ࡱ; uv  !"#$%&'()*+,-./0123456789:;<=>?@ABCDEFGHIJKLMNOPQRSTUVWXYZ[\]^_`abcdefghijklmnopqrstzxwy{|}~Root Entry F8޺@CompObjbWordDocument-ObjectPool5t޺5t޺ 4@   !"#$%',./012346:;<=>?@ABCD FMicrosoft Word 6.0 Document MSWordDocWord.Document.6; ࡱ; LE lࡱ; E:  .1  &` & MathTypeSymbol-2 |ܥe= e-['   ! ! &&&((((((":((,E(:("))))))+ + + +)2++,%-Ty-t, &)v{)))),)~#j#L)()))) &) &)+*&&&!X"&4$$&)+))Preliminary Draft (6-29-95) Please Do Not Cite Copyright 1995 Valuing Beach Renourishment: Is it Preservation? Rebecca P. Judge, Laura Osborne, and V. Kerry Smith* June, 1995 Abstract: A proposed plan to preserve beach and beach access along the Cape Hatteras National Seashore through beach re-nourishment offers the opportunity to apply contingent valuation techniques to explore the implications of three sources of preference heterogeneity for measures of peoples willingness to pay. Preferences are modeled as functions of: (a) attitudes, socio-economic and demographic characteristics; (b) past experience with the Cape Hatteras area; (c) anticipated future use; and (d) knowledge about the problem and proposed plans to address it. We find that while income and environmentalist attitudes have limited indirect effects on respondent willingness to pay for beach re-nourishment, past and future use of the resource, as well as respondent knowledge about the re-nourishment plan, directly affect willingness to pay. I. Introduction The last election resulted in a dramatic change in the congressional perspective on U.S. environmental policy. Those favoring large transformations in the legislation defining these policies argue that current mandates are inconsistent with most peoples preferences. That is, protection of environmental resources is only one of a number of objectives people want policy to address. The task of reconciling conflicting objectives in the presence of scarce resources is a central motivation for using economic analysis to inform policy. Most past uses of these methods focused on policy questions involving marketed goods and services. Beginning in the seventies, there was growing acceptance of indirect approaches to non-market valuation. While capable of providing measures of the economic value of some environmental resources such as recreation sites, these approaches generally focused on estimates for an average, or a representative, individual. Over the past decade the economic approach to measuring peoples valuation of environmental resources has given increasing attention to contingent valuation. This change has broadened the range of environmental policies that can be evaluated. It has also focused attention on how an individuals socio-economic characteristics, knowledge, attitudes, and constraints influence his or her willingness to pay for changes in these resources. To date few studies have attempted to consider cases where a resource with distinctive features that cause it to mean different things to different people is to be transformed.  In such cases we might expect that people would differ in their willingness to pay for any change. Indeed, some people might pay to avoid any changes. This is important because public policy decisions increasingly involve complex resource management questions that affect people with heterogeneous preferences for how environmental resources should be used. We have selected one such issue beach re-nourishment for the Cape Hatteras National Seashore, a national park located along North Carolinas barrier islands. Initially beach re-nourishment was an integral part of the management plans for the park. The last two decades have seen a different regime, one which emphasizes natural processes over beach protection. Both approaches have advantages and disadvantages for different people and as a result we expect they would be valued differently. Using a contingent valuation survey of a sample of North Carolina households we evaluate the feasibility of disentangling these diverse preferences and what the results imply for this type of policy. While beach re-nourishment is a fairly specialized situation, it is also a good example of the difficulties facing resource managers seeking to maintain or enhance environmental resources held in public trust.. By intervening to preserve a natural resource, as it happens to exist at one time, its natural state is necessarily transformed. While certain individuals may value the stream of benefits generated by the preserved resource, others will consider the changes to such resources part of a natural process. As a result, they would opt for continuation of those natural processes. This course of action may well be preferred, even in situations that threaten the existence of the resource itself. By choosing to study beach re-nourishment, we hope to understand how active attempts at resource preservation are viewed by various segments of the public. II. Policy Context Erosion causes North Carolina barrier island beaches to recede an average of four feet each year. The areas experiencing the highest rates of beach erosion are along the Cape Hatteras National Seashore. While beach erosion is a natural process, it poses some particular problems in the Cape Hatteras area. The only land access to the Cape Hatteras National Seashore and to village enclaves surrounded by the National Seashore is NC12, a state highway which is itself threatened by beach erosion at several locations. Furthermore, the Cape Hatteras lighthouse, a National Historic Monument, is also in danger of being damaged or destroyed as the shoreline around it erodes. The state of North Carolina, together with the National Park Service, has had a long history of attempts at erosion control along the North Carolina barrier islands, dating back to 1936 when Congress passed legislation to provide assistance for the protection of the beaches along the shores of the United States.  That same year, the National Park Service commenced work on what was to be the restoration of the Outer Banks. The following year, Congress passed legislation authorizing the creation of the Cape Hatteras National Seashore,  but local opposition delayed the parks creation for another fifteen years. The park was finally created in 1952, after the National Park Service agreed: (1) to reduce the size of the strip of ocean beach included in the parks boundaries to 500 feet in communities included in the Cape Hatteras National Seashore; and (2) to protect and control the sand dunes, to reestablish them when necessary, and hold them to protect the communities from the intrusion of the ocean. For the next twenty years, the National Park Service pursued a strategy of continued investment in beach stabilization within the park, with cumulative costs totaling over $20 million (1971 dollars).  