Valuation and use: traffic safety

1
Valuation and use traffic safety

• Rune Elvik
• Institute of Transport Economics
• (re_at_toi.no)

2
Overview of lecture

• Why value transport safety?
• Theoretical foundations
• Components of valuation
• Overview of current official values
• Overview of research and main findings
• Norwegian studies
• Ethical problems and paradoxes

3
Curriculum

• Mishan 1971 Evaluating life and limb a
  theoretical approach
• Jones-Lee 1983 The role of safety in highway
  investment appraisal
• deBlaeij et al 2003 The value of statistical
  life in road safety a meta-analysis
• There is a huge literature in this area

4
Why value transport safety?

• Road accidents are regarded as a problem its
  solution will increase welfare
• Transport policy has multiple objectives trade
  offs are needed
• Resources to provide for safety are scarce we
  need to know how far to go and how to set
  priorities between safety measures

5
Two extreme positions

• Laissez faire
• Why bother? We only die once, and why should the
  time and manner of death matter?
• Vision Zero
• It goes without saying that human life cannot be
  traded off against other goods
• The only ethically acceptable number of traffic
  fatalities is zero

6
Road accidents are a problem because

• Most deaths are untimely (involve a loss of many
  future years of living)
• Risks are disproportionately high (compared to
  other activities of daily life)
• Some groups of road users impose risks on others
  (risks are partly external)
• The system is unforgiving

7
(No Transcript)
8
(No Transcript)
9
(No Transcript)
10
Theories and approaches

• Historically the value of life the value of
  earnings (production)
• This valuation has no theoretical foundation
• Valuation of safety should be based on individual
  preferences
• Preferences manifest themselves as
  willingness-to-pay for a good

11
Willingness-to-pay for safety

• Refers to the valuation of changes in risk
• Values are elicited ex ante (ex post valuation
  makes no sense for the risk of death)
• Values are stated as willingness-to-pay for
  specific changes in risk
• Specific WTP is aggregated to the value of
  statistical life

12
Demand function for safety
13
The value of statistical life

• Initial risk 6 in 100,000
• Reduction of risk 2 in 100,000
• Specific WTP NOK 1,000
• Value of statistical life NOK 50,000,000
• Estimated as 1,000/(2/100,000)
• (the marginal rate of substitution)

14
Approaches for eliciting WTP
15
Components of valuation
16
Total valuation

• For fatalities
• Direct costs net lost productive capacity
  loss of welfare
• For injured road users
• Direct costs gross lost productive capacity
  loss of welfare
• For property damage
• Direct costs only

17
The basis of current Norwegian values of safety

• A systematic review (meta-analysis) of studies of
  willingness-to-pay for road safety (no Norwegian
  studies)
• Basis for estimating loss of welfare
• Estimates based on Norwegian sources of data for
  other cost items

18
(No Transcript)
19
Value of preventing a fatality
20
Some common misconceptions

• There are huge savings to be made by improving
  road safety (hospital bills etc)
• The savings are likely to be close to nil
• A tremendous profit awaits harvesting
• We do not get richer, but our welfare increases
• Life is priceless no amount of money can bring
  the dead back to life
• Life has value even it does not have a price

21
Is the valuation too high?

• Norwegian values are higher than for most other
  highly motorised countries
• But are these values also higher than empirical
  studies can justify?
• Evidence from some recent meta-analyses
• In principle, it is possible to assign too high a
  value to life-saving

22
(No Transcript)
23
Evidence from meta-analyses
24
A closer look at deBlaeij et al 2003

• It is the only meta-analysis confined to
  estimates of the value of road safety
• It tries to identify sources of variation in the
  value of statistical life
• It is comparatively updated

25
(No Transcript)
26
(No Transcript)
27
(No Transcript)
28
(No Transcript)
29
(No Transcript)
30
(No Transcript)
31
Norwegian WTP-studies

• Kvinge 2000
• Strand 2002
• Zhu 2004
• Tofte 2006

32
Results from Kvinge 2000
33
Results from Strand 2002
34
Results from Tofte
35
Safety policy objectives

• Maximum reduction of total fatalities
• Reducing differences in accident risk
• Reducing likelihood of major accidents
• These policy objectives may in principle be
  weighted by assigning monetary values to them

36
Implications of policy objectives

• Reducing total number of fatalities
• All fatalities prevented are valued the same
• Reducing differences in risk
• Preventing fatalities resulting from high risk is
  valued more highly than preventing fatalities
  resulting from low risk
• Reducing likelihood of major accidents
• Preventing several fatalities in one accident is
  valued more highly than preventing the same
  number of fatalities in single-fatality accidents

37
Practical implications

• Current road safety policy priorities are not set
  strictly according to cost-benefit analysis
• Basing policy priorities on cost-benefit analysis
  would lead to faster progress in improving road
  safety
• There are many reasons why policy priorities are
  not based on cost-benefit analysis

38
(No Transcript)
39
(No Transcript)
40
(No Transcript)
41
Key concepts of the demand for road safety
42
(No Transcript)
43
Can ethical dilemmas arise?

