Dr Richard Tooth

LECG is a global expert services firm with highly credentialed experts and

professional staff with specialist knowledge in regulation, economics, financial and

statistical theories and analysis, as well as in-depth knowledge of specific markets

and industries. The company's experts provide independent testimony, original

authoritative studies and strategic advice to both public and private sector clients

including legislative, judicial, regulatory, policy and business decision-makers.

LECG is listed on the NASDAQ Stock Exchange and has approximately 700

experts and professional staff worldwide. These experts are renowned academics,

former senior government officials, experienced industry leaders and seasoned

consultants.

Level 6, 39 London Circuit

PO Box 266

Canberra City ACT 2601, Australia

Ph: (61 2) 6263 5941

Fax: (61 2) 6230 5269

Level 14, 68 Pitt Street

GPO Box 220

Sydney NSW 2001, Australia

Ph: (61 2) 9234 0200

Fax: (61 2) 9234 0201

Level 2, 65 Southbank Boulevard

Southbank VIC 3000, Australia

Ph: (61 3) 9626 4333

Fax: (61 3) 9626 4321

Level 9, Axon House, 1 Willeston Street

PO Box 587, Wellington 6001

Ph: (64 4) 472 0590

Fax: (64 4) 472 0596

Level 17, West Plaza Building

3-5 Albert Street

Auckland 1010

Ph: (64 9) 913 6240

Fax: (64 9) 913 4650

3505-06 Tower Two, Lippo Centre

89 Queensway

Admiralty, Hong Kong

Ph: (852) 2106 3800

Fax: (852) 2106 3999

Name: Dr Richard Tooth

Telephone: (02) 9234 0216

Email: rtooth@lecg.com

3.1 Overview of road crash costs .................................................................... 8

3.2 The loss of life and quality of life ............................................................. 10

3.3 Extensions to existing estimates ............................................................. 14

3.4 VSL approaches and values used around the world............................... 18

3.5 The cost of road crashes using the WTP approach ................................ 23

4.1 Internalisation of costs by the driver at fault ............................................ 25

4.2 Other external costs associated with vehicle use ................................... 27

Road crashes are a significant issue for Australia. Each year over a thousand people die and tens of thousands of people are seriously injured on our roads.

An understanding of the road crash costs is important in guiding policies towards greater road safety and influencing planning and policy across all transport modes. The Bureau of Infrastructure Transport and Regional Economics (BITRE) recently published a study (BITRE study) that estimated the social road crash costs at $17.85 billion in 2006. Like the similar study conducted 10 years before, the study acknowledges the headline estimate is conservative and offers alternative estimates. However there is a risk that, as with the previous study, the conservative estimate will become the basis for analysis of road crash costs in Australia.

This paper re-evaluates, from a public policy perspective, the costs of road crashes and the extent to which these amounts are borne by the users who contribute to them.

The loss of life and quality of life provides a challenge to estimating the costs of road crashes. A vehicle can be replaced (at a market determined price) — a life cannot.

Policy analysis is forward looking. As such we do not need to value lives lost but rather the benefit of reducing further risk to life. This is achievable. Everyday people make decisions that trade-off risks to their lives against other benefits and in doing so, exhibit a willingness to pay (WTP) for risk reduction. With information on this WTP, policy makers may estimate the value of preventing a fatality (VPF) also more commonly referred to as a value of a statistical life (VSL).

Government agencies around the world have adopted alternative approaches to estimating VSL. The human capital approach, a variant of which is used in the BITRE 2010 study, essentially involves estimating the value of productive output of people over their remaining lifetime. Applications of the human capital approach (including proposed hybrid alternatives) have largely excluded benefits of life (e.g. leisure) not associated with production and thus underestimated the value derived from life.

More direct estimates of VSL are obtained from studies using a ‘WTP approach’ based on peoples’ stated preferences (i.e. surveys) or revealed preferences (i.e. observed behaviour) on WTP for reduced risks. The results of these studies have confirmed that the human capital approach has led to a significant underestimate of the VSL.

While, the WTP approach has been criticised for having great variation in results, it is largely accepted as the preferred basis for policy analysis. Transport and other government agencies in the United States, UK, Canada, New Zealand and Europe now all promote the use of the WTP approach to estimating VSL.

Transport departments have historically tended to use fairly low estimates for VSL. This is changing, with significant recent revisions made in the US and Canada and pressure for change in New Zealand. Recent values and approaches used and recommended are shown in the table below.

Sources: See Table 3 for details.

Recent Australian and international evidence suggest that a VSL of around $6 million (in 2006 dollars) should be used to reflect road users’ WTP to avoid risks to themselves.

This amount lies between European and North American values used for policy analysis.

Given the availability of WTP estimates, there is no justification for using of a human capital approach that underestimates VSL in policy analysis.

When the recent estimates for the VSL ($6m in 2006) are used, the total costs of road

crashes rises significantly from $17.85 billion to around $28 billion in 2006. This may

also be conservative due to other costs not borne by individuals.

For example, the above amount may underestimate society’s WTP to avoid risks to life,

as loss of life causes great pain and suffering to friends and family. Empirical evidence

supports expectations that people are willing to pay significant amounts to protect

friends and family. While research on this social WTP is limited, the existing research

supports making an upward adjustment to individual WTP and rejects existing

arguments put forward to exclude the effect.

Road crashes, relative to other causes of death, affect a young section of the population

and thus have a much greater impact in terms of life years lost. Individual WTP to avoid

risks also varies with age, first rising as wealth increases but eventually declining with

decreased life expectancy. When coupled with the WTP of others (particularly parents)

for risk reduction for an individual, it seems likely that the social WTP for risk reduction

will be largely declining with age of the individual. In such case the VSL for prevention

of road crash fatalities would be relatively high compared to other causes of death.

The external costs of an activity are those not borne by the party undertaking the activity.

External costs of road use have generally been estimated as being a small proportion of

the total social costs. An increase in the human costs of road crashes would cause a

significant upward revision to the estimate of external costs as much of the human cost

is not borne by the road-users responsible for crashes.

An additional issue for road crash policy is that people are overconfident. On average

people overestimate their driving ability and underestimate their risk of an accident. The

implication of this is that drivers do not internalise all their own costs. Given that the

human costs of road crashes can be estimated in tens of billions of dollars, a small

underestimate of risk can have significant implications for the size of the costs that are

not internalised.

In a two vehicle collision, the occupants of the smaller vehicle are far more likely to be

killed or seriously injured than those in the larger vehicle. A key implication is that by

their mere presence, heavy vehicles impose external costs on other road users. These

costs include the additional severity of accidents, the additional discomfort to drivers of

light vehicles and distortions in people’s choice of vehicle and road use.

Recent increases in the size of vehicles appear have led to increased concerns about the

safety of users of conventional vehicles and other road users such as pedestrians, cyclists

and motor-cyclists. Furthermore there is some evidence that increased vehicle size is

associated with more aggressive driving behaviour.

The presence of larger vehicles encourages consumers, in the interest of their own safety,

to themselves purchase larger vehicles. This effect has been described as an ‘arms race’

for larger vehicles which results in increased private spending on larger vehicles and

running costs and greater social costs associated with emissions and congestion.

The problem of vehicle size is an increasingly important issue that demands further

attention. It should be possible to estimate the impact of replacing large vehicles with

smaller vehicles or other forms of transport. Such analysis has been undertaken for the

United States where it has been estimated that the safety benefits of replacing sports

utility vehicles with smaller cars is in excess of A$300 per vehicle per year.

The key conclusions from this review are:

No-one would deny that road crashes are a significant issue in Australia. Each year there

are over a thousand people killed and tens of thousands of people seriously injured on

our roads. However, there is substantial debate as to the social costs of road crashes and

the extent to which these are borne by the road-users who contribute to them. An

understanding of these costs is important in evaluating the policies aimed at improving

road safety.

This paper examines issues in the debate over road crash costs. The objective of this

study is not to provide an estimate of the costs or to recommend policies, but rather to

review and clarify a number of important issues relating to road crash costs and the

extent to which these are incorporated into people’s decisions.

The paper is divided into four additional sections. Section 2 provides some background

as to the costs associated with road crashes and how costs are used in policy formation.

Section 3 examines the costs of road crashes in more detail. A key focus of this section

is the debate on costs associated with loss of life, including the approaches and estimates

used in other countries.

Section 4 examines the extent to which the costs of road crashes are internalised, that is

borne by the road users who contribute to these costs and incorporated into their

decision making.

Section 5 concludes with a brief discussion of implications.

Road crashes have a number of negative impacts on society. A summary of the type of

impacts that are commonly considered is listed in Table 1 below. As described in this

table, costs can be broadly divided into human costs, property damage and other general

costs.

rehabilitation & long-term care)

Lost labour (workplace and household)

Loss of life and quality of life

Legal costs

Premature funeral expenses

Publicly funded treatment

Correctional services

Coroner, funeral

Workplace disruption

Additional local air pollution

Pain and suffering of friends and relatives

Vehicle, towing repairs and

replacement

Unavailability of vehicles

Non-vehicle property damage

Travel delays

Police, fire and emergency services

Source: Adapted from BITRE 2010

Of note, in addition to direct costs to the road-users involved, the costs include a number

of indirect costs to others. These include costs borne by other road users (e.g. impacted

by travel delays), Government institutions (e.g. police and emergency services), the

workplace (included in ‘workplace disruption’) and other household members (whose

lives are disrupted).

The division between human costs and property damage is an important one. Property

can be replaced, human life cannot. Furthermore, there are market determined prices for

the replacement or repair of most property. Whereas there are market prices for many

treatment costs, there is no market price for the loss of life and quality of life. It is these

later costs which pose the greatest challenge, generate the most significant debate and

are the focus of this paper.

The appropriate method for valuing costs depends on the purpose. There are a number of

reasons why we might be interested in the costs of road crashes. These include:

For some categories the valuation will be similar regardless of the purpose of valuation.

This is the case for goods and services that are easily replaced at established prices in

the market. For example, the pre-crash market value of a vehicle lost during a crash

provides a good estimate of the insurance cover a vehicle owner obtains, the loss to

productivity and the benefit to society of preventing the vehicle’s loss.

In the case of loss of life or quality of life, the purpose for the analysis is important. For

example, people with no dependents may pay significant sums to reduce risk to their

lives but have no need for compensation once dead.

This paper examines the costs of road crashes for the purposes of policy analysis i.e. so

as to evaluate the costs and benefits of alternative policies. For this purpose we are not

interested in the cost of road crashes that have occurred but rather the benefits of

reducing the likelihood and severity of road crashes in the future.

Taking a forward looking approach to costing simplifies the problem of estimating

human costs. We do not wish to value life but rather to value a change in the risk to life.

This is achievable. Every day people make decisions that trade-off risk to their lives

against other benefits. For example, people may choose to pay more to fly on a safer

airline, or take more time crossing a road by crossing at a safer location. In making these

decisions, people exhibit a willingness to pay (or willingness to accept) for a change in

the risk to their life.

Thus, as Kip Viscusi, a leading academic and policy adviser on valuing of life, says

most succinctly:

Viscusi (2008)

The challenge of determining what society would be willing to pay (WTP) for the

benefit of reducing the likelihood of accidents is discussed in the following section.

For the purposes of policy analysis we are also interested in the extent that costs of road

crashes are incorporated into road-users’ decision making. If the costs were fully

incorporated then, assuming well-functioning markets, there would be little justification

for government intervention into road safety. However, many costs are external to the

decisions of road uses who contribute to them. As shown in Table 1 there are many costs

not borne by the individuals involved (and their insurers).

Although this paper focuses on the cost of road crashes and not specific policies, some

comment on policy analysis is appropriate.

The existence of external costs or other market failures do not by themselves justify

policy intervention. The benefits of any intervention need to be weighed against the

costs, particularly in light of changes in behaviour resulting from the policy.

For policy analysis we are often interested in costs associated with an activity (e.g. an

additional road use) and in particular the marginal cost of the activity. The marginal cost

of an activity is the additional cost caused by a small increase in the level of that

activity. Generally it is much easier to measure total and average costs than marginal

costs. It is not clear to what extent this is an issue with regards to road crashes. For

example, increased road use could lead to an increased rate (per level of road use) of

road crashes but a decreased rate of fatalities due to lower speeds.

For most transport policy analysis, it is also necessary to consider alternative transport

modes. Some of the issues raised, particularly in Section 3, also apply to other costs of

transportation and other transport modes. For example, similar arguments are also

applicable to issues related to air pollution caused by transport and the cost of incidents

that occur at railway level crossings. Due to the relatively high incidence of road

crashes, the issues raised are far more significant to road use than other modes such as

rail.

The valuation approaches adopted should be consistent across all transport modes. No

one transport mode cannot be considered in isolation of other modes as, for example:

An important reference point for studying the costs of road crashes in Australia is the

study by Bureau of Infrastructure Transport and Regional Economics (BITRE) titled

‘Cost of road crashes in Australia 2006’ (“the BITRE study”). The study (BITRE 2010)

was a comprehensive examination of the costs of road crashes and has since been the

basis of many estimates of the total and external costs of road crashes.

