Estimating the Human Cost of Transportation Accidents: Methodologies and Policy Implications

By Jagadish Guria

Estimating the Human Cost of Transportation Accidents: Methodologies and Policy Implications discusses the estimation methods needed to determine the monetary value of loss of life and quality of life when evaluating transportation safety programs, policies and projects. In addition, it highlights how to overcome the many challenges researchers face in choosing the right values, including estimating loss of life and life quality, examining strengths and weaknesses, and critically analyzing social costs and implications. This book will allow researchers to better formulate accurate social costs, select safety improvement values, and understand limitations.

• Provides a comprehensive, theoretical, one-stop reference on non-market valuation methodologies, issues and policy implications for transportation health, safety and economics researchers
• Helps researchers better evaluate the actual total cost of road safety programs, policies and projects, including life quality valuation due to environmental impacts, such as harmful vehicle emissions
• Provides valuable case studies from around the globe

• Language: English
• Publisher: Elsevier Science
• Release date: Mar 5, 2020
• ISBN: 9780128126127

Jagadish Guria

Jagadish C. Guria is an Editorial Advisory Board member of Elsevier’s Accident Analysis and Prevention journal. He has taught Econometrics at University of Massachusetts-Boston, and is the former Chief Economic Adviser of New Zealand’s Ministry of Transport. Currently he is a Transport and Health Economics consultant, working with numerous private and public sector clients, including the World Bank.

Book preview

Preface

Jagadish Chandra Guria, Wellington, New Zealand

In my long career on road safety programme and policy evaluations, estimation of the value of statistical life (VSL) and its applications have been very important issues. On the estimation side, there have been improvements in methods of estimation over time and contingent valuation method is now generally accepted as the best way to estimate VSL. However, there has been new emphasis on estimating VSL and other attributes of road safety, especially value of time from the same surveys using the choice modelling approach. The problems have not been so much on selection of estimation method, but on how best to conduct these studies and even how best to present the survey scenarios to the respondents. The main question faced has been on what the level of risk reduction should be that the respondents would be able to comprehend and act on. It is generally accepted that to get realistic responses, survey questionnaires must be realistic, otherwise hypothetical questions may provide only hypothetical answers. The base-risk level and the possibility of achieving a lower-risk level are the two very important characteristics of such survey questionnaires. Unless these are well taken care of the likelihood of errors in the evaluation of safety programmes and policies cannot be ruled out.

The main task of these programmes is to determine the social costs of premature deaths. Along with this there are two other significant issues, namely the social costs of injuries and consequent risk of disabilities. One argument related to these cases is that at premature death, the life span is shortened, and/or the quality of life is reduced during the period an injured person lives before death. The second problem arises from the argument that death as such is not avoided, but only a few years of life are lost. So a premature death means a loss of a few life years. Since all life years are not the same and at least some of them are of low quality of life, it is well argued that life years should be measured by quality adjusted life years. There is no problem with this argument, but it does arise in determining the monetary value of these life years in order to value the loss of life to society, as well as the social cost of injuries. Another measure of cost of injuries is the disability adjusted life year.

Although the main objective in this book is to discuss human costs of transportation accidents, most of the discussions are equally valid for other areas, especially health improvement and environmental effects. So these areas are mentioned to emphasise the applications of human costs in these as well. The VSL, quality adjusted life years and disability adjusted life years and their monetary values are important factors in the evaluation of safety and health quality improvements in these areas.

Where available, the methods of estimation are discussed with a review of the literature on the subjects based on theory and empirical experience. The weakness and limitations are noted and scope of further developments is discussed. This short book will hopefully help researchers to become more aware of weaknesses in the estimation process and what distortions the method can create so that they can try to modify the method of estimation where possible. Hopefully this book will help policy makers be aware of the problems associated with different criteria and of postponing decisions when scientific research indicates the need for changes in the existing system.

