and normalisation operate as for the indicators, objectivesor problems described above. Where ‘any strategy’ is specified adefault score of 1 is used.
4.6. Calculating the instrument score
The scoring systems for each of the attributes: user type; areatype; objectives, problems or indicators; and strategies; have beendescribed above. The aggregate scoring system is described in thebox below. This provides a marked change compared to the oldfilter, where a simple additive formula was used. The result ofusing this formula is a score between�100 and 100 for each of thepolicy instruments in the database for each of the sets of criteriaselected. A score of 0 means that the policy instrument was not
suitable for one of the preferences of the user (e.g. user type), orhad no impact on the selected objectives.
The overall score for instrument I is calculated as
ScoreI ¼ 100� UserTypeI � AreaTypeI � ObjectiveI
� StrategyI (A)
where I is the instrument of concern. Note that depending onthe user’s selection, ObjectiveI can be replaced by IndicatorI
or ProblemI.The user type score is based on a simple binary selection:
UserTypeI ¼ ð0; 1Þ (B)
The area type score is based on a normalised scale:
AreaTypeI ¼ ð0pAreaTypeIp1Þ (C)
The objective (or problem or indicator) score is based on
ObjectiveI ¼X
o
WoOI (D)
whereX
o
Wo ¼ 1 (E)
and
�1oOIo1 (F)
where Wo is the weight for objective O and OI is the objectivescore for instrument I. The strategy score is weighted in asimilar way using weights WS for strategy S and scores SI forinstrument I.
4.7. A worked example (part 1)
Looking at real world examples illustrates how the filter selectspotential policy instruments for consideration. The user illu-strated in Fig. 4 is a local authority decision maker concerned witha town or city centre, who is not concerned to focus on anyparticular strategy, has chosen to focus on objectives, has selectedthe categories of efficiency, liveable streets, protection of the
ARTICLE IN PRESS
C. Kelly et al. / Transport Policy 15 (2008) 361–371368
environment, equity and social inclusion and safety, has weightedefficiency as of high importance (5), safety as of moderateimportance (3) and the others as of low importance (1). Thesechoices are summarised in the box below.
Stage 1 KonSULT webpage
User Group: Local authority decision maker
Area type: Town or City Centre
Objectives/problems or indicators: Objectives
Objectives being considered and weightings:
Efficiency (weight 5)
Liveable streets (weight 1)
Protection of the environment (weight 1)
Equity and social exclusion (weight 1)
Safety (weight 3)
Strategy: Any
Given this selection, the KonSULT option generator provides anorder list of suggested policy instruments, as shown in Fig. 5. Thissuggests that road pricing could be the most effective (since it hasthe highest overall score). Case studies included within KonSULTsuggest that road pricing is particularly effective in reducingcongestion and hence enhancing efficiency, as has been thecase in both London and Singapore examples (KonSULT, 2008).Consequently, road pricing will have been scored highly withinthe KonSULT assessment using the scales outlined above and
Fig. 5. KonSULT option generation output
weights in Fig. 4. Intelligent transport systems (ITSs) come next,followed by urban traffic control.
4.8. Outputs
As shown in Fig. 5, users are provided with a list of instrumentsranked in descending order of score. Each instrument has ahyperlink back to the relevant KonSULT pages. This provides a linkto the more detailed information about each policy instrumentcontained within the KonSULT database and in doing this utilisesthe full functionality of this library-based approach. Thus, to takethe example above, the user can link to the pages on road pricingand find summaries of the case studies in London and Singapore,as well as those from Norway and the United States.
An assessment has also been provided as to whether theinstrument is a high-, medium- or low-cost measure in terms ofimplementation and operating costs. This is made on thefollowing basis:
�
exa
high cost: infrastructure measures;
� medium cost: operational and regulatory activities, trafficmanagement and information systems;
� low cost: behavioural and other readily implemented measuresand
� neutral cost: pricing measures that will pay for themselvesover time.
One of the new features of the tool is the set of filters that havebeen included to specify the policy instruments that should be
mple (presented for scores X20).
ARTICLE IN PRESS
Table 2An interaction matrix.
These instruments Contribute to these instruments in the ways shown
Land use Infrastructure Management Information Attitudes Pricing
Land use
Infrastructure
Management
Information
Attitudes
Pricing
Key: benefits reinforced, financial barriers reduced, political barriers reduced, compensation for losers.
