Workshop Summary:

The scope of this workshop was to link policy and methodological issues, with a special focus on long-term effects and discounting. Long-term health effects and life expectancy effects are largely neglected in current evaluations of policy choices, since little evidence is available, and the methodologies for valuation are not yet stabilised. The workshop was divided into four parts, two of which were devoted to policy questions (general policy issues and a special issue on prevention) and the other two to methodological questions (discounting of future benefits and costs, and the role of 'context' in contingent valuation applied to health).

Health and life benefits & costs induced by environmental quality are of growing concern in most European countries. There is an increasing body of evidence showing that air pollution (e.g. particulates, CO2, SO2, ozone) significantly affects various health indicators in the short run (e.g. bronchitis, asthma) with large social and private costs (e.g. workdays lost due to illness, increased costs of medical care, willingness-to-pay to avoid illness episodes). Today, air pollution is the best-documented case, but it is suspected that other types of ambient pollution (e.g. soil and plant contamination, groundwater pollution) also significantly affect human health, with large potential economic damages. Evidence available for the case of air pollution is essentially based on calculating short-term costs due to high episodes of pollution. Chronic morbidity and reduced life expectancy caused by background pollution are even less known but should become of central concern in the near future.

Air pollution is of particular interest for the valuation of climate change policy options. Until recently models designed to predict the effects of greenhouse-mitigating strategies have not taken into account the indirect or ancillary effects of reductions in greenhouse gas emissions. There are, however, obvious local benefits. For example, the reduction of CO2 emissionssimultaneously reduces other emissions such as SO2, particulates and NOX. There maytherefore be important local benefits in reducing transport sector emissions in urban areas, especially in large populated cities of developing countries (e.g. Mexico, Delhi, Bombay) but also in developed countries (e.g. Tokyo, New York, Los Angeles, Paris, London, Athens).

The global warming issue confers a special interest on methodologies for measuring health and mortality costs induced by environmental quality. This particular field of environmental valuation raises many issues, ranging from basic ethical and philosophical questions to very practical issues, such as the choice of the relevant indicators or the appropriate questionnaire design for estimating private costs of symptom days, as well as policy issues.

POLICY ISSUES

Michael Jones-Lee argued in favour of willingness-to-pay (WTP)based monetary value estimates of safety in the context of air pollution and the use of such values in the appraisal of proposed air pollution abatement schemes. In view of the relative lack of empirical work on the estimation of such values, particularly in Europe, he focused on work that has recently been undertaken to estimate the Value of Preventing one statistical Fatality (VPF) in the contexts or transport (road and rail) and fires (domestic and public). Professor Jones-Lee summarised various suggestions that have been offered concerning the way in which a WTP-based VPF for air pollution risks might be expected to relate to the VPFs in other contexts. Finally, he outlined what he regards as the most promising route to the empirical estimation of a WTP-based VPF (and corresponding values for other non-fatal adverse health effects) in the air pollution context. Several European countries have recently commissioned empirical work aimed at estimating WTP-based values of the health benefits of reductions in air pollution, which will be of importance for future progress in this area of research.

Olivier Chanel, Pascale Scapecci and Jean-Christophe Vergnaud addressed the economic implications of pollution reduction policies in presence of long-term effects. Policy design must take account of the likely substantial delay between the reduction of the level of pollution and obtaining the resulting benefits. This affects the valuation of the policy options and therefore their ranking with respect to various criteria. In particular, policies vary with respect to the number of years required to reach a given level of pollution. Results for various discount rates as well as for other characteristics were given for France.

LONG-TERM EFFECTS AND DISCOUNTING

Per-Olov Johansson addressed two questions: the WTP for risk reductions; and the discounting of health gains. He presented the results of a study that attempted to measure the value of an increased survival probability at high ages (for adult Swedes). He found that this valuation is strongly correlated with the expected quality of life at advanced age. Furthermore, the (maximum) insurance premium the average person is willing to pay for a programme which increases the survival probability is less than $1,500. The WTP increases with a person's age, but at a low rate. The implied average marginal rate of time preference is about one percent.

The issue of valuing life expectancy was further explored by Ingemar Leksell; in particular, the question of the appropriate discount rate. For air pollution the choice of discount rate is particularly relevant because there may be a long time lag between exposure to and effects from air pollution. There is a considerable debate about the choice of the social discount rate, some economists recommending a high discount rate (i.e. more than 10 percent) for health effects while others recommend to set it equal to zero for ethical reasons.

Olvar Bergland proposed to go further into the issue of discounting, by examining the behavioural foundations postulated by economic theory. The standard exponential discounting method is usually compared to the more satisfactory hyperbolic discounting method. It is well known, however, that hyperbolic discounting contradicts dynamic consistency, a fundamental property for rational inter-temporal preferences. Assumptions about inter-temporal preferences have dramatic consequences upon the way future health benefits and costs are discounted.