The main results of the study are summarised in Table 2 below. The headline result of

this study was that road crash costs in Australia in 2006 were estimated around $17.85

billion (in 2006 dollars), a $1.45 billion (in real terms) decrease from the previous study

for 1996 (also summarised in Table 2). As with the 1996 estimate, the recent study noted

higher estimates. However, the experience of the previous study is that it is the

conservative approach that led to the headline estimate and that is the basis of most

4,016 5,690

Loss of life and quality of life

2,278 1,768

Treatment costs & associated costs

(including legal costs, funeral)

4,080 3,123

Other largely not borne by individuals/

insurer (e.g. workplace disruption)

425 113

Non-vehicle property damage

5,332 4,482

Police and fire and emergency services

1,969 983

Insurance administration

1,193 1,691

Source: BITRE 2010 and BTE 2000 reports.

valuing human losses (BTE 2000), presenting a conservative estimate of the losses to society

from road fatalities.’

for the Australasian Railways Association (PJP 2005), recent academic studies (e.g. Connelly and

Supangan 2006), many submissions to the 2007 Productivity Commission Review on “Road and

Rail Freight Infrastructure Pricing” and the Australian Transport Council (ATC 2006 V5, p. 76).

Most of the costs (such as property damage) are fairly non-controversial in their

estimation. The most controversial aspect of the study is the amount associated with

human costs and in particular that relating to the loss of life and quality of life. These

amounts are of greater interest because, as is discussed in this paper, they can be very

significant and can represent a large amount of the external costs of road crashes.

Many of the other costs recorded in Table 2 are costs borne by the individuals at fault or

their insurer. These include the property damage, some of the lost labour productivity

and the insurer’s costs for medical treatment and the loss of their own life and quality of

life.

As noted above, for policy analysis of road crashes, it is society’s willingness to pay

use, policy makers use an amount in terms of the expected value of preventing a single

common term VSL.

The VSL is then estimated by dividing the WTP by the likelihood of a risk. Thus, for

example, if we find people are willing to pay $50 to reduce the risk of death by 1 in

100,000, the VSL is estimated as $5 million (equal to $50 x 100,000).

Of note, for purposes of road crash analysis (and most other analysis) WTP and VSL

reflect small levels of risk to life. Thus, for example, the VSL cannot be interpreted as

the amount someone would pay to avoid a certain death. Rather VSL is simply an

amount reflecting the sum of the WTP of a large number of individuals to reduce small

risks to their lives.

The issues relating to the value of preventing injuries, loss of life, and loss of quality of

life are similar. Road crash injuries are far more frequent than fatalities and represent a

significant proportion of the costs estimated in Table 2 above.

It is common for researchers to estimate the value of risks to quality of life as a

proportion of VSL. For example, the BTE 2000 study estimated that, for loss of quality

of life, the value for risk reduction of a serious injury was 11% of that for a life. In

further analysis that examined variations to the VSL the study assumed that this

proportion was maintained.

There are a number of approaches to estimating VSL. The two most common

approaches are described as the ‘human capital’ approach and the WTP approach. These

are discussed below.

The ‘human capital’ approach essentially involves estimating the value of productive

output of people over their remaining life (see Box 1 below). The often quoted estimate

of road crash costs that appears in Table 2 is based on a ‘human capital’ approach to the

valuation of life supplemented with an amount for lost quality of life.

The common applications of the method (including that used in the recent BITRE study)

do not attempt to measure the WTP to reduce risks. There is, however, a revised

approach known as a hybrid human capital/WTP that is based on a formulation which,

The proposed application of the revised hybrid human capital/WTP, like the more

common approach, is however flawed in what it excludes. In theory the hybrid human

capital/WTP measure would include not just the easy-to-measure benefits from labour

and non-labour income but all the other benefits we receive from living, including the

enjoyment of leisure, any enjoyment of life we have while working and the avoidance of

pain and suffering of dying. In practice, due to the difficulty of measurement, it is

proposed that these additional benefits from life are excluded – that is, they are given no

value. As advocates of the approach acknowledge, this leads to a large underestimate of

the value of life.

Potentially a hybrid type of approach could lead to reasonable estimates of a VSL

consistent with the WTP approach. This would, however, require estimating the hard-tomeasure

benefits of life (e.g. leisure). Such an exercise may provide some useful

insights, however, the simplest and most robust way to estimate the additional benefits

may be using the WTP approach.

In the absence of other VSL measures, Government agencies in Australia have used the

human capital approach for the purposes of policy analysis. However, as is discussed

below, more direct estimates of VSL have been obtained and it is now widely accepted

that the human capital approach leads to an underestimate of VSL. Given the availability

of other measures, there is no justification for using a human capital approach that

underestimates VSL in policy analysis evaluating the costs and benefits of the risk of

crashes.

The BITRE study application of the human capital approach estimated some

aspects of the value created by people including the lost labour:

rate paid by employers; and

household and community work and average income levels.

Such an approach is reasonably simple to apply, based on reasonably reliable

data and produces fairly consistent results.

Unfortunately the benefits are only achieved by excluding the difficult-toestimate

aspects of the value of human life to society. The estimate does not

include the value people bring to each other or themselves in other activities.

Thus for example, the value of time relaxing by one self or spent with friends is

not counted. The result has some unfortunate implications including that the lives

of retired (not doing housework) people have no value!

Another measure that is sometimes considered for estimating human costs is based on

the level of compensation for loss (of life or quality of life) that may be found from life

insurance payouts or from court awards. However, for loss of life or quality of life, the

level of compensation is not appropriate as a guide to the value of risk reduction. For

example, we would not choose to insure the loss of a loved one, if we had no better use

of additional wealth after our loss. For similar reasons, court awards typically do not

include an amount that reflects VSL and so are not applicable to use in estimating

The WTP approach to estimating VSL, as the name suggests, involves attempting to

directly estimate WTP to reduce risk to life. Broadly there have been two methods to the

WTP approach to estimating VSL. These are based on stated preference and revealed

preferences. Examples of these are given in Box 2.

that can be used to elicit how much they are willing to pay for a small reduction

in risk to life. As with all consumer surveys, results can be very sensitive to

survey design and there have been a variety of different methods trialled to

address this challenge.

A recent Australian example of a stated preference study is that by Hensher et al

(2009). The researchers asked road users to choose between alternative routes

with different safety records, travel times, costs and other characteristics. Based

on the results, researchers were able to estimate how much road users value risk

reduction.

majority of these studies have been based on people’s willingness to accept

riskier jobs. A number of other studies have been based on people’s product

choices.

The most common revealed preference study is based on wage differentials

between jobs with different risks to life. It is commonly found that workers are

compensated for facing higher risks to life with higher wages. The level of

additional compensation and the additional risk is used to estimate the VSL.

There have been numerous (in excess of 200) studies that have attempted to estimate

VSL via a WTP approach in Australia and around the world. There are also a number of

meta-studies (studies of studies) which examined the ranges of VSL estimates. A recent

meta-study by Access Economics (2008) included an analysis of 99 Australian and

international studies which ‘yielded an average VSL of A$6.0 million in 2006

Australian dollars’ and a suggested range for sensitivity analysis of $3.7 million to $8.1

transport study (Hensher et al. 2009) conducted in Australia using a stated preference

approach.

There have been many other meta-studies with a range of results. The US Department of

Transport’s most recent VSL estimate is based on the average of five meta-studies

ranging from US$2.6 to 8.5 million. Four of these meta-studies produced estimates

Australian and international studies reviewed.

higher than the Access Economics estimate. For example, Viscusi and Aldy (2003)

international studies.

The estimated VSL has been shown to vary with a number of dimensions including age,

income and even cause of death. Of note, more recent studies tend to show higher results

suggesting the VSL is increasing at a faster rate than is generally assumed. This is

important for considering how the VSL should increase over time.

Concerns over the variation in estimates have declined over time with more research that

has led to better techniques, more estimates and greater understanding of the variations

between estimates. Reasonably tight ranges for WTP estimates are now being developed

and these help to confirm that the human capital method leads to a large undervaluation.

In their 2008 review, Access Economics, based on their sensitivity analysis,

recommended for Australia a minimum value well above the human capital valuation.

A second concern appears to be based on the policy implications of adopting WTP

estimates for public policy. Higher valuations of the cost of road crashes could be used

as justification for large increases in public spending or greater safety regulations. There

is significant concern that many such interventions have been wasteful and in some

valuation of VSL; rather they should prompt more rigorous analysis of policies.

The WTP based methods that are used to estimate VSL are largely designed to capture a

person’s WTP to reduce risks to his or her self. But the death or serious injury of a

person affects other people, particularly family and friends, who cared for the person. It

is thus natural to ask whether the existing estimates of WTP for reducing risks take into

There has been some debate as to how the concern of others should be incorporated into

values of VSL. Most discussion on this issue is centred around the belief that the

appears reflected in stated preference studies that find people are more reluctant to pay for public

investments that are claimed to deliver the same safety benefits as a private benefit; a finding that

has been interpreted as people’s concern relating to the effectiveness of public investment (see

Andersson and Lindberg 2008).

feels responsible.

concern for the safety of others is based on altruism (i.e. an unselfish concern for the

welfare of others). Bergstrom (2005) showed that if people are ‘pure’ altruists who care

for both the safety of others and the costs of improving safety, then no adjustment is

required to existing VSL estimates. This argument has been the basis of a recent

European decision to make no adjustment for VSL to reflect the concerns of friends and

families (HEATCO 2006 D5, p. 85).

However, our concern for others is more than just pure altruism. We invest time and

money in developing relationships, enjoy the company of good friends and family and

miss them when they are away. Our lives are enriched by our relationships with others

the few empirical studies that have examined the issue find that people are willing to pay

to enhance the safety of family and friends in a way that is inconsistent with pure

There would also be no need to adjust WTP estimates if people’s attitude to risk

incorporated concern for the impact on others of their death. This may be the case to a

degree; nevertheless it seems unlikely that this is fully incorporated. For example, it

appears unlikely that a young male’s risk-taking fully reflects his parent’s interests. As

such, in general, society’s WTP for risk reduction of an individual is likely to be higher

than the individual’s WTP.

The extent to which existing estimates of VSL should be supplemented will likely

depend on how estimates were constructed. Some WTP valuations may reflect, to a

degree, household decisions and thus incorporate concern for household members. For

example, spouses may influence their partner’s choice of job and thus their concerns

may be embedded in studies that examine wage-risk trade-offs. Similarly product

purchase decisions and amounts chosen in stated preference studies are likely to be

drawn from a household budget. Regardless, people’s concern for others clearly extends

to people outside of the household.

(2008). They find that the loss of a loved one has a significant impact on measures of people’s

psychological well-being and life satisfaction. They also, by way of illustration, estimate the per

annum financial compensation that would be sufficient to restore these measures to their pre-loss

level. These vary from around A$30,000 for a sibling to over A$200,000 for the loss of a spouse

in the first year.

people’s lives than cash equivalents (see Jacobsen et al. 2007). This suggests people are

concerned with others’ safety more than their well-being. This concern has often been described

as a form of altruism called safety-paternalistic.

Given these considerations, increasing the values of VSL to reflect societal WTP for risk

reduction is justified. By how much is debatable. Unfortunately, as there is only limited

people’s WTP to reduce risk to their children is greater than that for themselves. There

is similar evidence for other household members. People also have demonstrated a WTP

to reduce risks to other relatives and friends outside of the household. In their study,

Andersson and Lindberg (2008) estimated this WTP be on average around one third to

one half of the person’s WTP for their own risk reduction. Furthermore they found

people were willing to pay to reduce risks to multiple friends and relatives.

Given the level of grief we observe by parents, spouses and other relatives and friends

for car crash victims, it would seem hard to justify not making some adjustment.

Lindberg (2001), based on existing research, recommends increasing the VSL estimates

important area for further research.

Road-crash victims have a different profile to those killed or seriously injured by other

causes. Most notably road crash victims tend to be younger people. Figure 1 below

shows the age profile of fatalities caused by road crashes and that of all causes. In terms

of years of life lost, road crashes have a particularly significant impact.

While it is tempting to focus on the years of life lost, policy analysis should be based on

society’s willingness to pay for the reduced risk to life. Policies that generate benefits

and impose costs for a group would ideally reflect the characteristics of that group. It is

efficient that the cost-benefit analysis of safety measures reflect the WTP of the roadusers

impacted.

evidence.

appears consistent with subsequent studies (e.g. Andersson and Lindberg 2008).

0%

5%

10%

15%

20%

25%

30%

35%

0 years 1–14

years

15–24

years

25–34

years

35–44

years

45–54

years

55–64

years

65–74

years

75–84

years

85–94

years

95+

years

Road crashes

All causes

Source: Road accidents from calendar year 2007 Australian Road Deaths Database.

All causes from ABS 3303.0 Causes of Death, Australia, 2007

Our WTP to avoid risks also changes by age. It would seem intuitive that the WTP to

avoid a risk should decline with age as we would expect a rational person would pay

more to reduce risk to their life the more years of expectant life they have. However a

person’s wealth and spending also changes over their lifetime. It is quite rational that an

Both theory and empirical work suggest that an individual’s WTP to avoid risk initially

friends and family. In particular, as discussed above, empirical evidence suggest

purchase a safety device at a young age and instead invest the funds saved so as to spend more on

safety at a later stage in life.

initially relatively high and would then continually decline with age.

government agencies currently employ different VSL estimates for their analysis.

Consideration of the age profile suggests that for the purposes of costing road crashes

and for analysing road crash policy, relative to some other agencies, a higher VSL

should be used.

The above discussion has been in relation to society’s willingness to pay for increased

safety. In some circumstances a change being evaluated will lead to an increase in risks

to existing users (for example, whether to allow large trucks on the road). In such

contexts it is society’s willingness to accept (WTA) the increased risks that should be

relevant.

While we might expect that WTP and WTA for a small change in risk should be similar,

empirical studies frequently display a large discrepancy between them. A recent review

(Horowitz and McConnell 2002) of 45 studies found that the mean WTA/WTP ratio is

around seven to one. Transport safety research found the WTA-based VSL to be around

three to five times the WTP-based amount (see NZ DOL 2004, p. 89).

Adoption of WTA based measures for VSL would be very significant for policy

analysis. Although possible explanations, including problems of survey design and

behavioural issues, have been suggested, there appears to be no clear resolution for the

difference in WTA and WTP estimates. Given the lack of clarity, it is perhaps not

surprising that policy makers have been reluctant to use WTA based measures.

Regardless where applicable, the implications of using a WTA based estimate should at

least be considered and recognised.

The approaches and VSL values used in a number of countries with similar economic

development and Government agencies are provided in Table 3 below. Some

qualifications are necessary. This is by no means a comprehensive review. Typically,

there are not strict requirements imposed by Government agencies; agencies may just

recommend approaches or values to be used or simply provide analysis that may be used.

of distinctions [of VSL] by age set off a political firestorm.’

As shown in Table 3, the majority of VSL values used are based on a WTP approach.

There has been a clear shift towards use of the WTP over the human capital approach.

Many countries that had once employed human capital approaches have shifted to the

WTP approach.

A number of European countries have, up until recently, used a human capital approach.

In 2006 a project to develop Harmonised European Approaches for Transport Costing

(HEATCO) recommended that VSL be based on a WTP approach. Of the countries and

agencies reviewed, transport in Australia appears to be the only case where a WTP

approach is not actively promoted.

The VSL estimates reported in Table 3 vary (in 2006 Australian dollars) from less than

$1.5 million to over $9 million. European countries have tended to favour forming VSL

based on stated preference (as opposed to revealed preference) studies. This appears to

have led to relatively low VSLs compared to the results of broader international studies

and to those used in North America.

Transport

BITRE study: Human capital approach used for headline analysis. Value included lost labour

productivity plus A$0.45m for lost quality of life. Total value of a fatality in 2006 was about

A$2.4m (including some costs not borne by individual such as workplace disruption).

Other Office of Best Practice Regulation recommends A$3.5m (2007 dollars) see

Australian Safety and Compensation Council sponsored analysis by Access Economics (2008)

who, based on WTP studies, recommended a VSL of A$6.0m.

Transport

In 1994 Transport Canada adopted an amount of CAD$1.5m (1991 dollars) based on a WTP

approach. This rose to CAD$1.76m in 2002 and was reviewed again in the mid 2000s. Recent

recommendations include a primary VSL Estimate for Canadian Policy Analysis in 2007 prices

Other Health Canada – CAD$5.7m in 2002. A 2005 Regulatory Impact Analysis Statement used

Transport

Historically there was great variation by country. In 2005, North / West European countries

with exception of Germany and Denmark had adopted WTP approach, whereas other

countries (in South and East) tended to use a human capital or liability payments approach.

In 2006, a project to harmonise European approaches recommended WTP be used for all

countries and that the WTP be based on local stated preference studies. Values proposed for

loss of quality of life for Western European countries ranged from €0.76m (A$1.25m) in

Greece to €2.63m (A$4.34m) in Norway. An additional 10% is proposed to cover the loss of

production net of an individual’s consumption. Source: HEATCO (2006 D5).

Transport

Transport agency adopted the WTP method in 1991 and established VSL at NZ$2 million in

1991 following a survey in 1989-90. Following a second survey in 1997-98 a 45% increase

was recommended but this has not been adopted yet. The VSL is increased annually at the

rate of increase in average hourly earnings (ordinary time). Current VSL recommendation in

Transport

US Department of Transport (US DOT) guidelines are based on a WTP based approach. A

significant revision was undertaken in 2008, raising the VSL from US$3m to US$5.8m. This

component of the Employment Cost Index, in constant dollars, and the Consumer Price Index’

(US DOT 2009). Guidelines recommend that no adjustment is required for loss of productivity.

The value of a statistical injury is set as a proportion of VSL.

benefit of a fatality averted, when agencies have not supplied a different measure.

The Department of Agriculture and the Food and Drug Administration have used US$5m

Department of Labor, including Occupational Safety and Health Administration and the Mine

Safety Health Administration, “follows the lead of EPA.” (US DOT 2008).

Transport

The UK Department for Transport has used a value for the prevention of a road accident

fatality (VPF) since the 1950s. The department shifted from a ‘human capital’ to a WTP

methodology in 1988. The initial VSL measure was set at £0.5m, the lower end of range of

Other The U.K. Health and Safety Executive uses the Transport VPF as a base, but inflates the

amount for other factors.

Sources: Where not recorded: Australia – BTE 2000 Study; New Zealand – Government

websites, US DOT (2008) & US DOT(2009); Canada – Canada Gazette; UK DFT

(2007) & HEATCO (2006 D5); Mainland Europe – HEATCO (2006 D5)

Notes: *VSL amounts converted to Australian dollars at purchasing power parity prices and

to 2006 values based on changes in Australian GNP/capital.

transport departments have historically tended to use fairly low estimates. For example,

in the US, UK, Canada and Australia there is evidence that VSLs used or recommended

by transport agencies have been lower than that employed/recommended by

environmental and health agencies.

There appear to be a number of reasons for this. A key reason appears to be historical. In

general, transport agencies were early adopters of value-of-life measures and began

using the human capital approach before the WTP approach had taken hold. When the

switch to a WTP approach occurred, conservatively low values were chosen so as to

minimise the policy impact. An example of this is from the UK where the WTP

approach was adopted in 1987. Chilton et al. (1998) report that a range of values was

determined from a comprehensive review of the literature at the time and a value

selected ‘at the lower end of this range in order to temper a radical change of

methodology … with an element of caution.’

Another historical reason is that often current values are based on a VSL chosen some

time ago and updated over time at the growth rate of some index. International estimates

of VSLs based on WTP approaches have generally increased over time and as such early

adopters of the WTP approach tend to be relatively low compared to more recent

evidence. For example, the current New Zealand VSL is based on a survey conducted in

1989–90 inflated by increases in hourly earnings. A second survey report conducted in

1997–98 recommended a much higher VSL (around a 50% increase) but its findings

Finally, political interests may be a factor. Viscusi (2008) argues that the US

Department of Transport (US DOT) had adopted (between 1993 and 2008) a lower VSL

than supported by evidence as a ‘political compromise’ influenced by an amount

estimates at the bottom end of [the] range and sometimes below it’.

survey and the resulting [VSL] estimate, a decision has yet to be made as to whether or not to

adopt a new [VSL].’

(2009).

Although typically transport agencies were slow to adopt a WTP approach and chose

fairly low VSLs, there appears to be some recent change. Viscusi notes that in 2008 the

US Department of Transport ‘finally revamped its official VSL policy to bring it in line

with the literature’ and now uses a figure of US$6.0m. Similarly Transport Canada has

recently significantly increased the VSLs used in Regulatory Impact Analysis. Europe,

through its harmonisation plan appears to have uniformly adopted WTP with a material

increase in VSL measures.

Some agencies and policy papers have considered whether VSL amounts should vary

with age. The US Environmental Protection Agency proposed different VSLs for

different age groups. Environment Canada estimated a VSL that reflected the age mix of

deaths associated by air particulates.

None of the policies reviewed explicitly include an adjustment to VSL to reflect the

WTP by friends and family for risk reduction. As discussed above, HEATCO

considered an adjustment for relatives and friends but rejected it on the assumption that

It is common for the value of reduced risk of injury (value of a statistical injury, VSI) to

be based on the VSL. For example, the US DOT guidelines (US DOT 2009) provide

guidance based on valuing injuries of different severities, ranging from minor to fatal, as

a fraction of VSL (as shown in Table 4 below). These proportions were established in

1993 and have not changed as the VSL has been updated.

incorporates the ‘extra’ value placed on relatives and friends…” The document does not describe

the basis of the claim.

Source: US DOT (2009)

International comparisons are difficult because the definition of injuries and the

approaches used in valuing risk reduction can vary by country.

Of note, the BTE 2000 study assumed a ratio of 11%. New Zealand transport assumes

10% (NZ DOL 2004) and Europe uses a standard 13% (HEATCO D5 2006). IRAP

(2008) estimated that for serious injuries the VSI/VSL ratio for the United States was

15% based on a weighted average of the amounts in Table 4 above.

In addition to the headline analysis based on a human capital approach, the BITRE study

included estimates of the cost of road crashes using WTP-based approach for valuing

life (BITRE 2010 Section 7.3). The BITRE alternative estimates was constructed by

substituting a valuation of $2.4 million per fatality for estimates of VSL based on WTP

approaches. Injuries were valued using adjustments to quality of life based on quality

adjusted life years. When the a VSL derived from recent Australian research (Hensher et

al 2009) of $6.19 million is used, the estimated cost of road crashes in 2006 rises from

$17.85 billion to $27.12 billion.

While this increase is dramatic, there is reason to believe this estimate is still

conservative. The cost estimate excluded a number of costs that are not borne by the

individual (and thus not incorporated into an individual’s WTP).

Some amounts not borne by the individual include net transfers to government

associated with the lost income (i.e. future tax revenue less any future transfer) and

the recommended approach is to increase VSL values by an additional 10% to cover the

The WTP based estimate above also does not include a value for impact on friends and

family. What adjustment should be made depends also on the extent to which friends

and family are willing to pay to reduce risk to life and to injury. Unfortunately, the

extent of this is difficult to estimate given the limited research available. A sense of the

potential significance can be achieved with some assumptions. If the uplift of 40 per

cent that Lindberg (2001) proposes was applied to both life and injuries, the VSL used

would be around $8.5m (2006 dollars) and the total cost would increase by around $4

billion for fatalities (an additional uplift would be required for injuries). Clearly given

the limited research, such an amount needs to be treated with great caution. It does

however highlight the importance of more research in this area.

Furthermore, the cost of serious injuries in the WTP based estimate also appear to be

injuries is less than that of fatalities. In other countries that use a WTP approach the

reverse appears to be the case. For example, IRAP (2008) estimated that in the US the

ratio of the serious injuries to fatalities was 8.3 to 1 and the value of statistical injury

was on average 15% of a VSL implying that the cost of serious injuries to be around

example, a person’s WTP to avoid an accident will depend on his/her expectations as to

compensation in the case of an accident.

do not take account of the fixed cost to the hospital system of providing base capacity needed to

respond to road trauma.’

productivity and non-pecuniary costs (such as workplace disruption, correctional services etc)

The BITRE estimated these to around $100m (see the BITRE study Table7.1).

proportional to the WTP to reduce risk to life. That is, an increase in the estimate of the VSL

results in the same proportional increase in the loss of life and loss of quality of life from injury.

While this is a common and pragmatic approach, there is little empirical research that tests the

validity of this approach.

by country is Germany 1.1, UK 1.1, New Zealand 1.8, Sweden 1.6.

The question of who bears the cost of road crashes is a key issue for public policy

analysis. If those undertaking risky activities internalise all the costs and benefits of their

activity, then, acting rationally with good information, they will choose an optimal level

of risk. In such a world there would still be road crashes (and presumably fatalities) but

the frequency of these would reflect society’s choice.

However, it is well recognised that many activities related to road crashes incur external

costs to other parties, thus not giving sufficient incentive for people to reduce their risky

activity. The greater the external costs, the greater the risk that the costs of road crashes

is excessive.

external depend on the activity being examined. While, as previously discussed, some

road crash costs are not borne by road users, collectively the victims of road-crashes are

in the main road-users themselves. However, in examining the external costs of road use

we are interested in individuals’ decisions including how they use roads and how often.

A useful starting point for examining the external costs of road crashes is to consider the

extent to which costs associated with an individual’s driving behaviour are internalised.

road crash bears the costs of the crash. This section begins by reviewing this approach in

light of the analysis of the previous section. It then considers a number of modifications

to this approach, most notably the extent to which additional externalities exist due to

the relative size of vehicles.

Some road crash costs are external (i.e. not internalised) to the driver at fault in an

accident. These external costs include the increased burden on the public health system,

other system costs (such as travel delays resulting from the crash) and non-compensated

internalised. In particular, the legal system through enforcement of road traffic rules provides a

major deterrent to risky driving activity. Social programs and social pressure also contribute to

encouraging people to drive more safely.

It is common to view the internalised costs as including the liabilities of the individual

and the individual’s insurer to other parties. Under our current system, the parties at

fault for a road crash are liable for at least some of the costs of other parties. In doing so,

the party at fault internalises these costs on others. The liability for most claims is

passed on to the insurer and this may be viewed as diluting the extent to which costs are

internalised. However the insurer is able to modify insurance premiums for risk and

provide incentives for the insured to drive more carefully.

The BITRE study did not explicitly identify the external costs of road crashes. Martin

(2005) reports Australian estimates of road crash costs ranging from $2.5 to 7 billion per

annum.

The estimates of external costs increase dramatically once the revisions made in the

previous section are incorporated. This increase includes the following.

fault. Based on the assumptions in Section 3.5 above and an additional assumption

that half of the loss of quality of life is of parties not at fault, this is around $13

billion.

seriously injured. By construction this is an external cost to the driver. The

estimated increase would be over $4 billion (on fatalities alone) if the uplift of 40

per cent proposed by Lindberg (2001) was used.

Again, the purpose of the example is not to provide an accurate estimate of the external

costs, but rather to highlight the potential size of the external costs of road crashes.