Source: May et al. (2005a).
C. Kelly et al. / Transport Policy 15 (2008) 361–371 369
presented. Users can filter by the number of instruments, aminimum score achieved (e.g. 420), the type of instrument (e.g.pricing), and the cost or a combination of the above. For example,Fig. 5 provides the output from the KonSULT option generationtool using the filter criterion of scores X20.
It is possible to save each output file, so that the user can goback to the filter search, make different selections and comparethe different sets of results.
5. Approach to developing packages of policy instruments
The second stage of the development of KonSULT is concernedwith the generation of packages of instruments. This was one ofthe key restrictions of the old filter in that it was only possible topresent a list of individual policy instruments considered inisolation.
5.1. Development of packages
It has long been known that there can be benefits fromintroducing packages of measures. Most approaches to creatingpackages of policy instruments focus on two key principles, firstlythe pursuit of synergy and secondly the removal of barriers.Synergy occurs when ‘‘the simultaneous use of two or moreinstruments gives a greater benefit than the sum of the benefits ofusing either one of them alone’’ (May et al., 2006). The removal ofbarriers implies identifying factors that hinder the implementa-tion of an otherwise desirable policy instrument, and using asecond instrument to overcome them. Table 2 provides theexpected interaction effects if two policy instruments of differenttypes are combined and shows the expected synergy and barriereffects (May et al., 2005a). For example, management instrumentscan contribute to infrastructure instruments by reinforcingbenefits, reducing political barriers and compensating losers.
The second stage of the development of the KonSULT optiongenerator uses the approach of defining pairs of policy instru-ments from the KonSULT output based on the scores that theyachieved in stage 1 and either their potential synergy effects(using a synergy matrix) or their ability to reduce barriers (using abarriers matrix) to either policy instrument.
The barrier matrix has been developed using the interactionsin Table 2. There are three types of barriers in the table(acceptability, finance and compensation of losers). The informa-tion within this table has been used to develop a scoring system.Where there are no positive effects of removing barriers if twopolicy instruments are combined then a score of 0 has beenapplied (e.g. a policy package of a management measure and aland use measure). Where there is one type of barrier removed thescore is +5, where there are two barriers removed the score is +10and where there are three the score is +15. For example if a policy
instrument of type pricing is packaged with a policy instrument oftype management then +10 would be added to the total combinedscore.
The synergy matrix (described in Kelly et al., 2008) has beendeveloped along the same lines, using the indicator for reinforce-ment of benefits in Table 2. However, a review of case studies ofthe performance of combinations of policy instruments suggeststhat true synergy rarely occurs (May et al., 2006), since two policyinstruments that reinforce one another are likely each to havesimilar impacts on some users. For this reason, all the values inthe synergy matrix were negative.
It should be noted that these integration matrix scores aresomewhat arbitrary. As May et al. (2006) note, there is very littleevidence as yet of the impact of integrated policies on which tobase more dependable scores.
A package development tool has been developed using MSExcel to import the output from the first stage of the KonSULToption generator (policy instrument code, policy instrumentname, score and cost) based on the user’s preferences. Once inthis tool the user can decide which policy instruments to includein the packages and conversely which policy instruments toexclude, perhaps because they would not be acceptable for theirlocal area. It then uses a morphological box approach to presentall possible pairs of instruments to users. These instrument pairsare selected from the shortlisted policy instruments, theirrespective scores in stage 1 and the score from either the synergymatrix or the barrier matrix. The rating of instrument pairs iscreated using Eq. (G). The result is a ranked list of pairs that, giventhe user’s preferences, would provide options for further inves-tigation and development.
Total score ¼ instrument 1 scoreþ instrument 2 score
þ selected interaction matrix score (G)
5.2. Example of using the tool
Taking the same example as in Section 4.7, the user hasselected the top 10 instruments from Stage 1, as shown in Fig. 5.She has then sought a list of pairs of these 10 instruments,focusing on the synergy between them, as summarised in theStage 2 box. On this basis the option generation tool generates thelist of pairs of policy instruments as shown in Table 3.
Stage 2 (import the output from KonSULT)
Which policy instruments to include: top 10 scoring fromstage 1
Matrix type: Synergy matrix
ARTICLE IN PRESS
Ta
ble
3R
an
ke
dli
sto
fp
air
so
fp
oli
cyin
stru
me
nts
.