PREVENTION ISSUES

Ari Rabl put the valuation issue into the perspective of intergenerational equity. Taking the example of the intergenerational damages caused by nuclear power (cancers), he showed that in order to obtain intergenerational equity with respect to the costs of nuclear power, one must follow a two-step discount procedure. The argument starts with the premise that the purpose of the intergenerational discount rate is to determine the amount of money to be placed into a fund to compensate future generations for any damage imposed by the present generation. Distinguishing the two contributions to the interest earned, namely growth of the economy and preference for present over future consumption, he argued that only the growth component can be considered a contribution by the first generation; the remainder will have been paid by future generations. The two-step discount procedure takes the conventional social discount rate for the first generation and the growth rate of GDP per capita beyond (assuming stabilisation of world population). This argument applies only to costs, not benefits, because compensation is possible only from the present to the future. He concluded that the accounting of intergenerational net benefits is a matter of altruism, not equity.

Olivier Chanel, Serge Masson, Pascale Scapecci and Jean-Christophe Vergnaud discussed relevant concepts on which evaluations of reductions of life expectancy due to air pollution should be based. There are two possible concepts: (i) the value of preventing a fatality (VPF), discussed previously by Michael Jones-Lee, and (ii) the value of increasing life duration. From a 'physical' point of view the total number of years lost or gained within the population is the same for both concepts. However, there might be large differences between valuations based on VPF and valuations based on increased life duration, essentially because discounting is taken into account in a different manner. The presentation concluded that, based on French data, both approaches provide comparable results.

Philippe Quenel presented the principles underlying the French epidemiological monitoring system. He also proposed a clarification of the terminology used by epidemiologists, and showed how short-term effects relate to long-term effects. According to Quenel, high pollution episodes only marginally affect the population's health indicators. Therefore, preventive policies should be tailored to reduce background pollution, i.e. mean concentration. However, average concentration may be strongly affected by the frequency of high episodes, and therefore preventive policies should also reduce these episodes.

INFORMATION AND CONTINGENT VALUATION: WHICH IS THE RELEVANT DESIGN?

Two papers, one by Ståle Navrud & Richard Ready, and the other by Anne Rozan & Marc Willinger addressed the same methodological question about valuation but drawing opposite conclusions. Most valuation studies linking air pollution to health indicators are based on contingent valuation studies that ask, for example, for WTP for reduced symptom days. The question has been raised how and if the respondent's WTP is affected by the information provided about the origin of these symptom days, and whether or not such information should be provided by the questionnaire if it does have an influence.

Ståle Navrud and Richard Ready showed that the information about the cause has a negligible influence on stated WTP. Their results are based on a five European country survey, which applied similar contingent valuation techniques in each country. A split-sample experiment tested whether the cause of the episodes influenced the values generated. A similar test was also carried out for the ordering of the questions. It is common practice in health valuation questionnaires to ask several valuations questions to the same person. This has been sometime criticised because it might induce uncontrolled ordering effects on valuations. Using the same samples, the authors found that tests based on parametric regressions show little evidence that episode ordering influences stated WTP values. These results seem to suggest that the researcher need not provide information about the origin of the symptoms and that the respondents are able to evaluate each symptom in isolation, allowing for questionnaires with multiple valuation questions.

Anne Rozan and Marc Willinger studied the impact of providing information about the cause on the basis both of a within-subject test and a between-subject test. Their results show that additional information about the cause significantly affected the respondents' WTP (between 20 and 50 percent increase or decrease). Such large observed differences suggest that questionnaire design becomes crucial, because an incorrect choice may lead to under-provision or over-provision of pollution abatement, and an incorrect assessment of the social costs. The authors argue in favour of a questionnaire including the information about the cause, since the researcher will have better-focused values. There is a risk of embedding which cannot be ignored.

OUTCOME OF GROUP DISCUSSIONS

The final session of the workshop was organised into group discussions about policy and methodology questions, some of which had been raised during the workshop. In particular, the question of the relevance of individual preferences for designing policies for reducing health risks due to environmental pollution emerged:

• Should these matters be left to experts?
• How can valuation data be used for public decision-making?
• Are the valuations transferable across sites and over time?
• How do we choose among alternative policy options?

These questions are linked to those raised in the introduction about the role of health impact valuation studies, especially with respect to local effects of greenhouse mitigating strategies. One of the central conclusions was that long-term benefits of environmental quality on health, especially those involving several generations, are probably the most important. These benefits are highly uncertain today, not only because of the lack of data, but also because of methodological weaknesses for the assessment of long-term costs and benefits.

Contributions

Summary

Participants

Workshop Participants:

Contributions

Summary

Participants

Contact Details: