2.1.1 Multiple Account Evaluation Framework

Multiple criteria decisions have a three-dimensional complexity: the stakeholders are many, there are a number—sometimes a large number—of alternatives to consider, and the objectives are legion—reflecting the multiple stakeholders’ goals. Environmental concerns are but one group of criteria on which transportation initiatives are judged. As a justification for the approach proposed in this report for costing environmental impacts, it is helpful to elaborate on the different evaluation methods used within the multiple criteria appraisal frameworks. Several concepts need to be explained first.

The multiple criteria are derived from multiple objectives of a project or policy. The objectives are more concrete than the goals above them, but still may be difficult to measure. Consider, for example, the goal of transportation service for customer satisfaction. It can be translated into the objectives of acceptable travel time, comfort, safety and vehicle operation, parking and fare costs. The travel- time objective, in turn, is broken down into more specific criteria: travel time of people, holding time of goods in transit, and the time utilization of commercial vehicle stock.

A measure of merit, or an attribute, is developed for each criteria, which captures the differences between different alternatives proposed as a solution to achieve the objectives of the project or policy. The attributes must be meaningful with respect to the criteria, well defined and readily measurable. Some criteria, however, cannot be measured, thus the attributes are of necessity qualitative; for example, the low, medium and high flexibility of a transportation facility to adapt to future unknown requirements. For travel time, the attributes may be total hours of travel considered separately for people, goods and vehicles. Travel-time attributes may be the corresponding value of time for people and vehicles expressed in dollars, but expressed in hours for goods, based on the sensitivity of the shipment to time.

The British Columbia government’s policy is to use multiple account evaluation (MAE) to support decision making about public investments. In February 1993, the provincial government issued general guidelines for multi-objective decision making in the provincial government (CCS 1993). One of the major accounts in the guidelines concerns environmental impacts.

The MAE framework is based on a number of tenets of rational decision analysis:

The first group of approaches to incorporating environmental costs in planning are the methods of cost benefit analysis, which attempt to monetize non-market criteria. Cost benefit analysis (CBA) may be used when the benefits are described under more than one criteria and are monetizable. Cost benefit methodology for appraisals of public projects and policies considers only the net use of resources. The investment must produce a net saving in the use of resources, or a net gain in production, in order to be viable. To be economically efficient, the benefits must at least equal the costs of the project.

Cost benefit analysis is based on the concept of measuring the net impacts of projects on society and, where possible, monetizing these impacts to determine the maximum benefits to society as a whole. It does not consider the distribution of benefits among different sectors or regions of society. In other words, it considers efficiency but notequity. Transfer payments from one party to another, such as taxes and subsidies, or transfers of economic activity from one region to another, are manifestations of equity rather than efficiency, and so are not considered.

In theory, cost benefit evaluations should count only the net benefits in excess of opportunities forgone due to the project, assuming that resources would be otherwise employed if the project did not exist. The analogy for the environment is that the environmental impacts should be the difference between the transport project in question and the "average" alternative use of resources in other industrial or public projects (Waters 1995). However, in practice, the alternative use of resources is not known, and so the impacts of a proposed transportation project are normally considered relative to the impacts that would take place for a base case of "business as usual," "do minimum" or "do nothing," which are usually well known. Consequently, project evaluations assess the increments (negative or positive) in environmental impacts of a proposal compared with the "do minimum" base case.

Sometimes it is pointed out, particularly in connection with environmental disbenefits of a project, that many of the indirect economic benefits are disregarded, and therefore the additional environmental costs bias the analysis. Benefits of transportation service to the general economy, job creation in vehicle manufacture and repair, road construction employment, traffic policing, etc., are frequently mentioned by opponents.

It is true that transportation does contribute to economic output. However, many of these indirect benefits are covered by the appraisal of user benefits. In addition, jobs in road construction and vehicle manufacture can only be counted as creating employment if workers who fill these jobs were previously unemployed rather than employed elsewhere.

The standard CBA of transportation projects compares the costs of the transportation facility or service with travellers’ benefits of reduced travel time, fewer or less severe accidents, and savings in vehicle operating costs, which are projected over a future period. If the present value (the present equivalent of future values) of the difference between the benefit and the costs is positive, the project is economically viable (the net present value is positive). In theory, CBA should include all costs and benefits, and should reduce all of them to monetary units. Doing both is extremely difficult, and historically, monetizable agency costs and user benefits have been the chief criteria CBA has used to judge projects and policies. This means the analysis is incomplete.

Attempts have been made to reduce all types of costs and benefits to a monetary unit, but many non-market and intangible impacts defy such inclusion. The use of unreasonable monetization methods in some cases has discredited cost benefit appraisal in the eyes of decision makers and the public.

Social cost benefit analysis (SCBA) attempts to reduce all types of costs and benefits to a monetary unit. In principle, SCBA is no different from CBA, but the distinction is made here to draw attention to the often unrealized capability of CBA to consider a wider range of social costs than is normally practiced in transportation planning and policy. Labelling of SCBA is not an attempt to set up yet another arcane subdiscipline. Many organizations have attempted comprehensive CBA (that is, SCBA), but they ran into problems of insufficient data, time constraints and insufficient funding for additional research.

It should be noted that although CBA uses monetized values, even those impacts that are exchanged in the market are not necessarily valued at their market prices. CBA frequently uses shadow prices that differ from market prices, in order to correct for various externalities and distortions. This implies that the analyst is making judgments about the "correct" price. No evaluation method, including CBA, can be free of some degree of subjective judgment; the best that can be done is to use accepted methods in a consistent and transparent manner.

While certainly difficult, monetization of at least some environmental impacts can be done. For example, Swedish road planners have been using the Calculation Guide for Investments in Roads and Streets (SNRA 1986) for many years. The procedures are based on the cost benefit method and include monetized environmental costs of traffic noise, air pollution and community severance (Lindberg 1994). In New Zealand, cost benefit analysis has long been mandated as the analysis method for project and policy evaluations in government, including transportation and roads (Silvester 1993; Transit New Zealand 1994). The development of SCBA tools for the ministry was strongly influenced by the work done in these two countries.

Multi-Attribute and Rating Approaches

A second group of approaches to incorporating environmental costs in planning are the multi-attribute and rating approaches, which avoid monetizing non-market criteria. The multi-attribute utility (MAU) approach uses the concept of utility to assess value judgments on monetizable and other attributes of alternatives (Keeney and Raiffa 1993). In theory, this approach makes it easier to consider the full range of criteria (accounts). However, it is difficult for decision makers to fulfill the requirements of MAU to understand a large number of criteria, trade them off and ascribe relative utilities over the range of each attribute.

Trade-off questions are particularly difficult because they involve ascribing weights to the various criteria to represent their relative importance. Because the weights represent the subjective evaluations of decision makers, they are difficult to set, and yet may have a substantial impact on what decision is made. Some criteria may be weighed too heavily, in the opinion of others (Hobbs and Meier 1994). In many cases, there is also the difficulty of reflecting the conflicting value judgments of multiple stakeholders in the decision. For example, Lesser and Zerbe (1994) show how drastically the ranking of alternatives changes depending on the value systems of 10 stakeholders concerned with an environmental issue.

Another group of evaluation methods comprises a variety of schemes for weighting, rating and scoring multiple measures of merit (for example, Wedley 1990). The ratings and weights on which the final score of an alternative is based are necessarily subjective, as they depend on a particular stakeholder’s focus. In addition, rating weights often fail to represent trade-offs that people are actually willing to make. Weights cannot be assigned on an intuitive basis without determining whether they reflect the trade-offs that people would in fact be willing to accept (Hobbs and Meier 1994).

