In addition to costing money for the individual transport user, transport activities cause a range of external effects for non-users and society as a whole. These external costs, e.g. air pollution, noise, accidents, congestion, barrier effects and effects on nature and landscape, have become an important subject at national and international level in discussions about transport regulation and pricing schemes. In all policy-relevant cases, be it road-pricing systems for Danish cities or harmonisation of a European tax system for transport, the determination of the "right" prices and suitable methods for valuation of external effects are at the centre of the discussion.

The Danish Environmental Protection Agency has published an overview of Danish and international attempts at valuation of external effects of transport (Kveiborg, 2001). The overview showed huge differences between unit prices. The size of price estimates and the uncertainty attached to them are often dependent on the valuation method chosen. The primary intention with this report has therefore been to follow up on the previous overview by specifically evaluating the applicability of price estimates (i.e. unit prices in terms of DKK/km) for Danish transport situations. This requires, however, that prices are calculated using valuation approaches that take account of Danish transport data, e.g. annual amount of kilometres driven, distribution of cars, emission coefficients, size of population affected, and other country-specific circumstances that are responsible for variations of prices between countries.

Keeping these restrictions in mind, this study concentrates in Chapter 3 on evaluating the calculations of a recent European study on the external effects of transport, INFRAS/IWW (2000), and the latest Danish calculations of unit prices. Only cases where methods and case studies exist are included, e.g. barrier effects and the external effects of transport on nature and landscape in the results of a Canadian study. The literature evaluated in this report therefore does not include all recent attempts of pricing the external effects of transport activities, but mainly those studies that take Danish circumstances into consideration.

Chapter 2 of the report includes a short description of the main valuation methods employed in the determination of external costs of transport activities. Valuation methods include both, direct (stated preference methods, e.g. contingent valuation) and indirect methods (revealed preference methods, e.g. hedonic pricing methods). Table 0.1 provides a summary of the different valuation methods, the types of values these methods can determine, and the external effects to which they can be applied.

One of the most controversial aspects of transport pricing is the valuation of a statistical life. Not only because there is a range of ethical questions attached to this type of valuation, but also because, to a large extent, costs related to mortality determine the total external costs related to air pollution and traffic accidents. Three different valuation methods are discussed in the report; valuation of a statistical life (VSL), value of a lost life-year (VOLY), and a human capital approach. The VSL method is based on willingness-to-pay (WTP) measures in order to avoid a closer defined increase in the risk of death. The value of a statistical life saved is then calculated by dividing individual WTP amounts by the observed change in risk. The VOLY method, on the other hand, is often based on VSL calculations, converted to prices for a life-year lost by annualising the value of a statistical life over the remaining life-years. Both methods, VSL and VOLY, are thus based on the welfare-economic method of project evaluation, i.e. individuals’ own perception of the benefits and costs. The third method, the human capital approach, on the other hand, measures the potential loss in production associated with a reduction in the labour force, i.e. the individual’s contribution to a nation’s gross factor income.

The results in terms of unit prices from the different studies presented in the report can serve to paint a rough picture of the total and relative size of the external effects of transport in Denmark. The results are used to estimate the total external costs of transport for a basic scenario and for sensitivity analysis. The evaluation of their applicability to Danish transport situations has, however, entirely been based on the information provided in the respective literature sources, and the level of detail of these descriptions varies to a great extent. In most cases it has therefore been impossible to end up with a clear recommendation of using one price instead of another. In those cases where prices have been applied in the example calculations in chapter 4, the decision was mainly based on methodical considerations or the data sources applied in the calculations. This does not imply that the prices chosen reflect the true external costs of transport in Denmark. Any application of values in policy decisions, e.g. road pricing or the determination of tax levels, requires a more detailed analysis of data sources applied in the different studies than what has been possible within the limits of this study.

Table 0.2 provides an overview of the different prices used in the example calculations of external costs from transport activities in Denmark in the year 2000. For air pollution, noise, and nature and landscape effects, prices are calculated as marginal prices, i.e. the costs associated with one vehicle driving an extra kilometre in a specific location, e.g. rural or urban. For the external costs associated with accidents, congestion, and barrier effects, however, average prices have been chosen. In the sensitivity analysis for congestion costs, though, marginal costs associated with different traffic situations have also been applied.

Air pollution causes physical damage in the form of pollution and corrosion of buildings and artwork, forest acidification and increases in mortality and morbidity related to certain illnesses. In addition, climate change, caused by increasing CO₂ levels in the atmosphere, is likely to lead to substantial damage in the future. By applying different emission, exposure, and dose-response functions, the size and amount of these physical effects can be determined. Valuation methods are then applied in order to place a monetary value on these physical units. The two studies discussed in this survey both use a bottom-up approach to calculate the marginal external costs associated with air pollution and climate change, based mainly on the results of the European ExternE model and adapted to either Danish or German traffic situations.