However, the rising costs, the short-term nature of any stabilization effort, the increased private development subsidized by public expenditures on beach stabilization, and a growing sentiment that beach stabilization would itself jeopardize the island caused the National Park Service to abandon its commitment to erosion control at Cape Hatteras. Since 1974, the National Park Service has followed a policy of nonintervention. Areas barren of vegetation are replanted, but dunes are not rebuilt. There have been no beach re-nourishment activities within the park for the last twenty years. A policy of non-intervention is itself controversial. The cessation of the federally funded beach stabilization effort has shifted costs to the private sector and to the state of North Carolina. Private property, developed and acquired under the federal policy of intervention and stabilization, is now threatened. Recreational opportunities may be threatened. The cost to the state of maintaining NC12, the sole land access to the National Seashore, has increased as erosion slowly brings the ocean closer to the highway itself. The road has had to be relocated further inland, and sections of it are frequently washed out as the result of storms. The cumulative costs of the noninterventionist policy have caused the state to re-examine its position on beach re-nourishment. The state of North Carolina, together with the U.S. Army Corps of Engineers, is presently considering re-instating the pro-active beach stabilization policies of the 1950s and 1960s. A major beach re-nourishment project has been proposed to expand the Dare County beaches at Kitty Hawk, Kill Devil Hills, and Nags Head. The project, which has an estimated initial cost of $32.5 million, calls for the placement of 7 million cubic yards of sand along ten miles of beach on Cape Haterras, widening the beach by 50 feet. In addition to the initial costs, the Corps estimates that $4 million will be needed each year to replace sand lost from the re-nourished beach. While the federal government may contribute as much as 65 percent of the funds needed, the remainder must come from the state itself. Beach re-nourishment dredges sand from one location and deposits it in another, expanding the size of a beach area. It is a temporary response to beach erosion that will require periodic additions if the size of the beach is to be maintained. The exact nature of these replacements will depend on the pattern of erosion in each area. Use of sand to re-build or to maintain beach widths within the Cape Hatteras National Seashore poses special problems for other reasons. The area was initially included in the national park system as a primitive wilderness. The enabling legislation recognized that barrier islands, such as the sections of the three islands included in the Cape Hatteras National Seashore, move and shift in response to wind and waves. The National Park Services mandate is to protect these processes. The private nature of some of the benefits generated by beach re-nourishment creates controversy when the public is expected to pay for the undertaking. Beach re-nourishment protects property values and the industries which provide goods and services to those who recreate and live along a re-nourished beach. These private benefits are responsible for some of the controversy surrounding re-nourishment activities. I dont know why we think the government -- the taxpayers -- have the responsibility to help people who are so stupid as to build right on the ocean, argues a North Carolinian opposed to a state-subsidized beach re-nourishment plan.  Of course, not all of the benefits of beach re-nourishment are strictly private. The major output of a beach re-nourishment project is the beach itself, which, along the Cape Hatteras National Seashore, is publicly owned. The beach may in turn generate other types of public or quasi-public goods, from habitat preservation to historic landmark preservation such as protection of the Cape Hatteras Lighthouse. These goods could generate use and non-use values. III. Beach Re-nourishment as an Economic Commodity A. Background Measures of the economic value people place on anything arise from their choices. These estimates are derived by combining peoples selections with the properties economic models attribute to the tradeoff inherent in each choice situation. For simple decisions the circumstances of choice will include a clearly defined object of choice and a set of consequences that describe an individuals responsibilities with different decisions. For example, if the consequences specify that a person making a defined total payment will receive the object of choice, then that choice allows one to conclude, for those agreeing to the selection, that the payment must be a lower bound for their willingness to pay (WTP). Those who decide not to select the item also tell the observer something. The proposed payment is, in their case, an upper bound for their WTP. This recognition is so straightforward that it can easily be overlooked in the complexity of modern applied welfare economics.  For the most part, the focus of this research seeks to adjust for specific features of the circumstances of choice. Thus, the conditions required to recover measures of Hicksian consumer surplus from Marshallian demand functions are not relevant. These are the assumptions required to connect the tradeoff that can be observed in market choices with what is needed for measuring WTP (see Hausman [1981], Bockstael and McConnell [1993], Hausman and Newey [forthcoming]).  Contingent valuation offers the opportunity to simplify this process by allowing the analyst (within the limits of what will be credible to respondents) to specify the circumstances of each proposed choice.  This is important because for most environmental resources the object of choice is generally a complex, multi-dimensional, resource that may be valued by people for diverse reasons. An examination of the conceptual literature on the use/nonuse value debate offers clear examples of this diversity, both in motives people might have for deciding to maintain specific environmental resources and in analysts conclusions about their relevance for policy.  Unfortunately, most of this complexity and potential for heterogeneity in peoples responses to environmental policies must be assumed away in policy analysis. These evaluations have tended to rely on either indirect methods for valuation (e.g. travel cost, hedonic property value, or averting behavior) or simple benefit transfers using past contingent valuation or indirect studies involving other resources.  In this study, the contingent valuation object of choice was chosen to allow investigation of the implications of four sources of preference heterogeneity for measures of peoples willingness to pay. These include the differences in respondents:(a) attitudes and related socio-economic and demographic characteristics; (b) past experiences with the Cape Hatteras area; (c) anticipations for future patterns of use; and (d) knowledge about the problem and proposed plans to address it. A telephone-mail-telephone survey was designed to consider North Carolina residents choices (and the economic values they implied) for coastal issues that reflected the conflicting effects public policy can have in mitigating the impacts of coastal development. That is, policies that attempt to preserve coastal resources or protect them from degradation can promote development. As a result, an undesirable indirect effect of those policies can be a reduction in the quality of resources they seek to enhance. Two public programs which would, in the short-term, reduce development pressure on the coast -- beach re-nourishment and enhanced waste treatment -- were selected and presented in detail to survey respondents. The beach re-nourishment project, as described, provides the opportunity to investigate the nature of peoples preferences for a program that has conflicting implications. As noted at the outset, those individuals interested in maintaining the National Seashore in its original natural state may well prefer that the process of erosion be allowed to proceed. However, this choice would reduce access for recreation due to the effects of increased erosion on NC12, the only land access route. A non-interventionist choice would also limit the states role in assisting private citizens facing the consequences of natural hazards (i.e. residents with private property in the village enclaves also on the island). On the