Ne
wra
nk
Inst
rum
en
t1S
core
1C
ost
1R
an
k1
Ca
teg
ory
1In
stru
me
nt2
Sco
re2
Co
st2
Ra
nk
2C
ate
go
ry2
Co
mb
ine
dsc
ore
Ma
trix
sco
reT
ota
lsc
ore
1IT
S7
0.9
Hig
h2
Ma
na
ge
me
nt
Ro
ad
pri
cin
g8
0.0
Ne
utr
al
1P
rici
ng
15
0.9
1�
51
45
.9
2U
TC
63
.6M
ed
ium
3M
an
ag
em
en
tR
oa
dp
rici
ng
80
.0N
eu
tra
l1
Pri
cin
g1
43
.64
�5
13
8.6
3A
ccid
en
tre
me
dia
l5
0.9
Me
diu
m4
Ma
na
ge
me
nt
Ro
ad
pri
cin
g8
0.0
Ne
utr
al
1P
rici
ng
13
0.9
1�
51
25
.9
4Lo
rry
fle
et
ma
na
ge
me
nt
42
.2M
ed
ium
7M
an
ag
em
en
tR
oa
dp
rici
ng
80
.0N
eu
tra
l1
Pri
cin
g1
22
.18
�5
117
.2
5P
ark
ing
con
tro
ls4
0.0
Low
9M
an
ag
em
en
tR
oa
dp
rici
ng
80
.0N
eu
tra
l1
Pri
cin
g1
20
�5
115
.0
6Li
gh
t-ra
ilsy
ste
ms
43
.6H
igh
6In
fra
stru
ctu
reR
oa
dp
rici
ng
80
.0N
eu
tra
l1
Pri
cin
g1
23
.64
�1
011
3.6
7C
om
pa
ny
trav
el
pla
ns
41
.8M
ed
ium
8A
ttit
ud
es
Ro
ad
pri
cin
g8
0.0
Ne
utr
al
1P
rici
ng
12
1.8
2�
10
111
.8
8D
ev
elo
pm
en
td
en
siti
es
mix
49
.5H
igh
5La
nd
Use
Ro
ad
pri
cin
g8
0.0
Ne
utr
al
1P
rici
ng
12
9.4
5�
20
10
9.5
9Li
gh
t-ra
ilsy
ste
ms
43
.6H
igh
6In
fra
stru
ctu
reIT
S7
0.9
Hig
h2
Ma
na
ge
me
nt
114
.55
�1
01
04
.6
10
UT
C6
3.6
Me
diu
m3
Ma
na
ge
me
nt
ITS
70
.9H
igh
2M
an
ag
em
en
t1
34
.55
�3
01
04
.6
11IT
S7
0.9
Hig
h2
Ma
na
ge
me
nt
Pri
vate
pa
rkin
gch
arg
es
38
.2N
eu
tra
l1
0P
rici
ng
10
9.0
9�
51
04
.1
12
Co
mp
an
ytr
ave
lp
lan
s4
1.8
Me
diu
m8
Att
itu
de
sIT
S7
0.9
Hig
h2
Ma
na
ge
me
nt
112
.73
�1
01
02
.7
13
De
ve
lop
me
nt
de
nsi
tie
sm
ix4
9.5
Hig
h5
Lan
du
seIT
S7
0.9
Hig
h2
Ma
na
ge
me
nt
12
0.3
6�
20
10
0.4
14
Lig
ht-
rail
syst
em
s4
3.6
Hig
h6
Infr
ast
ruct
ure
UT
C6
3.6
Me
diu
m3
Ma
na
ge
me
nt
10
7.2
8�
10
97
.3
15
UT
C6
3.6
Me
diu
m3
Ma
na
ge
me
nt
Pri
vate
pa
rkin
gch
arg
es
38
.2N
eu
tra
l1
0P
rici
ng
10
1.8
2�
59
6.8
C. Kelly et al. / Transport Policy 15 (2008) 361–371370
Table 3 shows the top 15 packages generated in this way, andillustrates the impact of the (negative) synergy scores. It showsthat based on the preferences of this user the policy package thatscores the best is the combination of intelligent transport systemsand road pricing. This output can then be saved and comparedwith the results of other preferences.
