5. Estimation of Benefits

5.1 Overview

Individuals walk and cycle for recreation, for exercise or to access work, education, shopping or entertainment. Active travel gives rise to private benefits – users would not walk or cycle if they did not perceive it to beneficial. Active travel also causes social benefits, such as reductions in health system costs, noise and air pollution, and congestion.

The size and composition of initiative benefit streams will be a function of:

  • The type of initiative under consideration. An initiative to widen an existing cycle path through a park might reduce the risk of cyclist/cyclist and cyclist/pedestrian crashes but not attract new active travellers.
  • Initiative location:
    • The active travel market segment that the initiative is targeted towards. Improving the signals at a pedestrian crossing outside a school might yield benefits on school days, particularly in the morning and afternoon school peaks, but have limited impact on crash risk outside those times on school days and in general on non-school days. By comparison, extending a pedestrian/cycle path linking residential, education and employment precincts could generate benefits all week and among a broad spectrum of active travellers.
    • Building new active travel infrastructure in an area that already features high levels of active travel use might reduce crash risk if it takes active travellers out of road environments. The impact on health costs might be limited, however, because with an already high level of active travel use in the catchment, the initiative might be unlikely to significantly increase active travel.
    • Initiatives in rural towns are unlikely to generate congestion reduction benefits.

Benefits of active travel initiatives fall into three categories, described in the following sections.

5.1.1 Change in users’ willingness to pay for active travel infrastructure

Generally it will be uneconomic and impracticable to survey existing and potential active travellers about their willingness to pay. Instead, surrogate or proxy values are used, based on research into the impacts on users of active travel infrastructure improvements. The benefits that make up or are analogous to a change in willingness to pay include changes in travel time costs, vehicle operating costs and parking costs. As explained in Part T2 of the Guidelines, these costs make up the active traveller’s ‘private generalised cost of travel’, with generalised costs being calculated for the base case and project case scenarios.

The principle behind the use of these proxy benefit measures is that the active traveller would be prepared to pay at least the total value of these benefits for a given improvement in active travel infrastructure. For existing users (that is, those active travellers who will not change their trip making in response to an infrastructure improvement), the willingness to pay benefit of that improvement is equal to the difference between their base case and project case generalised cost of travel. For new users (those active travellers who would not make their trip in the absence of the proposed infrastructure improvement or who make additional active travel because of new infrastructure), the willingness to pay benefit is equal to half the difference between base case and project case generalised cost of travel[1].

This approach to estimation of willingness to pay is less meaningful if there is an infrastructure initiative with an aesthetic element or an environmental element that has a positive influence on the decision to walk or cycle. An example would be a pedestrian cycle path along a river as an alternative to on-road or in-road access. Estimation of those benefits requires more sophisticated cross sectional or discrete choice studies that will be impracticable for most projects. Recent research for Sydney (see Table 25) suggests that cyclists place a value on the opportunity to shift from on-road cycling to an off-road pathway, but it is not clear whether cyclists are valuing amenity or the reduced crash risk associated with off-road cycle paths.

5.1.2 Externalities

Externalities include:

  • Reduced health system costs because active individuals are less prone to illness and place less demand on health system resources
  • Sometimes reduced crash costs[2] if cyclists and pedestrians are diverted from the road environment
  • Reduced congestion costs if an active travel initiative causes a modal shift away from motor vehicles towards walking and cycling. Social benefits associated with new active travel are not subject to the rule of the half.

It is useful, as presented in Table 4, to distinguish between benefits associated with existing users and those associated with new users for two reasons: first, because only willingness to pay benefits for new users are subject to the ‘rule of the half’ (see Part T2 of the Guidelines); and secondly, because CBAs might be sensitive to assumptions made about the potential for an active travel initiative to generate new trips.

Table 4: Benefits according to active travel type and scale
Benefit categoryProject type/scale
  New off-road AT path – inner city location New off-road AT path – outer suburban or rural location On- road cycle lane Improved signals and channelization for cyclists/pedestrians End of trip facilities
Willingness to pay benefits for existing active travellers
Health
Vehicle operating costs
Parking
Travel time
Willingness to pay benefits for new active travellers
Health
Vehicle operating costs
Parking
Travel time
Externalities - existing active travellers
Change in crash risk
Noise reduction
Air pollution reduction
Congestion reduction
Externalities - new active travellers
Change in crash risk
Noise reduction
Air pollution reduction
Congestion reduction
Resource correction – new active travellers
Vehicle operating costs
Public transport operating costs

5.1.3 Resource adjustment/correction

The resource correction is the difference between the price paid by the user and the resource cost of a change in travel behaviour. For example, a bus user who decides to become an active traveller saves her bus fare. The resource correction in this instance would be equal to the reduction in bus system costs caused by her decision to switch to active travel less her fare saving. If this new active traveller is a former car driver, the resource correction would comprise the saving in depreciation on her vehicle, which she would be unlikely to perceive, whereas she is likely to perceive the more obvious changes in fuel, tyre and maintenance costs associated with her decision to become an active traveller.

Other parts of the Guidelines deal in detail with the resource correction. In the active travel context, however, this level of precision might not be warranted because the impacts on the overall quantum of benefits would be relatively small. Instead, it should be sufficient to rely upon changes in average car and public transport costs as measures of resource correction. For users converting from public transport, some over-counting of benefits could occur because fare savings are part of the active travel willingness to pay, but the effect would probably be relatively small once the rule of the half is applied to the fare saving.

5.2 Steps applicable to all benefit categories

5.2.1 Identifying active travel segments

Approaches to estimating the benefits of active travel initiatives differ according to the type of active traveller expected to use an initiative. The first cause of variation is the rule of the half, which is not applicable to all user segments. The segment cause of difference is ‘without project’ trip behaviour. The composition of benefits differs according to whether the initiative causes changes in trip characteristics. Benefit composition will be quite limited for ‘existing’ users whose trip characteristics do not change. For beneficiaries whose current mode is car/motorcycle travel, benefit composition will be more complex. Public transport users who convert to active travel fall in the middle of these two extremes.

Another reason to identify market segments is that not all segments will have the same trip frequency over the course of a year. For example, data described earlier (in section 2) suggests that work trips might account for only a minor proportion of walk-only and cycle-only trip making. This has potentially important implications for the estimation of initiative benefits.

5.2.2 Perceived versus unperceived costs

For all benefit categories, benefits per trip are estimated by multiplying the per km benefit in dollars by estimated trip length by number of trips per year to arrive at annual benefits. Benefits will need to be calculated separately for existing active travel trips and new active travel trips because the willingness to pay component of the latter is subject to the rule of the half.

The rule of the half only applies to those components of benefit that influence changes in travel behaviour – in other words, to the components of private generalised cost (hereafter referred to as perceived cost on the assumption that active travellers accurately perceive their own costs). The rule of the half applies therefore to the willingness to pay component of health benefits and to the vehicle operating cost, travel time cost and parking cost components of perceived cost.

Other parts the Guidelines describe the rule of half in more detail, but its logic is that the user whose change in trip behaviour is most strongly motivated by upgraded or new infrastructure perceives the full difference between their base case and project case generalised cost. The user who is very much ‘on the balance’ in choosing between their base case and project case trip making behaviours sees almost no difference between the base case and project case infrastructure and hence almost no difference in perceived cost. Assuming a linear demand curve, the average benefit is mid-way between these two extremes of perceived cost change (hence the rule of ‘half’).

For improvements to existing infrastructure

For existing active travellers using upgraded infrastructure, the benefit per traveller is the difference between their current (base case, pre-upgrade) perceived cost of using the existing infrastructure and their perceived cost of using the upgraded infrastructure. Perceived costs here include health benefits (a negative cost) and cost of additional crash risk.

The rule of half will apply to new active traveller trips, including trips diverted from other modes, totally new trips and additional trips made by active travellers who use the base case infrastructure. The benefit per traveller is half the difference between their current (base case, pre-upgrade) perceived cost of using the existing active travel infrastructure (whether travelling by car, public transport or active travel in the base case) and their perceived cost of using the upgraded active travel infrastructure.

For new infrastructure

If the new infrastructure replaces existing infrastructure, trips made by active travellers in the base case are not subject to the rule of half because the new facility does not influence their trip making; they were prepared to make their trip even on a lower standard facility. They should be treated as existing active travellers, with the benefit per traveller being the difference between their base case and project case perceived costs.

For new infrastructure, the benefit is the entire consumers’ surplus area under the demand curve. Assuming the demand curve is linear, the rule of half can be applied, but there is a problem in locating the point at which the demand curve touches the price axis.

For users of the new infrastructure who switch from travelling by car, public transport or other active travel, there is a way around the problem. It is to use the perceived cost of the travel alternative from which they are switching as the price at which the demand curve touches the price axis for each group of switchers. It is certain that they would not be willing to pay more than this for use of the new active travel infrastructure. The rule of half can be applied assuming a linear demand curve.

For users of the new infrastructure who do not travel at all in the base case, their willingness to pay must lie somewhere between perceived cost for the alternative with the lowest perceived cost (car, public transport, other active travel) and the perceived cost of using the new infrastructure. The recommended approach is to assume that the demand curve touches the price axis at the perceived cost of the cheapest alternative, and apply the rule of half. Perceived costs here include values of time, health benefits and cost of crash risk.

If the active travel trip takes longer than the alternative, it is conceivable that the perceived cost of using the new infrastructure will exceed that for the cheapest alternative for work travel, where the value of time is significant. It could not occur for non-work time (people travelling for exercise or recreation) for which a zero value of time is recommended. It could not occur if the demand forecasts were derived from a model in which demand is a function of perceived costs. If it does occur, it would be warranted to consider using a lower value of time on the grounds that it is people at the low end of the distribution of values of time who are making the switch.

5.2.3 Identifying the relevant trip length

For existing active travellers

‘Existing’ active travellers are those whose trip pattern is not expected to change when a proposed initiative is implemented. For them, the length of the proposed initiative is relevant in benefit estimation. By definition, the initiative does not change their trip behaviour. Assuming no significant difference in track characteristics (such as surface condition, width and alignment) that would materially alter trip behaviour, benefits typically will be confined to a reduction in crash risk - for example, because an off-road cycle path is provided.

Table 5: Identifying the relevant trip lengths for benefit estimation
 User type
  Existing Existing, AT more frequently(1) New active travellers Diverted from car or motorcycle Diverted from PT
Km base for estimating benefits Length of proposed initiative Total AT trip length Total AT trip length Total AT trip length Total AT trip length

(1) Includes existing active travellers who decide to walk or cycle more frequently due to provision of enhanced active travel infrastructure.

For new active travellers

For new active travellers, including those who convert from other modes, an active travel initiative has broader implications. Active travel benefits now accrue over the total trip length, not merely the portion that includes the proposed active travel initiative. It would be rare for an initiative such as dedicated pedestrian/cycle path to provide a complete self-contained link to all of the origins and destinations of all its users. Some users might need to use on-road paths or footpaths to access the new pedestrian/cycle path.

‘New’ active travellers are those walkers or cyclists who are expected to:

  • Make a walk or cycle trip because a proposed active travel initiative is implemented. This group could include people who decide to take up walking or cycling because of an improvement in active travel infrastructure (referred to as ‘generated users’)
  • Change their trip characteristics (route, origin, destination, mode or frequency) because a proposed active travel initiative is implemented. This group could include:
    • Current walkers or cyclists who change route to take advantage of improved active travel infrastructure
    • Current walkers or cyclists who walk or cycle more frequently because of an improvement in active travel infrastructure
    • Current car, motorcycle or public transport users who switch to active travel because of an active travel infrastructure improvement.

Locally or regionally specific expansion factors derived from surveys of active travellers may be used where available, but the source and basis of the expansion factors used should be stated in the appraisal report.

5.2.4 Choosing appropriate annual expansion factors

For work, school and tertiary education trips, use of annual expansion factors that reflect the standard work or education year would be appropriate – for example, around 220 days for work trips.

If an active travel facility features high proportions of recreational use, care is needed to ensure that trip making and therefore benefits are not overstated. Recreational trips are more likely on weekends and public holidays than on work days. For those trips, an expansion factor of 104 to 114 (depending on the number of public holidays that fall on weekends) would be suitable.

5.3 Health benefits

The 2007-2008 National Health Survey identified that physical inactivity is related to chronic health conditions including ischaemic heart disease, stroke, Type 2 diabetes, kidney disease, osteoarthritis, osteoporosis, colorectal cancer and depression (AIHW Cat No PHE 157, 2012).

Active travellers tend to be healthier than people who are relatively inactive or sedentary and suffer less from medical conditions that reduce their life expectancy. Healthy individuals place less demand on the health system for diagnosis, surgery and recovery.

Active travel, including walking and cycling, can contribute to minimising risks of cardiovascular disease, Type 2 diabetes, some cancers and osteoporosis. It can also assist with managing obesity, high blood pressure and high cholesterol. Mental health benefits have also been identified but not quantified. Physical activity can improve self-esteem and confidence, and reduce stress, anxiety, fatigue and depression (AIHW Cat. No. PHE 157 2012). Incorporating the health benefits of active travel into economic appraisals of transport initiatives requires a tool for valuing non-motorised transport. The discussion in this section will produce a monetary value for each of the walking and cycling active travel modes, applicable to the Australian population.

The types of health-related benefit attributable to active travel are:

  • Morbidity and mortality benefits because people who are active get sick less often and have a longer life expectancy than people who are inactive
  • Reduction in health system costs because active people are less likely to need medical and hospital care.

5.3.1 Morbidity and mortality

Morbidity and mortality benefits are relevant to those active travel trips that are generated by improvements in active travel infrastructure.

Willingness to pay benefits are typically based on the potential of active travel, as a form of physical activity to reduce the number of disability adjusted life years (or DALYs) lost as a consequence of inactivity. Reducing DALYs is analogous to increasing life expectancy. The value of additional reduced DALYs (or increased life expectancy) is based on a value of statistical life (VSL) derived using a willingness to pay approach. The benefit is expressed on a per kilometre basis, derived from the number of kilometres of walking or cycling needed to achieve a desirable level of physical activity. The more active a person is, the less additional activity needed to achieve a desirable state of health and the lower the benefit from that additional activity. Walkers generally need more physical activity than cyclists because walking is less vigorous than cycling.

5.3.2Health and physical activity

Available data on the health characteristics of Australians derive from the 2003 Burden of Disease study (Begg et al, 2007, AIHW Cat No PHE 82), the 2007-2008 National Health Survey (ABS Cat. No. 4364 2009) and the 2011-13 Australian Health Survey (ABS Cat. No. 4364.0.55.004 2013). The Burden of Disease study quantifies the extent to which healthy life is lost to disease manifested as prolonged illness, disability, and/or premature death. The disability adjusted life year (DALY) is the measure used to quantify the effects of individual diseases and injuries. One DALY is one year of healthy life lost due to disease or injury (AIHW, Australia’s Health 2010, Cat. No. AUS 122 2010).

The 2007-2008 National Health Survey identified that physical inactivity is related to chronic health conditions including ischaemic heart disease, stroke, Type 2 diabetes, kidney disease, osteoarthritis, osteoporosis, colorectal cancer and depression (AIHW Cat. No. PHE 157 2012). With the exception of kidney disease and osteoporosis, these conditions were amongst the 20 leading causes of burden of disease in 2003 (AIHW 2007, p 39). The AIHW report ‘Australia’s Health 2010’ (Cat. No. AUS 122 2010) identifies these diseases as continuing to be amongst the leading causes of burden of disease in 2010. Of the total burden of disease, cancer was projected to account for 19% while the second leading cause, cardiovascular disease, was estimated to account for 16%. Type 2 diabetes was projected to become the leading cause of disease burden by 2023, partly attributable to the worsening problem of overweight and obesity. In 2010, diabetes accounted for almost 7% of the total disease burden (Type 2 diabetes was estimated to account for 94% of the diabetes burden). Arthritis and musculoskeletal conditions accounted for 4% of the national disease burden in 2010 (Cat. No. AUS 122 2010).

Using the 2003 burden of disease data, Begg et al found that the proportion of health loss attributable to physical inactivity is 6.6% for all conditions, 5.6% for cancers and 23.7% for each of cardiovascular disease and diabetes mellitus (Begg et al, 2008, p.38).

The National Physical Activity Guidelines for Australians for adults recommend at least 30 minutes of moderate-intensity physical activity on most, preferably all, days of the week. Examples of moderate-intensity activity are brisk walking, swimming, doubles tennis and medium-paced cycling. For activity to be sufficient for accruing health benefits, criteria for both time and the number of sessions need to be met. The definition for sufficient time and sessions is therefore ‘at least 150 minutes of moderate-intensity physical activity accrued over at least five sessions per week, with vigorous activity counted as double’. Sessions of a minimum of 10 minutes can also be included in the weekly count (AIHW Cat. No. AUS 122 2010, p.93).

The 2007-2008 National Health Survey identified that 37% of Australian adults participated in sufficient time and sessions of physical activity to accrue health benefits; 55% were insufficiently active; and 8% were inactive (AIHW Cat. No. AUS 122 2010, p.93). In comparison, the 2011-2013 Australian Health Survey found that:

  • The proportion of sufficiently active adults increased to 43.5%.
  • The proportion of inactive adults decreased to 20.5%.
  • There was a greater than four times increase to 36% in the proportion of insufficiently active adults (ABS, 43640DO004_20112012, Data cube Table 4, in 4364.0.55.004 - Australian Health Survey: Physical Activity, 2011-12).

Appendix A provides definitions for the levels of physical activity.

In a review article, Rissel et al (2012) found that between eight and 33 minutes of additional physical activity was attributable to daily walking associated with public transport use in Australia. Rissel et al used statistical modelling to predict that the proportion of the adult population considered ‘sufficiently active’ would increase significantly if 20% of all ‘insufficiently active’ adults increased their walking by 16 minutes a day for five days a week (Rissel et al, 2012). While road injury can be associated with cycling and walking, Ker et al (2011, p.16) have identified that the health and fitness benefits more than offset the net road trauma increase.

In 2007, Brisbane public transport users walked on average 2.3 kilometres for approximately 28 minutes, thereby meeting recommended physical activity requirements (Burke & Brown in Ker et al, 2011, p.18).

Journeys of less than five kilometres are considered appropriate for active travel (WHO in Genter et al, 2008, p.22). In New Zealand, criteria for active travel distances are destinations within seven kilometres for cycling and two kilometres for walking (Genter et al, 2008, p.22). The intensity of one minute of walking is considered equivalent to 0.5 minutes of cycling (Genter et al, 2008, p.38).

5.3.3 Valuing active travel

These Guidelines adopt the methodology adopted by Genter et al (2008) in valuing the health benefits of active travel in New Zealand to the Australian adult population, using local data where available, to determine a per kilometre value of the health benefits of active travel modes.

Since inactive adults have most to gain by participating in physical activity, full benefits are allocated to those who were previously inactive, with only marginal benefits allocated to existing active adults (Genter, 2008, p.40).

Value of a Statistical Life (VSL) is a monetary value of a human life based on the willingness to pay concept. It is updated using the Consumer Price Index. The value of statistical life is an estimate of the economic value society places on reducing the average number of deaths by one. A related concept is the value of statistical life year (VSLY), which estimates the value society places on reducing the risk of premature death, expressed in terms of saving a statistical life year. The Australian estimate of the value of statistical life is $3.5 m and the value of statistical life year is $151,000 (OBPR, RSCH.040.001.0188 2008). Updating the 2007 VSL for CPI movements yields a 2013 VSL estimate of $4,084,027.

5.3.4 Benefits of reduced morbidity and mortality

Morbidity and mortality costs of inactivity are valued using DALYs associated with inactivity and undiscounted VSL (Genter, 2008, p.47). The steps are:

  • Estimate DALYs for years of life lost to both morbidity and mortality. The Population Attributable Risk Fraction (PAF) for inactivity identified by Begg et al (2008, p.38) in The Burden of Disease and Injury in Australia 2003 study (Begg et al, 2007, AIHW Cat. No. PHE 82) is 6.6% of total DALYs.
  • Determine the annual ratio of DALY per inactive adult by dividing the DALYs due to inactivity by the estimated adult inactive population in 2010, derived from Australian Health Survey 2007-8 and 2011-12 results (ABS 2012, 43640DO001_20112012 Australian Health Survey: First Results, 2011–12 — Australia.)
  • Multiply this ratio by the undiscounted 2013 value of DALY[3] to produce the per capita annual benefit of physical activity, shown in Table 6.
Table 6: Per capita annual value (morbidity and mortality) using DALYs
Projected total DALYs 20102 849 000
Population attributable risk fraction for inactivity 6.6%
DALYs attributable to inactivity 188 034
Adult inactive population 11 483 475
Ratio of DALYs attributable to inactivity per inactive adult 0.02
Willingness To Pay DALY 2013 $90 555
Per capita annual value $1483

Source: Economic Associates analysis based on Genter 2008, p.48. See accompanying text.

5.3.5 Health system benefits

Health sector costs of inactivity are estimated by calculating the proportion of costs due to inactivity and then dividing this value by the number of inactive adults to determine the per capita health sector costs, as shown in Table 7.

Table 7: Per capita annual health sector costs attributable to inactivity
Total health sector costs 2010 (1)$130 266 000 000
6.6% due to inactivity $8 597 556 000
Inactive population 2010 11 483 475
Per capita health sector costs 2010 $749
Per capita health sector costs 2013 $796

(1) See AIHW 2012, Health Expenditure Australia 2010-11, Health and Welfare Expenditure Series No. 47, Cat. No. HWE 56.

Productivity costs are excluded because Genter et al found inadequate evidence supporting the association between active transport and reduced sick days (Genter et al, 2008, p.49). Therefore, the total annual per capita value in 2010 of health benefits of active travel is $2131, comprising morbidity and mortality costs ($1382) and health sector costs ($749), attributable to inactivity.

5.3.6 Total health benefits

Genter et al (2008, p.50) cite research indicating that more health benefits accrue when additional activity is initially taken up, with benefits accruing at a lesser rate for those who are currently active. As mentioned in the ‘Health and physical activity’ discussion earlier, data on the extent of physical activity in the Australian adult population (aged 18 years and older) derive from the 2007-8 National Health Survey and the 2011-12 Australian Health Survey. Table 8 summarises activity data from these surveys. Appendix A provides definitions for the levels of physical activity.

Table 8: Physical activity levels in the Australian adult population
Physical activity levelPrevalence 2007-2008 (1)Prevalence 2011-2012 (2)
Inactive 8% 20.5%
Insufficiently active 55% 36.0%
Sufficiently active 37% 43.5%

(1) See AIHW Cat. No. AUS 122, 2010, p 93

(2) ABS, 4364.0.55.004 - Australian Health Survey: Physical Activity, 2011-12, Download Data cube Table 4 Sufficient physical activity measure by selected population characteristics, Persons aged 18 years and over (estimate) (43640DO004_20112012 Australian Health Survey: Physical Activity, 2011-12-Australia ), 2013

The 2011-12 split of physical activity levels is considered more appropriate to the 2010 population and therefore is used for calculating per kilometre health benefits.

The three physical activity weightings used and the rationale adopted by Genter et al (2008, pp.50-51) are:

  • Weighting 1 - Inactive – Shifting the inactive group into some moderate physical activity has most benefits in terms of reduced morbidity and mortality. This group can receive full annual benefits by walking at 5 km per hour for 30 minutes, five days per week. This is an annual walking distance of 625 km.
  • Weighting 0.85 – Insufficiently active – The insufficiently active group can receive most of the health benefits of increased activity, even though they already engage in some moderate activity. An additional 20 minutes’ physical activity per day for five days per week requires an annual distance of 450 km.
  • Weighting 0.15 – Sufficiently active – The sufficiently active group may receive ongoing health benefits and encouragement to maintain physical activity. An additional 14 minutes’ physical activity per day for five days per week requires an annual distance of 312 km.

5.3.7 Parameter values for walking and cycling benefits

Parameter values for walking and cycling benefits are shown in Table 9.

Table 9: Per-km weighted health and health system benefits of walking (2013)
Walking annual health benefitsBenefit weightWeighted sum = annual benefits per person
$2279 1 0.85 0.15 $1314
Inactive Insufficiently Active Sufficiently Active
20.5% 36% 43.5%
$467 $698 $149
  Km over which benefits received Weighted per km benefits
625 450 312 $2.77
$0.74 $1.55 $0.48

Based on Genter 2008

Since physical health benefits per minute of cycling are twice those of walking, only half the time is required to achieve benefits. Therefore cycling has half the benefit per kilometre of walking. (Genter et al, 2008 p.51).