• Same safety benefits, different monetary value
• Smaller safety benefits, larger monetary value
• Mean willingness to pay more than majority can
  afford
• Ex ante benefits smaller than ex post costs

44
Ethical dilemmas, continued

• Traps of sub-optimality neither changes in
  income nor in safety can be compensated
• Context-dependent and multi-dimensional
  preferences for safety
• Objective identify potential problems, not try
  to solve them

45
Hypothetical choice between A and B

• Population 100,000
• Risk reduction 3 x 10-5
• Lives saved 3
• Initial risk 20 x 10-5
• Value of a statistical life 60,000,000
• Option A chosen

• Population 100,000
• Risk reduction 3 x 10-5
• Lives saved 3
• Initial risk 10 x 10-5
• Value of a statistical life 30,000,000
• Option B rejected

46
Hypothetical choice between A and B

• Population 1,000,000
• Risk reduction 1 x 10-5
• Lives saved 10
• Initial risk 20 x 10-5
• Value of a statistical life 100,000,000
• Option A chosen

• Population 250,000
• Risk reduction 11 x 10-5
• Lives saved 27.5
• Initial risk 20 x 10-5
• Value of a statistical life 30,909,091
• Option B rejected

47
Problems in choosing option A

• Option B saves 2.8 times as many lives as option
  A (27.5 gt 10)
• Option B reduces by risk 11 times the amount of
  option A (11 x 10-5 gt 1 x 10-5)
• Option B provides a final risk half the level of
  option A (9 x 10-5 lt 19 x 10-5)
• Willingness to pay for option B is 3.4 times
  willingness to pay for option A (3,400 gt 1,000)

48
(No Transcript)
49
Skewness of willingness to pay
50
(No Transcript)
51
More income and higher risk

• Income 25
• Risk of injury 0.0
• Ex ante WTP 25.68

• Income 36
• Risk of injury 0.2
• Ex post WTP 44
• Compensation needed 64

52
Utility functions assumed
53
Can both A gt B and B gtA?

• Income for 1 10,000
• Survival prob for 1 0.75
• Income for 2 10,000
• Survival prob for 2 0.50
• Move to B 2 must pay 1 5,000 (which he cannot)

• Income for 1 10,000
• Survival prob for 1 0.50
• Income for 2 14,000
• Survival prob for 2 0.50
• Move to A 1 must pay 2 4,000 (which he cannot)

54
Will the paradox always occur?

• It is much less likely to occur if both the level
  of risk and the changes in income are smaller
• Small changes in risk and income would be typical
  for road safety
• Still the possibility of a paradox can never be
  entirely ruled out
• We may land in traps of sub-optimality

55
Is road safety a homogenous good?

• Most probably it is not
• Aspects that may influence the demand for safety
  include
• Voluntariness and control of risk need to
  protect children
• Disparities in risk seen as inequitable
• The potential for major accidents

56
(No Transcript)
57
(No Transcript)
58
To take account of contexts and dimensions .

• One would need to develop a comprehensive price
  list
• There would be one value for saving an adult life
  exposed to a low risk activity
• There would be a different value for saving a
  childs life
• And so on

59
A price list would be regarded as

• Macabre or ridiculous, depending on the
  perspective
• Ethical dubious, by being responsive to morally
  irrelevant considerations
• Highly speculative, as preference structures are
  imperfectly known

60
Can human life be valued too highly?

• Can too much resources be spent to save one life?
• Yes, in principle spending a disproportionate
  amount to save one life (or a few lives) could be
  counterproductive in terms of feedback effects on
  overall mortality

61
(No Transcript)
62
Loss of income increased mortality

• If mean income per capita drops, overall
  mortality may increase
• The relationship is, however, highly uncertain
  and likely to be modified by changes in
  expenditure
• We need to know on a broad basis how various
  policies influence the length and quality of life

63
Concluding comments

• An economic approach to road safety may lead to
  ethical problems
• Some of these problems are, in practice, quite
  unlikely to occur
• Other normative approaches to road safety, like
  Vision Zero, may also be associated with ethical
  problems