Visit the NZ Transport Agency website
www.nzta.govt.nz
Travel behaviour change evaluation procedures
Technical report
Print version: Travel behaviour change evaluation procedures (PDF, 1.57 MB, 98 pages)
3. Benefits
- 3.1 Introduction
- 3.2 Benefits to travel behaviour changers
- 3.3 Resource cost corrections and other benefit categories
3.1 Introduction
This section considers the benefits and disbenefits associated with TBhC projects and recommends appropriate values for each benefit category for use in New Zealand TBhC evaluations.
Most of the benefit categories identified in existing TBhC evaluations are currently included in one or more of Transfund’s evaluation procedures. Transfund has recently reviewed many of the benefit values used in its project evaluation procedures and it does not appear that any values used in other countries are sufficiently better founded than Transfund’s current values to justify a change.
Criteria that were used for assessing benefit types include:
- Relevance to TBhC projects
- Foundation and reliability of recommended values
- Applicability to New Zealand
- Data requirements
- Ease of quantification for TBhC projects
- Ease of application.
Some of the benefit categories in Transfund’s procedures demonstrate an acknowledgement that some types of projects entail special types of benefits that need to be included for those project types, eg the special benefits for sealing unsealed roads and providing passing lanes. This indicates that Transfund is likely to be receptive to including any benefit types that are specific to TBhC projects even if they are not applicable to other project types.
Benefit values are based on Transfund Project Evaluation Manual values where appropriate and also the Transfund Passenger Transport Funding report (particularly for resource cost adjustments).
Benefits are discussed in two separate sections as follows:
- Benefits to travel behaviour changers
- Resource cost corrections and other benefit categories.
3.2 Benefits to travel behaviour changers
Section 2.0 identified that the perceived benefits to TBh changers is a valid benefit to estimate and include in the evaluation of TBhC projects. This approach is consistent with theory incorporated in transportation planning modelling and in the assessment of benefits to induced traffic. This section discusses the investigations to obtain appropriate values for use in New Zealand TBhC project evaluations.
3.2.1 Winn estimate
In his 2004 ATRF paper, Winn described the method for determining the benefit to travel behaviour changers as follows:
- “For the valuation of the user benefits associated with changes in the mode of travel a logit based mode split model is used. This model form is commonly used in four stage strategic transport models to allocate trips between motorised and non-motorised modes and between public transport and motor car within the motorised category. The market share of one mode compared to another is a function of the difference in generalised cost, with a slope parameter governing the rate of change and a shift parameter included to take account of the attributes not included in the generalised cost formulation.
- Figure 1 illustrates the public transport share of motorised transport using typical parameters found in the Melbourne Integrated Transport Model (2000). Notice that where there is no difference in generalised costs public transport accounts for 35% (not 50%) of trips. This reflects the inherently superior comfort and convenience attributes of car-based travel.
- As a further illustration; if the current public transport mode share of motorised travel was 10% (a $10 difference between car and public transport costs), then a reduction in public transport costs relative to car of about $5 (500 cents) would be required to raise the public transport mode share to 20%.
- Figure 1: Public transport mode share of motorised travel
This information can be used in the following way to value the user benefits for those who switch from car to public transport for the same type of journey:
- in the example application considered (by Winn) a 15% increase in public transport trips and a consequent increase in the public transport share of motorised trips from 13% to 17% has been assumed; this is consistent with experience in Australia and internationally;
- applying the relationship shown Figure 1 suggests a change in the generalised costs of public transport compared to car of around $2 would be required to effect this shift, and
- the rule of a half should be applied to set the average benefit per user at half the full benefit (a downward sloping demand curve implies that some (marginal) car users would require only a small change in cost to switch while some would require the full $2 shift to move).
- A similar approach may be followed to value the benefits for those switching from car travel to walking and cycling. The example used in the next section assumes a 10% increase in walk trips and a 75% increase in cycling (from a low base) raising the walk/cycle share of all trips from 25% to 29%. The Melbourne Integrated Transport Model parameters suggest a change in relative generalised cost of about $1.50 is required to effect this change with an average benefit for those switching of $0.75.
- For the situation where participants change their trip making behaviour by linking trips or changing destinations there is no obvious source of information to estimate the net user benefits. It is recommended that these benefits are estimated as an effective cost reduction assuming that the trip or set of trips are identical in all other ways.”
3.2.2 New Zealand estimates
New Zealand valuations of travel behaviour changer benefits were derived using the same approach as Winn. Greater Wellington Regional Council and Auckland Regional Council were requested to provide the relevant mode choice relationships from the Wellington and Auckland integrated strategic transport models. We asked for these to be provided in the form of the proportion of those choosing public transport (and those choosing car) as a function of the difference in the (generalised) cost between car and public transport. The corresponding relationships between cycle and walking and car or slow modes/car were also requested.
We also sought the separate peak and off-peak relationships where the models included different functions for these different time periods. We asked for sufficient information to be able to determine the whole curve so that the implied benefit value could be determined from any initial mode share.
Wellington
Wellington Region provided graphs of the relationship for morning peak (7 - 9am) and inter-peak (9am - 4pm) from the Wellington Transport Strategy Model. Figure 2 shows the mode choice relationship between public transport and car mode share for the morning peak.