Table 10: Per-km weighted health and health system benefits of cycling (2013)
Cycling annual health benefitsBenefit weightWeighted sum = annual benefits per person
$2279 1 0.85 0.15 $1314
Inactive Insufficiently Active Sufficiently Active
20.5% 36% 43.5%
$467 $698 $149
  Km over which benefits received Weighted per km benefits
1250 900 624 $1.40
$0.38 $0.78 $0.24

Based on Genter 2008

Accordingly, the 2013 monetary value of the health benefits of walking is $2.77 per km and the monetary value of the health benefits of cycling is $1.40 per km for Australian adults aged 18 years and older.

Adjusting for local activity levels

The unit benefit values in Table 9 and Table 10 could be amended if information was available about activity levels in an active travel catchment. The activity weights are contained in the fourth row of Table 9 and Table 10 (for example 20.5% of the Australian population is estimated to be inactive). Once new activity weightings have been inserted the unit benefit values in the tables can be recalculated.

5.3.8 Indexing health benefit parameter values

Indexing of benefits is appropriate in cost benefit analysis because costs used by the cost benefit analyst will have been indexed by the cost engineer or quantity surveyor. Because initiative costs for budgeting and contract purposes are outturn or values of the day costs, the base year costs will reflect prices in that or in some other recent numeraire year[4]. For consistency, benefit parameter values established perhaps some years earlier will need to be indexed to the same base year.

The two components of health benefits should be treated differently.

Indexing morbidity/mortality benefits

Morbidity/mortality benefits are based on the value of statistical life (VSL) and should in theory be indexed by reference to the ABS Average Weekly Earnings series for Australia, with adjustment for income elasticity – that is, for the tendency for VSL to increase with increases in income and possibly to changes in wealth as well. The Office of Best Practice Regulation (OBPR, 2008) recommends that VSL be indexed to the Consumer Price Index (CPI). Implicit in the OBPR guidance is an income elasticity of one and, combined with the link to the CPI, the long term value of VSL would be maintained over the long term. Abelson (2008) - in a paper prepared for the OBPR - notes studies recommending an income elasticity closer to 0.5, but notes also that some studies link VSL to wealth, which will be affected by movements in asset prices as well as income. Hammitt and Robinson (2011, p.6) also note the prevalence of income elasticity estimates around 0.4 to 0.6. Their research suggests that this value might not be consistent across a wide income range as might be relevant when projecting incomes across the 30-year life of an infrastructure initiative.

In light of the controversies surrounding the valuation base for VSL and for changes in VSL over time, and the difficulties associated with income and wealth forecasting, the OBPR guidance that VLS be indexed to CPI with an (implicit) income elasticity of one is a conservative middle position. Adoption of that position would also have the advantage of consistency with Commonwealth regulatory assessments and, in particular, those that relate to health and safety.

Indexing health system benefits

The Australian Institute of Health and Welfare (AIHW) produces an annual national composite health cost index incorporating a range of health cost elements or areas of expenditure that include for example public hospitals, private hospitals, medical services, dental services, pharmaceuticals and capital expenditure. Sixteen expenditure areas are included in the index. According to AIHW (2013, p.108):

‘The national THPI [total health price index] provides the most useful available measure of overall health inflation in Australia. As such, it has been integrated into the indexation formula for payments in support of the National Healthcare Agreement under the Intergovernmental Agreement on Federal Financial Relations.’

The TPHI is shown in Table 11.

Table 11: Total health price index, Australia
Year Index
2001-02 76.1
2002-03 78.4
2003-04 81.1
2004-05 84.0
2005-06 87.5
2007-08 90.4
2008-09 94.9
2009-10 97.3
2010-11 98.3
2011-12 100.0

Source: AIHW (2013) p.109

5.3.9 Updating health benefit parameter values

Morbidity/mortality benefits

Morbidity/mortality costs estimated in these Guidelines are a product of research into links between inactivity and disease, the effects of inactivity on life expectancy and the quality of life, the potential for physical activity to reduce disease risk and the value of statistical life. Some of the supporting research, such as the 2003 burden of disease and injury study (Begg et al, 2007) and estimates of the value of life, is fundamental and likely to require considerable effort.

Because health benefits are central to active travel policy and to health policy generally, updating should be a joint venture between BITRE, Austroads and health agencies such as AIHW. Commonwealth and state regulatory agencies also have an interest because of the importance of health benefits in the broad suite of health and safety regulation including occupational health and safety, product safety, building safety and the like. With the extent of research involved, updating would be impracticable in anything less than a five- or even a ten-year interval.

Health system benefits

Health system benefits are based on the proportion of health costs attributable to inactivity. Updating that proportion would be a role for the related task of updating morbidity/morbidity costs. Total health care costs in Australia are regularly updated and reported by AIHW.

5.3.10 Application issues

The estimation of health benefits for individual active travel initiatives presents a range of challenges.

Variations in underlying physical activity levels

Underlying levels of physical activity may well vary between catchments and travel is only one form of physical activity available to people who want to be active. Others include gym or fitness centre work, team sport, swimming and the like. At the level of individual initiatives, information about the existing activity levels of the catchment is very unlikely to be available.

The implications of different levels of activity for the benefits of active travel infrastructure are potentially quite significant, as Table 13 and Table 14 illustrate, with benefits of better infrastructure for active people being only around one quarter of those for people who are sedentary or insufficiently active. Where data about underlying activity levels is unavailable, the per kilometre benefits recommended in Table 13 and Table 14 are weighted for the average distribution of activity levels in the Australian community. The data in those tables can also be used at a disaggregate level in the event that initiative-specific or catchment-specific information about underlying activity levels is available. Applying an upper limit to the amount of activity that generates benefits for inclusion in cost benefit analyses will be quite difficult for the reasons outlined below. Sensitivity testing is probably the only means to address the issue. For example, if the trips generated by an initiative are estimated to amount to 1000 km per walker per year and 2000 km per year for each cyclist, a sensitivity test might be carried out in which those benefits only accumulate over 625 km/year for walkers and 1250 km/year for cyclists (as per the limits in Table 13 and Table 14).

How important is trip frequency?

For most benefit categories, the dollar value of benefits is directly related to trip length and trip frequency. However, the picture is less straightforward for health benefits because for activity to be beneficial, it must be of sufficient frequency (trips per week) and sufficient length (a proxy for session duration).

In the methodology used to estimate health benefits presented above, there are limits to the amount of activity that is necessary for good health, meaning (by implication) that further activity will not accrue health benefits. Those limits are higher for inactive people than for those who are active – in other words, presently inactive people need to do more additional exercise than is needed by people who are active. Meaningful application of those limits in cost benefit analyses would necessitate quite detailed knowledge of the activity profiles of the market likely to use a proposed infrastructure improvement. Others who use the infrastructure once or twice a week may have no other physical activity sessions of sufficient duration to be effective for health. Some active travellers might only use the infrastructure once or twice or week as part of a wider physical activity regime. For others, active travel might meet all their physical activity needs

The active traveller and trip data that is needed to ensure accurate estimation of health benefits in these circumstances is unlikely to be available. As a partial alternative, sensitivity testing of cost benefit analysis results for the presence of recreational user health benefits would point to those initiatives for which there was a relatively high risk of benefit overstatement. Program planners could use that information to direct investment towards initiatives that supported high frequency active travel or collect information about active travel patterns that would allow more effective options development.

Will initiatives generate long enough trips?

[5] Secondly, initiative-specific or catchment-specific information about underlying activity levels is unlikely to be available. Project planners and assessors cannot then know whether a proposed active travel initiative will generate the increases in activity from which active travellers can benefit.

There is some limited evidence that the new active travel trips generated by better active travel infrastructure will be long enough to produce health improvements. Garrard (2009, p.7) reports Victorian data from 2007 that adults walking for transport walked 60 to 70 minutes per week for purposes other than recreation – that is up to 14 minutes per day, excluding recreation. Garrard (2009, p.6) also quotes 2006 Census data for Victoria that found that cycle trips to and from work on census day averaged 6.1 km total (about 20 minutes total). For walk to work trips, the equivalent results were 2.2 km and 27 minutes. Ker et al (2011, p.52) quoted results from a Bus Association of Victoria Survey that public transport users spend an average of 41 minutes per day walking and cycling (including recreational walking and cycling), which is five times the physical activity of commuters who use motorised private transport.

However, appraisals that include benefits based on small (sub-10 minute) increases in activity should set out the underlying assumptions and supporting evidence for the inclusion of those benefits. The appraisal should also include a sensitivity test in which the ‘sub-10 minute’ benefits are not incorporated in the benefit stream.

What is the retention rate for new active travellers?

