Oil Spills

TESTING AND CALIBRATING

THE MEASUREMENT OF NONMARKET VALUES FOR OIL SPILLS VIA THE CONTINGENT VALUATION METHOD

Principal Investigators: W. Michael Hanemann (UCB) and Jon A. Krosnick (Stanford University)

Summary of Research

Data on economic valuation are needed inputs in making policy and management decisions for coastal and ocean resources. Some of the economic aspects involve commercial activities such as fishing. Others involve what are known as non-market benefits, such as private recreation and aesthetics. For recreation and what are called use values, an economic technique known as the travel cost (TC) method can be used. For other aspects -- aesthetics, concern for wildlife and what are known as non-use values -- the approach used is to interview people directly and elicit from them an estimate of what they would be willing to pay (WTP) to prevent damage to these resources. This is known as the contingent valuation (CV) method. The 1990 Oil Spill Act (OPA) calls for measurement of non-use values in analyzing the damages from oil spills, and CV has come to be seen as the key way to collect this information.

The aim of this research is to extend and improve the contingent valuation (CV) method with particular reference to valuing California coastal and estuarine resources. The specific objective is to investigate how the results of CV surveys are affected by the form of the valuation question -- discrete response ("referendum") versus continuous-response -- and to analyze the implications for an extensive CV survey on oil spills in California that is currently in the field using the referendum format.

The CV method was first proposed by Ciriacy-Wantrup as a means of non-market non-valuation in the specific context of protection against soil erosion. Ciriacy-Wantrup argued that the appropriate analogy for this was not private choices as expressed through individual purchases of normal market commodities but rather collective choice through voting on the provision of a public good. He saw the CV survey as a surrogate for voting -- one would approach citizens to see if the item were worth what it cost to provide. The first CV survey was conducted by Davis to value recreation in the Maine woods. A decade later, Randall et al conducted the first major non-use value CV study on air quality and visibility in the Four Corners area. By the late 1970's, CV studies were being commonly performed to evaluate environmental and other non-market commodities. The 1980's saw several important methodological developments, including collaboration between economists and other social scientists with expertise in survey research. Two landmarks were an EPA conference in Palo Alto in 1984 that brought together leading CV practitioners, other distinguished economists and psychologists to assess the then state-of-the-art and the publication of what has become the standard reference on CV, Mitchell and Carson, placing it in the broader context of economics, sociology, psychology, market research and political science. This was a collaboration between an environmental sociologist, Mitchell, and an environmental economist, Carson, who had received his Ph.D under Hanemann at U.C. Berkeley.

One of the changes in CV methodology that occurred during the 1980's was a shift in question format. The early CV studies had used an open-ended question along the lines of "What is the most that you would be willing to pay for the item?" The alternative is to use an open-ended format: "If this item cost you $x, would you be willing to pay that much?" Different amounts, x, are presented to different respondents; their responses trace out a bid-response function that show the percent willing to pay an amount as a function of the amount. From this one can readily deduce the mean or median willingness to pay (WTP) in the population surveyed. If there were a referendum on the item, as with propositions in the California Ballot, the median would correspond to passage with majority voting. Hence, the approach is known as the referendum method. This was first used by Bishop and Heberlein in a study of duck hunting in Wisconsin. It was popularized by Hanemann who showed how the responses could be interpreted in terms of an economic model of utility maximization, and who developed the formulas to calculate estimates of mean or median WTP from survey responses. Since the late 1980's, it has become the standard approach for many CV studies. It has been extended in several ways by Hanemann, Loomis and Kanninen and Cooper and Hanemann to increase its statistical power. The evidence shows that subjects generally find it far easier to respond to the closed-ended valuation question, and often more meaningful -- it is, indeed, like voting in a referendum which, even in rare practise, seems like a natural thing. When NOAA recently created a Blue Ribbon Panel to advise it on the use of CV for measuring non-use values, the panel took this view and endorsed the referendum approach. However, the Panel indicated the need for further research on the implementation and calibration of the referendum approach, and that is the focus of the current proposal.

The empirical application of this study builds on existing CV research on California oil spills by Hanemann and Krosnick. In 1988 Hanemann was asked by the California Attorney General Office to assist it in seeking natural resource damages following an oil spill at the Shell Oil refinery in Martinez, CA. This work led to a $20 million settlement in 1989, the largest payment ever made in the US up to that time for a natural resource damages claim. As part of the settlement, $645,000 was set aside for a CV study to be directed by Hanemann aimed at valuing damages that might occur in a future oil spill in California (this sum was recently raised to $795,000 to cover additional survey costs). State and federal officials involved in the Shell spill felt that having such information available on the shelf would be very helpful for future planning and policy analysis as well as damage assessment. The study is being conducted under a contract between the California Attorney General Office and Hanemann. It involves a team of economists, sociologists and psychologists at several universities. Jon Krosnick is a key member of this team and was actively involved in developing the final survey instruments. The study is being conducted as academic research, not for litigation, and the results will be published in academic journals. Economists with OSPR in the California Department of Fish & Games and with the NOAA Oil Spill Office in the US Department of Commerce have served as peer reviewers for the California Attorney General Office, together with academic peer reviewers. The survey involves in-person interviews with a statewide sample of approximately 1,000 households conducted by a leading national survey company, Westat Inc, under sub-contract to Professor Hanemann.

During the course of developing the survey instrument, several prototypical oil spill scenarios were developed as candidates for evaluation. In the end, it was found feasible to include more than one of the scenarios in an interview while remaining within the time parameters specified in Westat's contract. The intention is to focus on one spill scenario for now, with the idea that OSPR or other agencies could subsequently replicate the survey with a different oil spill scenario if so desired. The basic structure of the survey instrument and the survey logistics would already have been developed and proven up in the current survey, thereby making subsequent replication both simpler and less costly. Thus, the current survey is viewed as a cornerstone for future state and federal work on systematic valuation of the effects of oil spills in California.

While the Shell settlement has funded an extensive survey, it provides no funds for methodology development or research on the effects of alternative survey designs. That is the focus of the current proposal. The aim is to investigate an important methodological issue which is currently the subject of much discussion among practitioners, namely the effects of question format and structure. The purpose is to determine how the results of the Shell-funded CV survey should be interpreted or adjusted in order to provide a robust estimate of WTP values for oil spills in California. It is anticipated that the results of this research will be of general benefit for CV research in addition to providing specific information for California coastal and marine resource planners.