other hand, actions to maintain access could promote future development, potentially increasing the size of the affected population within and around the area by conveying an expectation that the state would maintain land access indefinitely through continuing re-nourishment programs. Because of these conflicting interpretations of the program offered in our contingent valuation question, the object of choice is ideally suited to identify how differences in respondents preferences and experience influence their choices. These differences also can be related to their use and nonuse motives for this resource. As originally conceived, the beach re-nourishment project described to the survey respondents was entirely hypothetical. It was described as a plan under consideration by the state. During the course of the our survey activities, however, the state and U.S. Army Corps of Engineers announced that they were indeed considering a beach re-nourishment plan for the Cape Hatteras area that closely paralleled our descriptions. The plan was presented in newspapers across the state. For those aware of the newspaper accounts, these articles reinforced descriptions provided in the materials that were distributed as part of the survey. An individuals choice taken alone does not provide a basis for estimating use and nonuse values separately. To do so requires additional assumptions. The logic underlying this conclusion requires a somewhat more formal description of the choice process. As we noted, the model underlying the construction of monetary measures of economic value describes choice as an outcome resulting from a comparison of a tradeoff. That is, when an individual is offered a beach re-nourishment program with a specified increase in his or her taxes of t dollars, then in terms of the indirect utility functions, the decision is based on a comparison of two stateswith and without the program. This is given in equation (1) below: (1) EMBED Equation.2  where V() = indirect utility function m = income p = vector of prices of marketed goods and services bn = indicates the role of the beach nourishment program to the individual If the term on the left side of the equation exceeds the right, the individual would support the program and t provides a lower bound for his or her willingness to pay. If not, the decision identifies t as an upper bound for bn. If we use the estimates of this choice relationship to define  EMBED Equation.2  as the value that would equate both sides, then it is possible to describe how the willingness to pay varies with m, p, as well as the demographic and attitudinal variables found to influence these decisions. However, nothing in that process allows separation of use and nonuse values. To do so requires that the term bn be specified in more detail. It must be linked in some way to the other actions (i.e. purchases) an individual makes that would indicate a clear use of some aspect of the resource that would, in turn, be affected by beach nourishment. However, it is important to acknowledge that any separate measure of the use (or the nonuse) value from choices requires that this type of link be accepted as a maintained hypothesis. For our purposes, we propose to investigate the potential for use and nonuse values in a somewhat less restrictive way. We consider our survey respondents descriptive characteristics, classifying each according to those attributes we would expect to be associated with use values. With this classification, it is possible to evaluate how measures of the total willingness to pay vary with differences in both these characteristics and in peoples reported past behavior. While this does not measure separate use and nonuse values, it does offer suggestive evidence about their importance to stated choices. That is, by verifying a link between the factors in the theoretical literature as related to use and nonuse value and estimates of total willingness to pay, this process offers suggestive evidence of the influence of the use and nonuse dimensions of total value. B. Survey Design Between November 9, 1993 and January 14, 1994 a telephone-mail-telephone survey was conducted of households in North Carolina. The telephone interviews were conducted by professional interviewers using a random digit dial sample for North Carolina acquired from Survey Sampling Inc.  The initial interview was conducted with adult respondents (i.e. over 18 years of age) who identified themselves as having influence over the households budget. The interview was explained as a research project about issues in the news. Respondents were told at the outset that they could choose not to answer or indicate they did not have an opinion. The initial interview collected information about respondents opinions about a variety of issues including crime, improving education in North Carolinas primary and secondary schools, non-point-source pollution, commercial fishing, and attracting industry to North Carolina. It also collected demographic information about the respondent (and the household), and requested information about past visits to North Carolina beaches. Survey respondents were asked one contingent valuation question about a randomly assigned program to expand programs to enhance the appearance of highways or to use old tires in asphalt. At the close of the 11-minute interview, each respondent was asked if she or he would participate in a second interview about issues in managing North Carolinas coast. Respondents were told that they would each be sent a booklet describing these coastal issues. They were asked their names and addresses for the purposes of mailing the booklet, and assured that information would not be included in mailing lists or used in other ways. A program offering incentives for randomly selected respondents was explained. The response rate for the initial interview was 66%, for a total of 1012 initial survey participants. Among those who participated in the initial interview, 83% (836 respondents) agreed to a second interview. Of those who agreed to a second interview, 70% completed the interview after as many as eleven attempts to contact each respondent. This provided a sample of 580 responses. Both the first and second interview were conducted with the same person. Our survey questionnaire and information material were developed in an iterative process with focus groups (in Raleigh on June 3, 1993, and in Charlotte on August 11, 1993). The objective of these sessions was to evaluate the feasibility of presenting coastal issues within the context of other public programs. The sessions indicated that presenting these other programs elicited a wide range of reactions that distracted participants from both the specific features of program alternatives being presented and the coastal issues that were the focus of this research. Two conclusions were derived from these sessions. It appeared difficult to present choices among diverse public programs in a short telephone interview. Thus, our research was focused only on coastal quality issues. Equally important, it was clear that detailed information with photos would be important to respondents perceptions of the problems. To respond to these comments, a seven-page booklet (8 by 11 inch) was prepared with color and black-and-white photos depicting the effects of beach erosion on Cape Hatteras and the results of a re-nourishment program in Atlantic Beach.  A map was also included in the booklet which described erosion rates at different points along the North Carolina coast. This book was professionally printed in an attractive format. It was sent to all 836 individuals who agreed to participate in the second interview. During the second interview, virtually all of the respondents (i.e. 574 of the 580) indicated that they had read the booklet and either had it available for the second interview or remembered the descriptions. Multiple-choice questions about the booklets factual information were asked at the outset of the second interview. Eighty-two percent of the respondents provided answers for the beach erosion question that are consistent with the messages in the booklet. Table 1 compares the economic and demographic characteristics of the initial or Phase I sample and the follow-up sample with the characteristics of households in North Carolina based on the 1990 Census. This comparison is consistent with past experience with telephone surveys. The sample population is more highly educated, reports a greater household income than the North Carolina population at large, and is disproportionately female and white, relative to the states population. The telephone-mail-telephone (TMT) format serves to further accentuate this selection effect. While it is not possible to adjust for the selection effect inherent in the use of telephone surveys, the panel nature of the survey design permits testing of the selection effect associated with the TMT approach. Equation (2) below reports the probit estimate of a model seeking to explain the factors influencing whether respondents to the first interview would complete the second. The results reinforce the informal evidence in Table 1 comparing simple statistics. The numbers in parentheses below the estimated parameters are the asymptotic Z-statistics for the null hypothesis of no association. (2) Return = - 0.346 + 0.382 White (=1) + 0.131 Visit NC beaches (- 1.62) (3.44) (1.54) (=1) + 0.189 Self-Rated Strong + 0.229 Vote of Recent (1.97) Environmentalist (=1) (2.58) Referendum (=1) - 0.155 Would Vote for President Clintons - 0.277 Retired (=1) (- 1.82) Health Plan (=1) (- 1.89) + 0.004 Age - 0.223 Has Below High School Education (=1) (1.06) (- 1.69) - 0.128 Conditional Probability Favoring Environmental Projects (- 0.84) n=986  EMBED Equation.2 =0.041 With the exception of race, recent voting experience, and a self-rating as being very concerned about the environment, the other proposed determinants would not be judged as clear determinants of completing the second interview (labeled as RETURN for return interview). Nonetheless, the sign of their effect is consistent with both the existing literature and our a priori expectations from the qualitative research involved in developing the questionnaire. One variable in the model requires some further explanation. This is the variable described as the conditional probability of favoring environmental projects. Here we use the contingent valuation question from the initial survey as a gauge of the respondents support for environmental projects. As we noted earlier, this question randomly assigns one of two programs, poses a tax increase to undertake the program that has been described, and asks for a choice. Because the tax amounts varied across people, it is possible to estimate a choice model controlling for economic and demographic characteristics, the tax amount, and the specific environmental program assigned to each person. This term is the inverse Mills ratio from that modelthe conditional probability of favoring one of the two programs. While the term is not significant, it does imply a positive net effect on the likelihood of returning the survey. This outcome is consistent with our prior expectation that Phase I respondents who are willing to support environmental issues would be more likely to respond to a second interview about coastal issues. Before discussing the models and estimators using choice data to recover WTP estimates, we should note that several different specifications for the selection model were considered. Some of these models excluded the variable to gauge public support for these two environmental programs and included income and other attitudinal measures collected in the first interview. In most cases the results suggest the selection effect was not a significant factor in respondents decisions about a beach re-nourishment plan for the Cape Hatteras National Seashore. Only in some of the probit specifications was the selection effect found to be significant. The primary focus of our analysis is on the proposed plan for beach re-nourishment. The plan was described in two ways to those respondents completing the second survey. As we noted, the booklet included three photos, a map with the erosion rates and two and one-half pages of text describing the problem and the program. This description closed with a list of reasons for favoring and opposing the plan. In addition, the interviewer reminded respondents of this information and asked the following: The first plan described renourishing beaches. This plan involves adding more sand every four to six years to the beach areas most seriously affected by erosion. It is expected that the projects would be for places like Photo A, the view from a plane showing one place where NC Highway 12 is threatened by beach erosion. This would expand the beaches in those areas like the increase in the width of beach illustrated by the completed area of renouishment at Atlantic Beach in Photo C in the booklet. Each year different areas will have sand added based on how serious the erosion has been. To pay for these activities there would need to be an increase in North Carolina residents state income taxes. Because renourishment would be an ongoing process this would be added at the end of each year to the amount you pay now. It would be $_____ for your household. This is in addition to your current state income taxes and would need to be paid each year to maintain the program. Please keep in mind your current income and the things you now buy. If this renourishment plan to enlarge the beach areas where NC12 is threatened were on a statewide referendum and you could vote on it, would you vote for or against the plan? 1 For program 2 Against program 3 Dont know Respondents were assigned randomly one of seven possible prices for the beach re-nourishment activities, ranging from a $10 increase in state income taxes to $500 increase in state taxes each year. (Other price quotes included $25, $50, $100, $175, and $300.) Other than price, all respondents were given identical Phase II surveys. C. Estimation of Choice Functions Hanemann [1984] was the first to demonstrate the role for a theoretically consistent, random utility model in describing peoples choices to discrete (i.e. referendum style) contingent valuation questions. Initial research with this framework used conventional maximum likelihood estimators such as probit and logit. While these estimators, for linear and some nonlinear specifications, do allow the recovery of sufficient information to estimate willingness to pay, they do not constrain the estimates of WTP to be positive. There are a number of explanations that have been offered to describe why estimates of the conditional mean (or median) WTP might be negative. Often the tendency for probit (or logit) estimates of choice functions to yield these estimates is a reflection of responses that are inconsistent with the underlying probability structure. Probit and logit specifications imply thinner tails for the implied WTP distribution than has been found in several studies with the stated choices. These probability models are not sufficiently flexible to accommodate a thicker tailed distribution. To address this issue we have developed our analysis of the WTP estimates implied by peoples choices of the beach nourishment plans in two ways. The first of these uses the Turnbull [1976] non-parametric estimator of the WTP distribution implied by a set of choices. Proposed by Carson et al. [1994] as a robust, distribution-free method for analyzing discrete response contingent valuation questions, this approach imposes few additional restrictions on respondents stated choices for our proposed plan to re-nourish beaches in the Cape Hatteras National Seashore. By varying the subsamples used to estimate these WTP distributions it is possible to investigate the effects of respondent characteristics hypothesized to be associated with their preferences. The estimator treats the discrete responses as defining a censored random variable that estimates an upper bound for the WTP. In the case of single discrete choice questions, the probabilities for the upper bounds for WTP can be derived (as well as the implied cumulative distribution function) from the relative frequencies of no responses at each tax amount (see Turnbull [1976] and Haab and McConnell [1995]). To evaluate the influence of several independent variables on respondent choice functions we report both a probit model and a Weibull survival function to test the role of individual demographic and economic characteristics while controlling for the effects of other variables. III. Results Table 2 describes the results of Turnbull-estimated, lower-bound, mean willingness-to-pay for beach re-nourishment activities along the Cape Hatteras National Seashore. Note how the estimated willingness-to-pay varies with respondents experience with the resource. That is, the willingness to pay for beach re-nourishment among those who had never visited Cape Hatteras, $182.40, was significantly more than the willingness to pay for beach re-nourishment than among those who had visited the area, $162.64. This result is echoed in the Turnbull estimates among users; those who spent more than one day at the North Carolina beaches were willing to pay less than those who spent only one vacation day in the last two years at North Carolina beaches. While past use has a negative effect on estimated willingness to pay, future anticipated use has a positive effect on willingness to pay for beach re-nourishment. Among those who plan to visit Hatteras in the future, the lower bound estimated mean willingness to pay was $213.25, significantly greater than the $154.98 mean calculated from the subsample of individuals who indicated that they did not plan to visit Hatteras in the next two years. When individuals are grouped according to their past and anticipated future use as in Table 3, these results are even more pronounced. Those who have been to Hatteras, but have no plans to visit again have the lowest estimated willingness to pay for beach re-nourishment, at $131.53. Among those who have never been to Hatteras, and never plan to, estimated mean willingness to pay is $174.17. Those who have been to Hatteras, and plan to visit it again in the future have a Turnbull estimated mean willingness to pay of $177.71, while those who have never been to Hatteras, but anticipate visiting the island in the next two years have the highest estimated lower bound mean willingness to pay at $218.32. Experience with the resource seems to influence the estimated willingness to pay, and in this case, past experience with the resource appears to lower willingness to pay for beach re-nourishment. With the information available in our surveys, it is not possible to determine whether this result reflects a preference for letting nature take its course, as opposed to creating a manufactured beach, or whether it reflects an informed distaste for the National Seashore and the surrounding area as a result of a past recreational experience on the island. The results of the probit and Weibull survival models are shown in Table 4. Probit choice and Weibull survival models reflect a similar pattern in the determinants of the WTP estimates. In the probit case, the dependent variable takes on a 1 or 0 depending on whether or not the respondent voted for the beach re-nourishment project at the tax amount quoted by the interviewer. Table 4 includes two specifications for each model - a restricted form that includes income, a dummy variable for plans to visit Hatteras, a qualitative variable reporting each individuals evaluation of the effectiveness of the plan, and a variable measuring the number of articles relating to beach re-nourishment published in the respondents local newspapers around the time of the survey. The Weibull survival model is equivalent to modeling the log of WTP as a function of these variables, while the probit is consistent with a linear WTP function. We investigated probit models with the log of the tax amount and found comparable results to what is reported here. Our primary focus is on the consistency of the estimated effects with those developed using the non- parametric Turnbull estimates. Considering the most detailed specifications with probit and the Weibull specifications first, the probability that an individual would vote for beach re-nourishment is a negative function of the tax amount that each respondent is told he or she would pay for the re-nourishment program. The effect of past and anticipated future use of the resource is consistent with that observed in the Turnbull estimates of the lower bound willingness to pay. The probability of voting for beach re-nourishment is a positive function of an individuals plans to visit Cape Hatteras in the future. The negative sign on the dummy variable indicating past visits to Cape Hatteras echoes the same pattern observed in the Turnbull estimates; past experience with the resource lowers ones willingness-to-pay for beach re-nourishment. Distance from home to Cape Hatteras plays a significant role in this model only among respondents who have not visited Cape Haterras in the past. Within this group, distance reduces the probability of voting for beach re-nourishment. The effect of distance on the probability of voting for beach re-nourishment is insignificant among those who have visited Cape Hatteras in the past. This can be observed by looking at the combined effect of the estimated coefficient on the distance variable and the estimated coefficient on the variable created by multiplying distance to Nags Head times the dummy indicating past visits to the Cape Hatteras area. This composite variable equals the distance from home to Nags Head for those who have visited Cape Hatteras, and is equal to zero for those who have not visited Cape Hatteras. Comparable results can be found in the Weibull survival models. Income was dropped from this model because it was both insignificant and negative in its influence. However, we restricted the sample to those with reported incomes to be comparable to the probit model. The remaining terms in the Weibull allow for adjustment in the scale parameter of the distribution for heteroscedasticity. We investigated the selection effects (IMR) as a potential influence to the location (i.e., mean and median log (WTP)) and the scale parameters. It was not a significant influence. The term that is included is a qualitative variable for those reporting they had heard about beach re-nourishment plans for the area before receiving the survey. An interesting result concerns the impact of information on preferences as revealed through this survey. As mentioned earlier, our survey was inadvertently made more credible when the state and the U.S. Army Corps of Engineers announced at the outset of our field research that they were considering a plan to re-nourish parts of the beach along the Cape Hatteras National Seashore. The plan was described and debated in the media during the time that our survey was taking place. To evaluate the effect of this media attention on respondent preference, we created a variable measuring the count of the number of newspaper articles relating to beach re-nourishment that appeared in the respondents home newspaper between November of 1993 and the day the respondent completed the second survey. We can not distinguish between those who read the newspaper articles and those who did not, nor do we distinguish with respect to the general tone of the article (pro- or anti- beach re-nourishment). Despite these informational constraints, this count measure has a positive and significant effect on the probability an individual will vote for the proposed beach re-nourishment project.  Other respondent attributes which proved significant to our models included the educational level of the respondent, and whether or not the respondent indicated that he or she would vote for greater federal assistance to towns for maintaining beaches that experience heavy erosion due to storms. The lack of significance of income is a troublesome result in our simple and detailed models. The prices offered for beach re-nourishment ($10 - $500) seem non-trivial. Furthermore, income has been found to be a significant factor in other studies of the value of beach re-nourishment.  Income was, however, found to have a positive and significant effect on the probability that one would favor environmental projects. Income therefore appears, through the variable equal to the conditional probability favoring environmental projects, in the model describing the probability of an individual taking part in the second survey (see equation (2), above). Thus, income does enter models that include IMR, such as our detailed probit specification, through the inverse Mills ratio from the model of the probability of an individual taking part in the second survey. We conclude from this that the income effect is related in some yet-to-be-determined way to environmental preferences. Other results not reported here include an attempt to consider attitudes in support of environmental protection. We created a variable measuring the respondents identification as an environmentalist, and the lack of significance of this variable lends support to the hypothesis that beach re-nourishment is a multi-faceted environmental good, one that is not clearly an act of preservation or development. People who identify themselves as environmentalists may favor beach re-nourishment if they see it as a means of preserving unique habitat and recreational opportunities; they may vote against beach re-nourishment if they are interested in preserving the dynamic processes of beach creation and erosion along the barrier islands. IV. Implications Our survey results confirm the importance of preference heterogeneity in evaluating complex changes to significant environmental resources. Using the Turnbull lower bound mean, a non-parametric estimator for willingness to pay, we found that future plans and past experience in using the Cape Hatteras National Seashore significantly change WTP for programs to use beach re-nourishment to maintain the area. These conclusions were confirmed with multivariate parametric models, including both the probit models and the Weibull survival functions based on respondents stated preferences for the beach re-nourishment plan. These models tested the influence of economic characteristics, attitudes, knowledge, and ease of access to the Cape Hatteras area. While income was not a significant determinant of choices, those models with a significant term for the selection effect for our telephone-mail-telephone survey design indicated that preference for environmental commodities was important to respondent choices for beach nourishment. Because income was a significant positive influence for those choices, this seems likely to be the way income influences beach re-nourishment decisions. There also appears to be some evidence that the effects of income are even more complex, varying with the level of income, where the highest income households do not have proportionately higher willingness to pay for the resource. This may imply the range of substitutes that is available to this group for most resources is much larger than that available to others. It may also imply greater skepticism of the proposed program among this group. Unfortunately, our sample size was too small to investigate these issues. Our results offer another interesting insight supporting the reliability of contingent valuation surveys. The plan for beach re-nourishment presented to survey respondents was deliberately structured to be realistic, using detailed information about local conditions. It was not based on a real plan. During our survey a real proposal for a very similar policy initiative was announced in local newspapers. Our survey design kept records on the zipcodes of respondents home addresses. By recording the dates and newspapers with articles on the proposal, together with information about the circulation (by zipcode) of all major newspapers in the state it was possible to construct an event measure relevant to survey respondents who could have read about the proposal prior to answering our second interview. This variable was a count of the articles on the topic appearing in a newspaper relevant to their locations and consistent with the timing of their interviews. The variable was a positive and significant influence on the likelihood of voting for the program. One interpretation of this finding is that it supports the NOAA Panels [1993] call for maintaining realistic framing in the design of contingent valuation surveys. Finally, our results suggest that re-nourishment does appear to be an important management tool for maintaining the Outer Banks. By using the Weibull survival model and the distance access variables it would be possible to estimate conservative WTP measures for typical households in the primary population areas around the state. Moreover, the models allow that appraisal to illustrate the differences in support (as measured by WTP) with preference heterogeneity. Such micro-economic evaluation, indexed by the relative size of each population group, offers one way to provide information about the diversity of household preferences for complex environmental programs that nonetheless remains consistent with the basic principles of benefit cost analysis. References Behn, Robert and Martha Clark, The Termination of Beach Erosion Control at Cape Hatteras, Public Policy, Vol. 27, no. 1, Winter, 1979. pp. 106. Bockstael, Nancy E. and Kenneth E. McConnell, 1993, Public Goods as Characteristics of Non-Market Commodities, Economic Journal, Vol. 103, September, 1244-1257. Carson, Richard T., W. Michael Hanemann, Raymond J. Kopp, Jon A. Krosnick, Robert C. Mitchell, Stanley Presser, Paul A. Ruud and V. Kerry Smith, 1994. Prospective Interim Lost Use Value Due to DDT and PCB Contamination in the Southern California Bight. Report to National Oceanic and Atmospheric Administration, Natural Resource Damage Assessment, Inc., La Jolla, Ca., September. Carson, Richard T., Leanne Wilks, David Imber, 1994, Valuing the Preservation of Australias Kakaku Conservation Zone, Oxford Economics Papers, Vol., 46, (October) 727-49. Chestnut, Lauraine G. and Robert D. Rowe, 1990, Preservation Value for Visibility Protection at the National Parks, draft final report to U.S. Environmental Protection Agency, RCG Hagler Bailley, Inc., February. Cumming, Ronald G. and Glenn Harrison, 1994, Was the Ohio Court Well Informed in its Assessment of the Accuracy of the Contingent Valuation Method?, Natural Resources Journal, Vol. 34 (Winter) pp. 1-36. Diamond, Peter, and Jerry A. Hausman. 1994, Contingent Valuation: Is Some Number Better Than No Number? Journal of Economic Perspectives, 8 (Fall), pp. 45-64. Freeman, A. Myrick III, 1993, The Measurement of Environmental and Resource Values: Theory and Methods (Washington, D.C.: Resources for the Future). Haab, Timothy C. and Kenneth E. McConnell, 1995 Referendum Models and Negative Willingness to Pay: Alternative Solutions, unpublished paper, Dept. of Agricultural and Resource Economics, University of Maryland, April. Hanemann, W. Michael. 1984. Welfare Evaluations in Contingent Valuation Experiments With Discrete Responses, American Journal of Agriculture Economics, 66 (Aug.), pp. 332-341. Hanemann, W. Michael, 1994, Valuing the Environment Through Contingent Valuation, Journal of Economic Perspectives 8 (Fall), pp. 19-44. Hausman, Jerry A., 1981, Exact Consumers Surplus and Deadweight Loss, American Economic Review, Vol. 71, September, 662-676. Hausman, Jerry A. and Whitney Newey, forthcoming, Non-parametric Estimation of Exact Consumer Surplus and Deadweight Loss, Econometrica (in press). Jakus, Paul M. and V. Kerry Smith, 1992, Measuring Use and Nonuse Values of Landscape Amenities: A Contingent Behavior Analysis of Gypsy Moth Control paper presented at the Association of Environmental and Resource Economists Meeting New Orleans, La. Kopp, Raymond J. and V. Kerry Smith, 1995 Constructing Measures of Economic Value, The European Association of Environmental and Resource Economists Conference on Environmental and Resource Economics. NOAA Panel on Contingent Valuation, 1993. Report, Federal Register, Vol. 58, No. 10, January 15, 4602-4614. Pilkey, Orrin H. and William J. Neal, 1992, Save Beaches Not Buildings, Issues in Science and Technology, Spring 36-41. Silberman, Jonathan, Daniel A. Gerlowski and Nancy A. Williams, 1992, Estimating Existence Value for Users and Nonusers of New Jersey Beaches, Land Economics, Vol. 68 (May) 225-236. Smith, V. Kerry, 1987, Nonuse Values in Benefit Cost Analysis, Southern Economic Journal, Vol. 54 (July) 19-26. Smith V. Kerry, 1994, Lightning Rods, Dart Boards, and Contingent Valuation, Natural Resources Journal, 34 (Winter), pp. 121-152. Smith, V. Kerry, 1995, Social Benefits of Education: Feedback Effects and Environmental Resources, in J. Behrman and N. Stacey editors, Social Benefits of Education (in press). Smith, V. Kerry and Laura Osborne, 1995, Do Contingent Valuation Estimates Pass a Scope Test? A Meta Analysis. Center for Environmental and Resource Economics, Duke University, unpublished paper, revised, March. Smith V. Kerry, Xiaolong Zhang, and Raymond B. Palmquist, 1995, The Economic Value of Controlling Marine Debris, Center for Environmental and Resource Economics, Duke University, June. Stewart, T.K., P. Middleton, and D. Ely, 1983. Judgments of Visual Air Quality: Reliability and Validity. Journal of Environmental Management 3: 129-145. Stewart, T.K., P. Middleton, M. Downton, and D. Ely. 1984. Judgments of Photographs Versus Field Ovservations in Studies of Perception and Judgment of the Visual Environment. Journal of Environmental Psychology 4: 283-302. Turnbull, B., 1976. The Empirical Distribution Function With Arbitrary Grouped, Censored, and Truncated Data, Journal of the Royal Statistical, Series B., Vol. 38, 290-295. White, Halbert, 1982. Maximum Likelihood Estimation of Mispecified Models, Econometrica, Vol. 50 (January): 1-76. Table 1: Demographic Characteristics of Sample 1990 CensusPhase 1 SamplePhase 2 SampleNumber of Observations1012580Race (%)White75.682.287.6Black22.015.311.7Other2.42.50.7Sex (%)Male48.540.439.7Female51.559.660.3Household Size (Average)People/Household2.52.72.8Income ($1000)aPer Household32.835.136.8Education (%, Age > 25)> High School70.085.385.2> College17.430.132.2Employment (%)Labor Force67.669.670.1Unemployment Rate4.82.11.9 a These measures of income are for different years: 1989 for Census and 1992 for this survey, and are not adjusted for differences in the consumer price index. Table 2: Turnbull Lower Bound Mean (LBM) WTP Sample SplitNumber of Observations LBM WTP Standard Error1. Full Sample566178.4312.132. Plan to Visit Cape Hatteras National Seashorea. Plan236213.2519.54b. Do Not Plan322154.9815.383. North Carolina Beach Vacation Uses aa. One Day 32186.3266.46b. More than One Day534176.8512.434. Heard about Proposed Plans ba. Not Heard207140.3128.76b. Heard359160.2917.035. Visited Cape Hatteras Areaa. Visited340162.6418.75b. Not Visited226182.4021.216. Vote in Recent State or Local Electiona. Voted310132.5417.92b. Didnt Vote256207.0321.93 a Based on question in Phase I survey about the number of days spent at a North Carolina beach during the most recent summer trip. b At the outset of the second interview respondents were asked about whether they heard or read about the two proposed plans in the information materials mailed to them. They were described as independent proposals. The first and primary focus of the booklet was the beach re-nourishment plan. The second involved the construction of coastal outfalls at Nags Head and Atlantic Beach for disposal for treated wastewater. Table 3: Turnbull Lower Bound Mean WTP by Past and Planned Use Sample DefinitionNumber of ObservationsLBM WTPStandard Errora. Visited Cape Hatteras and Plan to Go Again in Next Two Years 66177.7137.42b. Visited Cape Hatteras and Dont Plan to Go Again in Next Two Years166131.5319.95c. Have Not Visited Cape Hatteras and Plan to Go in Next Two Years 41218.3252.19d. Have Not Visited Cape Hatteras and Dont Plan to Go in Next Two Years156174.1726.44________________________________________________________________________ Table 4: Multivariate Models for Choice Functions: Probit and Weibull Survival Modelsa IndependentProbitWeibull SurvivalVariablesSimpleDetailedSimpleDetailedEconomic VariablesIncome0.24x10-4 (0.08)0.13x10-6 (0.03)0.44x10-6 (0.07)Tax Amount- 0.003 (-7.85)- 0.004 (- 7.29)Performance of Plan0.743 (5.44)0.355 (1.79)1.699 (4.06)0.859 (2.03)Past and Planned UseDays in Last Visit- 0.001 (- 1.44)- 0.001 (- 1.25)Plan to Visit Hatteras.392 (3.22)0.519 (3.03)0.828 (2.51)0.866 (2.03)Been to Nags Head Area- 1.247 (- 1.77)- 2.697 (- 1.99)Information About PlanCount of Articles0.043 (2.46)0.061 (2.34)0.101 (1.98)0.118 (1.93)Access to AreaDistance to Nags Head- 0.003 (- 1.62)- 0.008 (- 1.93)Been to Nags Head * Distance0.004 (1.75)0.008 (1.84)Attitude/DemographicsVote to Support Beach Nourishment (First Interview)0.908 (5.07)1.88 (4.52)Education is at Least College Degree0.306 (1.92)0.892 (2.14)IMR0.002 (2.01)Intercept0.124 (0.90)0.686 (0.954)0.784 (6.35)5.556 (4.75)Sample Size 507286507286Heteroscedasticity AdjustmentIntercept______0.880 (3.79)Heard About Proposals______- 0.489 (- 1.86)_________________________________________________________________________________________________________ aThe numbers in parentheses below the estimated coefficients are the asympototic t statistics for the null hypothesis of no association. * Associate Professor, Department of Economics, St. Olaf College, Northfield, MN 55057; Assistant Professor, Department of Economics, Georgia State University, University Plaza, Atlanta, GA, 30303; and Arts and Sciences Professor of Environmental Economics, Duke University, Durham, NC, 27708 and Resources for the Future University Fellow. Thanks are due Kurt Schwabe for most capable research assistance with the project and to Paula Rubio for preparing and editing multiple drafts of this paper. Partial support for this research was provided by the UNC Sea Grant Program under project R#MRD-25. Footnotes PAGE  PAGE 12  An examination of the attitudes toward the environment using the NORC General Social Surveys from 1972 to 1993, based on 19 national samples, suggests a decline in support. These surveys asked the same questions about support for environmental and other programs. They involve independently constructed samples of English speaking persons 18 years of age or over living in non-institutional arrangements in the U.S. The decline in support for the environment seems to begin around 1991. See Smith [1995]. The Carson et. al. [1994] analysis of the Kakaku area in Australia is one notable case where this was done. The controversy about the policy implications of that study reflect the importance of evaluating the sources of preference heterogeneity and its implications for measures of economic value.  Public Law 834, Chap 849; 49 Stat. 1982 (June 26, 1936, S. 3505).  Public Law 311, Chap. 687; 50 Stat. 669 (August 17, 1937, H.R. 7022).  Conrad L. Wirth, A Letter to the People of the Outer Banks,The Coastland Times (October 31, 1952), p. 5.  Behn and Clark [1979] p. 106.  Raleigh News and Observer, page 18. December 15, 1993.  For a more detailed discussion of the relationships between direct and indirect approaches to non-market valuation, emphasizing the elements of this choice process, see Kopp and Smith [1995].  As a rule, market based applications involve pricing per unit and some link between the marketed good and the non-marketed commodity. The pricing per unit implies that the adjustment to a price or quality change will involve some change in consumption that must be considered in developing measures of the reduction in income a person would be willing to incur to acquire the change in the non-marketed resource.  The circumstances of a choice relate to all the factors that describe the assignment of rights, timing, degree of certainty and consequences of a choice.  Use values refer to the willingness to pay an individual would have for some change in an environmental resourceits quality, conditions of access, or amount that arises because a person engages in some in situ use. As a rule, this involves re_864816657F֠޺֠޺Ole PIC LMETA (Symbol-2 | )-_  Symbol-2 |V(Symbol-2 |)Times New Roman- 2 `V 2 `m 2 `1tk 2 `p 2 `bn 2 ` V 2 `Sm 2 `EpSymbol- 2 `-Symbol- 2 my > 2 y <Times New Roman- 2 `,` 2 `^,` 2 `a,` 2 ` ,` 2 `0 & "System-ࡱ;  FMicrosoft Equation 2.0 DS Equation Equation.2ࡱ; CompObj ZObjInfo Equation Native _865430696 F֠޺޺ࡱ; K͠?>d > Vm-t,p,bn() e  e>  e< Vm,p,0()ࡱ; Lࡱ; :  .creation, but it could also be amenity services provided to homeowners whose locations enjoy a clear view of some landscape, adjoin a river or lake, etc. Nonuse values are defined as a willingness to pay that does not entail observable actions by an individual to use the resource in any way. The precise nature of the definitions for each can be complex (see Smith [1987] and Freeman [1993a]).  Benefit transfers require the adoption of a metric for measuring the amounts of the environmental resource being valued. As a result, they rely on simplifications in the full characteristics of a resource that might be important components of the object of choice from an individuals perspective. An example of benefit transfer in the context of beach nourishment is the Corps practice of using a specific size of beach to derive a value per square foot of beach nourishment. See the March 1990 issue of Water Resources Research for discussion of benefit transfer practices.  The telephone interviews were conducted by HBRS, Inc. of Madison, Wisconsin. The coordinated mailings were processed through the Resource and Environmental Economics Program at NC State University using the addresses furnished at weekly intervals by HBRS, Inc.  This confirms research on the role of photos in perceptions of visibility effects of air pollution by Chestnut and Rowe [1990], research on trees and landscape amenities Stewart et. al. [1983, 1984] and Jakus and Smith [1992], and for marine debris Smith, Zhang and Palmquist [1995].  A copy of the information booklet which describes both the potential beach nourishment plan and a plan to construct coastal outfalls is available on request from the authors. There has been considerable debate in the literature concerning the validity and reliability of contingent valuation over whether people responding about resources or programs they did not know about could nonetheless have economic values for them. For a discussion of the pros and cons in this dialogue, see two recent exchanges: Cummings and Harrison [1994] and Smith [1994] and Diamond and Hausman [1994] and Hanemann [1994].  See Silberman, et al. (1992),  isaࡱ; Ole PIC  LMETA CompObj&Z1  `& p & MathType Times New Roman- 2 @~Times New Roman- 2 Ktk & "System--2 | )ࡱ;  FMicrosoft Equation 2.0 DS Equation Equation.2ࡱ; ObjInfo(Equation Native )<_864819046 F޺޺Ole *ࡱ; K ?>@> 2tࡱ; L4@,ࡱ; 4   .1  & & MathType Times New RomanF-PIC +LMETA -CompObj5ZObjInfo7 2 s~Times New Roman- 2 HR Times New RomanF- 2 W2p & "System-Fࡱ;  FMicrosoft Equation 2.0 DS Equation Equation.2ࡱ; ࡱ; Equation Native 8<SummaryInformation(9 t<@< 2R 2 Oh+'0 D h   @d (C:\MSOFFICE\WINWORD\TEMPLATE\NORMAL.DOT3II. Beach Re-nourishment as an Economic CommodityHospital AdministrationHospital Administration@޺@@޺@F#Microsoft Word 6.02>Ev"S T U irlmn***,,00222a3b3c3*5+5,5666FFFF2F3F4F5F'H(H;HuD3bceKuD3abvK uDbcbVbc EF%bc EF%bcPbcuDPbc]bcPbcUbcbcbcB;HHOOO [ [[[[ ] ]eeeeee6g7g8g>gssZu]u&_mИԘwxyßğşˣأ`hOR^b~{$4ɿ◳VU EFA%bc EF@%bc EF8%bc EFR%bcUbcVbc^bcuDf3bceKuDf3abcvKuDPbcPbcbc uDbcuDl3bceKuDl3abvK:4XSj9RrҼ&FQjLl24[ijk?@23#&stu UchVcch^ch^ccUcbcVZxdyjPZ]^defhioprstwxyz{wz{|:;yijo bc]VcuDPc uDPP uDPPbcch^ccTu/>?@ABCDEwxyz" T!$2(*,,,,2-58!!!!!!!!!!!!!!!!!!!!! !6+8;@HCFF6F7FbFrFFGGG'KNNNRUoWY^(acccc.dudvdddd'e?e@eeeee l,"(eeglnprrtt[u\uku~uuuvwyb{'֊zԘzklmn90*XYٶڶ_`./ѹҹijGH mnۿܿVW0,UV345AP_`abcde|~     l @ ," l4@ ," L0"         LL l @ ,"(45678IMQUVWXYZ[klmno}        LL l @ ,"( ,0489:;<=       LL l @ ,"#=>?67?@OPQRSTUeipvwxyz{| l :," l4:,"  l4@ ," | !-29?@VZagh8888v l h," l :,"h l  ,"#hijklm $ > VVVV l  ,"h l j," l :,"$$(/56789:;fghijtx12v l4:," h l h," l :," 2/FJNW]^_`abc        l x," l4x," "^cjpqrstuvQuv    "5 l x," 5        h ljI!&l@jI! 2lXjV ! "$%08@HQSUVjpw} !'            h ljI!)'.4;<ST\efnwx             h ljI!( )*GHNUV\cdz{|}~ "            ljI!)")*+,6<CIQW^dklxy}            h ljI!(iOPZ[\]fghtuvwxyz{9:hi!!!!!!!!!!!!!!!!!!!!!!!!!`% ljV!'ab!!!!!!!!!!!!!!K@Normala "A@"Default Paragraph Font+@ Endnote Text *@ Endnote Referenceh @ Header ! @" Footer !)@1 Page Number$'@A$Annotation Referencec @R Annotation Text@b Footnote Text &@q Footnote ReferencehZ]     ZZZZZZZZZZZZZZZZZZZZZZZZZZZZ] T m'-/b0+23L XXxĜ2y0QO ` i :        p!  !! !           !   !!!!!!$ 3 &M-4;BHUO+V;]cjr[y'[En9]EE5 ^  PUy@ &);H4vwxyz8e=|h$2' "{|}~C2C4C'E;E=Ebbb::: ")!!School of the EnvironmentA:\FINAL628.DOCAuthorized Gateway CustomerC:\TEXT\FINAL629.DOCAuthorized Gateway CustomerC:\TEXT\FINAL629.DOCAuthorized Gateway CustomerC:\TEXT\FINAL629.DOCAuthorized Gateway CustomerC:\TEXT\FINAL629.DOCAuthorized Gateway CustomerC:\TEXT\FINAL629.DOCAuthorized Gateway CustomerC:\TEXT\FINAL629.DOCSchool of the EnvironmentA:\FINAL629.DOCSchool of the EnvironmentC:\DOC1\MANUSCRI\BCH6-29.TXTHospital AdministrationJ:\HOME\ECON\FLS\TEMP\RENOUR.W6@HP LaserJet IID PostScriptLPT1:PSCRIPTHP LaserJet IID PostScript WD[ od,debris.ps.XZZeHP LaserJet IID PostScript WD[ od,debris.ps.XZZe ETimes New Roman Symbol &Arial"MS Sans Serif"hELFELFt2II. Beach Re-nourishment as an Economic CommodityHospital AdministrationHospital Administrationࡱ