How confident can we be that forecast levels of new active travel generated by improved infrastructure will be retained? A recent paper (Goodman et al, 2013, p.1) noted the lack of research into this question. However, their before and after study of three new active travel infrastructure initiatives in the UK noted very high retention levels two years after initiative opening. However, they did note that the new initiatives ‘may have primarily attracted existing walkers and cyclists'. A similar outcome is evident in the results of intercept surveys in Brisbane reported in SKM PWC (2011), presented in Table 12.

Table 12: Trip diversion rates from Brisbane intercept surveys
Trips fromDiverting to
  Cycling Walking
  Inner city Other areas Inner city Other areas
Car 10% 15% 5% 10%
Public transport 20% 0% 15% 50
Reassign 65% 55% 70% 50%
Induced 5% 30% 10% 40%

Source: SKM PWC (2011) p.33

Until better data becomes available, sensitivity testing should be carried out to simulate the effects of lower than expected trip growth outcomes; a ‘no growth’ assumption would be prudent, particularly if there is little usage history in the relevant initiative catchment.

5.3.11 Health benefits summary

Recommended parameter values for estimation of health benefits are shown in Table 13 and Table 14 for the two components – morbidity/mortality benefits and health system benefits. Cyclists are estimated to require twice the active kilometres of walkers because cyclists are assumed to travel at four times the speed of walkers[6] and cycling is assumed to have twice the physical intensity of walking[7].

Table 13: Mortality and morbidity benefits of active travel per km according to physical activity status 2013
 Mortality/morbidity benefit per km/activity level
  Inactive Insufficiently active Sufficiently active Weighted per km benefit
Walking
Benefits of additional activity per person $1483 $1261 $222  
Km over which activity benefits are received 625 450 312  
Proportion of population 20.5% 36% 43.5%  
Willingness to pay benefit per km $0.49 $1.01 $0.31 $1.81
Cycling
Benefits of additional activity per person $1483 $1261 $222  
Km over which activity benefits are received 1250 900 624  
Proportion of population 20.5% 36% 43.5%  
Willingness to pay benefit per km $0.24 $0.50 $0.15 $0.89
Table 14: Health system benefits of active travel per km according to physical activity status 2013
 Mortality/morbidity benefit per km/activity level
  Inactive Insufficiently active Sufficiently active Weighted per km benefit
Walking
Benefits of additional activity per person $796 $677 $119  
Km over which activity benefits are received 625 450 312  
Proportion of population 20.5% 36% 43.5%  
Willingness to pay benefit per km $0.26 $0.54 $0.17 $0.97
Cycling
Benefits of additional activity per person $796 $677 $119  
Km over which activity benefits are received 1250 900 624  
Proportion of population 20.5% 36% 43.5%  
Willingness to pay benefit per km $0.13 $0.27 $0.08 $0.48

5.3.12 Calculating health benefits

Health benefits need to be calculated in two parts because the willingness to pay component for new active travellers is subject to the rule of the half.

Each component of health benefit per kilometre is multiplied by trip kilometres to arrive at benefits per trip.

For existing active travellers

Health benefits are not estimated for existing active travel travellers who, by definition, are already accruing the health benefits of active travel. New or improved active travel infrastructure does not generate health benefits, but could well generate other benefits - in particular, crash benefits.

For new active travellers – conversions from other modes

Health benefits arise for this group because of their additional active travel relative to their existing trip. Where suitable trip data is available (zonal origin, zonal destination, route), active travel benefits would be calculated for the project case trip length, less any walk or cycle trip length in the base case trip. Assuming that information is available, the benefit estimation steps are:

Step 1: Estimate the average length of new (converting) trips in kilometres

Step 2: From the average trip length in Step 1, subtract the average active travel length for the base case and multiply the resulting value by the average number of daily trips per converting active traveller

Step 3: Multiply the net active travel trip length from Step 2 by the weighted average health benefit in Table 13 and then multiply by 0.5 (that is, apply the rule of half)

Step 4: Multiply the net active travel trip length from Step 2 by the weighted average health system benefit in Table 14

Step 5: Sum the benefits from Steps 3 and 4 to arrive at daily benefits. Use advice from other parts of the ATAP Guidelines to estimate annual health benefits in the base year and subsequent years of the analysis period.

For new active travellers – those new to walking and cycling

Some people might become active travellers because a new or improved facility removes a perceived barrier or sparks their interest. For whatever reason they become active travellers, their health benefits are estimated as follows:

Step 1: Estimate the average length in kilometres of new active travel trips

Step 2: Estimate the average number of daily active travel trips to be made by new users

Step 3: Multiply the active travel trip length from Step 2 by the weighted average morbidity/mortality benefit in Table 13 and by 0.5

Step 4: Multiply the active travel trip length from Step 2 by the weighted average health system benefit in Table 14

Step 5: Sum the benefits from Steps 3 and 4 to arrive at daily benefits in the base year. Use advice from other parts of ATAP Guidelines to initiative benefits over subsequent years of the analysis period.

These steps are equally applicable to existing active travellers who are anticipated to increase their active travel as a consequence of a proposed initiative.

5.4 Congestion reduction benefits

An active travel initiative that has the effect of causing private motor vehicle users to walk or cycle rather than drive could reduce congestion depending on the time their trip is made. Trips made during peak periods are much more likely to reduce congestion than say recreational trips made on weekends.

The extent of congestion reduction relates to the length and timing of the replaced trip. Hence for someone who previously drove 5 km to work Mondays to Fridays leaving home at 7 am and now uses a new cycleway to ride 7 km to work leaving home at 6:30 am, the relevant trip characteristics are those of the drive trip - that is, a 5 km drive trip commencing at 7 am[8].

The estimation of these benefits is covered in other parts of the ATAP Guidelines.

5.5 Crash benefits

As noted in section 2.2.4, active travellers face a high crash risk relative to the risk for car occupants[9]. It follows that car occupants who convert to active travel could experience a negative crash benefit that partly offsets other positive benefits of active travel. Interventions such as off-road pedestrian/cycle paths that attract new active travellers will attenuate the additional risk for new cyclists and reduce the risk for existing active travellers.

Active travel crash risk is not as well understood as motor vehicle crash risk. Related to that, information about the effectiveness of measures to reduce risk – such as provision of on-road and off-road paths - is relatively sparse. Finally, some commentators maintain that increases in cycling activity actually reduce crash risk – the so called ‘safety in numbers’ theory (see section 5.5.4).

Crash benefits are made up of the reduction in expected crash costs attributable to an active travel infrastructure improvement. They comprise two elements:

  • Current active travellers will probably experience a crash benefit to the extent that active travel infrastructure aims to separate motorised and non-motorised traffic streams including improving safety at intersections and crossings.
  • Active travellers who walk or cycle more frequently and new active travellers who divert from other modes are likely to experience an elevation in crash risk because walking and cycling have a much higher crash risk than public transport or motorised private transport. Improved infrastructure might reduce that expected crash cost, but trip conversion will still produce a negative crash benefit that partly offsets other active travel benefits.

While the crash risk of active travel relative to other modes is well known, there is less certainty about the contribution that specific remediation measures can make to reducing crash risk. In addition, because active travel in kilometres travelled terms is less than 1/300th of the extent of motorised private transport in Australia, the crash record at a particular site or a particular section of road might be the outcome of chance only[10], so that site- or section-specific risk reductions attributable to infrastructure will be difficult to estimate. Cost benefit analysts should endeavour to access advice from appropriately qualified safety specialists, but the data that is available suggests that in most cases a fairly coarse approach to estimating crash reduction benefits will be all that is possible.

Because the crash record is likely to be patchy on a site specific basis, an exposure based approach is proposed here in which benefits are calculated according to reductions or increases per kilometre of travel. This approach will be more satisfactory for initiatives made up of several elements (such as a pedestrian/cycle path through a park with road crossings at either end) but will be less satisfactory when applied to specific items of infrastructure, (such as a pedestrian/cycle bridge over a busy multi-lane urban road). Accident prediction models for different road elements - such as mid-blocks, roundabouts and T intersections - have been developed in New Zealand (Turner et al, 2006), but their application may be beyond the time and resource budgets for all but the largest appraisals.

5.5.1 Active travel crash risk

Relative fatal and serious injury rates for private motor vehicle users and active travellers are shown in Table 15 and Table 16. Active travel risk is up to eight times riskier than private motor vehicle travel. Pedestrians are at higher risk than cyclists. For cyclists, most fatal crashes involve a collision with a motor vehicle.

The data means, for example, that a car driver who decides to take some trips by bicycle increases their fatal crash risk on those trips by four times from 0.048 per 10 million km travelled to 0.2 per 10 million km travelled. A pedestrian making the same trip choices would increase their safety risk more by more than ten times from 0.048 per 10 million km travelled to 0.62 per 10 million km travelled.

Table 15: Fatality rates for motorists and active travellers, Australia (2002 to 2006)
User descriptionEstimated annual travel (10 million km travelled)Fatalities annual averageFatality rate per 10 million km travelled
Rail passenger (a) - - 0.00816
Bus passenger (a) - - 0.00816
Car driver 11 866 572 0.0480
Car passenger 5611 247 0.0440
Total car 17 477 819 0.0470
Motorcyclist 96 134 1.3960
Total car and motorcycle 17 573 953 0.0540
Bicyclist 124 24 0.2000
Pedestrian 271 168 0.6200
Total active travel 505 192 0.3800

Note: Excludes New South Wales and Queensland due to inconsistencies in data categories.

(a) Rail and bus passenger risk has been estimated separately by reference to BTRE (2003) which estimated fatality risk for bus and rail passengers as being 17% of the risk for car occupants.

Source: Estimated from Austroads (2010) p.8, other than for (a) and (b)

Table 16: Serious injury rates for motorists and active travellers Australia (2002 to 2006)
User descriptionEstimated annual travel (10 million km travelled)Serious injuries annual averageSerious injury rate per 10 million km travelled
Car driver (a) 7419 6264 0.844
Car passenger (a) 3471 2057 0.593
Total car 10 890 8321 0.764
Motorcyclist 61 1171 19.197
Total car and motorcycle 10 951 9492 0.867
Bicyclist 88 440 5.000
Pedestrian 152 926 6.092
Total active travel 240 1366 5.692

Note: Excludes New South Wales and Queensland due to inconsistencies in data categories.

(a) See note to Table 16.

Source: Estimated from Austroads (2010) p.8

Table 17: Cyclists killed in road crashes, Australia, 1997 to 2004
EventCounterpart% of cyclist deaths
Collision with Pedestrian 1
Pedal cycle or other motor vehicle 0
Car, pick-up truck, can or other motor vehicle 64
Heavy transport vehicle 22
Railway train or railway vehicle 1
Fixed or stationary object 4
Not a collision   5
Unknown   3
Total   100

Source: ATSB (2006) p.4

5.5.2 Effectiveness of interventions

Empirical evidence about the effectiveness of infrastructure in reducing active travel crash risk is limited, but what evidence is available suggests crash risk reductions in the order of 30 to 40% for interventions such as kerb lanes located in the roadway and up to 80% to 100% for interventions that achieve a high degree of active traveller separation from the traffic stream such as off road cycle lanes and grade separations.

For existing active travellers, measures such as these could reduce the crash risks associated with their existing active travel. For new users or users converting to other modes, these measures constrain the increase in risk as they shift from car or public transport to active travel.

Table 18: Effectiveness of active travel safety interventions - cycling
InterventionReduction in bicycle crash riskReduction in overall crash risk (all vehicles)Source
Cycle lanes (lane between kerb and parked cars) 10%   cited in Schramm and Rakotonirainy (2008) Europe
Cycle lanes (lane between kerb and parked cars) 28%   Lusk et al (2013) Montreal
Cycle lanes (lane between kerb and parked cars) 30%-62%   Lusk et al (2013) New York City
Mid-block cycle lanes 10% 30% Elvik and Vaa (2004) cited in Turner et al (2011)
Advanced limit lines (storage boxes) 27% 40% Elvik and Vaa (2004) cited in Turner et al (2011)
Adding cycle lanes through an intersection 12% (14%) Elvik and Vaa (2004) cited in Turner et al (2011)
Traffic calming 36% 40% Davies et al (1997) cited in Turner et al (2011)
Cycle lanes marked on-road, mid-blocks 29%   UK, Coates (1999) cited in Turner et al (2009)
Cycle lanes marked on-road, intersections 35%   UK, Coates (1999) cited in Turner et al (2009)
On-roadway cycle lanes 57%   York UK: Transport for London (2004) cited in Turner et al (2009)
Shared use footpath 28%   York UK: Transport for London (2004) cited in Turner et al (2009)
Signalled intersections 83%   York UK: Transport for London (2004) cited in Turner et al (2009)
Cycle track + ^ 100%   York UK: Transport for London (2004) cited in Turner et al (2009)
Cycle lane 70%   York UK: Transport for London (2004) cited in Turner et al (2009)
Advanced stop line at signals** ^ 100%   York UK: Transport for London (2004) cited in Turner et al (2009)

( … ) signifies an increase in crash risk.

(+) the definition of cycle track is unclear. See Turner et al (2009).

**Advanced stop line at signals includes storage box.

^ Turner et al (2009) indicates that reductions of this magnitude are unrealistic.

Table 19: Effectiveness of signalisation countermeasures – walking (US)
Countermeasure (s)% Crash reduction factor
  Crash severity Left-turn crashes Pedestrian
Add exclusive pedestrian phasing All   34
Improve signal timing (to intervals specified by the ITE Determining Vehicle Change Intervals: A Proposed Recommended Practice (1985)) Fatal/injury   37
Replace existing WALK / DON'T WALK signals with pedestrian countdown signal heads All   25
Modify signal phasing (implement a leading pedestrian interval) All   5
Remove unwarranted signals (one-way street) All   17
Convert permissive or permissive/protected to protected only left-turn phasing All 99  
Convert permissive to permissive/protected left-turn phasing All 16  

Note: US terminology, dimensions and spelling in original

Source: FHWA (2008)

Table 20: Effectiveness of signalisation countermeasures – walking (US)
Countermeasure (s)% Crash reduction factor
  Crash severity Left-turn crashes Pedestrian
Add exclusive pedestrian phasing All   34
Improve signal timing (to intervals specified by the ITE Determining Vehicle Change Intervals: A Proposed Recommended Practice (1985)) Fatal/injury   37
Replace existing WALK / DON'T WALK signals with pedestrian countdown signal heads All   25
Modify signal phasing (implement a leading pedestrian interval) All   5
Remove unwarranted signals (one-way street) All   17
Convert permissive or permissive/protected to protected only left-turn phasing All 99  
Convert permissive to permissive/protected left-turn phasing All 16  

Note: US terminology, dimensions and spelling in original

Source: FHWA (2008)

Table 21: Effectiveness of geometric countermeasures – walking (US)
Countermeasure (s)% Crash reduction factor
  Crash severity All crashes Pedestrian
Convert unsignalized intersection to roundabout Fatal/injury   27 (12)
Install pedestrian overpass/underpass Fatal/injury   90
All   86
Install pedestrian overpass/underpass (unsignalized intersection) All   13
Install raised median All   25
Install raised median (marked crosswalk) at unsignalized intersection All   46
Install raised median (unmarked crosswalk) at unsignalized intersection All   39
Install raised pedestrian crossing All 30 (67)  
Fatal/injury 36 (54)  
Install refuge islands All   56
Install sidewalk (to avoid walking along roadway)     88*
Provide paved shoulder (of at least 4 feet)     71*
Narrow roadway cross section from four lanes to three lanes (two through lanes with centre turn lane) All 29  

*Only applies to ‘walking along the roadway’ type crashes

Note: US terminology, dimensions and spelling in original

( ... ) signifies standard error. Estimates in bold signify a rigorous study methodology and a small standard error in the value of the crash reduction factor (CRF)

Source: FHWA (2008)

Table 22: Effectiveness of operational countermeasures – walking (US)
Countermeasure (s)% Crash reduction factor
  Crash severity All crashes Pedestrian
Add intersection lighting Injury 27*  
All 21*  
Add segment lighting Injury 23*  
All 20*  
Improve pavement friction (skid treatment with overlay) Fatal/injury   3
Increase enforcement ** All   23
Prohibit right-turn-on-red All 3  
Prohibit left-turns All   10
Restrict parking near intersections (to off-street) All   30

*Applies to night time crashes only

**Applies to crash reductions on corridors where sustained enforcement is used related to motorist yielding in marked crosswalks combined with a public education program

Note: US terminology, dimensions and spelling in original. Turn directions should be reversed for Australian application.

Source: FHWA (2008)

5.5.3 Estimating unit crash costs

Crash costs estimated on a distance basis are shown in Table 23. The crash cost estimates are not expressed on a net active travel distance basis, so analysts will need to estimate base case and project case trip length and trip generalised costs separately. The active travel crash cost estimates are based on crash exposure risk calculated separately for driver and passengers, and hence should not need to be adjusted for vehicle occupancy.

Two sets of crash costs are shown in Table 23: the hybrid human capital approach that has been used in Australia for some years and that derives from research carried out by BITRE and the more recent inclusive willingness to pay approach. The ATAP Guidelines favour the inclusive willingness to pay approach, but in the active travel context the hybrid human capital estimates might be more appropriate[11].

Table 23: Crash costs by mode per vehicle km (2013)
ModeCrash cost per veh km
  Hybrid human capital approach Inclusive willingness to pay approach
Car/ motorcycle $0.11 $0.21
Bicyclist $0.49 $0.95
Pedestrian $0.67 $1.44
Total active travel $0.59 $1.20
Train $0.02 $0.04
Bus $0.02 $0.04
Source: Estimated from Austroads (2012), Austroads (2010), BTRE (2006), BTRE (2003)

As noted in section 5.5.2, interventions that partially separate walkers and cyclists from the traffic stream can reduce active travel crash risk by around 40%, while reductions of 80% to 100% have been reported for interventions that separate active travellers from the traffic stream or that enhance protection of active travellers at intersections. Until further detailed Australian research is carried out, these values seem appropriate. For corridor initiatives – that combine a number of infrastructure elements – a large reduction might be appropriate, but there is insufficient research to provide guidance on this point.

5.5.4 Estimating crash reduction benefits

Crash benefits are estimated as follows:

For existing active travellers

Apply these steps for walkers and cyclists separately.

Step 1: Multiply the number of daily trips by estimated trip length.

Step 2: Convert daily trip km estimates to annual estimates. For transport trips, use expansion factors recommended in other parts of the ATAP Guidelines. For recreation trips, an expansion factor of 104 to 114 (weekends plus public holidays) would be suitable, but other expansion factors could be used if available and supported by evidence.

Step 3: Calculate the base case crash cost by multiplying base case trip km by the unit crash costs in Table 23.

Step 4: Multiply the annual base cost estimate from Step 3 by the relevant crash reduction factor in the tables in section 5.5.2 to estimate project case crash costs. Note that for initiatives that achieve complete separation of active travellers from the traffic flow, a crash reduction factor of 80% could be appropriate. For other initiative types, consider using a crash reduction factor of 40%.

Step 5: Subtract the Step 4 cost (project case) estimate from the Step 3 (base case) cost estimate to arrive at the base year crash reduction benefit.

Step 6: Use advice from other parts of the ATAP Guidelines to initiative annual crash benefits in subsequent years of the analysis period.

For new or converting active travellers

Apply these steps for walkers and cyclists separately.

Step 1: Multiply the number of daily trips by estimated trip length base case and project case trip lengths. In some instances the base and project case trip lengths will be similar, but not in others.

Step 2: Convert daily trip km estimates to annual estimates for the base and project cases. For transport trips, use expansion factors recommended in other parts of the ATAP Guidelines. For recreation trips, an expansion factor of 104 to 114 (weekends plus public holidays) would be suitable. But again, other estimates may be used if available and backed by evidence.

Step 3: Calculate the base case crash costs by multiplying base case trip km by unit crash costs shown in Table 23. For users converting from car, for example, the relevant cost will be $0.11 per km or $0.21 per km depending on which crash cost valuation methodology is being used.

Step 4: Multiply project case trip km by the unit crash costs in Table 23 for active travel. Then apply a suitable crash reduction factor from the tables in section 5.5.2 to estimate project case crash costs. Suitable crash reduction factors could be 80% for full separation of active travellers from the traffic flow and 40% for other initiatives.

Step 5: Subtract the Step 4 cost estimate from the Step 3 cost estimate to arrive at base year crash reduction benefit.

Step 6: Subtract the initiative case annual crash cost from the base case annual crash cost to obtain the annual crash benefit. In all likelihood, the result will be a negative benefit.

Step 7: Use advice from other parts of the ATAP Guidelines to initiative annual crash benefits in subsequent years of the analysis period.

5.5.5 Safety in numbers

The concept of safety in numbers emerges from a seminal paper by Jacobsen (2003)[12] which shows that

‘for pedestrians and cyclists, the fatality rate is inversely related to the amount of travel by that mode. The data is demonstrated with fatalities associated with walking and cycling data in Californian cities, injuries associated with cycling in Denmark and walking and cycling fatalities across European countries. All the data sets show diminishing rates of fatal and serious injury with increasing levels of walking or cycling.’ (Austroads, (2010) p.31)

Austroads’ review of the literature concludes that empirically demonstrated ‘safety in numbers’ relationships may provide evidence of correlation rather than causation, in that researchers do not control for improvements in infrastructure and other initiatives (safety awareness campaigns) that might encourage cycling at the same time as making cycling safer. Austroads comments that while safety in numbers might have a behavioural basis in that more cyclists on the road encourages motorists to look out for them, the research does not definitively prove causation.

Garrard (undated) reviewed the evidence and concluded that:

  • There is a safety in numbers association in some but not all situations.
  • The safety in numbers relationship is strongest and most consistent in European countries with high rates of cycling.
  • No studies have controlled for cycling infrastructure of driver/cyclist safety measures.
  • There is no evidence that increased cycling precedes injury rate reductions.
  • Limited Australian evidence, from Melbourne, is mixed.

Until the literature is more certain about the existence of a safety in numbers phenomenon, crash benefit estimates should not be adjusted to reflect a safety in numbers effect.

5.6 Savings in vehicle operating costs

Car and motorcycle users who decide to make their trip on foot or on a bicycle or who make active travel a bigger part of their trip making will incur savings in their motor vehicle-related costs. Austroads provides a simple model for estimating those costs on a per kilometre basis. The model includes fuel, repairs and maintenance, tyres and depreciation but excludes parking and personal travel time costs.

This component of benefit is part of the consumer surplus accruing to users who switch to active travel and is therefore subject to the rule of half.

Savings in reduced vehicle operating costs are estimated by reference to the motor vehicle trip that is being replaced by active travel. Therefore, if a 5 km car or motorcycle trip is being replaced by a 7 km active travel trip, the base for estimation of benefits is 5 km.

The estimation of these benefits is covered in other parts of the ATAP Guidelines.

5.7 Savings in parking costs

Users converting to active travel or active travellers making additional active travel trips might save in parking costs depending on when and where trips are made. Work and some education trips to inner city areas are the trips for which parking benefits are most likely.

The estimation of these benefits is covered in other parts of the ATAP Guidelines.

5.8 Savings in public transport operating costs

Trip conversions from public transport to active travel, whether by new active travel users or current active travellers who elect to walk or cycle more frequently in response to an infrastructure improvement, have the potential to generate savings in the costs of operating public transport services.

Savings are unlikely to arise in the weekday non-peak periods or on weekends, because at this time clock face scheduling will tend to determine public transport operating costs rather than demand. During the weekday peaks on the other hand, there is some potential for savings because service scheduling in those times will be demand driven.

Two recent Australian studies for Queensland and Sydney– by SKM PWC (2011) and PWC SKM (2010) - excluded this category of saving because of estimation difficulties engendered by independence between demand and service scheduling. On the other hand, Aecom (2010) in a Sydney network study estimated savings in bus and train operating costs resulting from diversions to active travel.

If public transport cost savings are to be included they should be confined to morning and evening peak trip conversions to active travel. In addition, the savings attributable to walk trips should be adjusted downwards to reflect shorter walk trip distances. This is because the relevant trip length for calculation of public transport operating cost savings is the in-vehicle component of the avoided public transport trip which may be shorter than the active travel trip that replaces it.

These benefits are covered in other parts of the ATAP Guidelines.

5.9 Savings in road infrastructure costs

Diversion of motorised transport trips to active travel could allow the deferral of road initiatives aimed at reducing congestion by increasing capacity. Cost savings from deferral would partly offset the costs of any active travel initiatives that encouraged trip diversion. This relationship can be looked at either from the perspective of deferring construction costs or reducing congestion, because a reduction in congestion allows deferral of measures to remediate it.

The estimation of these benefits is covered in other parts of the ATAP Guidelines.

5.10 Environmental benefits

Environmental benefits arise when car trips convert to walking or cycling. They might also arise when trips convert from public transport, but the likelihood is much less because of the indivisibility in public transport provision associated with clock face timetabling and the complexities of service scheduling in networks. Environmental benefits are only estimated for trips converting to active travel. The basis for calculation will be the in vehicle km of the motorised transport trip that is being replaced.

The method for estimation is the same as for a reduction in motor vehicle km associated with say a town bypass. The estimation of these benefits is covered in other parts of the ATAP Guidelines.

5.11 Travel time benefits

5.11.1 General

Typically travel time is a disincentive to travel. This means that initiatives that save time generate benefits. For active travel, however, time might be a positive feature of travel: for non-transport trips, time might be associated with enjoyment or with the positive outcomes of physical activity. In those circumstances, an infrastructure improvement – such as a pedestrian/cycle overpass that allows walkers and cyclists to travel faster – might not result in a trip time saving. Active travellers might simply do more exercise within the time available.

Hence for non-transport active travellers – those who are travelling for exercise or recreation- travel time savings are unlikely to generate time-related benefits, but they may generate additional health benefits because they allow faster, more vigorous travel. However, unless suitable data is available to support this possibility, it would be prudent not to include benefits for non-transport active travellers that arise for faster travel.

Time savings might be relevant for those transport trips – for work or education for example - that convert from private or public transport to active travel in response to an infrastructure improvement. An infrastructure improvement – such as a high speed cycleway might reduce trip times so much that some transport trips become feasible by bike[13]. In this situation, the time saving could take one or both of the following forms: a saving in trip time and/or a saving in time spent exercising at home or at the gym. This time saving is part of the consumers’ surplus motivating the mode shift. Unless specific empirical is available about the substitutability of cycling for other forms of exercise, it would be prudent to confine benefit estimation to travel related savings.

Existing and converting trip benefits estimation, including the application of the rule of the half, are outlined in other parts of the ATAP Guidelines. Unfortunately, there appears to be no research available about the weightings to be assigned to the travel time cost components of the active travel trip. Table 24 shows proposed weightings.

Table 24: Suggested travel time weightings
Travel time componentMotorised travelActive travel
In vehicle (including on bike) 1 1
Walk at trip ends 1 1
Wait time 3 -
End of trip time - 3

5.11.2 Weighting for cycling infrastructure quality

Australian and international research shows, as would be expected, that cyclists place a higher value on quiet street environments and off-road paths than they do on busy roads without any cycling infrastructure. Put another way, cyclists experience less disutility in using quiet streets or off-road paths than they do in using busy roads that lack cycling infrastructure.

The series of weights for Sydney shown in Table 25 imply that:

  • A cyclist would ride approximately 3 km on an off-road path to avoid riding 1 km on a busy road with no cycling infrastructure.
  • A cyclist would ride 1.26 km on an off-road path to avoid riding 1 km on-road in cycle lanes or in a quiet street.
Table 25: Travel time weightings for cycling
 PriceWaterhouse Coopers/SKM (2010)Wardman et al (2007)Tilahun et al (2007)Hunt and Abraham (2007)
Location Sydney UK US Canada
On-road 1.00 1.00 1.00 1.00
On-road with lanes or quiet street 0.43 0.48 0.55 0.24
Off-road 0.34 0.29 0.80 0.36

Source: Mulley (2014)

In principle, these weights could be used in comparing the perceived costs of cycling on different types of track infrastructure. A major impediment to their use is the potential over-counting of crash benefits. It is not possible to determine how much the relative preference for off-road paths is influenced by perceptions of safety and how much by other factors such as reduced stress from less interactions with motor vehicles, reduced exposure to road rage or other forms of abuse or the pleasure associated with cycling in a pleasant environment.

Until more research is carried out, it is recommended that these weights not be used in the estimation of generalised cost.

5.11.3 Active travel speeds

Values of travel time for use in transport cost benefit analyses are set out in other parts of the ATAP Guidelines. Methods for estimating speeds of other modes are also set out elsewhere in the Guidelines.

A range of estimates for active travel speeds are shown in Table 26.

Walk

Estimates of walking speeds in the literature vary according to age and, it would appear, the purposefulness of particular trips, so that someone walking with the intent of exercising walks faster than if they are walking for non-exercise purposes. A 6 km/h average speed is in the middle of the estimates shown in Table 26 and would be appropriate for CBAs. It is unlikely that analysts will have information about walker age, trip purpose and whether their purpose is mixed with exercise that would allow finer-grained speed estimates to be used.

Cycling

The data sources for cycling speeds are ambiguous as to whether they are cruising speeds, trip start to trip end speeds[14] or door to door speeds. The higher estimates in Table 26 appear to relate to cruising speeds on a dedicated cycleway, whereas the Copenhagen estimate reflects more of a door to door speed. The door to door estimates would be suitable for estimating the travel time effects of a dedicated cycleway replacing a section of street riding. The door to door speed would be used to calculate travel time effects for users who convert to cycling and whose trip involves a mixture of on- and off-street riding.

Table 26: Average speeds for active travel modes
SourceCommentSpeed descriptionSpeed (km/hr)
Parise et al (2004) Older adult males Walk –‘normal brisk walking speed’ 5.75 km/h
Older adult females Walk –‘normal brisk walking speed’ 5.5 km/h
British Heart Foundation (2014) Person with excellent fitness ‘Moderate’ walking pace 6.4 km/h
  ‘Fast’ walking pace 7.5 km/h
City of Copenhagen (undated)   Average cycling speed in 2012 15.5 km/h
Haworth (2011) Context suggests the data cited is cruising speed rather than average journey speed Median cycling speed 24 km/h
Aecom (2010, p.51)   Cycle 23 km/h
Aecom (2010 p.3) Dedicated cycleway Cycle 25 km/h

[1] Part T2, Section 6.3 of NGTSM outlines the rationale for the ‘rule of the half’.

[2] Modal shifts towards active travel will not always reduce crash costs. Because pedestrian and cyclist crash risk in the road environment is relatively high, a modal shift towards on-road and in-road active travel facilities could cause a net increase in crash risk.

[3] The annual value of a DALY was calculated by subtracting the average age (37 years) from the average life expectancy in 2010 (82.1 years) to get 45.1 years. The VSL was then divided by 45.1. (life expectancy and average life data sourced from ABS 4102.0 - Australian Social Trends, April 2013, and http://www.aihw.gov.au/deaths/life-expectancy/, AIHW 2013, accessed 22/7/2014).

[4] For example, a project costed in 2016 is likely to be costed from a base of 2016 input prices. Benefits estimated in 2010 values would need to be indexed to 2016 values to ensure consistency between cost and benefit estimates.

[5] The National Heart Foundation advises that its recommended levels of physical activity could be achieved by doing “30 – 40 minutes of moderate-intensity physical activity (like brisk walking) most days of the week. You can build up activity in shorter bursts, like in three ten minute walks”. See www.heartfoundation.org.au/active-living/get-active accessed 25 February 2016

[6] Meaning that cyclists would have to travel four times the distance to get the same activity benefit as walkers.

[7] Meaning that cyclists would only need to cycle half the distance to get the same activity benefit as walkers. Combining this intensity effect with the speed effect in the previous footnotes leads to the result that cyclists need to cycle twice the kilometres to get the same activity benefit as walkers.

[8] In some instances, base case trip length data might not be available. SKM/PwC (2010, p.37) reported that each active travel trip replaces 0.616 car kilometres but the source of this estimate is not described.

[9] Crash rates for motorcyclists are higher again than those for pedestrians and cyclists (see Table 16).

[10] For example, a site such as a pedestrian cycle road crossing with low active traffic volumes might have no crash history yet be inherently risky.

[11] The value of statistical life incorporated in the willingness to pay estimates in Table 23 derives from recent research into the value road users place on avoiding premature death in a range of trip choice situations (see Austroads (2015) and Volume 2 of NGTSM). Those values might not be readily transferrable to active travellers. The value of statistical life that is included in the health benefit estimates has a broader base, representing the willingness of individuals to pay to avoid a small increase in the risk of premature death.

[12] Jacobsen, PL (2003) ‘Safety in numbers: more walkers and bicyclists, safer walking and bicycling’ in Injury Prevention, Vol 9, No 5, pp.205-9

[13] There is evidence from Europe that for a trip length under 5 km, cycling can be quicker door to door than the private car and for trips under 9 km can be quicker than the train. Walking on the other hand has no time advantage over other modes (see EC, (1999), p.11). For walkers, travel time benefits might be negative depending on trip length.

[14] Trip start to trip end speed would exclude the walk trip to/from the bicycle storage point or shower/change facility.