6. Testing the product and further development
The developments described in Sections 4 and 5 are complete.The option generator within KonSULT is available free to view anduse from webpage http://www.konsult.leeds.ac.uk/new/private/level2/filter.php. The policy package development tool is availableon request from [email protected].
The option generation tool has been tested on a number oflevels. The detailed database that backs up the KonSULT tool hasbeen tested at the Institute for Transport Studies by a number ofresearchers and the programmer to ensure that the tool uses thecorrect formula and inputs. In addition to this the tool has beentested by a number of academics and local authority officers atvarious workshops linked to the project to determine whether therelative scores for the policy instruments appear consistent withthe criteria that are input. The general impression gained fromexpert users is that the results produced are consistent with theexperts’ expectations and knowledge.
Feedback from transport professionals has confirmed that thisis a potentially useful strategy option development tool. It is, ofcourse, only as effective as the information contained in theKonSULT knowledgebase itself, and continued efforts are neededto populate this with more policy instruments, and to keep thosealready included up to date as new case studies emerge. As part ofthis process there will be benefit in testing further the scoresincluded in KonSULT, and assessing their transferability todifferent contexts.
However, further development needs to focus on the policypackage tool (outlined in Section 5). The feedback from transportprofessionals stressed it needed to be incorporated as part of theKonSULT knowledgebase, so that it offers a more continuousprocess. One way of doing this will be to allow the user to identifythose policy instruments that would best complement a giveninitially selected policy instrument. This would, for example, be ofvalue in responding to the UK Department for Transport’sTransport Innovation Fund (DfT, 2006), which provides supportfor strategies that include congestion charging or related price-based methods for reducing car use in urban areas. Such anapproach could usefully expand on the current facility, whichidentifies complementary instruments for each selected policyinstrument.
One of the key reasons for considering packages of instrumentsand developing the pre-existing tool is that it is very rare thatpolicy instruments are considered in isolation, or even in pairs. Agrowing body of research is considering the effects that policyinstruments have when combined together. A further develop-ment will thus be to consider what the effects are when morethan two policy instruments are combined. To support this moreempirical evidence is needed of the effects of combinations ofpolicy instruments.
In 2009 all English local authorities will be issued withguidance on developing a third round of Local TransportPlans for the period 2011–2016. In preparing its guidance, theDepartment for Transport is already aware that option generationhas been a weak link in the second round of plans (DfT, 2004)and more generally (Eddington, 2006). It is anticipated that itwill take the opportunity to promote enhanced methods forgeneration.
ARTICLE IN PRESS
C. Kelly et al. / Transport Policy 15 (2008) 361–371 371
However, it would be wrong to suggest that option generationtools of the kind described here replace the need for professionaland political judgment in developing effective strategies. Theissues of public acceptability, affordability and feasibility will stillbe critical. What these tools offer is a stimulus to consider a widerrange of policy instruments, and ones that might prima facie bemore appropriate to the challenge faced than those developedsolely from professional judgment and past experience. KonSULTgoes further in providing objective information on each suggestedpolicy instrument, which can be used in advising politicians andin consulting stakeholders and interest groups. The next stage inresearch must be to test these tools as potential facilitators of thestrategy formulation process.
Acknowledgements
We acknowledge the support of the UK Engineering andPhysical Sciences Research Council in funding DISTILLATE, and theinput of our research colleagues and local authority partners. Weare particularly grateful for the contributions of Andrew Clark andGraham Tanner of the University of Westminster and SarahGawthorpe and Tony Horrobin of the University of Leeds to thedevelopment of the KonSULT option generation tool. KonSULTitself was developed with support from the UK Engineering andPhysical Sciences Research Council, the UK Department forTransport, the Rees Jeffreys Road Fund and a number of specificresearch grants led by the University of Leeds with inputs frommany international partners, whose support we acknowledge. Wetake sole responsibility for the opinions expressed in this paper.
References
ARTISTS, 2006. /www.tft.lth.se/artists/S.Atkins, W.S., 2007. Long term process and impact evaluation of the Local Transport
Plan policy: final report. Department for Transport, London.Babalik-Sutcliffe, E., 2002. Urban rail systems: analysis of the factors behind
success. Transport Reviews 22 (4), 415–447.Brannigan, C., Paulley, N., 2009. Funding for local authority transport and land use
schemes in the UK. Transport Policy, submitted for publication.de Bono, E., 1985. Six Thinking Hats. Little, Brown and Company, Boston.de Bono, E., 2006. /www.edwdebono.comS.Department for Transport, 2004. Full Guidance on Local Transport Plans.