Chapter 1

Values of losses of life and life quality in social costs: the basics

Abstract

Human costs, that is losses of life and life quality are the major costs of accidents to society. This important component of social cost is introduced in this chapter along with other components. This chapter briefly discusses the contents of the book and what is expected in other chapters. It briefly describes how the valuation of human costs started with valuing the productive value of individuals; recognition of the value of being alive and the cost of pain and suffering due to fatalities and injuries; and development of the concept of value of statistical life or the value of preventing a statistical fatality. This chapter describes the estimation methods, their limitations and how these can influence the social costs of accidents and, consequently, the resource allocation decisions. These components of human costs, though initially estimated for accidents in most cases, are applicable to determine the social costs of diseases. Therefore these costs are referred to for both health and safety areas.

Keywords

Social cost; value statistical life; quality adjusted life year; disability adjusted life year; human capital; revealed preference; stated preference; contingent valuation; willingness to pay; willingness to accept

1.1 Introduction

Transport accidents (crashes) cause huge losses of life and life quality around the world. Road transport is responsible for the major share of this social loss. A World Health Organization (WHO) report indicates that globally road collisions cause over one million deaths and between 20 and 50 million nonfatal injuries every year (WHO, 2013). Many of these nonfatal injuries result in long-term impairments, substantially reducing the life quality of the injured. (A transport collision is commonly referred to as an accident or a crash. There are people who have strong preference for the word ‘accident’ and others for ‘crash’. The use of a particular word does not make any difference to our discussion here.) The terms ‘accident’ and ‘crash’ are used interchangeably to mean an unintentional collision.

The consequence of an accident can range from minor property damage to several deaths and injuries, depending on the transportation mode and accident severity. Severe accidents incrementally increase the cost of trauma to individuals, their loved ones and to nations as a whole. The cost of an accident has many components. These include resource costs, such as emergency services, medical treatment, rehabilitation of the injured, repair or replacement of vehicles, traffic delays and, most importantly, losses of life and life quality. The total cost to a nation is commonly known as the social cost, which includes tangible and intangible components.

Transportation programmes have many objectives. Some examples in road transport areas include projects aimed directly at improvement of safety, such as the development of mid barriers, passing (overtaking) lanes on highways or improvement of safety quality of vehicles. Safety programmes and policies include development of vehicle and road safety standards, seat belt use policies, policies related to alcohol and/or drug impaired driving and their enforcement. In such cases, the main benefit is a reduction in the social cost of accidents. Similar examples can be presented for other modes as well. Road transport examples are referred to mainly for convenience as these happen in our everyday life and hence are understood by most people. A project may aim at improving nonsafety travel attributes such as comfort and travel time savings. Even in such cases, an evaluation must take into account the impact on safety along with benefits of these attributes, since safety is an integral part of travel. In some cases, safety improves along with achieving other objectives, while in others safety may not be affected or may even be compromised. In all cases, the impact of safety measures is an important part of economic evaluation. Any variation in non-safety attributes may have an impact on safety, since safety may improve or deteriorate as a consequence, affecting the social cost.

A major share of the social cost of an accident comprises losses of life and life quality. In New Zealand, for example the share of the estimated cost of losses of life and life quality accounts for about 90% of the total social cost of all injury accidents, and over 75% of the social cost of all accidents (MOT, 2016). These percentages, of course, depend on how the monetary values of losses of life and life quality are estimated. Due to this and other factors, the share of losses of life and life quality in the estimated total social cost of traffic accidents may differ between countries. This is also true for other areas of evaluating programmes and policies that cause changes in safety and health status.

The share of the cost of deaths and injuries in the total social cost increases with the severity of accidents. Proper valuation of this component is therefore a very important issue, particularly in cost–benefit analyses of transportation programmes and other areas such as health and environmental effects. The intangible nature of cost of deaths and injuries makes it difficult to measure in monetary terms, which leads to controversies and estimation errors, as well as distortion in resource allocations.

There are also programmes aimed mainly at improving nonsafety benefits, such as network expansion or bypass development that reduce travel time, but do not necessarily improve safety. However, due to higher level of travel and/or higher speed of travel, more accidents or proportionately more severe accidents causing losses of life and life quality can be expected. The loss to society due to these increases in risk of death and injury may be relatively high, as discussed in chapter 4: Value of loss of life quality. The evaluation process needs to ensure that an increase in social cost of accidents does not exceed the nonsafety gains to be achieved through these programmes and policies.

1.2 Optimal resource allocation

Since we have limited resources and the requirement of these resources to satisfy our needs far exceeds the total quantity available, it is necessary to allocate them efficiently. As in a household where the objective is to maximise the utility or satisfaction of household members from the limited income of the household, it is the objective of the economy to ensure that resources are allocated in such a manner that the maximum benefit to the nation is achieved. When this happens, the resource allocation is optimal. Optimal resource allocation is not only desirable for the economy as a whole, but also can be an objective in individual areas of operation such as transport, health and environment.

In all projects, programmes and policies (henceforth referred to as programmes), whether related to transport safety, health or environmental effects, where risks of death and injury are affected, the major benefits are likely to be reductions in losses of life and life quality. Therefore monetary valuations of these effects are very important. An improper valuation leads to nonoptimal allocation of resources. If safety is undervalued, it would not get its fair share in resource allocation and that would lead to more lives being lost or resulting in higher loss of life quality from injuries, or both. Because of the intangible nature of safety benefits, valuation of safety is not straightforward. The complexity of its valuation makes it less understandable and causes many problems for optimising resource allocation. Complexity also arises from preferences of decision makers. There are people who value safety strongly, while others may prefer a higher allocation of resources to alternative areas of the transportation system, for example network expansion or savings in travel time. Because of uncertainties in estimating the value of safety, some transport professionals argue that they should be cautious and take a conservative approach in monetising safety benefits. In other words, they would prefer to put a lower value on safety benefits, perhaps without realising that this attitude could result in more losses of life and life quality that would otherwise be avoidable. Such an approach distorts the economic optimisation process and, as a consequence, the overall welfare of society falls below the level achievable from its resources. The process of underallocation of resources to safety may also result in overallocation of resources to other areas beyond their optimal levels. Both underallocation and overallocation cause distortion and the total benefit to society is underachieved. This happens not only in areas of health and safety, but also in other valuations of intangible factors. One example in the transport area is the value of time. Often evaluations are made under the assumption that all amounts of time saved are equally valuable to society, which is not necessarily the case in real life. As in the case of health and safety, it is important that such intangible factors as time are also valued properly to ensure that the evaluation process leads to optimal allocation of resources. Though related to social cost, the valuation of time savings is not included in our discussion here. It is merely pointed out as an example to highlight the importance of all intangible costs and benefits in the social cost estimation. Our main purpose here is to point out the importance of safety improvement valuation, the observed problems associated with it, how these should be addressed and the need to reduce the gap in our knowledge.

There is, of course, the other side of resource allocation. Overenthusiastic promoters of safety may like to improve safety at any cost, or at least without proper evaluation, thinking it is their moral duty to ensure that safety is not compromised. Emotions, often arising from negative personal experiences (death of a loved one, for example), supersedes the rational expectation of economic optimisation. This may sound like an extreme case, but it is not unusual in real life. To avoid this, it is necessary to ensure that losses of life and life quality are appropriately valued and a comprehensive cost–benefit analysis is carried out without any bias towards a particular attribute, whether it is safety, value of time or another factor. Improvement of safety or health conditions at any cost, or at an exorbitant cost, may increase the allocation of resources for these improvements, but would distort the overall resource allocation.

1.3 Value of statistical life

Deaths and injuries cause pain and suffering, not only affecting the victims but also their families and friends who care for them. Injuries can vary in severity from minor cuts and bruises to much more serious ones causing long-term impairments and discomfort for the rest of life, or premature deaths. Here we first discuss the loss to society of premature deaths and then loss of life quality due to long-term impairments from injuries.

The obvious question is: What is the value to society of preventing a premature death? Many would prefer to consider life as priceless and hence should not be measured in monetary terms. As noted by Blomquist (2001), some find the concept of value of life vulgar and ethically wrong. Some researchers (Sen and Williams, 1982) even find vulgarity in valuing life for completeness of cost–benefit analysis. The views that life is priceless and that it is unethical to monetise the value of a human life have been debated for years. There has been considerable research over recent years on the valuation of losses of life and life quality. The objective here is not to measure the value of life as such, but to determine the value to society of a risk reduction. This is the main aim of developing safety programmes. As reduction in risks of accidents and injuries are the main benefit of a safety improvement programme, it is necessary to assess the loss to society of premature deaths and loss of life quality due to injuries to determine the value of benefit, which can then be compared with the cost of implementing the safety improvement programme. Even among various possible safety programmes, it is necessary to determine which ones are more worthwhile than others so that limited resources can be allocated accordingly. When the risk of death can be reduced either by reducing the risk of a fatal accident or reducing the risk of death in the event of an accident, eventually premature deaths and loss of life quality due to impairments are avoided. The value to society of these consequences is the subject of our discussion. Estimation methodologies of valuing statistical life and injuries are briefly discussed next. The methodologies and all related issues are elaborated in Chapter 3, Value of statistical life.

1.3.1 Human capital

At an early stage, economists considered the productive contribution by a person as the value of life of that person. This productive value used to be a basis to value loss of life. Like physical capital, the productive value of an individual was considered to be the human capital value of that person. This concept arose from the interests in economic growth and national income accounting. Investments in human beings were considered important as higher skills would increase the productivity of those individuals. Both skill development through education and good health were considered necessary for improving labour productivity. Without good health, one would not be able to perform to one’s full capability. This led to the thinking that if a premature death could be avoided through investment in human health, society would gain the labour contribution of that person and that would increase the national income. The expected income that would be lost with a person’s premature death became a measure of the value of life of that person at that time (Mushkin, 1962; Blomquist, 2001).

In this approach, it is assumed that the loss to society of a person’s premature death is the productive potential of that person (Landefeld and Seskin, 1982). It is commonly measured by the discounted present value of expected future output that would have been achieved if the person did not die. Questions have been raised on whether the human capital value of life should be the gross value of lost output or the net value after accounting for consumption. The loss of output, or future earning, indicates what the person would have contributed to society. However, the person would also have consumed goods and utilised services. In that sense, there would be some savings to society at premature death of the person. Therefore some researchers consider it appropriate to measure human capital value as expected loss of output net of consumption (Weisbrod, 1971). This approach considers the value of life of an individual to the rest of society and ignores the value of life to the person himself/herself. Because the individual belongs to a society, but consumes for his/her own enjoyment and maintenance of health, the net value is obviously not appropriate as a society’s value of life. Besides, the person would get satisfaction from consumption and hence that would increase the society’s welfare. Consideration of output net of consumption would imply that the values those who have passed the retiring age, or are not producing output otherwise, would be negative. In other words, society would be better off when the individuals were dead than when they were alive. As Jones-Lee (1985) observes, no society would accept such an outcome. There have been considerable developments since then. Economists generally consider it inappropriate to value loss of human life. However, some countries still use this method to value loss of life to society.

There are many shortcomings of this human capital approach of valuing statistical life. It ignores the values of those services which are not marketed, for example services of children, unemployed and retired elderly. That these are not marketed does not mean they have no value to society. Human capital value that ignores these contributions would result in underestimation of the total value to society of an individual’s contribution. Labour compensation of these nonmarketed services is more recently being imputed to account for their values, following the approach by Jorgenson and Fraumeni (1989).

Another very important issue is that this measure does not value the loss of life as such and does not take into account the cost of pain and trauma suffered by family members, friends and society. Some people may not like to put a monetary value on life, but life does have value. The loss of output measure of life only indicates what the person could produce, not the value of being alive to themselves and to others. Some countries, where the human capital approach is followed to estimate the value of statistical life (VSL), include additional values to the human capital value to take account of this loss, such as the cost of pain and trauma suffered by the person before premature death and by other members of society. Australia follows such a hybrid approach by adding some value for pain and suffering to the estimated human capital VSL (BITRE, 2009). While this improves the estimated VSL, there is scope for further improvement through a systematic statistical estimation of the value to society of pain and suffering due to a premature death.

1.3.2 Valuation of risk changes

A different approach related to welfare maximisation was developed based on the consideration that individual preferences for safety should be the basis for valuing the level of safety improvement. This approach was initiated by Drèze (1962), and later on became known as the willingness to pay (WTP) approach developed by Schelling (1968). Drèze looked at valuation of safety improvement from individual’s point of view as well as society’s collective decision making (Jones-Lee, 1976). Schelling (1968) introduced the concept of valuing small risk reduction instead of valuing life, which many felt uncomfortable with as its entails putting a monetary value to life whereas life should be priceless in their consideration. Schelling appears to be the first to relate potential change in fatal accident rate to the change in probability of death for an individual (Jones-Lee, 1974). This is very relevant for policy evaluations. Safety programmes aim at making small changes in risks of death and injury and Schelling’s approach leads to a trade-off between money and risk reduction, that is improvement of safety. This risk reduction–money trade-off indicates how an individual values small changes in risk. This led to the term VSL, which is different from value of life because it is the total amount of money society is willing to pay for reduction in probability of deaths, so that one premature death is prevented.

When a person is in a life-threatening situation, a society would possibly try to save the person with all its resources. This is different from the effect of risk changes in general. Think of a small township of 100,000 people and the average number of deaths during a year from traffic accidents is 5. Thus the average risk of death during a year is 5 in 100,000 per person. Now suppose something can be done which would reduce the average risk by 1 in 100,000. If people in the township are willing to pay on average $50 for reducing the risk by 1 in 100,000, then the total amount of money they are willing to pay is $5 million. It is expected that one life would be saved as a result of this risk reduction. In other words, people in the township are willing to pay $5 million to save one statistical life. It is not known in advance whose life would be saved; it is not the value of life of an identified person. Thus the VSL is $5 million for the township. It can also be considered the amount of money society is willing to pay on average to prevent one premature death (death of a statistical person). For this reason, the VSL is also known as the value of preventing a (statistical) fatality. As scarce resources are required to develop safety programmes to reduce the risk of death or injury, the total benefit accruing from such activities must exceed the total cost of carrying out these activities, otherwise there is net loss to society. If there are alternative resource allocations to be evaluated for public investment then Mishan (1971) suggests that these allocations should be compared on the basis of potential Pareto optimality. The theoretical foundation was further developed by Jones-Lee (1974) through a discussion on compensating variation for changes in the probability of death. Following Von Neumann Morgenstern’s expected utility maximisation under uncertainty, Jones-Lee (1976) developed an estimation method for the marginal rate of substitution of wealth for changes in survival probability. This pioneering work has helped develop many empirical studies. This can be explained in simple terms. Suppose the probability of survival for an individual is improved from P to P* or by ΔP (i.e. P−P*), and the individual is willing to pay V . Suppose there are n . Jones-Lee (1976) shows that the unweighted aggregation of WTP by n . This is the VSL based on WTP. There have been two streams of development for estimating VSL based on people’s preferences. Since there is no market for safety, a nonmarket valuation is necessary to put monetary values on safety, a necessary factor for carrying out cost–benefit analyses of the programmes. Even though there is no explicit market for safety, people do exchange wealth for safety improvement in real life. Economic analyses of actual behaviours to determine the trade-off people commonly make is known as the revealed preference approach. This analysis helps find the marginal rate of substitution of wealth in exchange of safety improvement or risk reduction. An alternative approach discussed earlier was developed to hypothetically create a market situation and asking people to state how much they would like to pay for any risk reduction or be compensated for an increase in risk. This approach is known as the stated preference