Marginal unit prices differ substantially between the two studies, however, mainly because of the different unit costs applied for CO₂ emissions. The Danish study (COWI, 1999), uses a price of DKK40 /tonne CO₂, which is based on estimates of the potential damage costs associated with a changing climate. These estimates, though, are taken from the Second IPCC Assessment report and do not reflect the latest results in climate change damage assessment. In the INFRAS/IWW (2000) study, on the other hand, a relatively high price of DKK980 /tonne CO₂ is applied. These costs are based on the calculations of shadow prices for reducing CO₂ emissions in European countries. Baseline example calculations are based on the marginal unit prices calculated in COWI (1999). The higher unit prices from INFRAS/IWW (2000) are, however, applied in the sensitivity calculations in order to show the impact on total external costs of higher CO₂ prices.

Higher noise levels increase the risk of heart disease, hearing disabilities, stress and sleeping problems, but can also be the cause of high blood pressure and headaches. Willingness-to-pay measures for noise reduction can either be determined with the help of contingent valuation studies, avoidance-cost methods, or hedonic pricing studies based on real estate prices. In addition to the welfare losses of affected individuals, society faces additional costs in the form of rising expenses for hospital and other medical treatment. For the example calculations in Chapter 4, the higher unit prices from a somewhat older Danish study (Ministry of Traffic, 1997) are applied. These unit prices include both WTP measures from a Danish hedonic pricing study and estimates of public costs associated with increases in morbidity. These unit-cost estimates for urban areas are, however, still substantially lower than those calculated in INFRAS/IWW (2000). Marginal unit prices in the later study are based on model runs and valuation studies for Germany. The higher urban costs from this study are applied in the sensitivity analysis.

For the example calculations of external costs from accidents, average unit prices for rural, urban and motorway driving are taken from the INFRAS/IWW (2000) study. In this study the application of average external costs is recommended because a recent literature survey of marginal-cost estimates could not show a clear relationship between traffic flow and changes in accident risk. This stands in contrast to the Danish estimates from the Ministry of Transport (1997), where marginal external costs are calculated, albeit without differentiation between types of road and location. In the European study from INFRAS/IWW insurance payments to injured persons and relatives are also subtracted which, together with lower welfare losses for injured and killed accident victims, might be one of the reasons for lower unit prices. The higher unit prices from the Ministry of Transport (1997) are, however, applied in the sensitivity analysis.

For congestion costs, two different studies were analysed: an older Danish study (Ministry of Transport, 1997) and the European study INFRAS/IWW (2000). The method applied by the Ministry of Transport (1997) is based on the costs of extending an existing road network in order to avoid an increase in congestion in the future. These costs, however, are not directly related to the external time losses and increase in operation and maintenance costs caused by driving an extra kilometre, as calculated in INFRAS/IWW (2000). In the example calculations, therefore, the average costs calculated in the European study are applied in the baseline scenario. Marginal costs for different traffic situations, taken from the same source, are applied in the sensitivity analysis.

INFRAS/IWW (2000) is the only study that uses model calculations (from individual sample cities) to calculate the costs associated with barrier effects (mainly loss of time for pedestrians) for different types of vehicles. The most recent Danish study (Road Directorate, 1999), calculates barrier costs as 50% of external costs related to traffic noise, based on the results of an old Norwegian study. Two Canadian studies, Bein (1997) and Litman (2001), however, calculate barrier costs as 100% of noise costs. For the calculations in this report the medium-sized unit costs of the Danish report are applied, while the lower unit costs from INFRAS/IWW and the higher Canadian prices are used in the sensitivity analysis.

Unit costs associated with changes in landscape management are only calculated in INFRAS/IWW (2000) and therefore only these prices are applied in the baseline calculations. In the European study a "repair-cost" approach is applied, where the costs associated with (fictive) re-establishment of the original natural situation of the existing road network are calculated.

Table 0.3 shows the results of the baseline calculations for the different external effects in Denmark for the year 2000. Total external costs from road transport are thus equal to approximately DKK33 billion in 2000. External costs related to accidents make up the biggest part, with approximately 53% of total costs, followed by air pollution (28%), noise (7.7%), and congestion (5.5%). Costs related to the often-neglected barrier effects and effects on nature and landscape, however, are only of minor importance at about 3% and 2.3% respectively. Sensitivity analyses show a substantial increase in external costs from air pollution (incl. CO₂ emissions), if the higher unit prices from INFRAS/IWW (2000) are applied. Also congestion costs are likely to rise substantially if it is assumed that the majority of rush-hour traffic (30% of total traffic per day) takes place under traffic-jam conditions.