Department for Transport, London.Department for Transport, 2006. The Transport Innovation Fund. Department for
Transport, London.Directorate General of the Environment, 2005. Expert working group on
sustainable urban transport plans. DGEnv, Brussels.
Eddington, R., 2006. The Eddington Transport Study: transport’s role in sustainingthe UK’s productivity and competitiveness. Department for Transport,London.
European Conference of Ministers of Transport, 2002. Implementing SustainableTransport Policies: Key Messages to Governments. OECD, Paris.
Hull, A., Tricker, R., 2005. An assessment of the barriers to the delivery ofsustainable local surface transport solutions. In: Proceedings of the EuropeanTransport Conference. PTRC, London.
Hoinville, G., 1971. Evaluating community preferences. Environment and Planning3, 33–50.
Jones, P., Lucas, K., 2006. DISTILLATE Project B Option Generation. /http://www.distillate.ac.uk/reportsS.
Jopson, A., May, A.D., Matthews, B., 2004. Facilitating evidence based decision-making—the development and use of an on-line knowledgebase on sustain-able land-use and transport. In: Proceedings of the 10th World Conference onTransport Research, Istanbul.
Jopson, A.F., Menaz, B., Page, M., 2007. Appraisal: a barrier to sustainabletransport? In: Proceedings of the 11th World Conference on TransportResearch, Berkeley.
Kelly, C., May, A., Jopson, A., Gawthorpe, S., Cook, A., Tanner, G., 2008. A KonSULT-based strategy option generator. Final product and tool report. /www.dis-tillate.ac.ukS.
Kocak, N., Jones, P., Whibley, D., 2005. Tools for road user charging scheme optiongeneration. Transport Policy 12 (5), 391–405.
KonSULT, 2008. /www.konsult.leeds.ac.ukS.Marsden, G., Kelly, C., Snell, C., 2006. Selecting indicators for strategic performance
management. Transportation Research Record 1956, 21–29.May, A.D., Still, B.J., 2000. The instruments of transport policy. Working Paper
WP545, Institute for Transport Studies, University of Leeds.May, A.D., Taylor, M.A.P., 2002. KonSULT—developing an international knowledge-
base on urban transport policy instruments. Presented at the 25th AustralasianTransport Research Forum, Canberra, October 2002.
May, A.D., et al., 2005a. A decision-maker’s guidebook. Institute for TransportStudies, Leeds. Also available at /www.konsult.leeds.ac.ukS.
May, A.D., Shepherd, S.P., Emberger, G., Ash, A., Zhang, X., Paulley, N., 2005b.Optimal land use and transport strategies: methodology and application toEuropean cities. Transportation Research Record 1924, 129–138.
May, A.D., Kelly, C., Shepherd, S.P., 2006. The principles of integration in urbantransport strategies. Transport Policy 13, 319–327.
May, A.D., Page, M., Hull, A., 2009. Developing a set of decision support tools forsustainable urban transport in the UK. Transport Policy, this issue, doi:10.1016/j.tranpol.2008.012.010.
Matthews, B., May, A.D., 2001. Initial Policy Assessment. Deliverable 4, PROSPECTS,Institute for Transport Studies, Leeds.
Shepherd, S.P., Pfaffenbichler, P., Emberger, G., 2007. Improving the capabilities anduse of strategic decision making tools. Presented at the 11th World Conferenceon Transportation Research, Berkeley, USA, 24–28th June 2007.
Sumalee, A., May, A.D., Milne, D.S., 2002. Optimal locations for road pricingcordons. In: Proceedings of the International Conference on Seamless andSustainable Transport, Singapore.
Townsend, J., Faviour, J.P., 1991. The Creative Manager’s Pocket Book. ManagementPocket Books Ltd., Alresford, Hampshire.
Victoria Transport Policy Institute, 2008. Online Transport Demand ManagementEncyclopaedia. /http://www.vtpi.org/tdmS, VTPI, Victoria.
WEBTAG, 2006. Web-based Transport Appraisal Guidance. Department forTransport, London.
Zwicky, F., 1947. Morphology and nomenclature of jet engines. Aero Engine Review,June 1947.