Problems similar to those of the MAU method arise because multiple and opposing stakeholders are involved in real-world decisions. A typical example is provided by Gylvar and Leleur (1983), who weighted agency and user costs, noise, barrier effects and air pollution in a model for prioritizing capital projects. Although a formula is provided for the weighting, no justification is given for the weights, and no mentionis made of the need to test the sensitivity of project rankings to the distribution of weights between different criteria. In another case (IBI Group et al. 1995), equal weights are assigned to aggregate ratings of strategic, economic, environmental, social and financial criteria. The equal weights are applied to criteria within a group, and to each group.

Goal programming finds wide application in the utility industry and the public sector. It ranks alternatives using a formula that chooses the alternative whose criteria values are closest to a set of predefined goals, which have been defined as an acceptable, desirable or ideal level of each criteria measure. Goal programming focuses on the achievement of goals, as opposed to additive value functions, which emphasize trading off criteria.

Some of those concerned about sustainability believe more attention must be paid to the following intangible factors:

Another criticism is that markets, and the traditional market-based economics, do not necessarily produce sustainable outcomes. Using market prices and valuations can be misleading, not only because money is confused with material and social well-being, but because the signals given by market prices do not account for critical biophysical realities, such as natural capital depletion. Market signals can serve as incentives to nonsustainability. Wackernagel and Rees (1994) make the following points concerning market valuations of natural resources:

Various approaches have been suggested to incorporate sustainability into evaluation frameworks. At one extreme, the unknowns are tackled by adopting an absolute standards approach (also called strong sustainability approach), which makes it the duty of the decision maker to avoid future harm. The strong sustainability approach assumes that futuredamage might be so significant that action is warranted regardless of quantification. Costs cannot be traded off with benefits, and the current generation has a duty to avoid harming other species and humans now and in the future. This approach is impractical for transportation planning, because transportation today uses energy and resources and produces pollution. Sustainability is a long-term goal, but society needs to make day-to-day production and investment decisions that are also responsive to near-term economic needs.

To avoid unacceptable costs to the current generation, the safe minimum standard (weak sustainability) approach could be adopted. This version of sustainability permits the current generation to use resources to the limit imposed by leaving future generations as well off as itself. The weak sustainability approach is not yet operational in transportation project and policy evaluation because it requires cost benefit trade-offs for which damage cost estimates are not yet well developed, and because an appraisal framework that could consider the necessary decision parameters is not yet widely implemented.

Current research indicates that some of these limitations are being addressed. This report attempts to make the weak sustainability approach operational. Table 2.1 summarizes the comparison between these different decision-making frameworks.

Monetization is the process of determining the equivalent non-market monetary value. Transportation project evaluation increasingly incorporates shadow prices of non-market costs and benefits, including valuation of travel-time savings, accident reductions and environmental impacts. These can then be compared with market costs and benefits, such as project capital costs and vehicle-operating savings. The use of shadow prices makes planning more effective and equitable by allowing a consistent application of non-market costs among projects, organizations and economic sectors (Bernow et al. 1991). If market and non-market costs are not expressed in common units, comparisons become difficult and the trade-offs between economic, environmental and other criteria become more subjective. Monetized values are also useful in road pricing, and in project design and operation to optimize resource use.

There are, however, a number of limits to the use of monetization. One set of philosophical questions concerns whether to monetize; another set of pragmatic questions concerns how to monetize. Some significant issues have been raised concerning ethical limits to monetization. Some argue that we cannot place economic value on human life, ecosystem functioning or aesthetic value. To do so disregards the idea that some things are sacred or "priceless." While this is a values position, which cannot be debated in scientific terms, humans frequently are forced to make monetary evaluations of such items: for example, how much a hospital should budget to pay for a life-saving piece of equipment.

Some fear that placing monetary value on things we view as sacred or priceless can lead to treating them as commodities, which can be traded or sold. It is true that one of the aims of monetizing the value of a wetland, for example, is to be able to compare or "trade" it on a common scale with other things, such as the value of accidents prevented. It may be argued that doing so will lead to the wetland being more highly valued, even if no dollar figure is included in the comparison.

Nevertheless, it is acknowledged that "trading off" may be inappropriate for some items. To "trade" two items implies that one can substitute for the other. Since capital stock cannot substitute for all of the natural stock, many non-use values do not have a price. This is equivalent to saying that their price is infinite. Also, under probable future conditions when the environment has been degraded, the price of environmental goods could become extremely high or infinite.

Pragmatic limits to monetization also exist even if it is agreed that monetization is at least in some cases philosophically or ethically appropriate. These limits are as follows:

The completeness of various cost estimates and the way of using them in analysis and policy making will always be debatable. Some methods for monetizing environmental values are innovative and not well tested (Gregory 1986; Kahneman and Knetsch 1992; Smith 1992; Edwards 1987). Non-use benefits, such as existence, option and bequest values can only be measured by inadequate proxy values such as "willingness-to-pay."

Lack of data is common, sometimes resulting in either over-extensive or too-modest conclusions based on limited surveys or studies. Different survey methods may provide significantly different estimates of how much society values an environmental asset. Limited data may make it difficult to determine how much an activity contributes to a particular impact, and if so, whether the impact is constant, increased or decreased by other influences, so marginal costing may be difficult. Consequently, different studies of valuations of the same impacts are likely to yield a wide range of results. There are always intangibles of some kind, and more research is needed to better determine the use value of specific resources. Many of the values used in this report are considered preliminary estimates for one or more of these reasons.

Uncertainty is especially significant in assessing the impacts of present-day activities on the loss of species and ecosystems, and on future generations. Inter-generational equity requires taking into account possible human and environmental trends, and estimating ultimate costs based on future human values, which are "unknowable." For these reasons, some costs may still need to be treated qualitatively, even though monetization is a goal.

Continued efforts to research currently unknown data, and to find methods to measure and to value effects, will reduce the number of unmonetizable effects but can never eliminate them. The unmeasured effects are collectively referred to as intangibles. Some authors reserve the word to refer only to intrinsically unknowable or unmeasurable impacts, but it is used here to refer to all decision variables, criteria and value judgments that currently cannot be determined regardless of the evaluation method used.

As well as being required by the provincial government, the MAE framework, discussed in Section 2.1.1, has several other advantages. It attempts to ensure that all alternative projects are included, and that all accounts (categories of impacts) are considered. This overcomes a shortcoming of some cost benefit analyses, which omitted items that were difficult to monetize. It also ensures that stakeholder values are considered in the decision process.

The established routine for a structured decision-making process is to identify stakeholders, find out where they disagree, separate facts from values, and focus information gathering on those factual or value-based questions that seem most important. King (1994) classified decision problems into four categories on the basis of underlying facts and values: (1) In the simplest case, a consensus exists on both facts and values. (2) If the stakeholders disagree on facts, more research is called for. (3) If they disagree on values, the best chance at a solution may be negotiation. (4) In many cases, there are so many unresolved questions about facts and so many different values involved that it is difficult to focus evaluations in ways that make sense.

Decisions involving environmental impacts fall into the latter category, and all transportation decisions involve trade-offs of human capital with natural capital. MAE serves as a practical tool to minimize disagreement of stakeholders on facts and values.

2. Monetize values wherever possible, while recognizing the practical limits to monetization.
In social cost benefit analysis, the impact of a transportation project or policy on resource consumption (such as construction and maintenance costs, travel time, accident damages and vehicle capital expenses) are first quantified. Then the quantity of each of these resources is multiplied by an appropriate unit cost. In principle, shadow pricing environmental and ecological impacts is no different from evaluating other costs to a transportation agency and road users. The quantity of each impact and the resulting damage must be assessed and multiplied by an appropriate unit cost.

The value of an environmental resource should consider the total benefits it provides, including use and non-use values. For example, preserving a wetland may provide water quality, wildlife habitat, recreation, and aesthetic, existence, option and bequest values. The value of the wetland is the sum of all these benefits. In many cases, beneficiaries may be unaware of the full extent of these benefits. Or the value may be difficult to translate into monetary terms.

Although not absolutely accurate, first-cut estimates of shadow prices are preferred to ignoring potentially significant impacts. Too many precise estimates have been made that exclude the significant costs of nature’s non-market goods and services. Monetized costs based on shadow prices, as well as underlying emissions and physical and ecological damage data, should be provided with the cost benefit analysis so that impacts are properly appreciated (Maslany 1995).

Several factors are currently motivating efforts to monetize environmental impacts. Public concern about the environment is increasing and decisions at all levels of government and the business are beginning to reflect this concern. Growing emphasis on efficiency and equity, public participation and more experience in incorporating environmental costs into decision making increases the demand for shadow pricing. Costing techniques are also improving. Public utility and energy resource planning have been using and refining this approach for some time (e.g., Harrison 1995; Harrison et al. 1994; Harrison et al. 1993; DPU 1992; Dodds and Lesser 1992; G.E. Bridges and Associates Inc. 1991; Chernick and Caverhill 1989; Reyier 1986).

In cases where environmental assets are difficult to quantify and value, the analyst should identify and describe the impacts as fully as possible, including a description of intangible and intrinsic aspects, so that decision makers can consider them. Trying to monetize everything may obscure the issues rather than support decision making. One of the criticisms of cost benefit analysis is that it has at times attempted to quantify the unquantifiable.

3. Follow sustainability principles.

Use the precautionary principle

The sustainability approach is characterized by the precautionary principle. "Precautionary" is defined in the 1992 Rio Declaration on Environment and Development as follows: "Where there are threats of serious or irreversible damage, lack of full scientific certainty shall not be used as a reason for postponing cost effective measures to prevent environmental degradation." The precautionary principle is so frequently invoked in international environmental resolutions that it has come to be seen by some as the basic normative principle of international environmental law (Costanza 1994). The precautionary principle is found in Article 3.3 of the United Nations Framework Convention on Climate Change, which obliges its parties to "take precautionary measures to anticipate, prevent or minimize the causes of climate change and mitigate its adverse effects. Where there are threats of serious or irreversible damage, lack of full scientific certainty should not be used as a reason for postponing such measures, taking into account that policies and measures to deal with climate change should be cost effective so as to ensure global benefits at the lowest possible cost".

In practice, applying the precautionary principle means that in cases where the precise value in question is uncertain but there are threats of serious or irreversible damage, higher shadow prices should be used, in order to avoid risk. Adoption of the precautionary principle filters out those project and policy alternatives that carry a high penalty for irreversible and potentially catastrophic impacts, such as global warming, loss of biodiversity and ozone layer depletion.

Use the weak sustainability principle

To avoid unacceptable costs to either the current or future generations, the safe minimum standard (weak sustainability) approach can be adopted. This is the Brundtland Report standard, which has been endorsed by almost all nations and governments.

Use the no regrets principle

A no regrets policy will be of benefit regardless of which uncertain outcome actually occurs. It means that efforts to reduce some uncertain impacts have an immediate benefit in another sphere of impacts. Reductions in greenhouse gases, for example, improve local and regional air quality and conserve fossil fuels for future use. If achieved by inducing more energy-efficient travel modes or by reducing travel, policies leading to such reductions would also decrease traffic noise and possibly reduce community severance. No regrets actions are win-win solutions, which can always be advantageously undertaken.

Use appropriate discounting

To maintain the sustainability principle of equity for future generations, the issue of discounting must be examined. Section 2.8.1 reviews the problems created by discounting and a possible approach to resolving them.

4. Use appropriate pricing methods.

Consider the level of planning used

This report develops shadow prices of environmental impacts for road transportation planning and project appraisals. We have assumed that these decision-making situations are subordinate to a pre-determined, socially optimal strategy concerning the direction of different sectors towards government goals. For lower-level applications (see Section 1.2.3), the average damage of transportation activities is less relevant than the marginal environmental damage from an additional unit of transportation. The latter is crucial to determining socially optimal transportation policy and to appraising different project alternatives. But marginal costs are also much more difficult to determine than average costs, and it has not been done in this report for a number of impact categories. The average cost estimates per kilometre of vehicle travel should not be used as a substitute for marginal costs in applications, such as cost benefit analysis of abatement measures, road pricing or other economic instruments for dealing with external environmental impacts. However, where marginal cost estimates are not available, average costs can be used tentatively until more complete information is available.

Consider the need for sustainability incentives

Prices found in the market act as signals for action. Shadow prices have the same function in the decision-making process. Where it may not be possible to arrive at a precise shadow price, a concern for the incentive effect of the price may lead the analyst to use a higher rather than a lower price. However, it is always necessary to provide an explanation of the range of prices possible, and the reason for the choice made, so that decision makers have complete information.

Consider the use of "conservative" prices

Professionals may resist using new and higher cost estimates that incorporate uncertain non-market values because of their preference for "conservative" estimates of value. In accounting and economics, it is desirable to avoid overestimating wealth and to incorporate unanticipated costs in economic analysis. However, "conservative" estimates of non-market costs result in less conservation of resources. A truly conservative analysis relies on low estimates of benefits and high estimates of costs, including non-market costs. In the context of environmental impacts of roads, it seems more appropriate to use rather than omit uncertain cost estimates, and to use the higher range of such estimates for the sake of conservative analysis in the precautionary spirit.

SCBA is a natural extension of CBA, which has been firmly established in transportation appraisals. Agency and user costs are standard in transportation appraisals. Many environmental impacts can be monetized as rigorously as the value of travel time or the costs of accidents. The major challenge is to include the non-monetizable values and intangibles.

The recommendation to develop defensible monetary values of environmental and social impacts for SCBA where possible and to evaluate intangibles and other criteria that are not amenable to monetizationin whatever terms they can be expressed, is not a new proposition; for example, the SACTRA report (SACTRA 1992) and the United Kingdom’s response to its recommendations recognized the wisdom of monetizing every criteria that could be reasonably monetized (Wood 1994). One important benefit of applying such a framework would be the process itself, which emphasizes a rigorous decision-making approach.

King (1994) pointed out that some of the most catastrophic environmental and economic impacts from past projects and policies were due to ecological factors that could have been foreseen, but were never addressed in narrowly focused cost benefit analyses. The baseline cause-effect relationships in transportation planning can be foreseen and are amenableto analytical standardization. General purpose sets of shadow prices that are assigned to specific types of emissions and to withdrawals of resources due to transportation are also feasible (see Chapter 4).

The MAE framework could provide a suitable basis for including tangible, monetizable environmental impacts side by side with unmonetized and intangible ones. With such an approach, existing environmental cost information, however imperfect, can be assembled and included in a monetary SCBA. This form of SCBA incorporates the weak sustainability approach by:

2.6.1 From Sources to Environmental Damage Costs

Transportation activities cause a variety of environmental impacts. Several steps may separate the original activity and the ultimate cost in terms of impacts on humans, materials and ecosystems (Figure 2.1). It is often necessary to understand and quantify each of these steps in order to measure the different types of impacts and the total costs. Ecosystem models play a central role in converting estimates of resource withdrawals from, and wastes dumped into, nature by transportation activities into estimates of ecological and economic damages. Various studies increasingly recognize that, by adapting to environmental change over time, socioeconomic systems and ecosystems can effectively lessen the potential impacts and resulting damages. It is also becoming clear that mitigation and adaptation may produce other benefits, in addition to the primary targets of reducing and minimizing impacts. Many of the elements in Figure 2.1 still involve uncertainty due to limited knowledge, or remain fuzzy because of present ignorance, so. analysis of a transportation activity contribution to its ultimate cost can be challenging.

Starting at the top of Figure 2.1, human activities or sources, produce changes in the environment through emissions of pollutants and habitat alterations. These changes, termed stressors, can be chemical and/or biophysical. The stressors can be mitigated before they reach the receptors (humans, other life forms and materials) which they may harm. The range and fate of a pollutant transported through air, water or soil must be determined in order to assess the exposure of the receptors to the stressors. Impacts are the chemical or biophysical consequences expected in a receptor after a change in exposure to stressors. Dose-response and physical damage functions describe the impact sensitivity of receptors to a given concentration of a stressor. Receptors may have an adaptation ability which lessens the magnitude of the impact with time. Adaptability and sensitivity to change in a stressor determine a receptor’s vulnerability.

The net quantities of impacts remaining after mitigation and adaptation are multiplied by the unit values of each category of the residual damage. The costs of mitigation and adaptation are added to the total cost of impacts. Possible side-benefits of the impacts and of the adaptations are subtracted for a full accounting. Non-monetizable impacts cannot be quantified in the same way, but they must be described in order to maintain a comprehensive representation of the activity-impact process. The elements in Figure 2.1 will now be elaborated.

The typical stressors and external environmental impacts arising from transportation activities are elaborated in Chapter 3. Some impacts result directly from road construction, maintenance and vehicle operation, while others result indirectly from activities that produce automotive fuels, vehicles, road construction materials and machinery. Direct impacts occur at the same time and place as the source activity (driving a car, roadway construction), while indirect impacts occur at a different time or place.

The sources of polluting activities can be classed as point, mobile and area sources. Point sources are stationary commercial and industrial establishments with significant impacts, such as power plants, landfill sites and refineries. Mobile sources are vehicles of all types: on-road vehicles, off-road vehicles and machinery, trains, ships, airplanes and military vehicles. Area sources are multiple-point sources dispersed over an area, where each point contributes a small amount of pollution, for example, residential and commercial furnaces. Leksell and Löfgren (1995) point out that roads and urban streets can be viewed as line sources of vehicle exhausts, or more strictly, thin and narrow box sources producing continuous, albeit fluctuating, emissions of exhaust components.

Volume-oriented Measures

Structural Change

Quality Improvements

Sink Enhancement

Current, Near Future, Distant Future

Local, Regional, Global

Critical Load, Target Load

Damage per Unit of Stressor
 

Spontaneous in Ecosystems

Planned in Socioeconomic Systems
 

Value of Material Damage/Replacement

Mitigation and Adaptation Costs

Value of Human Life and Health

Costs of Social Consequences

Ecological Resource Values

Lack of Knowledge

Irreversibility and Risk Aversion

Non-market Values

Some pollutants can also be mitigated by sink-enhancement measures. A sink is a natural medium that assimilates and stores a pollutant, thus removing it from those environmental media where it might do harm. Intentional augmentation of sinks in the environment rather than mitigation at the source is an attractive alternative for stressors with widespread range and long-term performance. Carbon dioxide assimilation into trees planted in large numbers, or seeding the oceans with minerals to enhance algal growth that would capture carbon dioxide, are examples of sink-enhancement measure currently considered to partially mitigate global climate change.

Many environmental impacts result from cumulative or indirect stressors, such as the accumulation of small amounts of non-point pollutants or land use changes. Indirect and cumulative stressors should be included in the environmental assessment process where they are considered significant. They have sometimes been ignored in the cost benefit framework, which traditionally has had a project focus and has not examined the accumulation of seemingly small impacts that become apparent after the project’s completion. Acid rain and ozone depletion problems have resulted from millions of small contributions, the latter one over several decades.

The fate and range of a stressor transported through air, water or soil must be determined in order to assess the exposure of receptors and the damage. Some impacts may range from local to global—air pollution, for example—and the underlying stressors can be created, dispersed or concentrated by agents, such as the weather or water currents. For example, on a warm sunny day nitrogen oxides react with hydrocarbons and form ground-level ozone locally. The local ozone may be blown by wind into one end of a region’s valley (regional effect). Nitrogen oxides, too, blown away by wind may come down as acid rain elsewhere in the region (regional effect) or outside the region (transboundary effect). Both ozone and nitrous oxide contribute to global climate change.

Once the pollutant is dispersed and its concentration is established in the environment, the dose-response relationships (for living receptors) and the physical damage functions (for non-living receptors) express how much damage of a given class is produced per unit of the impact of a given concentration. They summarize the complex physical, chemical and other processes separating the dose of a stressor from the exposure and the effect on the receptors. The dose-response relationships for wildlife and ecosystems are difficult to study, because the existing ecosystem dependencies and patterns of ecological succession and ascendancy change as a result of pollution. Differentiating the damage caused by exposure to a specific stressor, as distinct from other pollutants and natural background processes, may also be difficult.

For a given dose of a stressor and a given exposure, there could be a variety of responses from the receptors. For example, a given dosage of ground-level ozone can harm apples but not blueberries, and will cause different ailments in different people. Local conditions and likely exposure of receptor organisms and materials play an important role in the dose-response and physical damage relationships. A pedestrian walking for a brief period through a polluted downtown area will not be harmed to the same degree that a bicycle courier would be during a day’s work in the same conditions. A forest clear-cut in a remote area may affect the aesthetic experience of only a few visitors, but the same dose of forest clearing will produce a greater response in a populated area. A selectively logged forest, however, may not impose any visual impact.

Ecosystems have some assimilative capacity (also called pollution space), that is, some amount of a pollutant may be released in a particular airshed, watershed or soil reservoir before harmful effects begin to appear. A critical load must be exceeded for environmental harm to take effect. For acidification, the critical load limit is expressed in terms of the deposition of sulphur, nitrogen or hydrogen ions per surface unit and time unit. A similar definition applies to heavy metals. For ground-level ozone damage, the critical levels are described as the sum of the hourly ozone concentration values during a vegetation period. In the context of climate, an appropriate measure might consist of a critical rate of change in the mean global temperature per decade (Rodhe 1994).

Target load refers to a load that is set up as the goal at any given time on the basis of political and administrative criteria. From the ecological point of view, this limit should be either the same as, or lower than, the critical load for the most susceptible ecosystems or for human life. In reality, it is often higher for technological and economic feasibility. This means that a certain amount of environmental impact is accepted, at least on a temporary basis (Rodhe 1994).

The magnitude and rate of change in a stressor are important in determining sensitivity, i.e. the degree to which a system will respond to changes in stressors. Larger and faster changes generally lead to more severe effects. New stressors add stress to systems already affected by pollution, resource demands and nonsustainable management.

A particular stressor may cause many types of impacts, from human morbidity to social costs. Each type of damage is measured in physical units such as number of fatalities, days of restricted activity, square area of building facades soiled with dust, percent of crop affected, and so on. System adaptation may reduce the magnitude of an impact, but the faster and larger the change, the less effective adaptation might be in the short term. Adaptation is spontaneous in natural systems as they respond to changes in conditions, but human activities have limited the natural adaptation potential through fragmentation of habitats and elimination of wildlife migration corridors. Adaptation can be planned in socioeconomic systems in anticipation of the changes. Wealthy societies can adapt better than countries lacking capital, institutional arrangements, know-how and other resources. Extreme events tend to accelerate socioeconomic adaptation.

Vulnerability, or the extent to which stressors may harm a system, depends on both its sensitivity and ability to adapt to new levels of stressors. Vulnerability increases with the magnitude and rate of change. Vulnerability decreases with economic development, as wealthy societies are, in general, less sensitive and more adaptable than less developed countries. Vulnerability determines the net impacts of a stressor on a receptor. The net quantities of impacts are multiplied by the unit values of each category of the damage to produce the total value of residual damage, i.e. impacts that remain after mitigation and adaptation actions have been carried out.

It is clear from the conceptualization of the process that the total cost of impacts of direct and indirect transportation activities includes the costs of mitigation, residual damage and adaptation to the damage. To maintain integrity of environmental impact accounting, possible side-benefits of the impacts and any gains in addition to minimizing damages realized from the adaptations, must be subtracted from the total cost of negative impacts. This is the "integrated costing approach’ discussed in Section 2.7.6.

Unexpected damage may still materialize, since we do not have full ecological knowledge to be able to foresee all possibilities. Other qualitative accounts arise because some environmental values cannot simply be expressed in money. On the other hand, human ability to adapt through institutional arrangements, change in attitude and expectations, innovation, and technology development may counteract the extent and magnitude of socioeconomic damages foreseen at present. For these and other reasons outlined in Section 2.2, full accounting must include descriptions of the non-monetizeable aspects of impacts.Whereas Figure 2.1 applies generally to all types of environmental impacts, Figure 2.2 shows the logical relations among the different components specific to local human health consequences of traffic air emissions in urban areas. The ministry applied a similar methodology in 1994 to assess damage to human health and crops in the Lower Fraser Valley as a result of ozone and fine particulate pollution.

A better appreciation of the costs of environmental impacts can be gained by considering the goods and services that natural resources and systems provide. To distinguish them from human-produced goods and services, natural goods and services are called natural capital. Environmental costs arise from a loss of the following four categories of natural systems functions (de Groot 1994):

Environmental goods and services benefit individuals, society and nature, regardless of who owns the resource. Environmental benefits include a combination of the following socioeconomic and monetary values, as well as ecological and intrinsic values that can be attributed to the environmental functions of natural capital (de Groot 1994):

2.6.3 Defining Cost Categories

It is useful to begin by considering the meaning of the word cost. Usually, cost is used interchangeably with the word price to reflect market payments. But economists define the word more broadly as benefits forgone. Thus, anything that requires a trade-off of another benefit implies a cost. Cost could be used interchangeably with negative impact, resource loss or problem to reflect the negative trade-offs implied. Although any of these terms might be appropriate, cost is used in this report because it is the standard term in economics. In some cases, especially when considering a non-market cost that has not been widely studied, an appropriate beginning is to consider the value of the environmental good or service that changes as a consequence of a project or policy. The change can be in either direction: the good or service may be degraded or lost (the case for most of the environmental impacts of transportation) or it may be enhanced or created (for example, a bog built to compensate for a wetland lost due to road construction). A gain in the environmental good or service is a benefit, or negative cost.

Costs can be classified as either internal or external depending on whether they are borne by the provider and user of the road facility, or by others (either individuals, society, future generations or other species). An externality (external effect, impact or cost) is an effect that impacts third parties without invitation or compensation. A broad consensus exists among OECD countries on a basic definition of an external effect of transportation: it is a negative (or positive) consequence of a transport activity, but the person who generates it (or benefits from it) does not have to offset the cost in monetary terms (Bonnafous 1994).

External costs represent a failure of the market or decision makers to capture all costs (including non-monetary costs) of production and consumption. In the case of roads and motor vehicles, the direct beneficiaries of transport are typically the owners, drivers and occupants of vehicles. The "losers" may be the general taxpayer who subsidizes the infrastructure, ecosystems that suffer damage, or future generations that would have to deal with climate change induced by vehicle emissions today. Even if a road-user charge internalized an external cost, but compensation was not paid to the third party being harmed, or the traffic would not drop so low as to render the impact harmless, the external effect could continue.

Transportation externalities are produced directly by road construction, maintenance and use of roads, and indirectly during the production and disposal of fuels, vehicles, road materials and machinery. Indirect impacts also include changes to land use and resource use patterns induced by transportation. The impacts that are "embodied" in the direct and indirect activities include the use of energy and resources and the production of pollution. Where possible, this report incorporates the embodied externalities into its estimates, but more work is required to fully account for them. The externalities can also be classified as "upstream" (for example, emissions from blast furnaces used in making steel for vehicle manufacture) and "downstream" (for example, the impacts of the disposal of vehicle bodies, batteries, tires, engine oil and brake fluids). The direct externalities are usually "end-use," that is, they are associated with the consumption of a product (fuel, vehicle) or service (public transit). Because the number of such "systemic" externalities is potentially infinite, especially at the upstream end, it is impractical to trace all of them (Dobes 1995).

An alternative classification of externalities is by life cycle and system cycle (BTCE 1995). Vehicle life cycle emissions are those associated with vehicle energy end-use (tailpipe emissions), energy used to manufacture, maintain and dispose of a vehicle, and fuel production and supply emissions (part of full fuel cycle emissions). A study for the ministry estimated the vehicle life-cycle energy use to be 20% to 45% higher than the direct automotive energy used to propel the vehicles. Transit buses had the lowest figures, while cars had the highest (Sypher 1995). Emissions are proportional to the amount of energy consumed. The sum of life-cycle energy of all vehicles and energy used in the construction and repair of transportation infrastructure can estimate transportationsystem-wide emissions. In 1992 to 1993, Australian road transportation accounted for 21% more greenhouse gases (CO2 equivalent) on a system cycle, and 12% more on a full fuel cycle, compared to end-use (BTCE 1995). Inclusion of energy used by the urban and rural form as an indirect consequence of transportation (sprawled suburbs versus more compact towns) would provide transportation/land use system-wide emissions.

Table 2.2 presents a basic classification of the best-known external costs of road transportation. A distinction is made between externalities that road users impose on each other, on the ecological environment and on society at large. The external costs are divided into impacts that arise from direct transportation activities (including facility construction and maintenance, as well as the parking of vehicles), those arising from indirect activities (production and disposal of vehicles and production and distribution of fuels), and those related to the presence of the facility.
 

Most costs fall into one of four domains shown in Table 2.3, which lists examples of various motor vehicle costs. Costs of goods that are regularly bought and sold in a competitive market are called market costs. Costs involving goods that are not directly bought and sold, such as aesthetics, health and comfort, are called non-market costs. Non-market goods often affect market costs. For example, landscape aesthetics affect the market value of real estate, and the value placed on comfort affects many consumer-purchase decisions. Because non-market goods and services do not usually have market transactions in which one could see the price the market puts on them, non-market monetized values are also termed shadow prices.

The line between environmental and social costs is hard to define. Some literature reviewed for this report maintained a distinct division between these costs; others lumped them together. What some authors categorize as social costs, others consider to be environmental costs. This report focuses on environmental costs, as listed in the table of contents of Chapter 3. They are external non-market costs and would fall into the lower right quadrant of Table 2.3.

Typically, environmental costs have been excluded from transportation economic analysis. These costs may be local, regional or global and can include impacts on human health, lifestyles and comfort, material objects, such as buildings, and the natural environment. Many of these costs also have a time scale. For example, current CO2 emissions may alter the climate and result in damages in the future.

Verhoef (1994) classifies all monetization methods into:

2.7.1 Damage Cost Approach

The damage cost approach focuses on the residual damages, net of any gains, that might be realized by the change in a stressor level In relation to human health and mortality from degraded urban air, Hall et al. (1992) applied the method to the Los Angeles airshed. Many authors have applied it in the transportation sector (for example, SENES 1994; Small and Kazimi 1995), using improved explanations of the dose-response relationships on which the method’s accuracy relies. Harrison (1995) considers the damage-based methods to be the only valid ones for social costing of electric utility. Ontario Hydro supports the damage function approach to costing (Boone 1995).

The easiest costs to monetize are loss of market goods, such as reduced yields of timber from a decaying forest, damage to wood construction materials and clothing from increased ultraviolet irradiation, and agricultural crop loss. The market price (or a surrogate market price in the case of consumptive-use value) provides an accurate socioeconomic assessment of these direct values. Since damage costs include the impacts on non-market goods, techniques for estimating their value must often be used, but intangibles, such as loss of biodiversity, have to be left out of the quantitative evaluation.

The main challenge of the damage cost approach is to figure out the optimal level of mitigation expenditures and to estimate the net residual, unabated damage. This difficulty is also encountered in the control cost approach (Section 2.7.2) and in the integrated cost approach (Section 2.7.6).

2.7.2 Control Cost Approach

Control costs are often used to monetize non-market goods, but there is no reason to believe that control cost estimates approximate actual damage costs. Control includes mitigation measures necessary to reduce pollution or other stressors to a "reasonable" level, adaptation to make the consequences less severe, and, in rare cases, elimination of the cause of harmful impacts. A relatively easy technique to use, the control cost approach, is understandably a popular choice in many situations. Unfortunately, some studies choose the least-cost control measure arbitrarily, without regard to the residual externality. As in damage costing, this approach must consider how to determine what level of control is optimal before the costs exceed the benefits, and how to value residual impacts.

In most situations, increasing levels of protection can be purchased at a progressively increasing price, and a residual impact often remains despite the control. For example, the impacts of road or airport noise on nearby residents can be reduced by installing insulating windows. The cost of providing such windows to all houses that experience a certain noise level is sometimes used to calculate noise-control costs. Such costs, however, are affected by the level of noise exposure used to determine which homes are to receive protection, and residual noise impacts remain at night, even with the best windows, and when residents are outside their homes or open the windows. Thus, expenditures on insulated windows may provide a rough minimum estimate of specific levels of traffic-noise reduction, but not the absolute value of noise impacts.

Control cost estimates are based on the assumption that control standards rationally reflect society’s values. In practice, such standards are seldom applied consistently and may reflect many influences besides rational valuation. Control costs can be overestimated, since the cost of an externality might be less than the cost of preventing it. Determining the most appropriate control level ultimately requires the valuation of non-market resources. Relying on control costs as a determinant (rather than an indicator) of non-market values implies cyclical reasoning. For these reasons, control, cost estimates are usually inappropriate substitutes for the total environmental impact costs.

2.7.3 Hedonic Pricing Methods

Hedonic pricing infers values for non-market environmental goods from consumer decisions in the market. The methods belong to the class of revealed preference methods, since consumers reveal their values indirectly to the analyst. This is done by finding associated market-priced goods that correlate to the non-market good under consideration. A common method analyzes the effects of an impact on property values or wages. For example, if houses on streets with heavy traffic consistently have lower property values than otherwise comparable houses on low-traffic streets, the cost of traffic noise (or, conversely, the value of quiet, clean air, safety and privacy) can be calculated (Mackie 1991). Hedonic studies also use the higher wages paid in jobs that involve environmental discomfort, or risk, to calculate the monetary value of such disbenefits. The hedonic methods are suitable for monetizing localized impacts only, and are also sensitive to the specification of the tolerance threshold of the disbenefit.

Hedonic pricing studies must be carefully structured to identify and adjust for explanatory variables, and the hedonic prices may not reflect the full impact. For example, hedonic studies that use the impacts of traffic on property values to indicate noise costs must consider the possibility that other traffic impacts, such as accident risk, air pollution and loss of privacy also influence property values, and that home buyers may not have complete information on each house’s noise exposure. Also, such studies can only capture noise impacts on residents, not the total impacts including other occupants, non-motorized traffic and motorists themselves. Finally, property values do not measure sleep loss, stress and other impacts on human health.

For these reasons, hedonic methods should be interpreted carefully, and they should not be construed to indicate the total costs in an environmental impact category. Also, the impact of the environmental variable may be present in a number of markets, for example, housing and employment. In principle, all would have to be considered, and if they are not, the results may be inaccurate (Poldy 1993).

2.7.4 Contingent Valuation and Conjoint Analysis Methods

Contingent valuation and conjoint analysis methods are used to survey a representative sample of society on how much they value a particular non-market good. Because contingent valuation relies on respondents’ statements rather than on actual behaviour, it is also called the stated preference method. Questionnaires based on this method give respondents a hypothetical choice of investments in environmental protection or improvements. For example, they may be asked how much they would be willing to pay through increased electrical utility charges to save a threatened species of salmon (WTP: willingness-to-pay method). Or respondents may be asked what is the lowest value they would accept as compensation for giving up the use of a particular recreation site (WTA: willingness-to-accept compensation).

Contingent valuation is considered the only method available for monetizing option and existence values (Winpenny 1991). Verhoef (1994) notes that contingent valuations essentially measure the psychological cost associated with a public amenity’s option and existence values. As such, the cost can be added to the estimates obtained by other methods, without the risk of double-counting. If anything, these costs understate the non-use values. They are also biased because of the human tendency to value larger and "prettier" animal and plant species more highly than other species that may play a far more important role in ecosystems.

Contingent valuation surveys have methodological problems. Often, surveys seek to explore trade-offs that the respondents have never experienced and thus have difficulty imagining. Valuations of environmental assets vary over time for a given person, changing with pollution levels and societal attitudes, income level, scientific knowledge of environmental deterioration processes, and the degree of knowledge of environmental issues. Respondents may also attempt to influence the selection of project alternatives by distorting their answers. Conversely, the answers depend on the way scenarios are described and questions are formulated. In any case, what people say about their values is not necessarily reflected in their actual choices in everyday life.

Willingness-to-accept compensation is often several times higher than willingness-to-pay. WTA is more difficult to determine and many analysts opt to use WTP instead, which can seriously bias assessments and undermine policies based on them (Knetsch 1994). The demands for valuation numbers are so great that the shortcomings of the contingent valuation methods are often ignored. Expressing this concern, Knetsch (1994) writes: "Rather than a continuation of present practice, perhaps socially desired interests and objectives might be better served by a more even-handed regard for empirical findings..., and a greater willingness to explore alternatives to current practices."

The methodology of conjoint analysis (or similar techniques known as discrete choice modelling) is an improvement over the contingent valuation method, but many of the limitations still remain. Conjoint analysis assesses environmental benefits relative to each other and to other benefits of transportation, and not in absolute terms. Instead of relatively simple and direct questions concerning a single environmental amenity, respondents choose among hypothetical, mutually exclusive alternatives; for example, trips, which are characterized by a number of attributes, such as fare or fuel cost, local air pollution reduction, seating time in a public transit vehicle, noise, and dust and dirt from the road (Sælensminde and Hammer 1994). A logit modelling technique is used to analyze the responses and derive the WTP function of the respondent group. An explanation of the theory can be found in Hensher (1991), for example.

2.7.5 Travel Cost Method

The travel cost method of costing is a revealed preference method designed to evaluate parks and sites of natural importance. It uses the cost of travelling to and visiting a natural area to estimate the consumer benefits of a site. In visiting a campsite, for example, an individual incurs costs related to admission fees, goods such as camping supplies, transport and time spent travelling to the site. People travelling a greater distance reveal the greater value of the site they are going to, compared to another site located at a smaller distance.

Numerous travel cost studies have been undertaken in the United States. In the United Kingdom, applications have mainly related to forestry projects and nature reserves (Mackie 1991). This technique requires a large amount of unbiased survey data. The results may be skewed by the positive value that some visitors may obtain from travel itself (ITS and Beca 1992; Beca 1992). It is also difficult to incorporate multi-purpose trips into the analysis. Intangibles such as option and existence values, and benefits enjoyed by people who never actually visit the site, cannot be measured with this method (Winpenny 1991).

2.7.6 Integrated Cost Approach

Many authors pointed out that in order to produce a reasonable estimate of environmental impact costs, a single technique should not be used alone. A meaningful result cannot be expected from such analysis and, consequently, shortcuts should not be adopted for the sake of analytical expediency (for example, Schechter 1991; Ayres and Walter 1991; Verhoef 1994).

The single techniques do not address three key issues in environmental impact costing adequately. First, there are many functions and values that may be affected by an environmental stressor, not just the ones covered by one single technique. Second, attempts to respond to a stressor by controlling its levels (or abating its consequences), and by adaptating to reduce damages once impacts occur, represent partial costs due to the stressor. These costs should be added to the costs of residual damage remaining after application of the response actions. Third, mitigation and adaptation may produce beneficial effects other than the primary impact reductions from a particular stressor. The benefits should be subtracted from the costs of response actions. Similarly, any positive impacts produced by a stressor must be subtracted from the negative consequences of residual impacts. The "integrated cost approach" attempts to address the three key points, although it is not an easy task.

Table 2.3 reveals that most of the costs due to loss of environmental amenity can be approximated by adding up results obtained with different methods. Control cost plus residual damage cost plus contingent valuation of non-use values should capture the monetizable part of the total cost, provided any significant multiple-counting is filtered out. It would not be sensible to make disproportionate efforts to avoid double-counting quantified and costed impacts, while at the same time ignoring non-monetizable impacts, which are frequently orders of magnitude larger. Errors arising from double-counting are usually inconsequential in the big picture, and time and effort spent on discussions of the topic detract from more substantial issues.

Economic theory requires that the response costs of mitigation and adaptation be determined at an economically efficient optimum level. It is commonly assumed that the higher the impact reduction desired, the more it costs per unit of impact to achieve it, i.e. the marginal cost of response actions is expected to increase at an increasing rate. By contrast, the marginal value of gains from the response actions per unit of impacts avoided decreases. According to the economic theory, the optimum is reached at a point where the added expenditure on a response action equals the corresponding impact cost reduction, i.e. the rate of gain in reduced damage equals the rate of increase of response costs. Outside of the equilibrium point, either the expenditures on unit reduction of impacts could still be increased without "overshooting" the gains, or the gains in averted damage become smaller than the outlays on the response actions.

The concept of economically efficient equilibrium is difficult to operationalize, because the marginal cost curves are not easy to determine. To calculate the optimum, all damage reduction resulting from a curtailment of a particular environmental impact must include present and future market and non-market costs. This is a difficult if not impossible task. In fact, the economic damage cost function may only be guessed (global warming, ozone layer depletion), and non-market impacts remain largely unknown or unpriced.

The cost of response action at any level of the impact reduction must also be known to determine the cost-effective measures, but pollution control and abatement technologies and other social and economic factors change over time. Future adaptation measures and their costs and effectiveness are not knowable. The marginal cost curves do not usually have the shape assumed by the economic theory and may be discontinuous. Mitigation costs may exhibit a declining (not increasing, as commonly assumed) marginal cost curve due to factors such as economies of mass production of pollution control devices. In the real world, response actions are not chosen strictly on the cost-effectiveness basis (Harrison 1995).

The integrated costing approach recognizes the theoretical and practical limitations. Whether an economically efficient or inefficient response action is carried out is irrelevant for environmental impact accounting, as long as a reasonable approximation of the actual response cost is used to evaluate the total impacts. Many response costs are internalized in other accounts, such as vehicle capital and operating costs (engine emission control devices), infrastructure costs (anti-noise barriers along transportation arteries, insulating windows in buildings), or fuel and energy costs (industry costs of pollution prevention and abatement). Future social adaptation is thus the primary item among the response measures that should be considered in the integrated costing. Because adaptation is largely uncertain regarding timing, costs and effectiveness, the case is made for the precautionary approach to environmental impact valuation.

Finally, the assessment of impacts should net out any secondary benefits accomplished by the response measure as a by-product of the primary impact reduction. For example, fossil fuel conservation to slow down global warming also serves to improve local air quality, reduce acidification and toxicity, and perhaps also alleviate the general traffic noise level, congestion and severance in urban areas if it is achieved by reduction of travel demand. To assess these impacts and to cost them is a formidable task by itself. The secondary benefits could sometimes exceed the response costs in the lower range of emission reductions. Because both the response costs and the secondary benefits are uncertain, subtracting one from another may not be meaningful.

Gains produced by a stressor should also be subtracted from the losses it causes. The most obvious are the economic gains realized from the change in physical conditions brought about by the stressor. In climate change assessments for example, it is recognized that losses in the ski industry may be offset by increases in recreation activities that take advantage of warmer, drier summers in the Northern Hemisphere. Heat wave mortality may be accompanied by reductions of cold-weather fatalities. Losses in agricultural yields in some regions might be compensated by better productivity in previously unproductive or marginal farming areas.

The integrated costing approach does not claim the ability to accurately and fully provide all costs and benefits that determine the total net monetized costs of an environmental impact. Rather, it is a checklist of the most important elements that need to be considered in full-cost environmental accounting in order to reflect the activity-impact process outlined in Figure 2.1. Depending on data availability and confidence in the cost data, some items on the list may be elaborated in detail while others remain best guesses. The important feature of the approach is that it forces the analyst to retain a full picture of the costing problem instead of focusing on a single monetization techniques and incomplete set of values.

Since environmental costs and benefits from current actions often extend over many years or even generations, the choice of discount rate used in economic analysis of environmental impacts significantly affects the results. Typical economic analysis models, such as the ministry’s SCBA model, which was introduced in 1992 to evaluate highway investments on the basis of road user costs (MoTH 1992a, Bein 1994), use a real rate of discount (net of inflation) of 8% and an analysis period of about 20 years. With this rate, $1.00 twenty years from now will be worth only about one-fifth of its present value. With a longer time period, values collapse almost completely. At an 8% discount rate, damage of $1 billion 50 years from now is currently valued at only $21.3 million. In 200 years the present value is $200.

In transportation project and policy appraisals, the capital costs tend to be concentrated at the beginning of the analysis period, while social benefits are more evenly spread out throughout the period. Discounting may produce a false picture of a project’s social desirability, when capital projects compete for funding with maintenance and rehabilitation projects.

There are two main views of why discounting is appropriate. One view holds that since current resources can be used to increase wealth through investment, they have greater value than future resources, even after adjusting for inflation. This is the opportunity cost of capital (OCC) approach. The other view focuses on individuals’ higher preferences for consumption now rather than consumption later. This is the social rate of time preference (SRTP) approach. In theory, the approaches are equivalent, but in practice the consumption rate differs from the investment rate. Traditional cost benefit analysis, which was developed with a focus on single investment projects and a time horizon of usually 20 to 30 years, has tended to use the OCC as the basis for the discount rate.

Sustainability requires equity between this generation and future generations. The prospect of global atmospheric change and other long-term or irreversible environmental effects can only be assessed over long time horizons. Because traditional cost benefit analysis excludes intergenerational equity and uses standard discount rates that imply short horizons, it is inadequate for planning sustainable transportation.

For sustainable transportation planning, a discounting model is needed that includes intergenerational equity, appropriately values the environment, includes allowance for risk, uncertainty and irreversibility, and accounts for both the social rate of time preference and capital opportunity costs. Leading economists recently set out their recommendations on intergenerational equity and the discount rate (Arrow et al. 1995). This model equalizes the utility derived from consumption in each generation, by allowing for the high and rising value of the environment as perceived by consumers. It involves converting all effects to consumption-equivalents and then discounting at the SRTP. This includes calculating a consumption-equivalent value of the opportunity cost of capital. Risk and uncertainty, as well as environmental effects, are handled similarly by calculating their "certainty-consumption equivalents." The Arrow et al. (1995) model accomplishes the goals of sustainability by using a discount rate that is theoretically and empirically sound. It does not distort decisions by overvaluing the present, by unduly weighting the opportunity cost of capital, or by requiring several rates within a single analysis. Huston (1995) summarizes the current issues in discounting and discusses the Arrow model.

