Værdisætning af pesticidanvendelsens natur- og miljøeffekter

Economic Valuation of the Effects of Pesticide Use on Nature and the Environment

Summary

In spite of extensive literature on economic valuation, only few studies on changed pesticide use exist. This may be explained in two ways: firstly whether pesticides have been a policy issue in countries with a tradition to performing economic valuation studies, and secondly the complex preferences associated with preferences towards pesticides.

Economic valuation studies can contribute to environmental policy by providing a consistent monetary measurement of the value of different environmental goods. Thus, the goods can be compared with each other, and with marketed goods in cost-benefit and cost-efficiency analyses.

Valuation studies often distinguish between studies based on hypothetical markets or observed markets. By presenting a number of respondents with hypothetical scenarios and asking them to express their willingness to pay (or accept) the specified changes or rank the specified scenarios, data are collected for hypothetical markets. As a consequence of the hypothetical approach, all types of goods can be encapsulated in the studies as well as both use and non-use values covering recreational to existence values. A number of problems are related to the hypothetical methods e.g. whether respondents respect their budget constraint when expressing their preferences.

When relevant relationships exist between environmental goods and existing markets, observed market data can be used. One example could be the demand for goods produced with and without pesticides, or differences in the price of hunting permits on farms with or without pesticide-free buffer zones. Only use values are reflected in valuation studies based on observed market data. For the purpose of testing the methodological validity of valuation studies, combined use of hypothetical and observed marked data can be recommended.

Valuation studies of biodiversity require a description of relevant indicators for the goods to be valued. Looking at Danish natural science studies that can provide such indicators based on driving forces/effect models, these are scarce. Thus we discuss if such a quantitative relationship is necessary for carrying out valuation studies, and two approaches have been drawn up: the “ideal” and the “pragmatic”. With the first approach the description of the biodiversity effects meets the requirements of cognitive unambiguous indicators, as well as the requirement for a direct reference to the underlying natural science studies. With the second approach, a certain amount of best judgement is used when quantifying the effects on the indicators. General indicators understood unambiguously by the respondents are used (the number of birds, quality of drinking water, etc.) but they do not necessarily refer directly to natural science studies.

If the “ideal” approach should be used, the natural science studies should reflect the ecological effects (in casu) of pesticide use under actual Danish circumstances. Insofar as the effects of pesticide use can be isolated based on the studies, their relevance will increase. The literature survey has focused on narrowing down the number of Danish studies that fulfil these requirements, and as mentioned only few seem to do so. The “pragmatic” approach does not require the same accuracy of the natural science descriptions. This approach is typically found in international valuation studies and allows for a broad field of application. Based on these findings the considerations of a stringent driving forces/effect-modelling do not necessarily limit the possibilities for performing valuation studies of the effects of pesticide use in a Danish context, as the “pragmatic” approach offers a useful alternative. Nevertheless, the use of the results in actual policies will be strengthened if a driving forces/effect-model exists, as the required policy measures are implicitly included.

A central issue using hypothetical market data is the design of the questionnaire. This is because the primary data for estimating the preferences are obtained by asking a representative selection of respondents to complete the questionnaire. In the case study performed in this project the questionnaire appeared to work well, as we obtained a high response rate and the respondents generally accepted the trade off between biodiversity and their consumption of food. However, it is necessary to investigate how respondents associate to the relative abstract indicators of biodiversity and the limitations of respondents’ budget constraint in order to improve the design of valuation surveys. This should be done in actual valuation studies by testing different designs of questionnaires in both focus group interviews and in full-scale surveys. Also the effects on the expressed preferences of including the term pesticides in the introduction text should be tested. This is motivated by the findings in several studies of risk perception, which indicate that pesticides are perceived as “stigmatising” as they are often considered as undesirable, even though they may actually have advantages. The occurrence of such preferences may have implications for the interpretation of the results and should be looked into further.

The statistical analysis of valuation data is subject to a large number of methodological econometric studies in the international literature. A general result is that results from valuation studies may be subject to considerable methodological uncertainty. Even if the results from a valuation study may be of methodological (research) relevance, the use of the results in practical policy analysis may be questionable. Therefore, proper validation in terms of sensitivity analysis and comparison to other relevant studies are crucial. Also validation by using alternative data sources should be done when possible.

In our case study encompassing valuation of the effects of pesticide-free buffer zones along field margins we found respondents willing to accept an increase in the price of bread of DKK 0.57 (4 percent) if the survival of partridge chickens increased by 10 percentage points. Similarly, respondents accepted an increase in the price of bread of DKK 0.07 (0.5 percent) if the number of wild plants increased by 10 percentage points. The results are statistically significant, but this does not implicitly indicate that they can be used for aggregate policy analysis, as the study was based on only 250 respondents. To demonstrate how the results could be used for calculating the welfare economic result of implementing pesticide-free buffer zones along field margins in a national context, the estimated benefits are compared to the estimated loss in income from agricultural cereal production. The benefit estimates show a willingness to pay for an increase of 10 percentage points in the biodiversity indicators, corresponding to an 8 percent increase in household outlays on bread. This equals DKK 245 per year for an average Danish household adding up to a total of DKK 690 million per year. The aggregate benefit should be compared to a loss in income from cereal production from implementing pesticide-free buffer zones along fields cropped with cereals of DKK 240 million per year.

We stress that the calculation above is an example, and we do not advise to use the result in policy recommendations without further analysis, e.g. by expanding the number of respondents.

In general we recommend that results from valuation studies should be used with caution in policy analysis because of the methodological uncertainty. Uncertainty is known for all empirically based science and is not determinative for the scientific validity of the results, but for practical policy analysis it will often be of importance. One aspect, which should be stressed when interpreting results from valuation studies, is if the context of the study is compatible with the actual policy issue. This may be especially important in studies related to pesticides, as effects of pesticide use and preferences seem to show considerable variations between countries. When using benefit transfer it is therefore recommended to transfer benefit functions instead of simply the price estimates and to calibrate the functions with site-specific data. This approach may, however, be as resource demanding as performing an actual valuation study. Also the limitations of valuation studies should be observed, as some policy issues (e.g. extinction of species) may be too controversial to be analysed in this context.

Based on our findings we conclude that economic valuation studies of the effects of pesticide use can be performed based on the current knowledge and methods. However, there is a need for further empirical work with respect to validating study methodology and price estimates in order to discuss if the results can provide a meaningful input to policy analysis. First, the effects of pesticide use are complex and therefore difficult to present to lay people in an economically rational context. Secondly, our case study indicates that respondents find it difficult to distinguish between the role as “rational consumer” and “political citizen”, and that their willingness to pay was influenced when the word pesticides was introduced in the questionnaire. Therefore the survey material, i.e. interview form, introduction, questionnaire and follow up questions, should be designed carefully so that the respondents understanding of the problem and acceptance of their budget constraint can be tested.