The interest in distinguishing between external and internal costs is due to the belief that

individuals will incorporate their own (i.e. internal) costs into their own decision

making. Unfortunately this does not always appear to be the case and particularly not so

in the case of driving.

There is increasing research on, and recognition of, behavioural biases, which may cause

people not to act in their own best interests. An important bias for driving behaviour is

overconfidence.

Evidence of overconfidence with regard to driving has been found in numerous

seems that on average we have reasonably good perceptions of societal risks, but that we

overestimate our own driving ability and underestimate the likelihood that we will be

involved in a crash.

The implication of the problem of overconfidence is that the traditional view of what

costs are internalised needs some revision. A bias of overconfidence means that much

less of the expected cost of road crashes is truly internalised into driver behaviour. This

includes costs claimed via insurance (e.g. property damage) and loss of one’s own

quality of life.

Insurers do not suffer from bias of overconfidence in estimating the crash risk of their

insured clients. Arguably the costs incurred by insurers are thus appropriately

internalised via the setting of insurance premiums and associated activities (such as ‘pay

by the mile’ insurance and rewards for safe driving). However, there are limits to the

costs covered by insurance (e.g. claims do not pay for loss of quality of life), and

restrictions on the costs of insurance premiums (e.g. limits on how Compulsory Third

Party insurance premiums are applied).

The lack of internalisation of one’s own loss of life is significant. Using recent WTP

estimates, the costs of loss of life and quality of life is estimated at over $20 billion. A

significant portion of this is borne by the driver. If drivers underestimated their risk of

having a road crash by 10%, then the additional cost not internalised could exceed a

billion dollars.

A topical issue is the effect that additional vehicles have on the incidence and severity of

road crashes. Intuitively it would seem that the likelihood of any vehicle being involved

in a road crash increases with each additional vehicle on the road. As such, it would

appear that the marginal (i.e. additional) cost of road-use would be much higher than the

average cost and that the external costs of any road-use are greater. Thus from the point

of view of road crashes, other drivers are better off the less roads are used.

The size of this effect is debatable. More congested traffic can prompt people to drive

more slowly and carefully. While this reflects costs to those road users, the effect can be

to lower the incidence and severity of road crashes. Greater road use can also motivate

and fund greater investment in infrastructure, which can both improve road safety and

work.

The effect of additional road use is clearer when it applies to heavy vehicles. When it

comes to road crashes, size matters. The severity of an accident is closely linked to the

relative size of the vehicles involved. In a two vehicle collision, the occupants of the

smaller vehicle are far more likely to be killed or seriously injured than in the larger

vehicle. This is the case for standard passenger cars involved in collisions with heavier

sports utility vehicles (SUVs) and for all passenger vehicles involved in accidents with

trucks and buses.

A key implication of the importance of relative size is that by their mere presence, heavy

vehicles impose external costs for other road users. These costs include the additional

severity of accidents, the additional discomfort to drivers of light vehicles and

distortions in people’s choice of vehicle and road use. Simply put, people are better off

driving on roads surrounded by lighter vehicles.

The concern about relative size of passenger vehicles has increased in recent years

(particularly in the US) due to the increased prevalence of SUVs and even heavier

vehicles such as Hummers. It wasn’t clear at first whether increased passenger vehicle

size was necessarily bad for vehicle safety. Intuitively it would seem that if everyone

used larger vehicles then we might even be safer. However, not everyone chooses to

drive the larger cars and so large disparities in vehicle size exist. Furthermore, there are

additional reasons to be concerned with vehicle size.

Firstly, pedestrians, motor cyclists and cyclists suffer as a result of larger vehicles. All of

these parties are more likely to be killed or injured if hit by a larger vehicle.

Secondly, there is some evidence that the driving behaviour of passenger vehicles

heavier passenger vehicles are more likely to be involved in a crash and drivers of equal

sized vehicles are more likely to be killed or seriously injured when the vehicles are

larger.

There is an additional problem of relative size. The presence of larger vehicles

encourages consumers, in the interests of their own safety, to themselves purchase larger

vehicles. This effect has been described as an ‘arms race’ for larger vehicles which

results in increased private spending on larger vehicles and running costs and greater

social costs associated with emissions and congestion. Both these additional private

costs and public costs are external costs associated with the presence of large vehicles

on the road.

The issue of relative vehicle size is also gaining greater recognition with regard to road

terms of road crashes, they are over represented in terms of road-fatalities and serious

equal, more trucks on the roads increases the risks to other road users.

It should be possible to make reasonable estimates of the relative size externality

associated with a large vehicle such as a truck or SUV. Given available data on the

accident rates and fatalities by vehicle types it is possible to estimate the impact of

replacing large vehicles with smaller vehicles or other forms of transport. Using such an

analysis, White (2004) estimated that the safety benefits of replacing a million SUVs

In addition to the increased accident externality, the external costs include the increased

discomfort for road users and the ‘arms race’ effect. An indication of the former could

potentially be obtained through stated preference surveys. Estimating the latter, would

require an estimate of how consumer choice of vehicle size changes in response to the

presence of other vehicles.

The ‘relative vehicle size’ effect has now been accepted as an issue. Available research

suggests that it is sizeable and at least some of its effects quantifiable. Clearly

quantification of the relative size effect could lead to changes in estimates of the

external costs associated with heavy vehicles and road crashes.

This paper supports the case that the cost of road crashes is much more significant than

is commonly reported, both in terms of overall costs and the extent to which these costs

are internalised. The key issues raised in this paper are not new. They have generally

been accepted by experts in the field, but largely have not made their way into effective

analysis of policy.

reasonable to argue that heavy vehicles impose an external cost on other road users.’

vehicles ‘at least 4.3 additional fatal crashes involving others occur’.

The recent BITRE study reported that its headline estimate of the road crash costs in

Australia in 2006 of $17.85 billion (2006 dollars) was conservative, as it was obtained

using a human capital approach to valuing the most significant human component of

costs. The alternative approach based on the WTP approach should be adopted. This is

the approach that is supported by the vast majority of experts and comparable

Government agencies around the world.

There should be no debate as to which method should be used for public policy analysis.

The human capital approach (or at least how it is applied) because of what it excludes,

underestimates the social costs of road crashes. In addition, it is not as theoretically

sound as is the WTP approach. The debate between using the WTP method or the

human capital method is more about whether it is better to be broadly right or more

precisely too low.

Recent Australian research suggest a value of a statistical life (VSL) to a private

individual in the order of $6 million (2006 dollars) is appropriate for general application.

An upwards adjustment to this amount is also appropriate to reflect the value friends and

relatives place on an individual’s life. Furthermore a relatively high VSL may be

justified for road crashes, given the age profile of those killed and injured.

Adoption of the WTP approach is particularly significant for the analysis of internal and

external costs. The increase in cost estimates reflect loss in life and quality of life, much

of which are borne by road-users not at fault and by friends and relatives of victims. As

discussed in this paper, the problem of over-confidence of drivers means even less of the

true cost of road crashes is effectively internalised by drivers.

The problem of vehicle size is an increasingly important issue that demands further

attention. Although the issue has been recognised, there has been little research that has

attempted to quantify the effects in a way that may support using them in policy

analysis.

A large revision to the external cost of road crashes and greater consideration of the

other factors discussed would have important implications for cost-benefit analysis of

transport policy. It is worth recognising the concern by some that an increase in the

external costs of road crashes could drive inefficient investment and regulation. Properly

applied this should not be the case, but regardless, rigour and caution should be applied

in further policy analysis in this area.

In summary, the key conclusions from this review are:

underestimated, both in terms of the overall cost and the extent to which these costs

are incorporated (i.e. internalised) in road-users’ decisions.

has been commonly reported. This increase relates to costs of loss of life and

quality of life, much of which is not borne by the road-users responsible.

values used should reflect the WTP-based approach rather than the human-capital

approach that undervalues life. The WTP-based approach is now common practice

in other areas and other developed countries.

considered in policy analysis.

vehicle size and the external costs that larger vehicles impose on other road users.

they underestimate their likelihood of being involved in a road crash.

Access Economics 2008, ‘The Health of Nations: The Value of a Statistical Life’, A

report for Australian Safety and Compensation Council. Available from

Andersson, H & Lindberg, G, 2008, ‘Benevolence and the value of road safety’,

Swedish National Road & Transport Research, Working Papers, Revised June 4, 2008

http://swopec.hhs.se/vtiwps/abs/vtiwps2007_004.htm

Ashenfelter, O 2005, ‘Measuring the Value of a Statistical Life: Problems and

Prospects,’ Working Papers 884, Princeton University, Department of Economics,

Industrial Relations Section.

ATC (Australian Transport Council) 2006 V5, ‘National Guidelines for Transport

System Management in Australia: Volume 5 Background Material’ available from

Bellavance, F, Dionne, G & Lebeau, M 2007, ‘The Value of a Statistical Life: A Meta-

Analysis with a Mixed Effects Regression Model’, Canada Research Chair in Risk

Management Working paper 06-12.

Bergstrom, T 2005, ‘Benefit-cost in a Benevolent Society.’ UC Santa Barbara:

Department of Economics, UCSB.

BITRE (Bureau of Infrastructure Transport and Regional Economics) 2010, Cost of

Road Crashes in Australia, Research Report 118, available at

http://www.bitre.gov.au/Info.aspx?ResourceId=748&NodeId=58

BTE (Bureau of Transport and Economics) 2000, Road Crash Costs in Australia, Report

BTRE (Bureau of Transport and Regional Economics) 2006, Submission to the

Productivity Commission road and rail freight infrastructure pricing inquiry, August

2006.

Chilton, S, et al. 1998, ‘New research on the valuation of preventing fatal road

Stationery Office, London

Connelly, LB & Supangan, R 2006, ‘The economic costs of road traffic crashes:

93.

Dahdah, S & McMahon, K 2008, ‘The true cost of road crashes – valuing life and the

cost of a serious injury.’ iRAP, UK. available at

www.irap.net/documents/pdf/iRAPValueoflifeseriousinjurypaper.pdf

Dickie, M & Messman, VL 2004, ‘Parental altruism and the value of avoiding acute

HEATCO (2006 D5) Developing Harmonised European Approaches for Transport

Costing and Project Assessment, ‘Deliverable 5 Proposal for Harmonised Guidelines’,

Hensher, DA, Rose, JM, Ortúzar, J & Rizzi, LI 2009, ‘Estimating the willingness to pay

and value of risk reduction for car occupants in the road environment’, ITLS working

Horowitz, John K. and McConnell, Kenneth E., 2000, A Review of WTA/WTP Studies

(October 2000). Available at SSRN: http://ssrn.com/abstract=257336 or

doi:10.2139/ssrn.257336

Jones-Lee, MW 1992, ‘Paternalistic Altruism and the Value of Statistical Life.’

Landefeld, JS & Seskin, EP 1982, ‘The economic value of life: linking theory to

Martin, L 2005, ‘External Accident Costs of Motor Vehicles Revisited’, Staff Paper,

Department of Transport and Regional Services, 28th Australasian Transport Research

Forum, September 2005. Available from www.bitre.gov.au.

NZ DOL (New Zealand Department of Labour) 2004, ‘Measuring the Costs of Injury in

New Zealand’, available from http://www.dol.govt.nz/PDFs/COIMeasuringCostOfInjury.

pdf

Oswald, AJ & Powdthavee, N 2008, ‘Death, Happiness, and the Calculation of

37(S2), pages S217-S251, 06.

Parish R, no date, ‘The Valuation of Human Life’, available from author on request.

Parry, I, Wells, M & Harrington, W 2007, ‘Automobile Externalities and Policies’,

PJP (Port Jackson Partners) 2005, ‘The future for freight’, Report published by the

Australasian Railways Association.

Sandroni, A & Squintani, F 2004, ‘A Survey on Overconfidence, Insurance and Self-

Assessment Training Programs’, unpublished report, available from author on request.

Schwab C, Nathalie G & Soguel, NC 1996, ‘The Pain of Road-Accident Victims and the

UK DFT (United Kingdom Department for Transport) 2007, Highway Economics Note

No. 1, 2005: Valuation of the Benefits of Prevention of Road Accidents and Casualties,

available at http://www.dft.gov.uk/pgr/roadsafety/ea/pdfeconnote105.pdf.

US OMB (United States Office of Management and Budget) 2003, ‘Regulatory

Analysis’ Circular 17, A- 4:13.

US DOT (U.S. Department of Transportation) 2008, Re: Treatment of the Economic

Value of a Statistical Life in Departmental Analyses’

US DOT (U.S. Department of Transportation) 2009, ‘Treatment of the Economic Value

of a Statistical Life in Departmental Analyses – 2009 Annual Revision’, Memorandum

to secretarial officers, modal administrators, Dated: March 18, 2009, available at

Viscusi WK 2008, ‘How to Value a Life’ Vanderbilt University Law School Law and

Economics, Working Paper Number 08-16/

Viscusi WK and Aldy J 2003, ‘The Value of a Statistical Life: A Critical Review of

Market Estimates Throughout the World’ Journal of Risk and Uncertainty, vol. 27(1),

pp. 5-76.

Viscusi, WK forthcoming, ‘The Flawed Hedonic Damages Measure of Compensation

Vanderbilt Law and Economics Research Paper No. 08-05

White, M 2004, "The 'arms race' on American Roads: The Effect of SUV's and Pickup

Zhang A., Boardman, AE, Gillen, D & Waters, WG 2004, ‘Towards Estimating the

Social and Environmental Costs of Transportation in Canada’, A report for Transport