Many transportation projects have impacts on the social and natural environment that will be felt past the current generation, but at the same time it would be impractical to operate with project appraisals spanning very long time horizons. For these reasons, using lower discount rates for potentially irreversible or catastrophic impacts and for loss of future lives was contemplated for this report. However, most scholars and practitioners in the economic appraisals of projects that were consulted have pointed out that this would be an analytically corrupted approach, and they advised against it (personal communications: Cline 1995; Clough 1995; Lake 1995; Mackie 1995; Poldy 1995; Waters 1995). Collectively, the advice is:

Uncertainty is the lack of a point estimate of a variable. Uncertainty may be due to lack of knowledge, insufficient information, or a result of the statistical character of a variable (for example, incidence of weather patterns conducive to the formation of ground-level ozone). Risk is the combination of uncertainty about an outcome occurring and the magnitude and severity of the outcome. An average driver’s risk associated with hundreds of "fender-benders" occurring on roads and parking lots is low compared with traffic fatalities that occur less frequently but have more dramatic consequences. In environmental impacts, risk is a combination of probability of outcome, its potential magnitude and irreversibility. The precautionary principle was developed in recognition of the existence of risks of irreversible and dramatic outcomes.

Hardly any appraisals based on an analysis period comparable to the length of the life of a transportation project can be free of uncertainty. Inclusion of environmental impacts that would be felt generations from now poses even more questions about risks of potentially harmful decisions and about the associated uncertainties. Even for impacts that are taking place now, knowledge is often insufficient to use point estimates for quantities of impacts and for their unit costs.

Uncertainties should not be considered as a reason for disregarding existing information, particularly if risks are high. In policy and project appraisals, there may be a danger of environmental aspects being given less attention than economic aspects because the latter seem more definitive. Until recently, this has been the main reason for using CBA rather than SCBA in road project appraisals.

Sensitivity analysis is recommended and widely used in cost benefit appraisals with uncertain estimates. The drawback of sensitivity analysis is that it is only able to assess one variable at a time. It is the combinations of uncertainties of all relevant variables that matter in complex evaluation problems. Where available, probability distributions can be used to do uncertainty analysis of the results of appraisals. Such facility already exists in the ministry’s road cost benefit UBCS software, and it can be implemented in analysis of environmental impacts if it is programmed with a leading spreadsheet software such as Lotus 123 or Excel.

Commercial software is available (for example, @Risk, Crystal Ball) that can perform the uncertainty analysis. Using the appraisal spreadsheets as a base, the uncertainty software applies specified probability distributions of all or selected variables in Monte Carlo simulations. This approach has been used by SENES (1994) to estimate the distributions of mortality from particulates in a study for the ministry. The output from the simulations would be input into an appraisal framework, allowing the input uncertainty to carry through to the evaluations of the final results, again using software such as @Risk or Crystal Ball.

Within the proposed framework, however, sight of the overall analysis should not be lost. Monetization and numerical analysis will only cover part of the overall evaluation, particularly in policy appraisals and strategic transportation planning. The non-monetized values will largely escape quantification, and many of them will simply be intangible. By definition, the greatest risk and uncertainty are attached to the non-monetizable variables. Uncertainty analysis on "quantifiables" and "monetizables" could be counterproductive at the policy and high-level planning levels where "non-monetizables" are prominent. Sensitivity analysis may be more justified. For project appraisal and other tactical analyses, such as program planning, uncertainty analysis would be more warranted.

2.8.3 Value of Statistical Life

Economists widely accept the principle that in market economies the social cost of a change in economic outcomes is most usefully measured by the sum of individuals’ willingness-to-pay for that change, given their current economic circumstances. If public policy and project appraisal consistently uses such valuations, most people will find that their overall welfare is improved (for example, Hicks 1941). For cost benefit analysis, groups ranging from the US Office of Management and Budget to the National Safety Council to academics agree that health and safety benefits should be based on "willingness-to-pay." The willingness-to-pay method has been adopted for traffic-accident costing in the US, New Zealand, Scandinavia and the UK. In the environmental appraisal area, the Intergovernmental Panel on Climate Change (IPCC) economic analysis group is also using the method to account for mortality due to global warming. In air quality assessments of traffic in North America, the method received widespread exposure in the work on air emissions in the Los Angeles region (Hall et al. 1992). The electric utility companies also subscribe to the willingness-to-pay method for the value of life, as well as for other damage to human health.

The willingness-to-pay concept considers the average amount people actually pay for small reductions in risk to be the amount society should spend to reduce the public’s risks by a small amount. It examines, for example, how much people pay for optional safety features on their cars, trade-offs people make between time and safety when they choose a driving speed, or the extra wages paid to induce workers to take risky jobs. Generally, analysts use willingness-to-pay, revealed through choices people make in everyday situations, toestimate the value of a statistical life (VOSL)—the amount to spend for an expected value of one life saved. An overview of the concepts can be found, for example, in Viscusi (1993). Miller and Guria (1991) found that fatal risk reduction values are comparable between countries provided one adjusts for personal incomes per capita in the different countries. The values have large variability, meaning the average value across studies probably is more reliable than the value from a single study.

The VOSL in the ministry’s SCBA appraisals of projects was adapted from US willingness-to-pay estimates (Miller et al. 1993; Miller 1993) by switching to the 22% average income tax rate in British Columbia from the 15% rate in the US (Miller 1992). Miller recommended an after-tax value of C$2.83 million per statistical life, the mean and median from an analysis of 47 values. The standard deviation of the values is C$0.84 million. Subtracting mean lifetime after-tax earnings, fringe benefits and household production yields the value of pain, suffering and lost quality of life per statistical death prevented. Values from Canadian studies (Cousineau et al. 1992; Martinello and Meng 1992) range from C$2.8 million to C$4.6 million (in 1991 after-tax dollars, computed at a 27% national plus provincial tax rate). Fisher et al. (1989) makes a case for valuation between $US 2.1 million and $11.3 million at 1992 prices.

For policy reasons, the ministry adopted a lower value than recommended by Miller. The ministry had previously been using a C$0.5 million value of statistical life. The administrators responsible for highway safety evaluations felt that a VOSL of C$2 million, that is, one standard deviation below the mean, would smooth the transition. To maintain integrity of cost benefit analysis, values of travel time were lowered accordingly (Bein et al. 1994) or else safety would effectively have a lower weight, contrary to the ministry’s mission statement, which gives traffic safety the highest priority. The ministry’s value of statistical life is currently under scientific review. Possibly, the value will increase. The Ontario Ministry of Transportation estimated the value at over C$5 million in a recent study based on the willingness-to-pay method (Vodden et al. 1994). The Ontario value was likely overestimated and should have been C$3.7 million (Miller 1995). VOSL of C$3 million is adopted in the present report.

In valuing damages from global environmental impacts, questions of international equity arise. Pollution produced in wealthy countries might eventually cause the deaths of people in poorer parts of the world, where life has a much lower value by comparison when measured by willingness-to-pay. If low values of statistical life were adopted in appraisals of global impacts on human mortality, it would imply that pollution could occur at the expense of the lives of individuals in other parts of the world. Any appraisal that quantifies global damage in terms of national wealth (such as GDP and willingness-to-pay for life saved) in effect embodies an assumption that the human impact is less significant if it is poor people, or people in poorer countries, who suffer (Banuri et al. 1995). The result of using the willingness-to-pay measure based on incomes in different countries can be seen as ethically unacceptable.