Figure 2: Wellington public transport/car mode choice relationship - morning peak
To produce the graphs Greater Wellington Regional Council extracted the difference in generalised cost between car and public transport for each origin-destination pair in the network. This determined the point on the horizontal axis. Origin-destination pairs with similar generalised cost difference (car minus public transport generalised cost values) were grouped together within bands of five minutes generalised cost difference. The “count” figure in the graphs represents the number of origin-destination pairs with generalised cost differences within each five minute band. The Wellington Transport Strategy Model records costs in terms of equivalent minutes of travel time. Because most of the calibration and application of the model involves analysis of travel times it is more convenient for the model to work this way than for it to record all costs directly in dollars. This is a common transport modelling approach.
The public transport share of trips for each origin-destination pair was determined from household travel survey data. The average public transport mode share of all the origin-destination pairs in each band was calculated to determine the point on the vertical axis for that generalised cost difference band. The resulting points form a fairly well defined relationship in the range covered by most of the origin-destination pairs. More than 99% of public transport trips are between origins and destinations represented in the graphs. Intrazonal trips and trips to/from external zones have been excluded.
The public transport generalised cost includes various weights on components of the public transport trip, such as a factor of two for access/egress and wait times whereas the car generalised cost does not have any weightings applied (it could be argued for instance that congested travel time has a higher perceived cost than uncongested travel time and should be weighted accordingly). This could partly explain why people choose public transport even when the generalised cost difference in the graph appears to strongly favour car. Note that adjusting for this might shift the curve to the right but would not necessarily change the slope which is what is used to determine the behaviour changer benefit.
From Figure 2, over the range of 8% to 15% public transport mode share (which accounts for the majority of origin-destination pairs), a one percentage point change in mode share corresponds to approximately 4.2 minutes change in generalised cost difference. Based on a $7.96/hour value of time savings for car driver (2004 $) this gives an inferred cost difference of 56 cents, which gives a behaviour changer benefit of 28 cents after applying the rule of half. A four percentage point change in mode share as assumed in the example by Winn would equate to a behaviour changer benefit of $1.12. It is not clear whether the calculation should be based on the car driver value of time or an average of the car driver and public passenger values of time. The latter would result in a slightly lower benefit value.
Figure 3 shows the mode choice relationship between public transport and car mode share for the interpeak period. A small number of outlier values at each end of the horizontal axis were omitted because they unduly influenced the mode choice relationship curve but only represented a very small number of counts. Ideally the mode share values should probably be weighted by the number of counts when fitting the curve.
Figure 3: Wellington public transport/car mode choice relationship - inter-peak
In Figure 3, for a three percentage point change in public transport mode share between four and seven percent (which covers the majority of origin destination pairs), the change in generalised cost difference is approximately 68 minutes (103-35). Therefore a one percentage point change in mode share corresponds to approximately 22.7 minutes change in generalised cost difference (68/3). This implies a behaviour changer benefit of $1.50 (0.5 x $7.96 x 22.7/60) which appears too high when compared with the peak period benefit of 28 cents. Intuitively it might be expected that the benefit from changing would be greater in the peak than the off-peak. It is possible that the inter-peak relationship is being influenced by a base proportion of trips that have no choice but to use public transport. If these could be excluded the curve might be considerably flatter. Another consideration is that a one percentage point increase in public transport mode share from 4% to 5% is a 25% increase, which is quite high and possibly could be expected to correspond to a high benefit. Nevertheless it is concluded that the peak period benefit value is likely to be more robust and is used to determine the recommended values below.
Auckland
Instead of completed graphs, Auckland Regional Council provided model relationships as formulae covering:
- The probability of travel by a particular mode as an exponential function of the utility of travel by that mode and other modes
- Relationship between utility and generalised cost
- Generalised cost of different modes and time periods as a function of time, distance, parking charges, and fares (these were derived from network skims, giving cost and distance between each pair of zones for each mode).
Maunsell developed a spreadsheet model incorporating these formulae. The mode choice relationship indicated by this model for different car generalised costs is shown in Figure 4.
Figure 4: Auckland public transport/car mode choice relationship - peak
The information provided by Auckland Regional Council did not enable derivation of a unique car-public transport mode choice curve. In the absence of information about the overall average generalised cost for all car trips the $8.00 car generalised cost curve (800 cents) was selected as being representative. Based on this curve a four percentage point change in public transport mode share from 13% to 17% corresponds to approximately 150 cents change in generalised cost difference. Applying the rule of a half gives a behaviour changer benefit of 75 cents for a four percentage point mode share change, which is slightly less than the benefit derived for Wellington.
Based on the formulae provided by Auckland, the relationship for off-peak has a similar gradient to peak and therefore produces similar behaviour changer benefit values.
Conclusion
The higher behaviour changer benefit value obtained for Wellington peak mode changers compared with Auckland might reflect higher attractiveness of trains (in Wellington) compared with buses (in Auckland). On the other hand it might simply be due to variations in the method of deriving the benefit, which is subject to some uncertainty. The process to derive the mode choice relationship for Auckland was less direct than for Wellington and therefore greater weight should probably be placed on the Wellington result.
Table 3 shows the proposed benefit values to be used for travel behaviour changers following the above analysis. These values have been rounded down slightly from the value derived from the Wellington peak period mode choice relationship reflecting the degree of approximation in the methodology for deriving these values.
Table 3: Proposed travel behaviour changer benefit values ($ per trip)
| Mode change | $/trip |
|---|---|
| Car driver to public transport | |
| 1 percentage point mode change | $0.25 |
| 2 percentage point mode change | $0.50 |
| 4 percentage point mode change | $1.00 |
| Car driver to cycle/walk | |
| 1 percentage point mode change | $0.25 |
| 2 percentage point mode change | $0.50 |
| 4 percentage point mode change | $1.00 |
It is proposed to use the same benefit values for peak and off-peak periods because the Auckland model relationships produced little difference and the Wellington results for off-peak were higher than peak but not considered sufficiently reliable.
New Zealand model relationships for mode change from car driver to cycle/walk were not analysed. The benefit value for changers from car to cycle/walk, obtained by Winn from the Melbourne Integrated Transport Model, was less than the benefit for changes from car to public transport. Nevertheless, it was considered that there was insufficient evidence to adopt different values for car driver to cycle/walk than car driver to public transport in New Zealand at this stage.
Consideration was given to whether different benefit values may apply in cities with less extensive public transport choices but again evidence was not available to confirm or refute this. There appears little reason for the values for people changing from car driver to cycle/walk to vary by location as the perceived benefits are likely to be similar wherever they occur.
The method and detail of obtaining benefit values from mode choice relationships is still evolving. A number of issues need further consideration and investigation.
- It is likely that the mode choice relationship will change as a result of the TBhC project itself, ie the position and slope of curve may change from that represented in the strategic transport model and therefore indicate different benefits.
- We might not be measuring the slope at a point on the curve that is representative of TBh changers. Changers might tend to come mostly from a particular (flatter or steeper) part of the curve.
- When different values of travel time exist for the “from” and “to” modes, as in New Zealand, it is not clear whether the benefit should be based on the from or to mode value of time or an average of the two.
- More information is needed on how the mode choice models are constructed. The statistical validity of the curve will depend on how much detailed survey data on actual travel characteristics and choices has been used to calibrate the model.
- There is also a need to research benefit values for car driver to car passenger and car passenger to environmentally friendly modes.
In addition to the model related questions, if TBhC projects are successful in New Zealand and investment in such projects expands it may also be worthwhile to undertake market research surveys to determine more directly the factors that affect TBh changers’ perceived benefits.
3.3 Resource cost corrections and other benefit categories
It is important to note that because the evaluation framework is based on perceived benefits/dis-benefits and resource cost corrections plus external benefits, that a number of benefits will have different values than if they were expressed as straight resource values as in the Project Evaluation Manual. These differences are discussed in each section below.
The general principle with perceived benefits/disbenefits is that travellers will fully perceive time and comfort aspects and out of pocket costs such as fuel, parking charges, and public transport fares. These aspects/costs are taken into account in their choice of mode and hence in the benefit to TBh changers derived in the previous section. However in many cases the out of pocket costs are not an accurate reflection of resource costs, hence the need for resource cost corrections. Figure 5 shows the breakdown of costs associated with a car trip into perceived and unperceived components, including externalities.
Figure 5: Categories of costs associated with a car trip
Normally car drivers only consider the internal perceived costs described above and shown in dark blue in Figure 5. Other internal costs such as non-fuel variable vehicle operating costs, and accident costs are considered to be unperceived as shown by the light blue or dimension A in Figure 5. Externality costs such as environmental effects are also generally considered to be unperceived.
One of the effects of a TBhC project is to provide travellers with information that changes their perceptions of costs of different modes. This is illustrated by the two scenarios on the right hand side of Figure 5. The first scenario shows the situation if the TBhC programme corrects a proportion of the internal unperceived costs. The second shows the situation where all internal costs are perceived as a result of the TBhC project. In some situations TBhC projects may even make people aware of some of the externalities and take these into account in their travel choices but such effects are considered unlikely to be as durable as changes in the perceived internal costs and are ignored.
The change in perceived benefits/disbenefits resulting from the TBhC project causes people to make the travel behaviour change as they now perceive the cost of making the trip by car as being higher than the alternative. This is shown in Figure 6.
Figure 6: Change in perceived costs resulting from TBhC project
In the situation without the TBhC project Figure 6 shows that for a particular individual the perceived costs of travel by public transport are greater than by car so car is the preferred mode. The TBhC project causes the individual to become aware of a greater proportion of the costs of car travel and as a result the perceived costs of a car trip now exceed those of undertaking the trip by public transport and public transport becomes the preferred mode. The difference between the car total cost and public transport total cost represents the benefit of this behaviour change. Some of this accrues to the behaviour changer as savings in perceived and unperceived internal costs and some to society due to the lower externality costs associated with a public transport trip compared with a car trip. Note that there is no actual change in the total cost of either the car or public transport trip but that the behaviour change and resulting benefits arise solely from the change in perceived costs brought about by the TBhC project.
The net benefit to TBh changers that was derived in the previous section incorporates the perceived benefit of avoiding the car trip. Dimension B in Figure 5 shows the remaining unperceived internal costs of the car trip following the TBhC project. This is the required resource cost correction that is counted as a benefit in addition to the net perceived benefit if a car trip is removed by the TBhC project. As shown in the second scenario in Figure 5, the TBhC project may reduce this to zero. The next section discusses the estimated values of resource cost corrections that are proposed for TBhC project evaluations while the section after that discusses externality benefits and other unquantified benefits of TBhC projects.
3.3.1 Resource cost corrections
This section discusses the benefit and disbenefit components that are internal to (ie directly affect) TBh changers but that may not be fully perceived and included in the net benefit to TBh changers covered in the previous section.
Travel time savings
Travel time savings (or increases) do not need to be assessed directly when using the recommended perceived cost approach because travel time changes and related impacts are considered to be fully internalised in the perceived net benefit to behaviour changers estimated in the previous section. This includes effects such as differences in travel time by different modes, differences in the value of that time, other time costs such as waiting, transfers, changing etc, and trip time reliability. All of these tend to be quickly taken into account by users based on their experience and directly influence their mode choice and other travel behaviour decisions.
This is a key difference between the perceived cost approach and the alternative “difference in resource cost” approach used by Ker as discussed in Section 2.0. Proper application of the “difference in resource cost” approach would require the estimation of all changes in travel times on each mode for all TBh changers and the application of the appropriate mode specific values of travel time from Table 2.
Private vehicle operating costs
A proportion of vehicle operating costs are perceived by motorists and therefore taken into account in the net benefit to TBh changers. However it is normally considered that users only perceive the fuel component of vehicle operating costs and therefore that a resource cost correction is required for the difference between this and total resource cost avoided as a result of their mode change. Table A5.1 of the Project Evaluation Manual shows that the average value for car vehicle operating costs is approximately 16 cents/km. Project Evaluation Manual Table A5(a) shows that fuel and oil accounts for 30% of vehicle operating cost, which would equate to 4.8 cents of the 16 cents/km. The perceived cost is actually higher than this because of fuel taxes, which are part of perceived cost but are not resources costs.
Research indicates that the average fuel consumption of the New Zealand car fleet is approximately 9 litres per 100km. Assuming a market price for petrol of $1.00/litre (as may have applied when vehicle operating costs were last updated and when Auckland and Wellington transport models calibrated) implies a perceived fuel cost of 9 cents/km. Based on the assumption that only fuel cost is perceived, the required resource cost correction would be 7 cents/km (16 - 9).
One of the objectives of TBhC projects is to provide information that corrects peoples’ misperceptions of the costs of private car use. Winn assumed that TBhC programmes are successful in this and than consequently the required resource cost correction for vehicle operating costs in TBhC evaluations is zero. For TBhC project evaluations we have assumed that the TBhC project will provide sufficient information to make users aware of the difference between their perceived vehicle operating cost (9 cents/km) and resource cost (16 cents/km) and that the resource cost correction is therefore zero. Sensitivity tests were conducted with a 7 cents/km resource cost correction (the maximum value that would apply if the TBhC project had no effect on existing perceived vehicle operating cost.
Cycle operating costs
A Transfund paper derived a resource cost value of 10 cents/km for the variable component of cycle operating costs but noted that this is very much an approximation dependent on assumptions about annual usage. This value appears high compared with the values recommended by Ker for use in Australian evaluations (Aust 2.5 cents/km). We have therefore assumed a New Zealand resource cost value of 5 cents/km.
People changing to cycle are likely to be aware of the probable incremental cycle costs. However it could be argued that, as with cars, people do not fully take into account infrequent costs such as tyres and maintenance. The TBhC benefit values will assume that cycle costs are correctly perceived and taken into account in the net benefit to TBh changers and therefore the resource cost correction is zero. Sensitivity testing is conducted using a resource cost correction of 5 cents/km (which counts as a disbenefit) to assess the implications of people not perceiving cycle operating costs. The value of cycle costs may need further research if results are found to be sensitive to this.
Walking costs
In theory these would be treated the same as cycle costs but are ignored because they are likely to be small and because in any case they are likely to be correctly perceived and taken into account in the net benefit to TBh changers.
Accident costs — car
Accident costs can be considered in three parts: internal costs (ie affecting the behaviour changer) that are perceived and hence included in the net benefit to TBh changer, internal costs that are not perceived and hence require resource cost correction (discussed here), and externality costs borne by others/society covered in the next section.
There is little information available on the extent to which people perceive accident risks and costs and take this into account in their travel choices. If it is considered that people take full account of the accident risk and costs to themselves then this is already included in the net benefit to TBh changers and only the externality costs need to be added. However if people under-perceive their accident costs a resource cost correction is required. Intuitively it is likely that people do not perceive much accident cost and therefore most if not all of their internal accident cost will require a resource cost correction. If this is the case the resource cost correction plus the externality costs will equal the total resource value of accident costs. Accident resource cost values are available and it is therefore considered that these should be used for the combined value of resource cost correction and externality. The recommended values are discussed in the externality section below.
A sensitivity test is conducted to assess the effect on the results if half of the resource cost of accidents (internal and external) was actually perceived and therefore included in the net benefits to TBh changers.
Accident costs — cycle/walk
The same considerations apply in relation to cycle and walking accident costs as for car accident costs and are also discussed in the relevant externality section below. Two additional considerations with cycle and walk accident costs are that TBh changers to walking and in particular cycling probably have a fairly clear perception of the associated accident risk (so possibly some of it is included in the net benefits to TBh changers), and that an increase in the number of pedestrians and cyclists might actually lead to a fall in the average per kilometre accident cost per pedestrian or cyclist (referred to as the “safety in numbers effect”). This is also discussed further in the externality section below.
Health (fitness) benefits of cycling and walking
Health benefits of cycling and walking are considered in three components in the same way as accident costs but they are of opposite sign (ie a benefit rather than a dis-benefit). However unlike accident costs it seems implausible to expect that none of these benefits are perceived by TBh changers and included in the net benefit to TBh changers. One of the main selling points of TBhC projects that causes people to take up cycling or walking is the promotion of the health benefits.
As an approximation it is proposed that half of the total health benefits are internal to the TBh changer and are included in the net benefit to TBh changers. The other half are considered to be internal unperceived benefits to be covered by a resource cost correction or externality benefits to society (eg avoided hospital and other health care costs). For convenience both of these categories are considered together below in the externalities section.
Car parking
Reduced car usage results in a reduction in the demand for parking facilities. The resource costs of car parking include the opportunity cost of using land for parking, the capital cost of parking facilities, and the provision of adequate security.
Motorists are charged a fee for the use of parking. This charge differs depending on the destination of a journey and the time of day that the journey is made. Parking costs are higher for trips to the CBD and during peak times, and lower for off peak and non-CBD trips. The average parking fee paid by car users is generally less than the resource cost of providing parking. TBh changers are likely to consider only the parking fee that they actually save in their TBh changer net benefit. People are assumed not to be aware of or take account of any difference between the fee they paid and the actual resource cost of providing the parking facilities.
The Transfund Passenger Transport Funding and Evaluation Procedures (Stage 2) report, and other references, note that a resource cost correction is required for the difference between the parking fee and the resource cost of parking. If parking charges reflect full costs of providing parking then no resource cost correction is required. Appendix D of the Passenger Transport Funding final report provides estimates of the parking resource cost correction for different cities as shown in the third row in Table 4.
Table 4: Parking costs and resource cost corrections (dollars per car round trip)
| Peak period commuting trips to: | Off peak trips to: | |||
|---|---|---|---|---|
| Auckland CBD | Wellington CBD | All other destinations | All destinations | |
| Resource cost | $10.00 | $10.00 | $2.00 | $0.25 |
| Average parking fee | $2.50 | $3.60 | $0.00 | $0.25 |
| Resource cost correction | $7.50 | $6.30 | $2.00 | $0.00 |
| Resource cost correction after TBhC project | $2.50 | $2.50 | $0.50 | $0.00 |
The resource cost corrections shown in Table 4 are per round trip so they are halved to obtain the value per one-way trip, eg commuting trips to or from CBD.
The resource cost correction values used in the passenger transport funding policy (third row in Table 4) are quite large relative to most other benefit categories being considered for TBhC projects, particularly in the case of Auckland and Wellington commuting trips. One of the key strategies included as part of TBhC projects are measures to make car users more aware of the full costs of car parking so it is assumed that TBhC projects will reduce the misperception of car-parking resource cost and hence the required resource cost correction will be less than without the TBhC project. We have assumed that as a result of the TBhC projects car users will perceive 75% of the resource cost of parking and so the required resource cost correction is 25% of the resource cost. These resource cost correction values are shown in the bottom row of Table 4.
Public transport fares
A resource cost correction is also required for public transport fares. This is because fares are a financial transfer (from TBh changer to public transport operator) rather than an actual resource cost. This is regardless of whether the public transport provider incurs additional costs to provide additional services. However the fares are perceived as a real cost by the TBh changer and therefore form part of their assessment of net benefit to TBh changer. The person changing to public transport perceives fares as a cost but they are not a resource cost so it is necessary to make a resource cost correction by adding back (as a benefit) the (tax inclusive) amount of fare. Tax inclusive fare is used because this is the cost that the TBh changer perceives.
Average fares from Transfund Public Transport Statistics 2003/04, with GST included, for the main centres are:
• Auckland |
$1.83 |
• Wellington |
$2.15 |
• Christchurch |
$1.10 (assumed to be $1.50 for benefit calculation) |
These values are included as benefits for each additional public transport trip resulting from the TBhC project. For school related trips we use half the above values. A value of $1.50 is used for Christchurch trips because Environment Canterbury noted that the minimum fare for a student is 75 cents whereas the reported $1.10 value, when halved, would result in a value of 55 cents for school related trips.
3.3.2 Externality benefits
In addition to the internal perceived and unperceived benefits and disbenefits to TBh changers, TBhC projects also result in externality effects on other transport system users and on society. These are discussed in the following paragraphs.
Congestion
This is the reduced congestion costs (time and vehicle operating cost) experienced by remaining road users due to removal of a marginal vehicle - it does not include the saving to the TBhC changers themselves as this is part of their internalised benefit.
Estimates of decongestion benefits have been obtained from the Transfund Passenger Transport Funding and Evaluation Procedures (Stage 2) report, Table 2.5. The average marginal travel time saving to remaining road users per vehicle-km removed is Auckland $1.190/km, Wellington $0.911/km, Christchurch $0.085/km. Units are stated in the Transfund report as being per diverted passenger-km but in the derivation it appears to be based on removal of vehicle-km. We have taken it as a benefit per vehicle-km removed or per car driver-km removed. There is also a small decongestion benefit from diversion of car passengers because they cause additional vehicle kilometres in some cases.
The Transfund Passenger Transport Funding Report also notes that the analyses to derive the above values were undertaken for each centre in totality, to provide average peak period decongestion rates for the three centres. However in principle they could similarly be carried out for each corridor (or even individual bottlenecks) if payment rates disaggregated by corridor were required. Similar analysis could also be used to derive corridor or area specific benefit values for TBhC projects but this is probably not warranted in most cases. The values for Christchurch may justify further investigation as they are perhaps more significantly lower than those for Auckland and Wellington than might be expected.
Decongestion also includes the savings in vehicle operating costs to the remaining road users per vehicle-km removed. The Transfund Public Transport Funding Report Appendix F recommends that the vehicle operating cost externality benefits be assessed at 7% of the total travel time decongestion benefits. In other words, the travel time saving values above are factored up by 1.07 to account for the vehicle operating cost externality saving.
This results in the following total decongestion benefit values:
• |
Peak: |
Auckland |
$1.273/km |
Wellington |
$0.975/km |
||
Christchurch |
$0.091/km |
||
Other |
$0.00/km |
||
• |
Off peak: |
All regions |
$0.00/km |
From the Transfund Passenger Transport Funding report, Appendix E, it appears that value of time used to derive decongestion benefits is based on July 1998 Project Evaluation Manual values. The September 2002 values in Table A4.3 of the Project Evaluation Manual are slightly (5%) higher for base peak period travel time values and lower for the congestion increment so the overall difference is minimal and it is not considered necessary to update these values.
Private vehicle operating costs
As discussed above, this is the vehicle operating cost savings to remaining road users due to the decongestion resulting from the removal of a marginal vehicle. The Transfund Passenger Transport Funding Report Appendix F recommends that the vehicle operating cost externality benefits be calculated as 7% of the total travel time decongestion benefits. The decongestion benefits above incorporate this benefit.
Induced traffic demand
The reduction in congestion resulting from TBhC projects is likely to make car travel more appealing for other potential road users, leading to increases in car use by other individuals which has the effect of partially reducing the first round decongestion benefit. The Transfund Passenger Transport Funding Report Appendix J concludes that this induced traffic effect should be valued as a disbenefit equivalent to 50% of the decongestion benefit and the same valuation is proposed for TBhC projects.
Table 5 shows the net effect of decongestion savings (travel time and vehicle operating cost) and offsetting induced traffic effects based on this factor and the average marginal congestion cost per vehicle-km reported in Table 2.5 of the Transfund Passenger Transport Funding Report.
Table 5: Peak period marginal decongestion benefits after adjustment for induced traffic effect
| Auckland | Wellington | Christchurch | |
|---|---|---|---|
| Average marginal cost/vehicle-km | $1.190 | $0.911 | $0.085 |
| VOC factor | 1.07 | 1.07 | 1.07 |
| Average cost including VOC factor | $1.273 | $0.975 | $0.091 |
| Induced traffic factor | 0.50 | 0.50 | 0.50 |
| Net marginal decongestion benefit/veh-km | $0.637 | $0.488 | $0.046 |
Road systems
Road system benefits include the benefit of reduced road maintenance and deferral of road capacity increases. It would be valid to include road maintenance savings but these are negligible for the numbers of car trips and/or car vehicle kilometres that are likely to be removed by TBhC projects. Deferral of capacity is not included because it is also considered to be negligible. Furthermore it would not be correct to include both the value of deferring improvements and the full decongestion benefit discussed above. If the capacity improvements were undertaken rather than deferred, the congestion levels would be less and the decongestion benefit theoretically slightly lower.
Public transport operating cost
Public transport operating costs would need to be included if the increase in demand resulting from a TBhC project was sufficiently great to require the operation of additional services. It is assumed that for off-peak travel there is generally spare capacity to handle the likely mode changes to public transport resulting from TBhC projects in New Zealand and that there will be no additional public transport operating costs associated at off peak times. Some larger scale household-based projects may result in greater increases in public transport demand but these tend mostly to influence off-peak trips when spare capacity is greatest.
For peak period trips it is considered that increases in patronage may lead to marginal increases in operating costs (given that existing services are at capacity in peak periods). Therefore in principle, additional public transport operating costs should be included for new peak period public transport trips resulting from TBhC projects. They have been included in the interim evaluation procedures dated 16 November 2004 but as a disbenefit (numerator), rather than as a cost (in the denominator) which would be more appropriate.
However it is considered that if the demand for public transport increases by sufficient to require additional public transport services, these would probably be evaluated by an ‘Alternatives to Roading’ or patronage funding assessment undertaken as part of a composite evaluation with the TBhC evaluation. In this case the additional public transport operating costs would be included as a cost in this parallel evaluation and should be ignored in the TBhC evaluation. If the increase in peak period public transport demand from the TBhC project is only small the additional costs could also be ignored. It is recommended that when the interim evaluation procedures are reviewed, consideration should be given to resetting additional peak period public transport operating costs to zero if it has been common to account for these in an ‘Alternatives to Roading’ or patronage funding assessment undertaken as part of a composite evaluation with the TBhC evaluation.
Public transport Mohring effect
In theory the Mohring effect may be a benefit of TBhC projects but it could only occur if the behaviour change was sufficient to require substantially more frequent additional public transport services. It is likely that this will only occur in the same situations where additional public transport operating costs need to be allowed for, ie where there is substantially increased public transport demand in peak periods. This is only likely to be the case with workplace travel plans for some large workplaces or clusters of workplaces. As discussed above, in this situation increased public transport operating costs are likely to be accounted for in the parallel ‘Alternatives to Roading’ or patronage funding assessment undertaken as part of a composite evaluation with the TBhC evaluation and the same consideration would apply to the Mohring effect. The Mohring effect benefit has therefore been set at zero in the interim TBhC evaluation procedure.
Accident costs — car
Externality accident costs are those imposed on other road users and society (hospital, lost productivity etc). The Transfund Passenger Transport Funding report contains peak and off-peak marginal accident cost values for Car as Driver and Passenger and Taxi. The marginal cost for car as driver is 1.8 cents/km for peak period travel and 2.9 cents/km for off-peak travel. The report notes that these are total accident resource costs that include both externality and internalised costs.
Booz Allen Hamilton has also done more recent research for the Ministry of Transport as part of the Surface Transport Costs and Charges Study (unpublished) that estimates peak urban marginal accident costs at -6 cents/km of which -3.0 cents/km is externality, and off-peak urban marginal accident costs at +7 cents/km of which +2.3 cents/km is externality. The negative marginal cost in peak periods reflects the “traffic calming” effect of more traffic in congested conditions. This means that accident costs actually increase with removal of a marginal vehicle-km. However, as this research has not been formally accepted and published it was considered that the TBhC evaluation should use the same values as in other Transfund policies. At such time as the values in the Surface Transport Costs and Charges Study are officially accepted it is expected that they would be incorporated in all relevant Transfund evaluation procedures at the same time.
The current marginal resource cost values in the first paragraph will be correct to use as the total accident benefit if we consider that the net benefit to TBh changers does not include any perception of accident risk or cost avoided. If this is the case the marginal resource costs will be equal to the total of the resource cost correction and the externality costs. Intuitively it is likely that people do not perceive much accident cost so this approach is likely to be reasonable and consequently the above values are proposed for accident cost savings.
The Surface Transport Costs and Charges Study estimates that 50 - 66% of accident resource costs are internal so it is theoretically possible that up to this amount could be perceived and incorporated in the net benefit to TBh changers. This possibility is considered less likely but a sensitivity test was conducted to assess the effect on the results if half of the resource cost of accidents was actually perceived and therefore included in the net benefits to TBh changers. This would mean that only half of the above benefit values could be included as additional resource cost correction and externality benefits.
As recommended by the Transfund Passenger Transport Funding report Car as Passenger accident costs are estimated at 1.8 times Car driver reflecting assumptions about some car trips being undertaken specifically for the passenger and others where the passenger goes to the same place as the car driver.
Accident costs — cycle/walk
The Transfund Passenger Transport Funding Report provides the following values for cycle and walk accidents:
• |
Cycle: |
Peak |
$0.336/km |
Off-peak |
$0.435/km |
||
• |
Walk: |
Peak |
$0.176/km |
Off-peak |
$0.189/km |
The report notes that these are total values that include both externality and internalised costs. These accident costs are quite high compared with car accident costs and could be considered to disadvantage cycling and walking. The Technical Working Group has suggested that it should be assumed that these costs are offset in the longer term by the traffic calming effect of increased numbers of cyclists and pedestrians (the safety in numbers effect) and hence should be assumed to be zero.
As noted earlier, TBh changers to walking and in particular cycling probably have some perception of the associated accident risk so possibly some of it is included in the net benefits to TBh changers. If their internal costs are perceived and already included in net benefit to TBh changer the required resource cost correction is zero. Therefore rather than having to assume that the entire accident resource cost has to be offset by the safety in numbers effect only the externality component has to be offset by safety in numbers effect and this is a less demanding assumption. Based on these assumptions cycle and walk accident costs are assumed to be zero.
The above assumptions are convenient but may be considered slightly optimistic. Sensitivity testing was undertaken based on half of the above total cost accident resource cost not being perceived or offset by safety in numbers and hence having to be included as a cost of increased cycle and walk trips.
Environmental
Environmental externalities are estimated as a total value of all effects including local air, noise, and water pollution, and greenhouse gas emissions. Environmental benefits are obtained from the Transfund Passenger Transport Funding report, which estimated values from a review of New Zealand and international evidence. Australian evaluations to date have tended to use an average cost for environmental externalities. The review considered that it is more appropriate to use a marginal cost value and that this is estimated to be approximately twice the average of 5 cents/km obtained from the international review.
TBh changers may initially include some of this cost in their perceived benefit as a result of the information provided as part of the TBhC project but any such perception is considered unlikely to be durable. Therefore we propose to assume no long-term inclusion of any of the environmental costs in the net benefit to TBh changers and include the full value associated with reduced car and passenger km as an externality benefit.
Environmental externality costs are:
• |
Car as driver: |
Peak |
$0.10/km |
Off-peak |
$0.05/km |
||
• |
Car as passenger |
0.8 times Car as driver |
|
Health (fitness) benefits of cycling/walking
Health benefits of cycling and walking are an area of considerable uncertainty. The 1999 Beca Carter Hollings and Ferner document on which current Transfund values for health benefits of cycling are based, reported that various estimates put health benefits of cycling between 5 and 40 cents/km, and settled for 15 cents. Research in the Netherlands indicated that a 9% increase in the amount of kilometres cycled in Amsterdam results in savings on absenteeism and medical treatment equivalent to 17 cents/km. A recent evaluation undertaken by the Victorian Department of Infrastructure uses 54 cents/km Australian. The Victorian evaluation uses a value of A$1.48/km for health benefits of walking. The basis of this value is considered questionable as it is based on the cost of a death times the number of deaths considered to be due to inactivity, and a number of other assumptions. However, it does suggest that benefits could be higher and may justify further investigation.
The October 2003 Transfund Working Paper on Travel behaviour change and cycling and walking recommended a cycling health benefit of $0.16/km. The Transfund paper estimated walking benefits as 2.5 times cycling benefit or $0.40/km based on the increase in calories consumed in walking over cycling. These are total values that include the benefit to the TBh changer.
As discussed earlier, some of the above benefit may well be perceived by the TBh changer. It can be assumed that TBh changers do not take into account the savings to society (other than themselves) of their improved health - eg hospital cost savings. However a review of the derivation of health benefits suggests that most of benefit is internal to the TBh changer rather than being external so then it becomes a question of how much is perceived and not perceived as a result of the TBhC programme. We assume that half is perceived as a result of the TBhC project and hence already reflected in the net benefit to TBh changer - health benefits are one of the main selling points used in TBhC projects. Consequently the only the remaining half of the benefits are explicitly included in the evaluation as resource cost correction and/or externality benefit. The (unperceived) health benefit values are therefore:
- Walk: $0.20/km
- Cycle: $0.08/km
A sensitivity test was undertaken based on the (unlikely) scenario of none of the health benefits being perceived by the TBh changer and hence the full resource cost value of the health benefits being included as additional resource cost correction and/or externality benefits, ie Walk: $0.40/km and Cycle: $0.16/km.
Other — not quantified
A number of other potential benefits are identified in some evaluations of TBhC projects but have not been quantified to date. These include:
- Reduced community severance
- More sustainable land use/urban form
- Community cohesion
- Improved security/safety to the community
- Less dependence on fossil fuels
- Viability of local shops and businesses
- Synergy with other marketing initiatives
These impacts are generally harder to quantify and include in evaluations, but some of them may be as worthwhile as some of the other quantifiable impacts. Some of these benefit types could also apply to other types of transport initiatives and some of them can be achieved possibly more effectively by more targeted non-transport policies.
3.3.3 Summary of resource cost corrections and other benefit categories
The resource cost corrections and benefit values identified in the above sections are summarised in Table 6.
Table 6: Summary of resource cost corrections and other benefit unit values (cents/km and cents/trip)
| Peak | off-peak | |||||
|---|---|---|---|---|---|---|
| Auckland | Wellington | Chch/other | Auckland | Wellington | Chch/other | |
| Car driver (per km) | ||||||
| VOC resource cost sorrection | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 |
| Congestion externality | 127.3 | 97.5 | 9.1 | 0.0 | 0.0 | 0.0 |
| Induced traffic | -63.7 | -48.8 | -4.6 | 0.0 | 0.0 | 0.0 |
| Accident costs | 1.8 | 1.8 | 1.8 | 2.9 | 2.9 | 2.9 |
| Environmental costs | 10.0 | 10.0 | 10.0 | 5.0 | 5.0 | 5.0 |
| Total per km | 75.5 | 60.6 | 16.4 | 7.9 | 7.9 | 7.9 |
| Car driver (per trip) | ||||||
| Parking resource cost correction | ||||||
| • trips to/from CBD | 125.0 | 125.0 | 25.0 | 0.0 | 0.0 | 0.0 |
| • trips to/from other destinations | 25.0 | 25.0 | 25.0 | 0.0 | 0.0 | 0.0 |
| Car passenger (per km) | ||||||
| VOC resource cost sorrection | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 |
| Congestion externality | 101.8 | 78.0 | 7.3 | 0.0 | 0.0 | 0.0 |
| Induced traffic | -50.9 | -39.0 | -3.6 | 0.0 | 0.0 | 0.0 |
| Accident costs | 3.2 | 3.2 | 3.2 | 5.2 | 5.2 | 5.2 |
| Environmental costs | 8.0 | 8.0 | 8.0 | 4.0 | 4.0 | 4.0 |
| Total per km | 62.2 | 50.2 | 14.9 | 9.2 | 9.2 | 9.2 |
| Car passenger (per trip) | ||||||
| Parking resource cost sorrection | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 |
| Public transport passenger (per km) | ||||||
| Accident costs | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 |
| Environmental costs | 1.0 | 1.0 | 1.0 | 1.0 | 1.0 | 1.0 |
| Total per km | 1.0 | 1.0 | 1.0 | 1.0 | 1.0 | 1.0 |
| Public transport passenger (per trip) | ||||||
| Fare resource cost correction | -183.0 | -215.0 | -150.0 | -183.0 | -215.0 | -150.0 |
| Additional PT operating cost | 163.0 | 191.0 | 98.0 | 0.0 | 0.0 | 0.0 |
| Total per trip | -20.0 | -24.0 | -52.0 | -183.0 | -215.0 | -150.0 |
| Cycling (per km) | ||||||
| Cycle resource cost correction | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 |
| Accident costs | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 |
| Health effect | -8.0 | -8.0 | -8.0 | -8.0 | -8.0 | -8.0 |
| Total per km | -8.0 | -8.0 | -8.0 | -8.0 | -8.0 | -8.0 |
| Walking (per km) | ||||||
| Accident costs | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 |
| Health effect | -20.0 | -20.0 | -20.0 | -20.0 | -20.0 | -20.0 |
| Total per km | -20.0 | -20.0 | -20.0 | -20.0 | -20.0 | -20.0 |
Note: These values are costs per kilometre or per trip. They become benefits if a trip is avoided, or costs if a trip is added. A negative value indicates that the effect is a benefit of that trip, eg health effects are benefits of cycling and walking trips.
Page created: 28 October 2008