Beneficial Use Value Database - California Water

Transcript

THE BENEFICIAL USE VALUES DATABASE Daniel K. Lew Douglas M. Larson Hiro Suenaga Rodrigo DeSousa Department of Agricultural and Resource Economics University of California, Davis April 2001 Additional assistance provided by Robert J. Oxoby Elyse Niemann Stephanie Jamelske THE BENEFICIAL USE VALUES DATABASE Executive Summary This report documents the development of the Beneficial Use Values Database (BUVD). The purpose of the BUVD is to provide an educational and informational tool to the general public and interested specialists, documenting the economic values for beneficial uses of water identified by the State Water Resources Control Board (SWRCB). It is envisioned that the BUVD be a companion to the Water Quality Standards Inventory Database (WQSID), which currently provides information to the public on water quality standards for, and beneficial uses of, water bodies throughout California, but no information on the value of those beneficial uses. In preparing this alpha version of the database, the literature on economic values of water was consulted widely but not exhaustively, so in its current form the BUVD should be considered as (1) a representation of many though not necessarily all economic values for beneficial uses of water available in the current literature; and (2) a template which can be added to as more studies are identified as suitable for inclusion in the database over time. Section I explains the background and purpose of the BUVD as an educational and informational tool, and its potential role in benefit-cost (B-C) analysis of stormwater management alternatives, which require knowledge of the benefits and costs of proposed alternatives. The ability to do this type of analysis has been hindered by a lack of basic information on what the expected beneficial use values of water are. The BUVD is designed to help remedy this deficiency by canvassing the technical literature on economic valuation to identify quantitative economic values of water in its many beneficial uses, storing them in a readily-accessible and searchable database form. Section II provides an overview of the basic structure and contents of the database. At the present time, the database contains information from 131 studies, consisting of 8 books or book chapters, 101 journal articles, 17 reports, and 5 working or unpublished papers, nearly all from the mid-1970s to the present. The information from these sources was transcribed, coded, and keypunched into a relational database with a series of cross-linked tables, and carefully error-checked for veracity. Separate tables contain information about (1) documents; (2) authors; (3) geographic areas the study is applicable to; (4) beneficial uses maintained by the SWRCB; (5) a modified list of beneficial uses (“searchable beneficial uses”) that are more convenient for analyzing the literature and are linked to SWRCB beneficial uses; (6) valuation methods used; and (7) value estimates and their characteristics. The cross-linked tables allow one to search any one of these pieces of information and quickly retrieve all studies and value estimates with requested characteristics. Twelve different categories of valuation methods were identified to represent the range of techniques used in studies found during the literature search, and brief descriptions of each are provided. The reader is also referred to Appendix A for a more thorough discussion of valuation methods, their strengths, weaknesses, and suitability for different types of valuation questions. Section II also notes some limitations of the database, including the scope of coverage of economic values in the literature, limitations on the information that can be provided for each study, limitations on the ability to compare value estimates across studies, and variations in the quality and reliability of the estimates themselves from the 1 original literature. No attempt was made to rate the quality of information from individual studies, though the discussion of valuation methods does indicate more preferred and less preferred methods from a theoretical standpoint to derive economic value estimates. Some considerations in integrating the BUVD with the WQSID are also discussed. Section III describes the literature review and search process used in constructing the BUVD. The basic objectives were to develop a search methodology that easily accommodates the conventions for reporting economic values in the literature while maintaining a close correspondence with SWRCB beneficial uses, and to provide as much coverage of those beneficial uses as was possible in the time frame of the project. A list of 12 “searchable” beneficial uses were identified from the 26 SWRCB beneficial uses, which principally involved grouping SWRCB categories into broader categories. Searches conducted were quite extensive, encompassing on-line bibliographic databases such as AGRICOLA, ECONLIT, and MELVYL, several University of California libraries, and a variety of printed bibliographies. These were searched by keywords corresponding to beneficial uses and valuation methods. Only studies containing quantitative economic value estimates were included; discarded were purely qualitative value discussions, conceptual analyses with no empirical value estimates, and studies that reported non-economic values, and studies that did not report original results. Section III also discusses a number of difficulties that arose and were resolved during the search process. These include difficulty in tracking down “gray literature” studies for which authors often had to be contacted directly, the limits on our ability to fully examine all sources within the project’s time frame, and the difficulties posed by inconsistent or incomplete reporting within the original studies. Section IV describes the data entry, evaluation, and quality control procedures used in constructing the database. The three criteria used for including a value uncovered in the literature search were that it: a) represents the value of a SWRCB-designated beneficial use; b) be quantified in monetary terms; and c) be an indicator of economic value. The variety of different economic value indicators encountered in the literature are discussed. The types of information recorded about each economic value, and reference information about the study that generated it, are discussed. A detailed discussion of the challenges involved in constructing the database, including variations in the quantity and quality of information provided in underlying studies and the issues involved with classifying values, is also provided. The procedures for quality control, which involved checks for accuracy, appropriateness, consistency of information reported, and readability of the database, are elaborated. Section V presents a summary of the contents of the BUVD. Some of the 2,034 unique value estimates are applicable to single SWRCB beneficial use categories, while more are applicable to multiple categories; in all, counting multiple entries, there are 3,322 beneficial use values represented. These range from a low of zero for several individual SWRCB categories to 962 for Recreation. Among the SWRCB beneficial use categories that had 100 or more value estimates are RARE, WET, AGR, WQE, REC, COMM, MUN, and POW. It was common to see value estimates applicable to water quality enhancement (WQE) and one or more other uses of water, such as recreation (REC1 and REC2) in particular. Section V emphasizes the point that most of the values 2 in the database are for beneficial uses involving water to some degree, rather than being a value of water per se. The remainder of Section V presents a more detailed breakdown of the characteristics of values in the database for each searchable beneficial use. The principal characteristics examined are the occurrence of single-use versus multi-use values, whether the value is an average/total or marginal value, units in which value estimates are presented, and the number of studies and individual values for each combination of characteristics. Detailed cross-tabulations of characteristics of values for each beneficial use category help to convey a sense of the coverage of beneficial use values provided by the BUVD, and the discussion of these tables highlights the types of value questions asked, research methods used, and specific examples of each amenity which are analyzed in studies contained in the database. It is not uncommon for values within a single beneficial use category to vary substantially, based on differences in the degree of aggregation inherent in the estimate, the methods and data used, geographic scope of the amenity, and the units in which the estimates are presented. These differences underscore the need to use caution in comparing values, to ensure that values of amenities with similar characteristics are being compared. Each value must be interpreted in the context of the study which generated it. Extensive comment fields and a large number of characteristics about the values and the studies which generated them will aid users in making these comparisons. Section VI discusses characteristics of the BUVD in relation to two other databases that have recently been constructed for other purposes, the ENVALUE on-line database maintained by the Environmental Protection Authority of the New South Wales government in Australia and the Sportfishing Meta-Analysis Database constructed by the U.S. Fish and Wildlife Service (FWS). ENVALUE has a wider scope of coverage in terms of resource values, as it is not focused on water per se, and includes conceptual as well as empirical studies. The FWS database is focused more narrowly, concentrating only on recreational fishing studies from the mid-70s to present. The focus of ENVALUE and FWS is more toward benefits transfer research applications by technical specialists, while the BUVD is more oriented to the general public. Both the ENVALUE and FWS databases express opinions about the quality of individual value estimates and their suitability for benefits transfer, whereas the BUVD refrains from making such judgments. Both the BUVD and ENVALUE have extensive comment fields which provide additional qualitative information about the values generated by each study. Section VI concludes with a discussion of further work that can be done to enhance and maintain the database. In addition to developing a web-based version to make the database more broadly accessible, more effort can be expended to track down hard-to-acquire references and to incorporate new studies as they become available. Several beneficial uses have little current representation in the database, and concentrated effort targeted to these beneficial uses may turn up additional studies in the extant literature. The Appendices are intended to provide non-specialists with a better understanding of the different approaches to determining economic values and the methods used in the process. Appendix A provides a detailed discussion in lay terms of the 12 categories of valuation methods represented by values in the BUVD, and the different beneficial uses to which each valuation method was applied. Appendix B 3 provides a definition of all acronyms found in the report, a description of the different currency units used for presenting values, and a glossary of terms that provides thumbnail sketches of the meaning of technical terms that occur repeatedly. In sum, the BUVD as currently constituted provides a useful framework within which economic values for beneficial uses of water can be collected and cataloged. Additional efforts should be made to update the database to cover a larger portion of the literature, especially the “gray” literature where the majority of the under-represented beneficial use values are likely to be found. Well-trained analysts with expertise in the area of non-market valuation should conduct these updates. 4 Table of Contents INTRODUCTION .........................................................................................................7 Providing Information to the Public about Economic Values of Water Quality Improvements ..............................................................................................................7 Benefit Transfers and the Database .............................................................................8 Organization of the Report ..........................................................................................9 II. OVERVIEW OF THE DATABASE ...................................................................10 Basic Structure and Contents of the Beneficial Use Values Database.........................10 Limitations of the Database .......................................................................................19 Limits on Coverage of the Literature........................................................................19 Limits on the Content of Database ...........................................................................19 Limits on the Usefulness of Comparing Values ........................................................20 Integrating the Beneficial Use Values Database with the WQSID ..............................20 III. LITERATURE REVIEW....................................................................................22 Goals of the Search Process ......................................................................................22 The Search Methodology ...........................................................................................22 Step 1: Identify Searchable Beneficial Use Categories ............................................23 Step 2: Search for Each Searchable Beneficial Use Category .................................24 Step 3: Identify Studies Containing Useful Information on Beneficial Use Values ...25 Step 4: Analyze Selected Studies .............................................................................25 Databases Used .........................................................................................................26 AGRICOLA ............................................................................................................26 ECONLIT ...............................................................................................................26 MELVYL ................................................................................................................26 Other Sources............................................................................................................27 Annotated and Selected Bibliographies....................................................................27 References in Studies ...............................................................................................27 Agricultural and Resource Economics Library ........................................................28 Difficulties Experienced in the Search Process ..........................................................28 IV. DATA ENTRY, EVALUATION, AND QUALITY CONTROL .......................30 Examining Studies and Values ...................................................................................30 Extracting the Value Estimates ..................................................................................31 Challenges in Constructing the Database ..................................................................32 Quality Control Steps to Ensure Data Reliability.......................................................34 V. A SUMMARY OF REPORTED VALUES FOR BENEFICIAL USES ............37 A Summary of Values by Beneficial Use Category .....................................................38 Agriculture ..............................................................................................................39 Aquaculture .............................................................................................................40 Commercial Fishing and Shellfish Production .........................................................42 Groundwater ...........................................................................................................43 Habitats and Ecosystems .........................................................................................45 Municipal ................................................................................................................49 Navigation...............................................................................................................51 Power ......................................................................................................................52 Rare.........................................................................................................................54 5 Recreation ...............................................................................................................55 Water Quality ..........................................................................................................65 Wetland and Floodplain ..........................................................................................67 VI. DISCUSSION AND FURTHER WORK ............................................................71 Other Databases with Economic Values for Water-related Amenities ........................71 Scope of Coverage ...................................................................................................71 Focus and Intended Audience ..................................................................................72 Specific Content.......................................................................................................73 Further Work.............................................................................................................74 REFERENCES ............................................................................................................76 APPENDIX A ..............................................................................................................77 Measuring the Economic Value of Water: Valuation Methods ..................................79 Contingent Valuation Methods.................................................................................80 Conjoint Analysis.....................................................................................................82 Travel Cost Models..................................................................................................83 Averting Behavior Approaches ................................................................................87 Hedonic Methods.....................................................................................................87 Replacement Cost Method .......................................................................................88 Damage Function Approach ....................................................................................89 Market Valuation.....................................................................................................89 Simulation Models ...................................................................................................90 Optimization Models................................................................................................90 Opportunity Cost Methods .......................................................................................93 Other Methods.........................................................................................................93 APPENDIX B ..............................................................................................................97 Abbreviations used in the database:...........................................................................98 Currency ...................................................................................................................99 Technical Terms ........................................................................................................99 REFERENCES ..........................................................................................................101 6 Introduction The Water Quality Standards Inventory Database (WQSID) was created to provide water quality standards information for California surface water bodies to the public. This inventory includes data from Regional Water Quality Control Plans and the State Water Resources Control Board’s Ocean Plan and organizes it in a Geographic Information System (GIS)-oriented database. It contains both numeric and narrative information about water quality objectives and beneficial uses associated with specific bodies of water within the State of California. The web-based database is currently being maintained by the Information Center for the Environment (ICE) at the University of California, Davis, and can be found at http://endeavor.des.ucdavis.edu/wqsid. This report describes the construction of an additional database that contains information on the economic values for beneficial uses, which can be integrated with the existing WQSID. Its primary function is to act as a starting point for providing information to the general public and interested specialists about the economic value of beneficial uses for water and the decreased (or increased) value from degrading (or enhancing) water quality. A potential secondary purpose is to serve as a resource for water quality specialists to access studies that provide beneficial use economic values necessary for implementation of benefits transfer used in benefit-cost analyses (BCA). Providing Information to the Public about Economic Values of Water Quality Improvements The principal goal of the database project is to compile a set of studies that provide estimates of the economic values of different beneficial uses of water, and illustrate the range of value estimates, types of beneficial uses, and methods of valuation that can be found in the economics literature. As this literature is being added to constantly, the database is not a static entity, but one which can be expected to grow and evolve over time. Our objective is to create a framework for organizing information about beneficial uses of water and to help fill it out initially with information currently available. In recent years there has been a substantial increase in public interest in the value of natural resources managed by state and federal authorities on behalf of all citizens. This has occurred partly as a result of highly-publicized environmental problems with air and water quality, and partly because of recognition that with increasing population growth in California, the value of natural environments must be known in order to achieve the proper balance between environmental protection and development. Thus, it seems natural to provide a centralized repository for economic values of water in its different uses, as a means of providing information to the public. 7 Benefit Transfers and the Database Another potential use for the database is as a resource for specialists interested in benefit-cost analyses of policies or programs affecting uses of water (e.g., Kalman, et al. [2000]). In general, BCA is used as a means of quantitatively comparing the advantages and disadvantages of public policies designed to increase societal well-being. In a typical BCA application, the costs of implementing a public policy or project are compared to the expected benefits. By the BCA criterion, if the benefits outweigh the costs, the policy or project is desirable. Conversely, if the costs are greater than the benefits, the policy or project is not desirable. Unfortunately, directly estimating all values necessary for a complete accounting of the benefits and costs arising from project alternatives is often too difficult and costly to obtain for preliminary analyses, particularly for projects with a limited time frame. This has prompted the use of “benefit transfer” methods in benefitcost analyses as a way of assessing the economic value of project alternatives. Benefit transfer methods attempt to use pre-existing information collected for other purposes, to make reasonable inferences about changes in economic values for a policy problem of interest. This "benefits transfer" process and some implementation issues are described in greater detail in a special issue of Water Resources Research (Vol. 28, no. 3, 1992).1 Although its application is fraught with difficulty, there is a precedent for its use in benefit-cost analyses (Federal Register, 1986; Economic Analysis, Inc., 1986; Walsh, et al., 1989; Sorg and Loomis, 1984). For analyses of changes in water quality (particularly for those applied in California), there are two potential problems when limited information on economic values is used for the benefit transfer. First, an abbreviated literature review may not identify values for all the affected beneficial uses in a study. Second, the economic values that are available for a specific beneficial use may be too limited or inappropriate, either because the available values represent an amenity or use that is not sufficiently close to the one being examined or the quality of the study and the resulting estimates may be questionable. Together, these potential problems suggest the need for a comprehensive database of economic values that: • • • Includes beneficial use values that span the entire set of State Water Resource Control Board (SWRCB)-defined beneficial uses; Contains enough values within each beneficial use category so that the probability of encountering a relevant and useful value is relatively high; and Is readily accessible. Although the database presented in this report does not represent an exhaustive accounting of the valuation literature, it does cover a majority of the set of beneficial uses and represents a convenient framework in which information from studies can be entered 1 A variation of this benefit transfer process involves transferring benefit functions instead of point estimates (Loomis, 1992; Downing and Ozuna, 1996; Kirchhoff, Colby, and LaFrance, 1997). 8 and maintained. The studies it does contain provide a good starting point to look for economic values of beneficial uses for use in basic benefit transfer-based benefit-cost analyses. Since the database was not constructed with benefit transfers as its primary focus, caution must be exercised if the database is to be employed in this fashion. We attempt throughout the discussion to identify potential limitations of using the database as a basis for benefit transfers are discussed in this report. Organization of the Report The remainder of the report describes the manner in which data was collected, screened, and input in the beneficial use values database; the content of the database; and suggestions for integrating with the WQSID. Section II provides an overview of the database, including its underlying structure, brief descriptions of the types of information it contains, how it will interface with the WQSID, and suggestions for how it should and should not be used. Section III provides a detailed description of the methods of collecting economic values from the literature. Section IV describes the steps taken to ensure both reliability and consistency of the reported information. In Section V, a brief summary of what was found for each beneficial use is provided. Finally, in Section VI, the overall literature review process and resulting database is evaluated and discussed. This section includes descriptions of other databases containing economic values for water. 9 II. OVERVIEW OF THE DATABASE Currently there are no economic values in the WQSID. One of the key considerations in designing the database and structuring the information content was the flexibility of searching the database for economic values or other information by endusers. Thus, it was important for the database to be structured in a way that accommodates searching by numerous criteria including by beneficial use, study location, valuation method, and other study attributes. Basic Structure and Contents of the Beneficial Use Values Database The beneficial use values database is a Microsoft Access relational database with nine underlying tables. These tables contain the beneficial use values that will appear in the web-based version and the documents in which they were reported. The nine tables are linked. The relationships between tables and field names within tables are displayed in Figure 2.1. Editing fields such as finalcheck and editcheck, which were used in quality control checks, are not for display and will be discussed in a later section. In this report, only those fields containing information for display will be discussed. The database design centers on the Documents table, which contains reference information (pubyear, title, and refinfo), a field describing the type of publication (pubtype), and general information specific to each document (amenity, sitedesc, and comments). The fields in this table are listed and described in Table 2.1. Documents were classified as one of the following publication types: • • • • • journal article, book/book chapter, report, unpublished/working paper, or other. Of the 131 studies included in the edited database, there were 8 books or book chapters, 101 journal articles, 17 reports, and 5 working or unpublished papers. Because 6 of the book chapters were separate chapters in 3 books, there are actually 128 distinctly different studies reported on in the database. The vast majority of the studies were conducted in the last 25 years, which reflects the relative paucity of studies valuing water and its uses throughout the valuation literature prior to the 1970s. The distribution of these studies by publication type and by date is displayed in Figures 2.2 and 2.3, respectively. 10 Table 2.1. Description of Fields in Documents Table Field Name Title Description Full title of journal article, paper, report, or book. If it is a book chapter, the title of the chapter is entered here. Year the document was published or released. Full reference information: Journal articles—journal name, volume, number (if available), and page number. Books—publisher name and location. Book chapter—title of book, editors (if applicable), publisher name and location. Reports—funding or parent agency and report number. Unpublished/working paper—where presented (conference papers), working paper number (working papers), or institutional affiliation (if internal working paper). Type of publication: journal article, book/book chapter, report, unpublished/working paper, or other. Type of good being valued in the study. Description of the study site being valued or populations being surveyed. General comments relating to the document. In most cases, a brief summary of the main thrust of each document is given here. Pubyear Refinfo Pubtype Amenity Sitedesc comments Although most of the other fields in the Documents table can be interpreted in a straightforward manner, one field deserves a few explanatory remarks. In the literature, valuation studies rarely (if ever) set out to value a specific beneficial use per se. Instead, the goal is to value some other amenity for which water is an important element, such as recreation at a specific site or drinking water quality. Thus, the amenity field describes the good being valued. This good may encompass a single beneficial use or many. For instance, a study valuing swimming in the Chesapeake Bay could be thought of as valuing a REC1 activity. On the other hand, if the study sought to value the change in value to swimming resulting from changes in bacterial levels in the water, the resulting values would be associated both with REC1 and WQE. The remaining fields are fairly straightforward. Complete bibliographic information, excluding author information, is contained in the title, pubyear, and refinfo fields. The sitedesc field provides a general description of the study site and/or the population under investigation, while the comments field contains information important to understanding the purpose of the document. This may include a brief synopsis of the paper or other pertinent information that will assist the reader to understand the context within which the values were generated. 11 Figure 2.1. Structure of the Beneficial Use Values Database Documents Authors lastname firstname middleinit (author_order) (authorid) (docid) (spellcheck) (contentcheck) title pubyear refinfo amenity sitedesc comments pubtype (docid) (finalcheck) (contentcheck) docid docid StateDocID (docid) (states) docid Combo_benuse (combo_benuse) (desc) (benuseid) benuse combo_benuse Values states Crosswalk_benuse orig_benuse (combo_benuse) (benuse_idnum) orig_benuse $value units year$ margavg methodid benuse comments1 comments2 comments3 (valueid) (docid) (finalcheck) (contentcheck) methodid States statename (stateID) Methods methname methdesc (methodid) SWRCB_benuse desc_benuse benuse_name NOTES: (orig_benuse) (ID_origbenuse) (1) Bolded field names are display fields. (2) Parenthetical field names are internal fields not for display. (3) Italicized fields link tables. 12 Figure 2.2: Distribution of Studies by Document Type 120 101 100 80 Number of Studies 60 40 20 8 5 0 Book or Book Chapter Journal Article Working or Unpublished Paper 17 Report Type of Publication Figure 2.3: Distribution of Studies by Year of Publication 12 11 10 10 9 9 8 8 7 6 5 4 4 3 2 2 1 0 Pre1972 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 11 10 Number of Studies 5 4 3 4 4 4 5 4 4 1 1 0 1 0 1 Year 13 The possibility of a single document possessing multiple authors, values, and geographic areas (locations studied) necessitated use of the cross-link tables Authors, Values, and StateDocID, respectively. The inherent value of using cross-link tables is the ability to enter multiple entries for a field associated with a given document. This allows the end-user to conduct specialized searches for values that conform to certain search parameters for these fields (searching by authors’ name, by a field within the Values table, or by the location of the study). The Values table, whose contents are described in Table 2.2, contains specific information about the values reported in each study, including the value estimate itself ($value), its units (units), the year in which the dollar value was reported (year$), the valuation method used (methodid), and comments related to estimation techniques, data used, and other important information for understanding what the reported value represents. Additionally, each value reported in the Values table is categorized as either an average/total value or marginal value (margavg). An average or total value estimate refers to a point estimate of the value of the amenity at a given level of water quality and other exogenous factors. For example, the per-day value of swimming for the average visitor to a river at a given water quality level is an average total value of swimming. On the other hand, the change in the per-day value of swimming at the river when water quality changes from a baseline level to a cleaner level is a marginal value. In general, marginal values represent the change in a total value of the amenity with a small change in some factor such as water quality. Section V provides a detailed breakdown of the reported values by beneficial use category. Table 2.2. Description of Fields in Values Table Field name $value units year$ margavg methodid benuse comments1 comments2 comments3 Value estimate. Units of value estimate, including currency used if not U.S. dollars. Year in which value is reported. Indicates whether the value is a marginal value or average/total value. Valuation method used to obtain value estimate. SWRCB beneficial uses embodied in the value estimate. Comments related to the data and its collection. Specific comments regarding estimation techniques used. General comments that explain the value. Description The $value field contains the actual monetary value estimate for the amenity being valued. Since several values from studies conducted in other countries are reported, not all value estimates are in U.S. dollars. However, unless otherwise noted in 14 the units field, U.S. dollars should be assumed. In addition to currency information, the units field contains the specific units in which the value estimate is reported, such as “millions of $,” “$ per household,” or “$ per person.” The year$ field indicates the year in which the value estimate is valued. It should be noted that many studies did not report the year in which the value estimate was valued. For studies in which it was not clear what year the estimate was applicable, the field is blank. However, for studies that used survey data to estimate values, it was assumed the year the data was collected was the applicable year of the value estimate unless there was information to the contrary. Using the year$ field, adjustments to a specific base year amount can then be made using price indices, such as the Consumer Price Index and the Producer Price Index, to compare value estimates across studies for similar amenities. The techniques used to generate each specific value estimate are reported in the methodid field, and Table 2.3 provides a brief description of each. The basic classification of valuation methods was: • • • • • • • • • • • • market valuation methods, contingent valuation methods, conjoint analysis, damage function approaches, replacement cost methods, hedonic methods, averting behavior approaches, simulation models, optimization models, opportunity cost methods, travel cost methods, or other methods. More detailed descriptions of each of these approaches are included in Appendix A. The methodid field is linked to the Methods table, which contains brief descriptions of the twelve valuation methods. Three additional tables are linked to the Values or Documents table. The first is the Combo_benuse table, which is linked to individual value estimates in the Values table. This table contains a list of the SWRCB beneficial uses and combinations of these beneficial uses (e.g., AGR and WQE) and allows each individual value to be categorized under multiple beneficial use categories. This is necessary because any given value could represent the value of multiple single SWRCB beneficial uses. Thus, each value in the Values table is linked to one of the beneficial uses or combinations of beneficial uses in the Combo_benuse table. To link the combinations to SWRCB beneficial uses, the Combo_benuse table connects to the Crosswalk_benuse table, which contains links to the combinations in Combo_benuse to the SWRCB beneficial uses. Finally, the description of each SWRCB is contained in SWRCB_benuse. A description of each of the SWRCB-designated beneficial uses is given in Table 2.4. 15 Table 2.3. Valuation Methods in the Methodid Field Method Name Market Valuation Description When market data is available, market price and quantity information can be used to estimate demand, supply, or production relationships. These relationships provide a means for directly measuring economic value. Uses survey questions to obtain direct estimates of WTP. This method is frequently used to value goods for which there is little or no behavioral (market) data. It is also the only method that can obtain estimates of nonuse values. A survey-based approach in which people are asked to rate or rank several different scenarios, each with different levels of attributes taken from a common set. Statistical methods are then used to estimate the WTP of individual attributes. These methods seek to determine a "dose-response" relationship between an environmental quality change and some physical effect, and then use market values for the estimated marginal effect to determine a monetary value for the overall effect. Hedonic methods assume that the price of a good is a function of its attributes. Thus, the price of a good is regressed on its characteristics to find the marginal value of the characteristics. Averting behavior approaches infer the value people place on an amenity by what they spend to prevent its removal or degradation. Optimization models are mathematical representations of an economic problem and include mathematical programming, calculus of variations, and optimal control models. This approach views the opportunity cost associated with using an amenity for one use as the value of the amenity used in its next best alternative use. Simulation models used in valuation of beneficial uses are typically used to determine the biological or physical response to economic stimuli. Although there are many variants of this approach, all TCM studies utilize expenditure and trip visitation data for visitors to a natural resource to extrapolate the associated value of a resource. The replacement cost method uses the monetary cost of replacing or restoring an amenity as a measure of the value of the amenity. Valuation methods falling into this category represent a variety of approaches to valuing beneficial uses that include three cost-based valuation techniques, an energy analysis approach, and an agricultural yield comparison approach. Contingent Valuation Methods Conjoint Analysis Damage Function Approach Hedonic Methods Averting Behavior Approaches Optimization Models Opportunity Cost Method Simulation Models Travel Cost Methods Replacement Cost Method Other Methods Since most studies have multiple authors, a separate table for author information was necessary. This is the Authors table, which contains the names of authors associated with each document and is linked to the Documents table. A field indicating the order of authorship of the document, author_order, ensures that the proper order of the authors appears in document information displays. Although the document table contains information on the location of the study, it was recognized early on that end users may wish to locate documents and values that apply to specific geographic regions. Since the 16 sitedesc field was not suitable to be used to create custom queries, a separate table (StateDocID) was created which links each study to the states for which the study and its values apply. This table is linked to the States table, which lists the geographic regions that can be searched, including all 50 states, the entire U.S., U.S. provinces, and other nations. 17 Table 2.4: State Water Resources Control Board-Designated Beneficial Uses. Beneficial Use AGR AQUA Description Uses of water for farming, horticulture, or ranching including, but not limited to, irrigation, stock watering, or support of vegetation for range grazing. Uses of water for aquaculture or mariculture operations including, but not limited to, propagation, cultivation, maintenance, or harvesting of aquatic plants and animals for human consumption or bait purposes. Uses of water that support designated areas or habitats, such as Areas of Special Biological Significance (ASBS), established refuges, parks sanctuaries, ecological reserves, or other areas where the preservation or enhancement of natural resources required special protection. Uses of water that support cold water ecosystems including, but not limited to, preservation or enhancement of aquatic habitats, vegetation, fish, or wildlife, including invertebrates. Uses of water for commercial or recreational collection of fish, shellfish, or other organisms including, but not limited to, uses involving organisms intended for human consumption or bait purposes. Uses of water that support estuarine ecosystems including, but not limited to, preservation or enhancement of estuarine habitats, vegetation, fish shellfish, or wildlife (e.g, estuarine mammals. waterfowl, shorebirds). Beneficial uses of riparian wetlands in flood plain areas and other wetlands that receive natural surface drainage and buffer its passage to receiving waters. Uses of water that support freshwater habitats necessary, at least in part, for the survival and successful maintenance of freshwater plant or animal species. Uses of water for natural or artificial recharge of ground water for purposes of future extraction, maintenance of water quality, or halting salt water intrusion into fresh water aquifers. Uses of water for industrial activities that do not depend primarily on water quality including, but not limited to, mining, cooling water supply, hydraulic conveyance, gravel washing, fire protection, and oil well repressurization. Uses of water that support marine ecosystems including, but not limited to, preservation or enhancement of marine habitats, vegetation such as kelp, fish, shellfish, or wildlife (e.g., marine mammals, shore birds). Uses of water that support habitats necessary for migration, acclimatization between fresh and salt water, or other temporary activities by aquatic organisms, such as anadromous fish. Uses of water for community, military, or individual water supply systems including, but not limited to, drinking water supply. Uses of water for shipping, travel, or other transportation by private, military, or commercial Vessels. Uses of water for hydropower generation. Uses of water for industrial activities that depend primarily on water quality. Preservation of Rare, Threatened, or Endangered Species. Uses of water for recreational activities involving body contact with water, where ingestion of water is reasonably possible. These uses include, but are not limited to, swimming, wading, water-skiing, skin and scuba diving, surfing, white water activities, fishing, or use of natural hot springs. Uses of water for recreational activities involving proximity to water, but not normally involving contact with water where ingestion of water is reasonably possible. These uses include, but are not limited to, picnicking, sunbathing, hiking, beachcombing. Uses of water that support inland saline water ecosystems including, but not limited to, preservation or enhancement of aquatic saline habitats, vegetation, fish or wildlife, including invertebrates. Uses of water that support habitats suitable for the collection of filter-feeding shellfish (e.g., clams, oysters, and mussels) for human consumption, commercial, or sports purposes. Uses of water that support high quality aquatic habitats suitable for reproduction and early development of fish. Uses of water that support warm water ecosystems including, but not limited to, preservation or enhancement of aquatic habitats, vegetation, fish, or wildlife, including invertebrates. Uses of water that support wetland ecosystems, including, but not limited to, preservation or enhancement of wetland habitats, vegetation, fish shellfish, or wildlife, and other unique wetland functions which enhance water quality, such as providing flood and erosion control, stream bank stabilization, and filtration and purification of naturally occurring contaminants. Uses of water that support terrestrial ecosystems including, but not limited to, the preservation and enhancement of terrestrial habitats, vegetation, wildlife (e.g., mammals, birds, reptiles, amphibians, invertebrates), or wildlife water and food sources Beneficial uses of waters that support natural enhancement or improvement of water quality in or downstream of a water body including, but not limited to, erosion control filtration and purification of naturally occurring water pollutants, streambank stabilization, maintenance of channel integrity, and siltation control. BIOL COLD COMM EST FLD FRSH GWR IND MAR MIGR MUN NAV POW PROC RARE REC1 REC2 SAL SHELL SPWN WARM WET WILD WQE 18 Limitations of the Database In general, the beneficial use value database allows users to discover what types of values have been estimated and used in the literature for various beneficial uses. It should prove to be a helpful tool to inform users of the types of techniques that are employed to value uses for water, and find studies and values for specific beneficial uses of water. The database can also aid analysts by acting as a source of economic values for beneficial uses similar to those being valued in benefit-cost analyses. Still, despite the wealth of information such a database provides, there are limits to its usefulness. Users should be aware that the database is for informational purposes only. It can act as a starting point for understanding the economic values of water uses and as a source for finding studies valuing beneficial uses, but due to its limitations it should not at this point be treated as the definitive collection of economic values of water. Its limitations can be divided into three types—limits on coverage of the database, limits on the content of the material in the database, and limits on the comparability of the information. Limits on Coverage of the Literature Our experiences suggest that the approximately two thousand values contained in the database may represent a relatively small portion of the values for water generated in the economics and engineering literatures. We make this statement for several reasons. First, we have collected numerous additional studies that contain useful economic values, but which have not yet been entered into the database because of time constraints. Second, our searches for values associated with several beneficial uses, particularly recreation (REC1 and REC2) and water quality improvement (WQE), yielded large numbers of references that have yet to be explored. And finally, our personal experiences conducting research in this field indicate that studies are being added to the literature on a regular basis. These observations imply that the database should be viewed as organic, continually growing and evolving to accommodate both the growing literature and the potential gaps in the existing database. This process of growth and evolution can likely be facilitated by inviting input from users, who may have knowledge about appropriate studies that are not contained in the database. No matter how the process is perpetuated, growth and evolution of the database will require additional time and resources to conduct subsequent rounds of literature searches and to periodically update the database with the results from these searches and to incorporate feedback from users. Limits on the Content of Database Any database can contain only a relatively limited amount of information explaining the source document and the economic values therein. Although comment fields convey enough information to interpret the economic values, users should take care to consult the original documents should any confusion arise about what a value represents. This remark is especially a propos in light of the field size constraint of 255 characters in the Access database program. This restricts the amount of information 19 contained in any single field, so while most entries are concise they may at times be cryptic. Information in the database was taken directly from the original documents. Although care was taken to ensure that the information in the database is both accurate and useful, users should be aware that there is variation in the quality of the underlying studies and associated values contained within the database. Studies without suitable information were screened out early in the literature review process. Those remaining were subjected to a detailed editing process. Inappropriate or questionably useful studies were removed from the final database. Still, the quality of the remaining studies varies due to the following factors: estimation techniques, theoretical rigor, and data.2 No attempt was made to grade or rate the quality of the documents. Limits on the Usefulness of Comparing Values The values in the database are not directly comparable for several reasons. First, not all values share a common currency, as several studies report values in foreign currencies. Second, values reported are in nominal dollars and depend on the year in which they are reported. For some values, the year in which the currency is valued was not reported, which makes comparisons inherently difficult. A third issue with comparability is differences in units of measurement of amenities being valued. For example, economic values for swimming measured in dollars per person per year cannot be directly compared to economic values for swimming measured in dollars per household per day, though one can make conversions based on assumptions about household size and recreation days taken by each person during the year. An added difficulty arises with determining which amenities are similar and can be compared. For example, are saltwater beach values in New England sufficiently similar to freshwater beach values in Michigan that they can be compared or substituted for one another? This is an example of the types of judgments users must make when considering the values reported in the database. As the above discussion makes clear, the database should be viewed as a work in progress, as perhaps it always will be. Given the time frame for our project, it has been possible to develop a complete framework for compiling, screening, and reporting economic values for beneficial water uses, and to make an initial pass through the literature to obtain economic values. Integrating the Beneficial Use Values Database with the WQSID Despite the cautions about its limitations, an abundance of useful information is contained in the database. The final remaining task is integration of the beneficial use values database with the on-line WQSID, which is outside the scope of this project as originally conceived. This will be done cooperatively with the Information Center for the Environment (ICE). 2 This is an important point. Although all values reported in the database are indicators of economic value, not all values can be interpreted as theoretically-consistent measures of economic value. 20 Two alternative approaches have been discussed to incorporate economic values into the WQSID. The first integrates the information in the beneficial use values database with the WQSID for a unified website. The second approach would involve constructing a separate website for the beneficial use values database. The latter, website-based, approach involves linking elements within the WQSID with an independent website, which would be a dedicated site for the beneficial use values data. Although specifics of the independent on-line version of the database have not been worked out, an example of an on-line economic values database is the ENVALUE environmental valuation database (http://www.epa.nsw.gov.au/envalue) constructed by the Australian government. The ENVALUE website has attractive features that should be considered when developing the user interface, or front end, of the on-line beneficial use values database. Like the ENVALUE site, users accessing the beneficial use values database should be confronted with an introduction to the database, delineating its purpose and its features, as well as instructions for navigating the site. Additionally, a disclaimer such as the one used by the New South Wales (NSW) Environmental Protection Authority (EPA) is useful for limiting liability of use. 3 To integrate economic information into the website, partial replicas from the Access database will be provided to ICE. Partial replicas contain a subset of all the information in the master database that meet given criteria. For the beneficial use values database, the partial replicas will be comprised of those documents and values that have passed quality control checks discussed later in this report. Use of partial replicas allows us to control what will be displayed in the user interface while maintaining the ability to update the information contained in the database. Additionally, because the database uses abbreviations, foreign currencies, and technical jargon (econometric and economic terms) that may be unfamiliar to the user, the site should include a glossary such as the one compiled in Appendix B. This appendix consists of definitions for abbreviations and technical terms used in the database, as well as a list of currencies used and their country of origin. 3 The NSW EPA uses the following disclaimer: “The NSW EPA has taken care to ensure the data presented in ENVALUE are an accurate representation of the source studies, however the EPA accepts no responsibility for results obtained from use of the database by third parties.” 21 III. LITERATURE REVIEW Given the time frame for this project, the time allowed for gathering data for inclusion in the database was limited. The overall search process was undertaken with several goals in mind. These goals helped define the search methodology and the type (and extent) of information to be obtained. This section begins by outlining these goals. After enumerating the goals, the methodology developed to achieve these goals is described. We describe the process by which the literature was searched, documents were found and screened, and information was collected from the screened documents. Next, the information sources used in the search (including databases, libraries, and additional bibliographical sources) are discussed. Finally, problems that arose during the literature search, how they were resolved, and how the search methodology was adjusted in response are discussed. Goals of the Search Process Four basic objectives were established to guide the literature search process. They are: 1. Develop a search methodology to gather economic values corresponding to SWRCB beneficial uses that covers as much of the known literature as possible and is sensitive to the way economic values for different beneficial water uses is presented in the literature. 2. Explore and report on the extent of coverage of economic values of SWRCB beneficial uses within the literature. This is necessary to determine which beneficial uses have been valued, which have not, and the availability of these estimates when they exist. 3. Cover the entire set of SWRCB beneficial uses within the time allowed for data collection. 4. Collect and organize information on economic values of beneficial uses for water and water quality according to SWRCB designated uses. The Search Methodology With these goals in mind, a carefully structured search methodology was constructed to maximize the quality and consistency of studies extracted from the literature. The search methodology consisted of four basic steps: • • • • identify searchable beneficial use categories; search literature for studies containing values for each beneficial use category; identify studies containing useful beneficial use values; and analyze selected studies. 22 Together, these steps provided the basic framework for searching the existing literature for beneficial use values of water. Use of the search methodology based on these basic steps allowed us to locate, collect, screen, and analyze the valuation literature in a consistent way. Step 1: Identify Searchable Beneficial Use Categories The first step in the search process was to group the SWRCB beneficial uses into searchable beneficial use categories. This was necessary because of the way beneficial use values are presented in the literature. Many SWRCB beneficial uses cannot be distinguished from each other in the literature because the amenity being valued embodies multiple beneficial uses. Multiple beneficial uses are often found being valued in the same study because of the similarity or related nature of the amenities being valued. Therefore, it is both inefficient and difficult to search for each SWRCB beneficial use in isolation. This highlights an important point—the focus of valuation studies is almost always amenities, like wetlands, irrigation water, beach recreation, etc., and not beneficial uses per se. Two important implications follow from the focus in the literature on amenities, not beneficial uses. First, the amenities being valued in studies may have uses that fall in multiple SWRCB beneficial uses categories. For example, many wetlands provide flood protection benefits. As a result, studies that value wetlands often value the flood protection services the wetland provides. Thus, the SWRCB beneficial use categories WET and FLD often cannot be separated. Second, in the valuation literature, some amenities are not described in enough detail to be put conveniently into a single SWRCB beneficial use category. One important category of this type is habitat valuation. The SWRCB classifies habitat beneficial uses into numerous categories ranging from coldwater habitats (COLD) and warm-water habitats (WARM) to estuarine (EST) and marine habitats (MAR). However, it is rare that studies provide enough detail about the habitats being valued to adequately categorize the habitat into one of the SWRCB beneficial uses. Thus, it is more efficient to search for a more broadly-defined amenity, such as habitat irrespective of specific type. Then, if additional information allows the type of habitat to be characterized more specifically, this can be done following the broad-based search. In summary, grouping beneficial uses into “searchable beneficial use categories” that take into consideration the difficulties just described allowed us to more efficiently search the literature for useful beneficial use values, reducing the time and effort involved. In effect, this “grouping” approach takes account of the way information is presented and organized in the existing literature. A total of twelve searchable beneficial use categories were identified and used in the search process. Each of the 26 SWRCB beneficial uses were accounted for in one of these searchable beneficial uses. Seven searchable beneficial use categories contained more than one beneficial use, while five consisted of a single beneficial use. The latter case applies to SWRCB beneficial uses that have few (or no) similarities with others. The searchable beneficial use categories and the corresponding SWRCB beneficial uses are shown in Table 3.1. 23 Table 3.1. Searchable and SWRCB Beneficial Use Categories Searchable beneficial use category SWRCB beneficial uses Habitat /Ecosystem BIOL, COLD, EST, WARM, MIGR, SPWN, MAR, SAL, WILD, FRESH and RARE Wetland /Floodplain WET and FLD Agriculture AGR Aquaculture AQUA Water Quality WQE Recreation REC1 and REC2 Industrial IND and PROC Commercial SHELL and COMM Groundwater GWR Municipal MUN Navigation NAV Power POW Step 2: Search for Each Searchable Beneficial Use Category Searches for economic values corresponding to each searchable beneficial use category were conducted separately. Each searchable beneficial use category was researched for a pre-specified amount of time, ranging from four to six weeks. During that time, many sources were used to access the literature and find studies with economic values corresponding to the searchable beneficial use category of interest. These included several on-line bibliographic databases (AGRICOLA, ECONLIT, and MELVYL), University of California library resources (including UC Davis’ Shields 24 Library, Physical Sciences Library, and Agricultural and Resource Economics Library, and other UC campus libraries), and a number of printed bibliographies, both published and in the “gray” literature. These information sources vary both in content and emphasis. The specific sources used in each search depended upon the specific searchable beneficial use category being investigated. To locate relevant studies in the literature, we searched databases for keywords in titles and abstracts for these studies. The information provided by document titles and abstracts were used as indicators of their usefulness in meeting our objectives. This approach generally led to studies with relevant economic value estimates of beneficial uses. Relevant studies were also identified through the use of keywords like “value,” “valuation,” and “economic benefits,” particularly in studies related to recreational, environmental or consumption benefits related to water use. This title, abstract, and keyword approach formed the basis for identifying relevant studies that could be further reviewed for useful values and selected for inclusion in the database. At times, searches through the available sources turned up “hard-to-find” or “gray literature” documents. These documents included conference papers, working papers, and unpublished studies that could not be found in libraries accessible via the University of California or California State University library system. When possible, useful studies of this type were procured directly through their authors. Step 3: Identify Studies Containing Useful Information on Beneficial Use Values Once the documents identified as potential candidates for inclusion in the database were located and physically acquired, they were skimmed for valuable information such as monetary values of amenities related to water, valuation methods commonly used to value beneficial uses, and references that may contain leads to other useful studies. Only those studies containing economic value estimates were kept. Strictly methodological or conceptual papers were excluded at this stage in the literature search process. Moreover, value estimates also had to be derived in the study. Therefore, studies reporting values estimated in another study and not used in a new way in the present study were excluded, though the source documents were pursued. During this screening, many publications without monetary value estimates were excluded. Some of these documents valued amenities related to the SWRCB beneficial uses qualitatively rather than quantitatively. Like the strictly methodological and conceptual studies, these studies were disregarded after obtaining potentially useful bibliographic references contained in them. Step 4: Analyze Selected Studies Documents making it to this point were photocopied and analyzed to extract useful information for the beneficial use values database. The analysis and data entry procedures for values selected from these documents are described in Section IV. The four steps of this literature search process were used to locate and acquire useful studies from the literature containing economic values for each of the searchable 25 beneficial use categories. Although the steps involved are straightforward, implementation of the search procedure was a constantly changing, organic process. Each step was revised in light of new information turned up in the search process, to improve the overall methodology as new types of studies, amenities, and values were encountered. Databases Used A large number of sources were used to identify and locate useful studies. The databases used included AGRICOLA, ECONLIT, and MELVYL. The remainder of this section provides detailed descriptions of each of these databases. AGRICOLA The AGRICOLA (Agricultural Online Access) database contains bibliographic records of materials acquired by the National Agricultural Library (NAL) and cooperating institutions in the agricultural and related sciences. Ninety percent of the records describe journal articles and book chapters, and the remaining ten percent describe monographs, series, microforms, audiovisuals, maps, and other types of materials. Together they provide worldwide coverage of the agricultural literature. In addition, AGRICOLA contains subfiles of related bibliographic citations that have been prepared by sources other than the National Agricultural Library, such as the Food and Nutrition Information Center (FNIC) and the American Agricultural Economics Documentation Center (AAEDC). These and other information centers and cooperators contribute subfiles to AGRICOLA covering special subjects. The database is updated quarterly and covers studies from 1970 to 2000. ECONLIT ECONLIT is a comprehensive, indexed bibliography with selected abstracts of the world’s economic literature. It is produced by the American Economic Association and includes coverage of over 400 major journals as well as articles in collective volumes (essays, proceedings, etc.), books, book reviews, dissertations, and working papers. Over 99% of the articles are in English or include English summaries. Generally, all articles from all publications are indexed, including notes, communications, comments, replies, rejoinders, and the like. The database is updated monthly and covers economic studies from 1969 through the present. Abstracts are included for a significant portion of the studies published on or after 1987. MELVYL The nine UC campuses share materials through special interlibrary loan arrangements, jointly purchase many Web-based reference tools and materials offered via the UC-operated Digital Library of California, and together maintain the MELVYL online catalog of books and journals. Using this system, one can access the expansive collections at UC Davis and the other eight UC campuses. At UC Davis alone, the 26 campus libraries collectively house close to 3 million volumes and subscribe to approximately 49,000 periodical and journal titles annually. Other Sources In addition to the databases mentioned above, other sources were used to find studies in the literature relating to the economic valuation of water and water quality. These sources include selected and annotated bibliographies on topics relevant to the search, references in other studies, and documents obtained from the Agricultural and Resource Economics Library at UC Davis. Annotated and Selected Bibliographies The following documents either contained useful bibliographies or was comprised of annotated or selected bibliographies useful for the literature search: • Douglas, Aaron J. “Annotated Bibliography of Economic Literature on Wetlands”, U.S. Department of the Interior, Fish and Wildlife Service, Research and Development, Washington DC, 1989. Carson, Richard T., Jennifer Wright, Nancy Carson, Anna Alberini, and Nicholas Flores, “A Bibliography of Contingent Valuation Studies and Papers”, Natural Resource Damage Assessment, Inc, La Jolla, California, 1994. Frederick, Kenneth D., Tim Vanden Berg, and Jean Hanson, “Economic Values of Freshwater in the United States”, Resources for the Future, Discussion Paper 97-03, 1996. Leitch, Jay A. and Herbert R. Ludwig, “Wetland Economics 1989-1993: A Selected, Annotated Bibliography”, Greenwood Press, U.S.A., 1995. Polasky, Stephen, Michael Jaspin, Susanne Szentandrasi, Nancy Bergeron, and Robert Berrens, “Bibliography on the Conservation of Biological Diversity: Biological/ Ecological, Economic and Policy Issues”, Agricultural Experiment Station, Oregon State University, Project 143 and the Environment, Infrastructure and Agriculture Division, Policy Research Department, World Bank, RPO # 67940, 1995. • • • • References in Studies As mentioned above, references contained in the documents found in the literature search were screened both for relevant beneficial use value information and for leads to additional studies that may be contained in the study’s references. In some cases, screened studies did not contain useful economic values, but did cite studies in their references that contained beneficial use values. Using these references, a number of studies with useful economic values were discovered. All studies found in bibliographic 27 references from other studies went through the same screening process as the studies found through other sources. Agricultural and Resource Economics Library The Agricultural and Resource Economics Library, a unit of the Department of Agricultural and Resource Economics, houses over 8,300 book and journal volumes, receives about 1,000 current serials, and maintains a pamphlet file containing approximately 150,000 pamphlets and government documents. The publications are indexed and arranged according to the Department’s interests and needs. The library’s catalog includes numerous pamphlets, papers, and documents not found in MELVYL. These documents primarily focus on California, western, and U.S. food and natural resource economics issues. Several studies containing beneficial use values for water were found in this library’s collection. Most of the documents obtained through this source are concerned with valuing recreation and agricultural uses for water. Difficulties Experienced in the Search Process Several difficulties arose during the search process. Two principal problems arose due to the limited time scope of the search process. First, gray literature sources, such as papers presented at conferences, reports produced by public or private institutions, working papers and unpublished documents, as well as Master’s and Ph.D. theses, were often extremely hard to procure. In most cases authors had to be contacted to acquire these studies. This was a relatively costly and less productive activity to engage in during the limited amount of time available for searching for studies related to each searchable beneficial use category. As a result, the database does not include many studies of the type mentioned above. This had a significant impact on beneficial uses that lacked significant coverage in the mainstream literature (refereed journals, books, etc.) and instead were found primarily in the gray literature (conference papers, working papers, etc.), such as COLD, IND, and NAV. Therefore, to some extent these beneficial uses are under-represented in the literature collected. However, these beneficial uses represent a small portion of the total. Values for most beneficial uses are readily available in the published literature. Another consequence of the short time for the literature search for each searchable beneficial use category was that not all sources used to find studies could be examined. In instances when this occurred, the difficulty was not related to the scarcity of studies, or the difficulty locating studies. Instead, it was a problem with an excess of available studies for a beneficial use category. This was most common for beneficial uses like REC (REC1 and REC2), WET and RARE. Even though some highly relevant values and studies could not be included in the “first round searches,” these beneficial uses are still well represented in the database, and the additional values can be incorporated in subsequent searches. Another difficulty concerned studies that relied on secondary sources for data, models, and results. It was not uncommon to run across studies that used data from secondary sources. However, some studies simply reported value estimates from other studies. These latter studies were eliminated to avoid double-counting or repeating 28 information already in the database. Studies that used secondary data in models to calculate original estimates were not eliminated. Finally, some studies did not report some important pieces of information. These studies often did not specify the valuation method used, the data used, or the year in which values were measured. Failure to include these pieces of information was viewed as a relatively minor problem. Despite their incompleteness, these studies were collected and included in the database because of the relevant information they contained. 29 IV. DATA ENTRY, EVALUATION, AND QUALITY CONTROL Section III explained the four steps used to identify studies containing suitable beneficial use values. The fourth step, analyzing the procured studies for economic values, is explained in more detail in this section. Implementation involved: 1. Examining the studies and values for relevance to the goals of the database project; 2. Extracting the value estimates identified as suitable, along with information necessary to interpret them; and 3. Updating and quality-checking the information stored in the database. Examining Studies and Values In the search for beneficial use values, each value reported in the collected studies was evaluated using three primary criteria. Only those values satisfying all three criteria were included in the beneficial use values database. The criteria were: 1. The reported value represents a value of a SWRCB-designated beneficial use. 2. The reported value is quantified in monetary terms. 3. The reported value is an indicator of economic value. The first criterion requires that the value represents a use of water directly, or a use whose value is primarily attributable to the water. Water may be valuable directly as an input to production or as a consumable good, such as for drinking. In addition, water may make an activity or amenity possible or accessible to individuals. Examples include water-based recreation and fishing opportunities. Without water, individuals would not be able to participate in these activities. The second criterion requires that the values reported in the database be quantified. Qualitative discussions of the value of beneficial uses were not particularly useful for our purposes as they cannot be reported in numeric form. Similarly, values measured in non-monetary terms were not included into the database.4 Studies that discussed methodological or conceptual issues without reporting quantitative measures of value were also excluded from the database. The final criterion relates to the economic concept of value. It is important to clarify what is meant by “economic value.” Water has long been recognized as valuable for its ability to enter as an input into production processes, as an input in recreation experiences, and as a consumable good. Water is used to generate power, is an essential input in agriculture, has numerous industrial and commercial applications, creates recreation and commercial opportunities, and is needed for drinking supply. Additionally, in recent decades economists have begun to recognize the valuable services 4 The exception is a study that measures the value of water in terms of days of labor foregone (Hardner, 1996). This study was included in the database because a monetary conversion factor was provided that allows the reader to easily determine a money equivalent for the days of labor. 30 water provides relating to its ecological and habitat-based functions. From this perspective, water is valuable because of the essential role it plays in habitats of all types. The basis for measuring economic value is individuals’ preferences for particular goods and services. As Freeman points out, "the trade-offs that people make as they choose less of one good and substitute more of some other good reveal something about the values people place on these goods" (Freeman,1993, page 7). The economic welfare of individuals and of society is highest when the use of goods and services is arranged so that the difference between total value and total cost, or net economic value, is the largest possible. The principal measures of net economic value are consumer’s surplus (the difference between a consumer’s total valuation of a good and what s/he must pay for it) and producer’s surplus (the difference between what the provider of a good or service receives and the cost of providing it). Both of these measures are based on the straightforward notion of the difference between total benefit and total cost. Where well-functioning markets act to guide the distribution and allocation of goods and services, researchers can analyze market transactions to obtain estimates of economic value. For many uses of water, such markets do not exist, so researchers typically resort to survey research methods to collect the information needed to determine net economic value. Less theoretically robust indicators of economic value found in the literature include changes in revenues earned and changes in costs spent on water and water-related amenities. Revenues or costs, considered in isolation, are at best only rough indicators of economic value because they do not provide enough information.. For example, individuals who pay $2 per gallon of water for 100 gallons may have been willing to pay considerably more in total than the $200 actually paid. This excess of value over what is actually paid is a surplus to the individual (hence the name consumer’s surplus). The cost itself provides no information on net economic value. Studies employing both the theoretically-sound measures of economic value, like consumer’s surplus and producer’s surplus, and rough indicators of value, like total costs, were included in the database. The specific methods employed by studies in the database are discussed in the next section and in Appendix A. Extracting the Value Estimates The information gathered from studies fall into two general groups: information on the reported values, and how they were determined within the underlying study, and reference information about the document containing the values. Value information includes the actual monetary value estimate, its units, the year in which the monetary unit was measured (to account for inflation), the amenity being valued, and specific information on the technique used to obtain the value estimate, the data used, and other pertinent information to help the reader understand the value. Reference information included general comments about the basic thrust of the document. This is important for users to be able to differentiate between values and understand the context within which they were generated. For details on the specific value and document information extracted from each study, see Section II. 31 The majority of studies provided multiple values. All values that satisfied the criteria above were reported in the database. Multiple values being reported in the same study could be a consequence of any of a number of factors, including differences in: • • • • • Data used for estimation Specific activities or amenities being valued Valuation methods Estimation techniques used for valuation Units and types of economic value measures used to present the value of water Some multi-value studies generated different values by estimating values using subsets of the total dataset. For instance, a number of studies provided beneficial use values for users and nonusers by using data from each group separately. Other studies provided distinct values for different scenarios, such as for differing water quality levels. In the case of different water quality scenarios, each water quality level represents a different good being valued. Similarly, values corresponding to different time periods and locations represent different goods as well, even if the same amenity (e.g. beach recreation) is being valued. A number of studies also report values for different activities or amenities at the location(s) being studied. In some cases, researchers generate distinct economic values through the use of different valuation techniques. These can be differences in valuation methods used, such as TCM and CVM, or in specific modeling assumptions and approaches, such as the use of ordinary least squares regression (OLS) and maximum likelihood estimation (MLE). Finally, multiple values often occur in studies contained in the database because researchers report both different types of individual value estimates and aggregate value estimates. For example, it is not uncommon for authors to report both a per-household or per-person value and an aggregate value calculated by multiplying the per-household or per-person value by the number of households or people in the study region. In addition, many studies reported both marginal values, which measure the change in value due to a small change in conditions, and average or total values, which reflect the value of a beneficial use at a given set of conditions. Challenges in Constructing the Database Because of the diversity of the literature and the range of information sought, several difficulties arose both when analyzing the studies for appropriate beneficial use values and when extracting information from the studies. This section briefly describes the problems encountered and the quality control mechanisms that were instituted to avoid them. The problems encountered can be classified as one of three main types: insufficient information in the studies, quality of information between studies, and classification problems. 32 • Variation in Quantity of Information Across Studies The amount of information provided to explain the economic values presented varied significantly across studies. Some provided a great deal of detailed information, while others supplied very little information on important items, like methodology, amenities valued, and data used. In some studies, key items, such as valuation methods used to estimate beneficial use values, were absent. In cases where key pieces of information were missing, reported values could not be fully characterized. Studies with an insufficient amount or depth of information required to differentiate values were discarded from the database. The absence of individual pieces of documentation was not necessarily grounds for expunging studies and values from the database. For instance, when data and methodologies were not given, the entries corresponding to this information noted that the information was not given. It was only when this information was required to differentiate multiple values within the same study that the study and its values were removed from the database. An item that was consistently absent from many studies was the year in which value estimates applied. For economic analyses, this is a critical piece of information required to express dollar values in a common base year, which is needed for comparability of estimates from different studies. Inflation, or changes in the value of a dollar from year to year, can make comparisons between values determined in different years misleading if not corrected for. The absence of information on year was not grounds for rejecting studies, since is was so common in the literature. An implication of missing or inadequate information in the underlying studies is variations in the depth of information provided in the database. These variations are noted in the comment fields corresponding to the different pieces of information. • Variation in Quality of Information Across Studies In addition to variation in the quantity of information provided in the studies, there was significant variation in the quality of the information across studies. The quality of a study can be judged from numerous perspectives. We define a high quality study as one that employs rigorous valuation methods and good data to yield theoretically-consistent measures of value for a carefully articulated water-related good or amenity. Unfortunately, this is an ideal that is seldom achieved in practice. As noted in other parts of this report, the reported values range from theoretically-consistent measures to rough indicators of the economic value of beneficial uses. Even when theoreticallyconsistent measures of value are estimated, other factors can influence the quality of the resulting value estimate. These include the type of amenity being valued, the valuation methods used, the estimation techniques used, and the data used in the study. Since the database was constructed, in part, to provide the user with an idea of the values contained in the literature, studies and values of varying quality were included in 33 the database. All values satisfying the criteria noted earlier in this section were included irrespective of the quality of valuation methods, data, or other research assumptions made by researchers. No effort was made to rate or judge the quality of studies. • Classification Problems As noted earlier, we did not encounter any study that explicitly valued SWRCB beneficial uses. Instead, studies estimated the economic value of water used for some specific activity or amenity that it provides. Since these specific activities or amenities often do not have a one-to-one relationship with SWRCB beneficial uses, it was necessary to sort values into the SWRCB beneficial use categories. In many cases, amenity values did not fall conveniently into any single SWRCB beneficial use category, because the amenity itself spanned multiple categories. When this occurred, values had to be cross-classified under more than one SWRCB beneficial use. For example, there were a large number of estimated values representing benefits of water quality improvements. A large proportion of these were measured by the effect of the water quality improvement on the value of recreation activities. Values of this type necessarily fall into both the water quality enhancement (WQE) and recreation (either REC1 or REC2) SWRCB beneficial use categories. Although this example is fairly straightforward, other classifications were more complicated because of the lack of information required to gauge the beneficial use categories into which values should be placed. Quality Control Steps to Ensure Data Reliability Several personnel were involved in the various facets of the literature review and database development process. To ensure that the information collected from studies was consistent across the personnel involved in data entry and analysis, and was consistent with the information in the original documents, a series of quality controls were implemented at every stage of the literature review and data transcription process. These quality controls included checking contents of the database for: 1. accuracy of the data extracted from the literature, 2. appropriateness of the data (including whether information was categorized correctly in the comments), 3. consistency in the types and depth of information reported in the database, and 4. readability of the database contents (especially the comments). Table 4.1 indicates when each of the above quality control checks were conducted. The post-literature review portion of the project consisted of three stages, data entry, completion of the first round search, and report preparation. As Table 4.1 shows, quality control checks for each of the four elements listed above were performed after the first round of the literature search was complete (after values were found for each searchable beneficial use) and while preparing this report. In addition, the accuracy and readability of the data were checked during and after the information was input during data entry. 34 Table 4.1. Quality Control Checks Stage Data entry Completion of first round search Report preparation Accuracy Appropriateness Consistency Readability X X X X X X X X X X During and after data entry, the database was checked for grammar, spelling, and omissions of information. The original documents were checked to ensure that all fields had the correct information in them. Empty fields were also checked to determine whether there was information that was missing in the database but could be found in the original documents. Editorial checks were also conducted at this stage—grammatical and spelling errors, ambiguous expressions, and unexplained abbreviations were corrected. More extensive quality control checks were conducted after the first round of the literature search was completed. At this stage, the database contents were checked for accuracy, appropriateness, consistency, and readability. First, the accuracy of the information contained in the database was checked both again with the original studies to ensure that the data had been correctly entered into the database. Second, the appropriateness of the reported values as beneficial use values were checked. That is, the values were checked to ensure they satisfied the three criteria defining an appropriate beneficial use value as described above. Values that failed to satisfy the criteria were discarded from the database. For the information associated with the value estimates, we checked whether the information was entered into the appropriate field and whether each field in the database had the appropriate information. Fields with blank entries were also checked to see whether the absence of information was due to coding errors or a lack of information in the original studies. Third, consistency in the information coded was checked across the studies and values. Since the types and amount of information input into the database could vary across coders (individuals who analyzed and input data), it was important to check for consistency in reported values and the associated document and value information. To ensure consistency across coders, individuals were restricted to analyzing studies they had not previously analyzed. This ensured that the re-coded information could be compared and checked for consistency across individuals. Although all fields were checked for consistency, particular attention was paid to the comment fields since these fields are open-ended. Finally, the readability check consisted of checking for grammar, spelling, and flow. This readability check was meant to ensure that the information in each field was conveyed in a manner that could be easily understood. Each of these elements was checked before the report was begun and after the second set of quality control checks were completed. During this round of controls, values corresponding to individual beneficial uses categories were analyzed separately. 35 Thus, the values and the studies from which they were drawn were examined for the appropriateness, accuracy, and consistency of the information. During these control checks, mis-classified values were correctly classified. Thus, one of the primary objectives of this process was to identify inconsistencies among the values classified in the same beneficial use category. 36 V. A SUMMARY OF REPORTED VALUES FOR BENEFICIAL USES The current version of the database contains 2,034 economic value estimates of beneficial uses of water collected from 131 studies. Table 5.1 summarizes the number of values and studies reported in the database for each beneficial use category. Table 5.1. Number of Values and Studies Reported in the Database by Beneficial Use Category Searchable beneficial use category SWRCB beneficial category Studies reporting at least one value Number of values for this SWRCB category only Number of multiplecategory values applicable to this SWRCB category 33 39 42 0 0 0 11 0 18 58 26 178 14 195 18 651 630 40 4 0 46 85 159 1 99 2347 Total number of values for this SWRCB category Habitat /Ecosystem Rare Species Wetland /Floodplain Agriculture Aquaculture Water Quality Recreation Industrial Commercial Groundwater Municipal Navigation Power Total Total multi-use values Total in database * BIOL COLD EST WARM MIGR SPWN MAR SAL WILD FRSH RARE WET FLD AGR AQUA WQE REC* IND PROC SHELL COMM GWR MUN NAV POW 5 1 3 0 0 0 3 0 2 5 16 19 6 17 3 39 50 6 1 0 6 13 23 6 9 131 5 0 0 0 0 0 0 0 3 0 192 39 4 61 14 71 332 40 0 0 57 8 87 32 30 975 1059 2034 38 39 42 0 0 0 11 0 21 58 218 217 18 256 32 722 962 80 4 0 103 93 246 33 129 3322 REC1 and REC2 were treated as a single category, REC, since studies rarely distinguished between contact and non-contact recreational uses. The table is organized by searchable beneficial uses, which appear in the first column of Table 5.1. In the second column, the individual SWRCB beneficial use categories associated with them are shown. 37 Next is the number of studies that report at least one beneficial use value estimate falling into each of the corresponding SWRCB beneficial use categories. It shows the large variation in the number of studies available for different beneficial uses. For example, economic values of water in recreational uses were collected from 50 studies, while values for many beneficial uses contained in the habitat and ecosystem searchable category, such as freshwater ecosystems, marine, and wildlife ecosystems were collected from a few studies. There are no studies (and hence, no values) associated with warmwater, migratory, spawning, and inland saline habitats. The large number of studies for recreation reflects the relatively extensive literature on valuation of water in recreational uses. The database also includes a large number of values for water quality. Of the total 722 values classified as WQE, about 90% (651 values) also fall into other beneficial use categories in addition to WQE. This indicates that while a fairly large literature exists for water quality itself, valuation of water quality is more often conducted in the studies evaluating water quality benefits to other beneficial uses. Columns four through six present the number of values included in the database for each beneficial use category. A distinction was made between values that fall into a single beneficial use category (column four) and those that fall into multiple categories (column five). These are termed “single-use” and “multi-use” values, respectively. Summing columns four and five yields the total number of values applicable to each SWRCB beneficial use category; this is shown in column six. Of the 2034 values reported in the database, 1059 values fall into more than one SWRCB beneficial use category. The cross-classification occurs for two reasons. First, many studies estimated values for specific amenities that fall into more than one SWRCB beneficial use category. The WQE studies noted above are an example. Second, some studies provided the total economic value of some particular water body or system by aggregating values across various beneficial uses. Although values are sorted by SWRCB beneficial uses in the database, they should not be interpreted as representing the value of water in the beneficial use as defined by the SWRCB. It must be remembered that in many cases the values represent the worth of some specific function, activity, or amenity that water provides access to or enhances. The amenities, functions, and activities valued within any given beneficial use category are themselves often quite diverse. The SWRCB beneficial uses are used primarily for grouping the reported values in the database. It is also important to note that while types of amenities and activities are the most significant factors determining the economic values of water, large differences among the values within the same beneficial use category are also created by other factors. These factors include differences in data and methodologies used for valuation and the types of economic value measures and units used in presenting values. A Summary of Values by Beneficial Use Category The remainder of this section contains brief descriptions of the values reported for each searchable beneficial use category, the types of activities and amenities for which economic values of water were estimated, and other components of both values and studies that differentiate them from one another. 38 It is hard to develop a classification of values that applies to all searchable beneficial use categories because of the wide diversity of amenities, methods used for their valuation, specific valuation questions addressed, and so on. The summary discussions that follow attempt to highlight the relative numbers of value estimates within each category by the type of amenity valued, the type of value estimate (marginal values vs. average or total values)5, and the degree to which they overlap other beneficial use categories. Since this classification scheme is primarily oriented to value estimates, and individual studies often contain several types of value estimates, the counts of number of studies in the tables for each beneficial use are not mutually exclusive (though counts of value estimates are). Agriculture A total of 256 agriculture-related beneficial use values were collected from 17 studies. There are 61 values from 7 studies that reflect estimates of the value of water uses exclusively for agricultural purposes, and 195 multi-use values taken from 13 studies. The most common methods used to value water in these studies were optimization models and market-based valuation methods. Table 5.2 outlines the type of values, units, and number of values and studies for each of the main groups contained in the agriculture searchable beneficial use category. Table 5.2. Type and Number of Values in the Agriculture Searchable Beneficial Use Category Group 1) Single-Use (Agriculture Only) Type of Value (Units) Average values (dollars/ acre; dollars/ acrefoot of water) Marginal values (dollars/ additional acre-foot of water; dollars/ acre-foot of water; dollars/ acre-foot of water/ year; dollars) 2) Multi-Use Values Average values (dollars; millions of dollars; millions of dollars/ year; dollars/ acre-foot; dollars/ acre) Marginal values (dollars/ ton salt discharged; thousand dollars; dollars/ acre/ year; dollars/ acre foot; dollars/ megawatt; million dollars) 5 Number of Studies 3 Number of Values 28 Other Beneficial Uses 5 33 6 91 FRSH, WQE, MUN, WET, FLD, GWR, POW, REC1, and REC2 FRSH, WQE, MUN, WET, FLD, GWR, POW, REC1, and REC2 9 104 An average or total value represents the economic value of water given a specific resource condition, and a marginal value represents the incremental value to associated with changes in resource conditions. 39 In single-use studies, beneficial use values are represented in a variety of ways. These include the marginal and average value of quantities of water used in irrigation, water rights prices, and losses to farmers due to decreased water availability. Typically, the studies aimed at valuing water as a means of irrigating crops, determining its value by calculating the change in profits or net returns to the farm from using different levels of water under various water availability scenarios. Studies using the optimization/programming approach generally valued water as an input by estimating the production function to determine the value of marginal product of water as an input to production. Farm-level data were commonly used by these studies, and most apply to arid areas where irrigation is a key component of agricultural production, such as the Four Corners Area (Colorado, New Mexico, Arizona and California), Oklahoma, Nebraska, Kansas, and Utah. Information about year of estimation is often excluded in these studies; the ones including the year for which the value estimates apply are mainly from the 1970’s. The multi-use studies consist of three principal types: estimates valuing water quality changes and its effect on agricultural use; values arising from using wetlands or groundwater as water sources for agriculture; and values for water measured as the opportunity cost of competing uses. The first group calculates the value of agricultural damages caused by salinity levels in water sources. These salinity damages to agricultural production are estimated by translating physical effects from high saline content in water into economic costs for crop production (represented by a yield reduction). The second group of multi-use studies contains values for wetlands as a direct or indirect (through groundwater recharge) source of water for crop irrigation. Groundwater sources are also included in this sub-group. Values are estimated either as changes in farm profits calculated in mathematical programming models under different scenarios for water provision or by estimating a production function dependent on the level of water supplied (to get the marginal value of water to crop production). The third group of multi-use studies approximates the value of water in agriculture by the opportunity value of water in alternative uses, such as hydropower generation, habitat maintenance, and domestic (municipal) consumption. These values were obtained primarily using opportunity cost methods, optimization models and replacement cost methods. Chronologically, the studies in this sub-group were spread out, reflecting the fact that water allocation between competing uses has been and continues to be a major issue. Geographically, these studies were conducted primarily in the western United States, where water scarcity is common and allocation issues are continually in the spotlight. Aquaculture Thirty-two estimates of economic values of water for aquaculture and mariculture operations were collected from three studies. Table 5.3 summarizes the types and number of values reported in the database that fall into the AQUA beneficial use category. 40 Table 5.3. Types of Number of Values in the Aquaculture Searchable Beneficial Use Category Amenities valued Type of Value (Units) Sturgeon, roe, and trout production Average Value ($ per year) Visitation to fish hatcheries for educational and recreational purposes Average Value ($ per year per hatchery site; $ per year per visitor) Number of Number of Studies Values Other Beneficial Uses 2 14 1 18 REC The values in this category are classified into two groups. The first group consists of 14 single-use values of water in the production of aquatic animals and plants. Values in this group were obtained by either market valuation methods or simulated models, which use information on production costs and the market price of the aquatic plants and animals produced by aquaculture operations to generate value estimates. In particular, these studies estimated the net income or return to aquaculture operations (trout production, sturgeon meat, and roe production) using price, production, and harvesting information. These values are typically obtained by subtracting costs of physical inputs from the total revenues. They reflect the economic benefits created through aquaculture operations that are attributable to all the inputs for which market values were not well defined. Provided that water comprises a large share of these non-marketed inputs, these values are thought to represent the economic values of water used for hatchery operations. The second group, consisting of 18 multi-use values from one study, measures the recreational value of aquaculture operations. Values of this type represent the benefits that hatchery visitors gain from traveling and spending time at a hatchery without consuming aquatic animals and plants produced by the hatchery. Since markets are not well defined for this type of amenity, the values in this group were estimated either by contingent valuation method (CVM) or travel cost methods (TCM). In the CVM approach, the value of the time spent at the hatchery is estimated from data on visitor’s willingness to pay for the entrance permit to the hatchery site. In the TCM approach, the demand for trips to the hatchery is estimated from the data on the time and monetary costs that hatchery visitors spend in travel. Although values in both groups measure the beneficial uses of water in aquaculture and mariculture operations, each group’s values represent the economic value of water to different economic agents. More specifically, the values in the first group reflect the benefits to owners of production (producers) whereas those in the second group reflect the benefits to visitors of a hatchery (consumers). There are large differences in values between the two groups. These are due to differences in which economic agents are considered, the type of aquatic animals or plants produced at the hatcheries studied, valuation methods, types of data used for valuations, and units in which the value estimates are presented. 41 Commercial Fishing and Shellfish Production A total of 103 values were collected from 6 studies relating to the value of commercial harvesting of fish and shellfish. Of these values, 57 are single-use values for commercial harvesting of fish and shellfish (beneficial use category COMM) while the rest (46 values) are multi-use values that also encompass wetland, cold water ecosystems, and recreation uses. Table 5.4. Types of values reported for Commercial fishing and Shellfish production Group Type of Values (units) Commercial Fishing Only Marginal value ($; $ per acre; $ per acre year) Commercial and Recreational Fishing Average value ($ millions per year; $ millions; $ per lb. of shrimp; $ million per 30 years) Number Number of Studies of Values Other Beneficial Uses* WET 8 16 3 73 COLD, REC 1 Marginal value ($ per fish) * Indicates beneficial uses represented by some of the values in this group. 14 COLD, REC Cross-classification occurs because the SWRCB beneficial use categories of ecosystem functions and habitats are also places where commercial fishing sometimes occurs. Some studies estimated the economic values of water for commercial and recreational fisheries combined, which are each a beneficial use category. In Table 5.4, the commercial fishing values are divided into two groups depending on whether recreational fishing is also included in the values. The commercial fishing values are summarized here, while the recreation values are discussed in the Recreation values section. Values reported under commercial fishing include the value of commercial catch of seven fish and shellfish species: blue crab, brown and white shrimp, oyster, menhaden, pelt, salmon, and steelhead. These values were commonly derived from data on catch or harvesting rates and market prices of each species. These values were typically estimated using market valuation methods and simulation models. In the market valuation approach, values of water in commercial fishing are derived from estimated harvest or production functions and market values of fish. Both total (or average) and marginal values were derived using this valuation approach. Average or total values were commonly measured as the net return or annual rent to the fishery sector. The marginal value of an additional harvesting area was also calculated by multiplying the marginal product of a harvesting area for production of a certain species of fish by its market price. Values were typically reported in dollars per 42 acre of harvest area. While these values vary by the types of species and water ecosystems considered, they are largely dependent on market prices of the fish or shellfish species. Many studies provided more than one estimate by assuming different prices for the fish or shellfish species being harvested by the commercial fishery. In the simulation approaches, water ecosystems supporting habitat for fish and shellfish are valued using bioeconomic models and harvest-production functions. In particular, bioeconomic models were used to link habitat conditions, such as water quality or aquatic vegetation level, to carrying capacity. Harvesting functions and market prices of fish and shellfish are then used to derive the economic values of fish and shellfish harvests at a certain population level. While the economic values of fish harvest evaluated at certain habitat conditions provide total values, marginal values of water in commercial fishing are also obtained by use of simulation models. Marginal values in these models typically represent the changes in the economic values of a fish harvest given by a specific change in the habitat conditions. For example, one study estimated the marginal value of restoring seagrass beds that support habitat for hard-shell blue crab from the associated changes in surplus to consumers and producers of blue crabs. Regardless of the valuation methods, the reported values vary widely both within and between studies. This variation in reported values arises from several factors including differences in fish and shellfish species and ecosystem characteristics such as water quality, size, and aquatic vegetation considered by the studies. Moreover, economic values generated by the fishery sector are largely determined by harvest effort level and the market prices of fish and shellfish harvested. Recognizing this, many studies conducted sensitivity analyses and provided more than one value estimate under different assumptions about market prices and effort levels. Groundwater This category is comprised of 13 studies with a total of 93 values that measure the economic benefits from groundwater and related water use categories. Table 5.5 summarizes the type and number of values in the groundwater searchable beneficial use category. 43 Table 5.5: Type and Number of Values in the Groundwater Searchable Beneficial Use Category Group Type of Value (Units) Average values (millions of dollars; French francs/household/year) Number of Studies 2 Number of Values 8 Other Beneficial Uses 1) Groundwater only 2) Multi-use Values Municipal uses (GWR & MUN) Average values (dollars/year; dollars/household/year; dollars/ 1000 gallons; million dollars/ 1000 households) Marginal values (dollars/month/household; dollars/household/week; dollars/household/year) 3 12 WQE 3 19 WQE Agricultural uses (GWR & AGR) Average values (dollars/ acre foot) Marginal values (dollars/ hectare/ year; dollars) 2 37 2 15 WET, FLD Water quality enhancement (GWR & WQE) Average values (dollars/ household/ year) 1 2 Less than 10 percent of these values (8 values) measure water uses exclusively falling under the groundwater SWRCB beneficial use category. The remaining 85 multiuse values represent the value of uses falling under groundwater and other beneficial use categories. These include values of groundwater for agricultural production (which were already discussed in the Agriculture section), groundwater as a source of drinking or other municipal water supply, and benefits from groundwater quality preservation and enhancement. Several different approaches were used to value groundwater values. Values embodying only water uses for the natural or artificial recharge of groundwater (as per the definition of GWR by the SWRCB) were generally estimated using mathematical programming models (optimization models). CVM was used to determine the value of the quality and quantity of drinking water supply from groundwater sources. Other valuation methods were also employed in estimating the value of groundwater for consumption and for the preservation of groundwater quality. For example, the value of drinking water quality was approximated by the expenditures made 44 by households on measures to protect themselves against trichloroethylene contamination. Geographically, these studies took place in numerous locations throughout the United States, including Cape Cod (Massachusetts), Honolulu (Hawaii), Georgia, Pennsylvania, California, and the Northeast. All of the studies containing information about the year in which estimates were valued took place between 1989 and 1993. Habitats and Ecosystems The beneficial use values for habitat and ecosystems found during this search not only represent the environmental benefits of water associated with habitat and ecosystem functions, but also the benefits from consumptive or recreational uses of water supplied by these areas. In total, there are 209 values from 16 studies included in this broad category, which are summarized in Table 5.6. Table 5.6: Type and Number of Values in the Habitat/Ecosystem Searchable Beneficial Use Category Group or Sub-group Type of Value (Units) Number of Studies Number of Values Other Beneficial Uses 1) Ecosystems Only Average values (Australian dollars/ person/ year; Millions of Australian dollars) 1 5 2) Multi-Use Values Group # 1 Primarily ecosystem and recreation Group # 2 Consumptive, ecosystem, and recreation values Average values (millions of dollars/ year; dollars/household/year) Average values (millions of dollars/ 30 years; dollars/trip; dollars/day; dollars/person/day) Marginal values (dollars/person/year; dollars/trip; dollars/fish; dollars/acre-feet; dollars/cubicfeet second/ day; dollars/year ) Group # 3 Nonuse values Average values (dollars/person/year; dollars/person; dollars/household/year; English pounds/respondent; English pounds/ mailing) Marginal values (dollars/ megawatt of electricity; thousands of dollars) 2 16 REC, COLD, FRSH 4 49 COLD, COMM, REC, EST, WQE, FRSH, POW, MUN 6 76 5 48 MAR, RARE, BIOL, WILD, FRSH, RARE, 1 4 45 Almost all of these values measure the value of other beneficial uses for water in addition to its value in the habitat or ecosystem in question, with less than 3% of the values (5 values in one study) measuring benefits derived solely from uses of water that support the habitat or ecosystem. The one study that does value an ecosystem exclusively (i.e., where only one SWRCB beneficial use category is included) measures the benefit of preserving a conservation area in Australia without making reference to its uses other than as a habitat. In this study, the willingness to pay of Australian households to preserve a conservation zone in Northern Australia is estimated using CVM. The remaining values are considered multi-use values because they account for the value of habitats in their varied uses, such as the benefits they provide for recreation, water supply, and agricultural uses. These 204 values fall into 4 principal groups. Group 1 contains 16 value estimates from 2 studies measuring environmental and recreational benefits from water use. The second group contains 125 estimates from 6 studies that measure the consumptive benefits from water in addition to its use for recreational and environmental purposes. The third group includes 55 values taken from 7 different studies that value environmental benefits, present use values (e.g., recreation uses), and future use values (option and bequest values) from habitat preservation. The final group (not included in Table 5.6) is composed of 11 values representing benefits from wetland functions and is discussed in the Wetland and Floodplain searchable beneficial use section. Studies in the first group of values rely on travel cost and TCM and CVM to estimate the value of water in habitats. These valuation methods require use of survey data. In several studies, TCM models were used to measure the recreational benefits from water use in specific habitats as measured by the value of sport fishing at water bodies located in these areas. These types of values provide a measure of the recreational use value generated by the water-based habitats under investigation. An alternative approach used by some of these studies measures both use and nonuse values associated with recreation activities in certain habitats. These studies employ the CVM approach, which relies on individuals’ responses to survey questions asking them, in one way or another, to indicate their willingness-to-pay for an amenity. The second group of multi-use values includes the value of water from ecosystems used for consumption, commercial, and recreation purposes. In the 6 studies that belong to this group, the value of water in cold water and estuarine ecosystems is estimated. These values are calculated using a variety of methods, from non-market methods (for recreation uses), like CVM or TCM, to market-based techniques (for commercial uses). In addition, values associated with an increase in the quantity or quality of freshwater stream flow are estimated. The value of water for natural maintenance of surface water quantity and quality is expressed as benefits from the enhancement of recreational uses, consumptive and non-consumptive uses, and water quality. Again, use of non-market methods for valuing recreational uses and water quality improvements is common in these studies. However, other methods were also used, including replacement cost methods for measuring the benefits of stream flow increases necessary for salt dilution, hydroelectric energy production, and municipal and agricultural 46 consumptive purposes. The six studies from the third group primarily rely on CVM to measure both the use and nonuse value of ecosystems and species residing within the ecosystems. The lone exception is a study that measured people’s willingness-to-pay by the actual donations they made to a charitable organization that benefited bird habitat (Foster, Bateman, and Harley, 1998). Amenities valued in this group included several endangered or threatened species, whales, and a variety of ecosystems. For those studies that value representative species instead of the ecosystem per se, the assumption is that the value of the ecosystem can be approximated by the value people place on the species that reside there. The use of CVM allowed researchers to estimate both values associated with the use of a habitat or ecosystem and values independent of actual use (i.e., nonuse values). These nonuse values were sometimes termed existence value, option, or bequest values. All of the studies in the database representing the habitat and ecosystem searchable beneficial use category were undertaken in the last two decades, indicating the valuation of habitats, ecosystems, and species as a relatively new area of research. Industrial The industrial searchable beneficial use category includes values corresponding to PROC and IND beneficial uses. In these categories, there are 84 values from 6 different studies. Table 5.7 summarizes the number of values and studies contained in each group. Roughly half the values (44) represent economic benefits produced exclusively from uses of water for industrial activities that do not depend on water quality, but rather on the available quantity of water (IND). Also, four values represent the value of water in industrial uses where water quality is an integral part of production (PROC). Industries that do not depend primarily on the quality of water include the paper, chemical, steel, mineral, and hydropower industries. The benefits of water used in these industrial activities are measured using a variety of valuation methods including market, simulation, and damage function approaches. The remaining 40 are value estimates that fall into at least one other beneficial use category besides the IND category. These multi-use values can be divided into 3 main groups: benefits from water uses to industrial and municipal consumers (17 values), values for wetland water that provides services to the production process (7 values), and values derived from the maintenance or enhancement of water quality to industrial and residential consumers (12 values). 47 Table 5.7. Type and Number of Values in the Industrial Searchable Beneficial Use Category Group Type of Value (Units) Number of Studies Number of Values Other Beneficial Uses 1) Industry only (IND) Average values (cents/1,000 gallons of intake water; millions of dollars; dollars/acre-foot) 2 9 Marginal values (cents/1,000 gallons of intake water; dollars/ 1,000 cubic meters; dollars/ acre-foot of water) 5 35 2) Process Water only (PROC) Average values (dollars/ ppm of hardness/ acrefoot) 1 4 3) Industry and Municipal Marginal values (dollars/ 1,000 cubic meters; millions of dollars) 2 17 MUN, WQE 4) Industry and Wetlands Average values (dollars/ acre; dollars/ 15 years; dollars/ acre/ 25 years; dollars/ acre/ 15 years 1 7 WET 5) Industry and Water Quality Enhancement Average values (dollars) 1 11 WQE Marginal values (dollars) 1 1 Additional treatment and operation costs are faced by industries that depend not only on water quantity, but on water quality, for their production processes, especially in areas where water sources are affected by pollution problems or high contents of sediments and minerals. These pollutants can be detrimental to production processes. To derive economic values of water quality for industrial uses, damage function approaches are often utilized. Using these methods, the value of water quality in production is measured by the cost of repairing damage or restoring the loss to industry resulting from damage from poor water quality. Studies in the industrial searchable beneficial use category provide economic values for industrial water use in Vancouver (Canada), Mexico, Arkansas, California, Arizona, Maryland, Louisiana, Texas, Oklahoma, Southwest and the Intermountain region, as well as the United States nationwide. Most studies do not include information about the year in which value estimates were measured. The year of publication for the studies contained in this category range 48 from the 1960’s to the 1990’s, with the majority being published during the 1970’s and 1980’s. Municipal In the MUN beneficial use category, there are 246 economic values of water collected from 23 studies. Estimates in this category represent the value of water for drinking and other water supply system uses. 87 of these values were estimated for municipal water supply uses only, while the remaining 159 values were multi-use values. The latter type of values represent other beneficial uses in addition to MUN. The crossclassification of these multi-use values occurs mainly for three reasons. First, the values associated with the municipal uses of water, especially drinking water, depend on the water quality level. Second, some studies estimated the economic value of water sources, such as wetlands, aquifers, and rivers. These values typically fall into MUN as well as the SWRCB categories since water from these places are used both for in situ functions and as a source of water for municipal use. Third, some studies provided aggregate values across various functions provided by water in addition to municipal water use. These other functions include hydropower generation, flood control, industrial, and agricultural uses. Table 5.8 summarizes the types of economic values and number of values reported for MUN. Due to the large number values falling into the category, values were divided into the following four groups: municipal supply uses only, municipal use values dependent on water quality, municipal water supply from groundwater resources, municipal water supply function of water ecosystems, water treatment cost for municipal supply. 49 Table 5.8. Summary of Values Reported for Municipal Water Supply Types of Values (units) Number Number of Studies of Values Other Beneficial Uses Specific Amenities Municipal water supply only Average Value (days worth of labor per week; Korean Won per household per month; $ per acre per year; $billion per year; $ per household per year; $ per respondent) Marginal Value ($ per household per month or year; Rupees per household per month; $ per 1000 cubic meters) 6 29 4 58 Water quality and municipal water supply Average Value ($ per day; $ per million 1000 gallons; $) Marginal Value ($ per household per month; $ per year; $millions) 2 27 WQE 3 6 WQE Groundwater Average Value sources of municipal ($ per year; $million per 1000 households, $ per 1000 gallons) water supply Marginal Value ($ per household per week, month, or year) Water supply functions of ecosystems Water supply and competing uses Average Value ($ per year; $ per respondent per year) 3 12 GWR, WQE 3 19 GWR, WQE 2 4 FLD, WET Average Value ($ millions per year) Marginal Value ($ per ton salt discharge; $ million; $ per acre foot; $ per 1000 cubic meters) 1 10 AGR, POW, WQE AGR, FRSH, IND, POW, WQE 5 81 Values in the first group represent the beneficial uses of water for municipal water supply uses. Values of this type reflect the benefits that users of municipal water supplies receive at a specific level of accessibility or stability (of water supply), which are average or total values, or the incremental benefits from improved accessibility or stability of supply systems, which are marginal values. This group consists of 58 marginal and 29 average values collected from nine different studies. These municipal water supply values were estimated using several different valuation methods including CVM, simulation methods, market valuation methods, averting expenditures approaches, damage function approaches, and opportunity cost methods. 50 In the second group, economic values of water for municipal water supply were estimated that explicitly consider water quality. They represent the values that users of drinking water place on the supply of water at a specific quality level or at an improved (or degraded) quality level. 33 values of this type were collected from four studies, of which 6 values are marginal values and 27 values are average values. Three valuation methods used to estimate these values of drinking water quality levels are CVM, damage function approaches, and simulation models. Under the CVM approach, the value of improving the quality of drinking water or avoiding potential contamination of household drinking water was estimated from individual responses to willingness-to-pay questions. The third group is comprised of 31 values from five studies. These values represent the value of municipal water from groundwater sources, such as aquifers. The majority of these values were estimated using contingent valuation methods, but other methods, including averting expenditure approaches, were used as well. Four values collected from two studies make up the fourth group of values. These measure the value of wetland ecosystems as sources of municipal water supply. The two studies in this group used different methods to derive the water supply benefits of wetlands and hence estimated different types of economic values. One study used CVM to estimate the use and non-use value of New England wetlands that provide water supply as well as other functions such as flood protection and pollution control. This study estimated individuals’ WTP for protection of a wetland. This value reflects both the values of the water supply function of the wetland and other functions that wetland provides. In another study, the water supply benefits of a wetland were valued by its proportional contribution to net operating income of the local water supplier. The final group of values consists of 91 values collected from five studies. The values in this group represent the value of water for uses of municipal water that compete with municipal uses. These competing uses include agricultural consumption, industrial use, power generation, and freshwater replenishment for instream flows. Four types of valuation methods are used to estimate values. They are damage function approaches, optimization methods, simulated methods, and market-based valuation approaches. Navigation The database contains 33 values of beneficial uses of water for transportation purposes. These values were collected from 6 studies. A brief description of the number of values reported by type of value is provided in Table 5.9. Values in this category are divided into four groups differing by valuation approach. The first group of studies calculates the economic value of water for transportation uses in terms of the cost savings over alternative transportation methods. These values reflect the money saved by using water navigation over other forms of transportation. Many studies used a railway system as the cheapest alternative form of transportation to calculate these cost savings. Operation and maintenance costs of waterways were subtracted from the cost savings. The remaining value was attributed to the transportation by water and presented either as the value of the entire river system or as the value per volume of water. The latter reported value was obtained by dividing the total cost savings by the volume of water necessary for navigation. 51 Table 5.9. Types of Values Reported for Beneficial Uses of Water in Navigation Group Type of Values (units) Cost savings Average value ($; $ per acre foot) Average value ($ thousands per year) Marginal value ($ millions; $ per acre foot) Average value ($ per year) Number Number of Values of Studies Other Beneficial Uses 7 2 Replacement cost 18 1 Reduced pollutants from agriculture 7 2 WQE Erosion value 1 1 WQE In the second group of values, the replacement cost method was used. In the single study in this group, the value of a navigation project was measured in terms of the hydropower energy that is lost due to the use of water for transportation instead of power generation. Values derived from this method reflect the cost required to use water systems as a transportation medium instead of for other beneficial uses. Two other studies, representing the final two groups, valued the quality of water required for navigation. In the third group, the values represent the benefits of reduced pollutant discharges from cropland to navigable waters. And in the final group, there is a value from a study representing the cost for dredging waterways damaged by erosion. Overall, values were spread over a wide range. The wide variety of values can be attributed to the different locations of the navigation systems, methodologies used for valuation, types of alternate transportation used to calculate cost savings, and units used to present the values. Power The database contains 129 values for water used in hydropower generation and related water uses. A breakdown of the groups of values and number of values in each group is presented in Table 5.10. 52 Table 5.10. Type and Number of Values in the Power Searchable Beneficial Use Category Group Type of Value (Units) Number of Studies Number of Values Other Beneficial Uses 1) Power only Average values (thousand dollars/ year; dollars/ acre-foot of water; million dollars/ year) 3 22 Marginal values (dollars/acre-foot of water) 1 4 8 23 WQE, AGR, MUN 2) Multi-use values Average values (millions of dollars; dollars/ year) Marginal values (dollars/ ton salt discharges; dollars/ acre-foot; thousand dollars; dollars/ megawatt) 3 76 WQE, AGR, MUN, FRSH The water held in dams built for hydropower generation can be used for agricultural and municipal consumption, recreation, power production, maintenance of instream flows, and flood control, among other purposes. Thus it is not surprising that most studies in this category present multi-use values. Only 30 of the 129 values in this category measure water values derived exclusively for energy generation. These values represent hydropower uses in a variety of locations inside and outside the U.S. and are estimated in a variety of ways, from residual imputation methods to cost savings methods. The remaining values are multi-use values that represent uses for water in power production and other uses—water quality, municipal uses, agricultural uses, and for freshwater replenishment. The sediment content in water used in hydropower generation has implications for the power production process. The value of water quality in energy generation is sometimes measured by the costs of cleaning water utilized in hydropower generation due to sedimentation and salt discharges. Municipal consumption benefits of water from dams are estimated using a optimization modeling framework for the Colorado River Basin. The final two groups of values contain the benefits derived from freshwater ecosystem preservation and stream flow increases for hydropower generation. The multi-use values associated with agriculture and freshwater replenishment are discussed in other sections. Overall, the studies contained in the Power searchable beneficial use category were conducted between 1972 and 1994, with the majority being conducted in the late 1980’s and early 1990’s. 53 Rare A total of 218 estimates from 16 studies are contained in the Rare searchable beneficial use category. Economic values of primarily rare, threatened, or endangered species make up this searchable beneficial use category. These values fall into three basic categories: values for species and their habitats, the public’s willingness-to-pay for the protection or preservation of species, and hunting values of species. Twenty-six values from six studies are for species and their habitats, which include freshwater, marine, and wetland ecosystems and other areas of special biological significance. These estimates have been discussed in other sections discussing these types of habitats. The remaining 192 value estimates for rare, endangered, or threatened species come from 9 documents and cover both preservation and hunting values. A brief summary of the values for each group is given in Table 5.11. The 26 values of species in special habitats are omitted from this table as they are included in other categories. Table 5.11. Type and Number of Values in the Rare Searchable Beneficial Use Category Group Type of Value (Units) Average values (thousands of dollars/ year; dollars/ person; millions of dollars; dollars/household/year; dollars/ person/ year; English pounds/ person/ year) Number of Studies 7 Number of Values 132 1) Benefits from preservation -Bald Eagle & striped shiner (Wisconsin) -Whale & monk seal (Hawaii) -Owl (Washington) -Whooping crane species (Aransas Refuge) -Bald eagle, coyote & wild turkey (New England) -Water vole & otter (Yorkshire, U.K.) 2) Benefits from hunting Average values (dollars/ person/ year; dollars/ person/ trip) 1 52 -Grizzly bear & bighorn sheep (Wyoming) -Elk (Montana) Marginal values (dollars/ person/ trip) 1 8 For both preservation and hunting values, CVM is the primary means for obtaining estimates of the value of species. The value of species is found either by measuring the willingness to pay for species preservation or protection directly or by 54 determining the willingness to pay for hunting of species and thus obtain a proxy for its preservation value. Some of the species for which preservation and protection values are estimated include spotted owls in California and the northwest, whales and monk seals in Hawaii, bald eagles, striped shiners, Atlantic salmon, coyotes, wild turkeys, otters, water voles, and whooping cranes. Hunting values, used as a proxy for the value of the species per se, are estimated for grizzly bears, elk, and bighorn sheep in Wyoming and Montana. Species being valued through their use in hunting are not Listed species under the Endangered Species Act. Recreation A wide variety of recreational activities take place in water-related environments. One useful way to classify these recreational activities is to divide them in terms of the level of contact with water associated with these activities. Beneficial use categories used by the SWRCB follow this classification scheme, dividing water-related recreational activities into those involving direct body contact with water (REC1) and those not normally involving body contact with water (REC2). This classification would be appealing if recreational values of water were easily classified by the degree of contact with water. However, we discovered it was not the most practical classification scheme for the database project for several reasons. First, none of the studies selected from the literature distinguished between contact and non-contact forms of recreation. While many studies evaluated water in specific recreational activities such as fishing, whitewater river rafting, and swimming, none provided sufficient information on the level of contact with water associated with the activities under investigation to be classified in this manner. Although the level of body contact with water can be inferred for some recreational activities, for others it is unclear. Fairly clear examples include swimming in a river or at a beach, which fall into REC1, and wildlife viewing or hiking in areas surrounding a river or wetland, which will generally fall into REC2. On the other hand, for activities such as fishing, recreational boating, water rafting, and general beach recreation, the extent of body contact with water will vary. It is difficult to classify values of this type as contact or non-contact without more detailed descriptions of the activities. Second, several studies estimated the economic values of water for general recreation purposes. In these studies, economic values of water in recreational uses were estimated for a particular site without specifying the types of recreational activities undertaken by individuals. For these reasons, recreational values of water were not categorized by the level of contact with water. Instead, all values were placed into a single category, REC, and divided into specific recreational activity values when possible. A feature common to most recreation values is that they are multi-use values, falling into one or more beneficial use categories besides recreation. This occurs for two reasons. First, recreational activities are conducted in a variety of water environments such as rivers, wetlands, beaches, estuaries, and lakes. Since the database includes categories for uses of water that support different water ecosystems, many values reported under recreation also 55 fall into one or more of these categories in addition to REC. Second, the levels and types of water-related activities are subject to the water quality level, and consequently values of this type fall under both REC and WQE. The database contains 962 estimates of economic value of water in recreational uses collected from 50 studies. Due to the large number of recreational water values contained in the database and the range of recreational activities considered, recreational use values are split into the following six categories: (1) recreational fishing (including water quality benefits for fishing), (2) wetland recreation, (3) hunting, (4) beach recreation, (5) water quality benefits to recreational activities (other than fishing), and (6) other recreational activities. Table 5.12 presents an overview of the six categories, which in turn are then discussed in separate subsections. Table 5.12. Beneficial Use Value of Water by Types of Recreational Activities Groups Types of Activities Number Number of Values of Studies 339 17 (1) Recreational fishing Recreational fishing of anadromous species, steelhead, rainbow trout, shore and private boat fishing, fish species not specified. Recreational fishing for a range of stream flow levels, Recreational and commercial catch of brown shrimp, Water quality benefits to recreational fishing. Beach recreation, Water quality improvement. (2) Beach recreation 246 6 (3) Hunting Deer and waterfowl hunting and congestion. 60 3 (4) Wetland recreation Recreational use value, visual cultural use value of wetland, Wetland habitat preservation. Water quality benefits to beach recreation, freshwater recreation, water-related recreational activities (types of activities not specified), Multiple recreational activities including boating and beach recreation. 37 8 (5) Water quality benefits to non-fishing recreational activities (6) Other recreational activities 184 10 Recreation benefits (types of amenities not specified), Whitewater rafting, Recreational benefits of increased stream flow, Protection of minimum stream flow, Visitation to fish hatcheries. 96 10 56 • Fishing Fishing is the category with the largest number of value estimates, 339, collected from 17 studies. They are summarized in Table 5.13.and classified further into 3 subgroups. The first subgroup includes values that measure the combined benefits of recreational and commercial fishing. Values in the second and third subgroups value recreational fishing only, but are distinguished from one another by whether or not the value explicitly considers water quality. In the second subgroup, water quality is not explicitly modeled to obtain the economic value. Values in the third subgroup are dependent upon water quality levels. Table 5.13. Summary of Values Reported in Database for Recreational Fishing Amenity Valued Type of Values (Units) Number Number of Studies of Values 2 14 Other Beneficial Uses COMM, COLD Commercial and Average Value ($ millions per year, $ millions per 30 recreational fishing years) Marginal Value ($ per fish) Recreational fishing Marginal Value only-- no water ($ per additional fish, $ per trip per quality person) Average Value ($ per person, $ per trip per person, $ per day per person, $ per fish, $ per total sample visits, $ per 12-hour Wildlife and Fish User Day per person, $ millions per season, $ per year, $ millions) Water quality Average Value ($ per day, $ per person per day) benefits to recreational fishing • Fishing benefits Marginal Value ($ per acre foot, $ per cubit feet second per at different day) stream flow levels • Water quality benefits to recreation fishing Average Value ($ per Capita, $ per fisherman, $ millions, $ per trip, $ millions per 30 years) Marginal Value ($ per trip, $ per angler per season, $ thousands, cents per fishing trip, $ per year, $ millions,) 1 14 COMM, COLD 4 66 4 105 2 18 FRSH, WQE 2 31 FRSH, WQE 3 51 COLD, EST, WQE 6 40 EST, WQE 57 While the methods used for valuation varied among the studies, values in the same subgroup tended to be estimated using similar valuation methods. For instance, recreational and commercial fishing water values (values in the first subgroup) were primarily estimated using market-based valuation techniques. On the other hand, recreational fishing benefits (values in the second and third group) are almost exclusively estimated by non-market valuation techniques, particularly CVM and TCM. These two methods were employed in the majority of the 17 studies reporting recreational fishing beneficial use values. In the studies using the CVM approach, the maximum amount anglers are willing to pay for access to fishing sites is elicited in questions contained in mail, telephone, or on-site surveys. In studies analyzing the effects of water quality change on recreational fishing values, the WTP for various fishing quality changes such as stream flow levels, catch rates, and water quality levels was estimated. In the TCM approach, a demand function for fishing trips is typically estimated. Like all demand functions, the TCM demand function depends upon travel costs, types of fishing activity, and socioeconomic variables such as respondents’ income, residence, and fishing experiences, and quality levels. The estimated demand functions were used to obtain estimates of the benefits that individuals receive from fishing at specific sites. Regardless of the methodology used, many studies estimated more than one value. In most cases, the different values reflected different model specifications (e.g., linear or log-linear functional form), statistical estimation techniques (e.g., OLS or MLE), and resource being valued. Values were also presented in different units. In the majority of studies, a value estimated by the same model specification, estimation technique, and under the same resource conditions was converted from one unit (e.g., value per day per person) to another using fishing frequency data (such as fishing days per trip, days per season, fishing hours per day) and demographic variables (such as average household size, number of fishing license holder per household, and number of visitors per season to the fishing site). • Beach Recreation The database contains 246 beneficial use values associated with beach recreation collected from 6 studies. These values can be divided into two groups, based on whether water quality changes are considered. Table 5.14 summarizes the types of values reported in the database for each of the two groups. In the first group, 47 beach recreation values that do not explicitly model water quality in estimation were provided by three studies. These values represent the benefits that individuals derive from using beaches for recreational activities and were derived using CVM or TCM. There is a wide variation in the range of the reported values, even those from the same study. For example, one study reported the average consumer surplus estimates of 0.9 to 255.3 cents per person per day for recreational beach users in the Boston area. 58 Table 5.14. Types and Number of Values Reported for Economic Values of Water in Beach Recreation Amenities Valued Type of Value (Units) Number Number Other of Studies of Values Beneficial Uses 3 Beach recreation Average Values ($ per person per day; $ per person per trip; cents per visitor day; cents per household; $ per user per month; $ per household per month) Water Quality Benefits to Recreational Beach Uses Average Values ($ per year; $ per person per trip) Marginal Value ($ thousands; Philippine pesos per month per household; $ per occasion; $ per season; $ per person per day) 1 8 WQE 47 4 191 WQE This variation in values, whether observed between values in the same study or between values in different studies, may be caused by differences in locations, methodologies, and data used for valuation. Unfortunately, neither the conditions under which recreational uses of beaches were evaluated, such as water quality levels, nor the types of recreational activities considered in valuation were clearly presented in many of these studies. This makes it difficult to address how the beach recreational uses values of water vary among the assortment of activities conducted on beaches and how they are affected by water quality and other factors. Values in the second group represent beach recreation values of water that were derived from a model incorporating water quality. Average or total values represent the economic value of water in beach recreation measured at a certain water quality level, whereas marginal values represent the incremental benefits to recreational beach users associated with some change in water quality or other condition. Due to the lack of well-defined markets for improved water quality at beaches, all studies in this group employed non-market based valuation techniques (TCM or CVM) that incorporated water quality variables into the models. In the TCM approach, demand for trips to a beach was specified as a function of travel costs, water quality, and demographic variables. Values of recreational beach activities were derived from the estimated demand function. Water quality benefits were derived as the difference in the recreational use benefits (change in value of a beach trip) evaluated at two water quality levels. In the CVM approach, respondents were asked to indicate the amount they would be willing to pay for water quality improvements (or equivalently, a water quality improvement program). The value of beach water improvement was then statistically estimated from the survey responses. There is a large range in the reported values both between the studies and within a single study. The variance between values in different studies is largely due to the differences in locations, methodologies, units that estimates are presented in, and data 59 used for valuation and the types of recreational activities considered by each study. The variation between values in the same study is attributable to the various types of water pollutants and the magnitudes of water quality improvements (including initial water quality level) considered by each study. • Hunting A total of 60 water-based recreational hunting values were collected from three studies. Table 5.15 summarizes the types and number of values reported in the database. Table 5.15. Summary of Values for Beneficial Uses of Water for Recreational Hunting Amenities Valued Type of Value (Units) Average Value ($ millions; $ per person per year; $ per person per trip; $ per refuge per year) Marginal Value ($ per acre foot) Deer hunting Average Value ($; $ per person) Number of Number of Values Studies Other Beneficial Uses WET Waterfowl hunting 32 2 12 16 1 1 All three hunting studies employed non-market based valuation techniques to estimate values. In particular, two studies reporting economic values of water in waterfowl hunting used the TCM, while one study estimated the value of water in deer hunting used the CVM. In these studies, data were primarily collected either through onsite, mail, or telephone interviews and used to estimate the benefits associated with the participation in recreational hunting activities. The reported values vary both between and within studies due to the differing population demographics of hunting sites considered, types of economic value measures used, and methodologies and data used for estimating these values. It is important to note that these values, like others in the database, represent the value of the activities undertaken by individuals that depend either directly or indirectly on the availability of water and its quality. Thus, they do not represent the values of water per se; they can loosely be considered as indirect estimates of the economic value of water. • Wetland Recreation Thirty-seven wetland recreation value estimates were collected from 8 studies. Table 5.16 summarizes the types and number of values, which are grouped by the amenities for which the beneficial uses of water were evaluated. The 23 values in the first of the two groups are contained in five studies and represent the non-use or existence value of wetlands. Non-use or existence values are the benefits that individuals receive 60 from knowing an amenity exists in a certain condition without directly using the amenity. Since values of this type typically cannot be isolated by observing market transactions, studies estimating these values employed a non-market based method, the CVM, to obtain individual WTP for preservation of wetlands under a specific set of conditions. The values estimated by this method are commonly presented as the individual or aggregate WTP for preservation of wetlands. Table 5.16. Summary of Values for Beneficial Uses of Water for Wetland Recreation Amenities Valued Types of Values (Units) Average Value (English pound per year; $ per household per year; $ per acre per year; $ per person per year; $ per year; $ millions per year) Average Value ($ per year; $ per acre per year; $ per person per season; $ per person per trip; $ per acre; $ millions) Marginal Value ($ per household per year) Number Number of Values of Studies Other Beneficial Uses WET Non-Use (Preservation) Value 23 5 Recreational Use Value 13 3 WET 1 1 Values in the second group, contained in three studies, represent the recreational use value of wetlands; that is, the benefits that visitors to wetlands receive from participating in recreational activities at wetlands. Like the first group, studies estimating values of this type rely on non-market valuation techniques. In particular, both CVM and TCM were used by these studies. None of the studies specified the types of recreational activities undertaken by individuals at wetlands. This is probably because the recreation data collected in these studies does not include activity-specific information. Consequently, the value of individual activities cannot be isolated from the estimated trips demand and WTP functions. As noted under the hunting values discussion, these recreational use values are indirect estimates of the economic value of water since it is the value of the activities being valued that are being used as a proxy for the value of the water itself. The observed variation in the reported values both between studies and within each individual study can be attributed to factors such as the locations of wetlands, economic value measures reported, and methodologies and data used for estimating these values. • Water Quality Benefits to Recreation Activities Of the 962 recreation values reported in the database, 523 were estimated with explicit consideration of water quality levels. Table 5.12 counts only the 184 water 61 quality studies that were not fishing-related to avoid double-counting; the discussion here pertains to all 523 water quality-related values. These values represent either the economic values of water in recreational uses at a specific water quality level (which are average or total values) or the incremental values to recreational waters associated with changes in water quality levels (which are marginal values). Because of the wide variety of recreation activities considered in these studies, we focus on four aggregate categories—fishing, beach use, other specified activities , and unspecified recreation. The latter two categories are necessary to account for those values associated with either specific non-fishing and non-beach activities or values corresponding to “general recreation” at a site. The number of values and studies corresponding to the four groups of recreational activities (fishing, beach use, boating, and unspecified activities) that explicitly account for water quality are summarized in Table 5.17. Table 5.17. Summary of Values for Water Quality Benefits to Recreation Recreational Activities Types of Values (Units) Number Number of Studies of Values Other Beneficial Uses COLD, EST, FRSH, WQE EST, FRSH, WQE WQE WQE WQE Fishing Average Values 5 69 Marginal Value Beach 9 71 Average Values Marginal Value 1 5 1 ($ thousands) 8 191 1 Boating Marginal Value Unspecified Recreation Average Value ($ per person; $ millions) Marginal Value ($ per person; $ per person per year; $ per day; $ per household per year; $ thousands; $ per person per season; $ millions) 1 7 WQE 10 176 WET, WQE Approximately 65 percent of all of values in Table 5.17 are for two specific water-related recreational activities, fishing and beach recreation. Recreational boating was the only other specific activity for which water quality-related recreation benefits were estimated. The remaining values were estimated for more broadly defined recreation; i.e., for general recreation benefits at a site without specifying the types of recreation activities available or participated in at the site. Brief descriptions of the fishing and beach recreation values are given in previous sections. Hence, the following discussion will focus on boating and general recreation values. 62 Table 5.18 lists the water quality changes and the economic value measures used to represent the water quality benefits to general recreational users and boaters. The first row of the table presents these characteristics for the water quality improvements to recreational boaters. This value represents the marginal value of water quality improvements defined in the second column as the incremental value that 496 boaters place on the use of water for recreational boating when water quality is increased by 20% (one value from one study). The second row of the table displays the water quality changes valued and the economic value measures used to present the water quality benefits accruing to general recreational users or to participants in specific non-beach and non-fishing activities. The database contains 183 value estimates of this type. One hundred seventy six of these represent the marginal value of water quality benefits, the change in recreational use value with an increase or decrease in the quality of water available. The remaining 7 average or total value estimates represent values that individuals place on recreational uses at current water quality levels. Table 5.18. Economic Value Measures and Water Quality Scenarios Considered in Valuation of Water Quality Benefits to Recreation Amenities Valued Boating Economic Measure Used Water Quality Improvement Scenarios 20% improvement in water quality Aggregate benefits to 496 boaters of water quality improvement of a bay. Average Value Total recreational benefits of maintaining current water quality, Mean WTP per current user for recreation trips to a bay, Marginal Value Change in use value associated with improvements in river water quality, PV off-site water quality benefits to freshwater recreation of the reduced pollutant discharges from croplands, Total recreational benefits of improved water quality, Total and household adjusted annual aggregate benefit of water quality improvement, Population and sample mean annual WTP for river water quality improvements, Mean WTP to avoid a decrease in river water quality, Mean WTP by recreational users and non-users for water quality improvements, Mean and total WTP for water quality changes. Multiple Recreational Activities or General Recreation Unspecified, Poor to fair, Fair to good, Good to excellent, Non-boatable to boatable, Non-boatable to swimmable, Boatable to fishable, Boatable to swimmable, Fishable to swimmable, Unswimmable to swimmable, Status quo to swimmable, Current to the level equal to the most locations outside the study area, Current to minimum swimmable, Current to boatable, Current to fishable, Current to swimmable. Many different water quality levels and changes in these levels were considered in these studies. In four studies, water quality levels were defined in terms of their suitability for water-related recreation activities, such as boating, fishing, and swimming. 63 In other studies, water quality levels and changes were defined in more general terms (e.g., fair to good and good to excellent water quality). All of these studies used either CVM or TCM to estimate these values. In studies using the CVM, the value of water quality improvements was estimated for users, and at times for nonusers. The TCM was used to estimate the demand for trips to recreational sites as a function of water quality levels. Every study estimating the marginal benefit of water quality improvement considered more than one improvement level. As one would expect, the estimated values vary across the levels of water quality changes being valued. Comparisons among the estimated values across the water quality levels suggest that larger values are associated with relatively large water quality improvements. • Other Activities (Including General And Multiple Recreation Activities) The remaining 96 recreational total or average values are economic value estimates of beneficial uses of water in multiple recreational activities, general recreational activities, and non-fishing and non-beach activities, which are referred to as “other” activities. These recreation values differ from the recreational benefits of water quality since they do not explicitly account for water quality levels. Table 5.19 provides a brief summary of the types and number of values for each of these 3 groups. Table 5.19. Number of Values and Studies Reporting Beneficial Use Values of Water in General, Multiple, and Other Recreational Activities Amenities Valued Type of Values (Units) General Recreation (Types of Average Value activities not specified) ($ person per day, $ per user group per visit, $ per visitor day, $ thousands, $ per person per year, $ per year, $ per person per visit, $ per household per year) Multiple Recreational Activities Average Value ($ per person per year, $ millions) Number Number of Studies of Values Other Beneficial Uses FRSH 5 35 1 24 Other Activities (Bear watching, White Water Rafting, Visitation of fish hatcheries) Average Value ($ per year, $ per person, $ per visitor) 4 37 AQUA, BIOL, RARE The first group of values came from five studies and were for beneficial uses of water in recreation that did not specify the types of recreational activities involved. Instead of focusing on particular recreational activities, these studies estimated sitespecific values, i.e., the recreational values of water systems at which recreational 64 activities are engaged. In particular, these studies provide estimates of the site-specific recreational value of three types of water resources—mountain reservoirs, lakes, and rivers. Most values were obtained using CVM and TCM methods. One study employed local economy analytic methods (e.g., income multiplier method) to obtain the economic impact of a recreation site by analyzing the effect of expenditures by visitors on the local economy. The second group of values is from a study that valued combinations of specific recreation activities. This study used a TCM model to estimate the economic value of water in three recreation activities—waterfowl hunting, wildlife viewing, and recreational fishing—at wetlands in California. The third group of values includes 4 studies that estimated the value of water in recreational activities other than those just described. Three specific activities for which recreation values were included in the database were whitewater rafting, bear-watching, and visitation to fish hatcheries for educational and recreational purposes. Like most of the recreation valuation studies, these studies used CVM and TCM to generate estimates of the value of these activities. Water Quality The set of SWRCB-designated beneficial uses includes a separate category for the uses of water for water quality enhancement. As many of the discussions of other categories of beneficial use values indicate, water quality and changes in its level affect a wide range of water uses. That is, the quality of water often directly affects other beneficial uses. For instance, an inordinate amount of debris in a river may preclude navigation or high bacteria counts in water may force the closure of beaches to waterbased recreation. In less extreme cases, it is reasonable to expect water quality levels to affect individuals’ valuation of activities and uses associated with water use. Consequently, the value people place on uses of water often changes as the quality level changes. Enhancements or improvements of water quality are evaluated in various contexts within the existing literature. The current database contains 722 values that fall into the WQE beneficial use category. Of these 722 values, more than 80% are water quality benefits to recreational activities and municipal water users, which were described in previous parts of this section. In this section, we focus on the 133 values not related to recreation or municipal uses of water. These values cover a variety of water bodies and water-related activities and amenities whose value is affected by water quality levels. In Table 5.20, these values are divided into 6 groups. Most of these values were multi-use values, representing values that spanned multiple uses. Of the multi-use values, a large proportion fall under recreational or municipal uses in addition to water quality enhancement, WQE. 65 Table 5.20. Summary of Values Reported for Water Quality Amenities valued Type of values (units) Number Number of Values of Studies Other Beneficial Uses Water Quality Impacts on Property Values Marginal value ($ per residence; $ per acre; $ per lot; $ per property) Marginal value ($ per person per year; $ per year; Philippine pesos per household per month) Average value ($ per household per year; $ thousand per year) 22 5 Nonuse Value of Water Quality Benefits to Various Water Systems 23 6 EST, FRSH 11 2 GWR Water Filtration/Purification Function of Ecosystems Average value ($ per riparian buffers; millions Swedish Krona per year; Swedish Krona per kg of nitrogen reduction) Average value ($ per year; $ millions; $ millions per year) Marginal value ($ per acre of cropland; $ millions per region; $ per acre foot) 15 2 WET Other Uses 36 4 AGR, IND, POW, NAV 26 2 AGR, FRSH, IND, POW, WQE The first group of values in Table 5.20 is the set of values for water quality impacts derived from property value studies. Five studies provided 22 estimates of the value of water quality changes as reflected by changes in the property values located on or near lakes, rivers, streams, estuaries, and bays. These values were measured by either the changes in average property sales values in one location or the differences in the property sales values of the two similar properties in similar locations with different water quality levels. Hedonic price methods were used to estimate these values. The general approach suggested by hedonic price theory involves regressing property sales values water quality variables, variables representing various features of properties, and demographics. The resulting hedonic price function relates how the price of the property changes with water quality, among other variables. Values vary among studies and even among those reported in a single study. Factors differentiating values included geographic locations of studies, measures of water quality, the degree of changes in water quality, and base water quality levels. Other factors contributing to the variation in reported values include types of properties studied (e.g. residential and rural land) and price variables used in the models. The second group of values was from six studies that provided 23 nonuse values of water systems evaluated at current or improved water quality levels. In this context, 66 nonuse values reflect the values that individuals or households place on a certain water body at the current or an improved water quality level without directly visiting or using water from the water body. These nonuse values were estimated by CVM, one of the few methods capable of generating estimates of nonuse values. Using this method, the willingness-to-pay for maintenance or improvement of water quality is elicited from users and nonusers by survey research techniques. The estimated WTP by nonusers is a direct measure of the nonuse value of the lake, river, estuary, groundwater resource, or other amenity. The third group of values consists of 15 values from 2 studies representing the value of water quality enhancement as an ecosystem function. These studies estimate the value of nitrogen purification functions of a wetland system and pollution control functions of riparian buffers using replacement cost techniques. Since the market for these functions is not well defined, the value of water purification and pollution control functions was deduced from the costs of attaining the same level of improved water quality by some alternative technologies. For example, the nitrogen purification function of wetland system was approximated by the cost of attaining the same nitrogen reduction level by a man-made filtration plant. In another study, the values of pollution control functions of riparian buffers were approximated by the costs required to reduce the pollution discharges from farm production. As mentioned earlier, water quality affects many uses. In the fourth group, values represent a variety of beneficial uses, including industrial and agricultural uses, hydroelectric production uses, and water storage uses. Values in this group were estimated by a variety of methods. The damage function approach was used by two studies to estimate the water quality benefits and salinity damages to water users. In this approach, water quality changes were first measured by their effect on physical or biological amenities and then converted into monetary values by using existing marginal value estimates of each amenity. Simulation models were used by two studies to characterize the economic behavior of water users and simulate the effect of various water quality and allocation scenarios. These studies provided estimates of the value of erosion control, industrial water pollution and salinity control, and water quality on hydroelectric production. Wetland and Floodplain There are 221 distinct values from 21 studies in the Wetland and Floodplain searchable beneficial use categories in the database. The Wetland and Floodplain SWRCB categories were included together in this searchable beneficial use category, and not in the Habitat and Ecosystem category, because of the considerable amount of research done specifically on valuing wetland and flood control benefits. These beneficial use values measure the value of waters that support wetlands and floodplains, and their functions, such as flood control benefits. Wetlands and floodplains provide both direct (e.g., ecosystem functions) and indirect (e.g., recreation) services. Individually, the WET category has 217 values, while FLD has 18. Summing the totals in each category leads to an inaccurate number of values because 14 of the values were categorized as falling into both WET and FLD. The values contained in the database under this beneficial use category, excluding 67 those that are discussed in the Agriculture section (50 values), are presented in Table 5.21. Table 5.21. Type and Number of Values in the Wetland/Floodplain Searchable Beneficial Use Category Group or Sub-group Type of Value (Units) Number of Studies Number of Values Other Beneficial Uses Direct Services 1) Wetland Only Average values (dollars/ person/ year; Australian dollars/ person/ year for 5 years; millions of Australian dollars/ year for 5 years; Australian dollars/ household/ year; English pounds/ household/ year; dollars/ acre of wetland; millions of Swedish Kronas (SEK)) 8 24 Marginal values (dollars/person/year; dollars/acre; SEK/ discharge reduction of 1 kg of Nitrogen) 5 15 2) Floodplain Only Average values (dollars/ acres /year) 1 3 Marginal values (millions of dollars) 1 2 1 9 3) Wetland and Floodplain (WET & FLD) Average values (millions of dollars; millions of dollars/year; dollars/acre; dollars/acre/year) Marginal values (millions of dollars) 1 2 Indirect Services 1) Recreation, commercial, consumptive, and water quality wetland benefits Average values (English pounds/household/year; dollars/ household/year; dollars/acre /year; dollars/refuge/year; SEK/kg N; millions of SEK/year dollars/person/season; millions of dollars/year) 9 78 REC, COMM, MUN, IND, WQE Marginal values (dollars/ ha of cropland/ year; dollars; dollars/person/year) 5 29 2) Wetland and other ecosystems Average values (dollars/acre/year; dollars/respondent/year) 2 10 MAR, EST, FRSH, RARE 68 Of the 221 values in this category, less than 25% (54 values) include only direct benefits for either wetlands or floodplains. These fall into three classes: values that measure direct benefits from uses of water to support wetland ecosystems or their in situ functions (WET), values of floodplains specifically as buffers for natural surface drainage (FLD), and values that measure the combined (multi-use) value of wetland and floodplain ecosystem services (FLD and WET). Estimates of the value of indirect services provided by wetlands and floodplains can also be divided into three groups. The first group includes values derived from wetlands as a source of recreation, commercial use, consumptive use, and water quality enhancement benefits. Values in the second group measure the environmental benefits associated with wetland ecosystems in combination with related ecosystems or species. The third group (not listed in Table 5.21) includes values for agricultural uses of water from wetlands, which are discussed in the Agriculture section. Direct benefits from uses of water that support wetland ecosystems and their functions are typically measured using energy analysis, CVM, or replacement cost methods. In the energy analysis method, gross primary production of wetlands is treated as an index of solar energy captured and then transformed into fossil fuel equivalent to obtain an estimate of the value of the biological energy (and hence the total potential biological value) available in the system. CVM is used to value wetland preservation or enhancement directly through survey questions, while replacement cost methods measure the value of a wetland by the cost of building, maintaining, and using substitutes that provide the same services as the wetland. For instance, the value of the filtration and cleansing functions of wetlands are approximated by the cost of using water treatment plants instead of the wetlands. Wetland value estimates in this group in the database are primarily values of preserving or restoring wetlands in various parts of the U.S. and abroad (particularly England and Australia). The two studies containing estimates of the economic value of floodplain benefits and the three studies that value both wetland and floodplain benefits all employ the damage function approach to value benefits. Flood control benefits are the value of preventing flood damage, using expected damage estimates calculated from data provided by the U.S. Army Corps of Engineers and the U.S. Department of Agriculture. Wetland and flood control benefits are measured for areas in Louisiana, Minnesota, South Dakota, and other areas in the U.S. using secondary data sources. Recreation can be viewed as a by-product of wetlands since wetlands often provide the settings and resources suitable for wildlife viewing, waterfowl hunting, and other recreation activities. The value of wetland recreation in the studies in the database is calculated by the replacement cost, TCM, and CVM, utilizing data from surveys and secondary sources. These studies value recreation at wetlands in England, Kentucky, Minnesota, California, and Louisiana. Commercial fishing and fur trapping activities occur in some wetlands. The value of these activities adds to the value of the wetlands and can be measured using conventional market-based analysis since price and quantity data for these markets are generally known. Some fisheries valued in these studies include brown and white 69 shrimp, menhaden, oyster, and blue crab. In addition, muskrat and nutria pelts obtained in wetlands contribute to the value of the wetlands and are measured as well. Wetlands are also sources of water for municipal and industrial uses. In addition to supplying water for consumptive municipal purposes, industries often need to treat wastewater. Since wetlands have the capability to filter and purify water, they have inherent value to industries that have access to wetlands for this purpose. The benefits of water supplied by wetlands for municipal and industrial uses are measured for wetlands in Minnesota, New England, and Louisiana. The value of wetland water for municipal uses is measured using the damage function approach, while the value of wetlands as substitutes for wastewater treatment plants is valued for various industries located near wetlands in Louisiana using the replacement cost method. Additionally, improvements in water quality provided by wetlands are estimated using CVM for wetlands in Illinois, Iowa, and Sweden. Some studies in the database valued environmental benefits from marine, estuarine and freshwater ecosystems geographically and ecologically related to wetlands. The values generated in such studies embodied both wetland benefits and the benefits associated with other ecosystem beneficial use categories. Marine habitats valued include salt marsh habitat, salt aquatic habitat, and coastal plankton habitat; estuarine habitats include brackish aquatic habitat, spoil banks habitat, and brackish marsh habitat; and freshwater habitats include fresh marsh habitat and fresh aquatic habitat. The value of species in wetland environments, particularly in New England, is estimated using CVM. The studies in this category were conducted between the mid-1980’s and the mid1990’s. CVM and TCM were the most common methods used from the studies from the 1990’s, while market valuation methods are mainly utilized in studies from the 1980’s. 70 VI. DISCUSSION AND FURTHER WORK The beneficial use values database (BUVD) is the first and only comprehensive database of beneficial use values for water. It is distinguished by its focus on the beneficial use categories defined by the SWRCB. Other databases have also been developed recently that have other purposes but do contain economic values associated with water uses and changes in water quality. One is an on-line database developed by the New South Wales provincial environmental agency in Australia, and the other is a recently completed database used by the U.S. Fish and Wildlife Service. To better understand the role of the BUVD, it is useful to briefly describe these two economic value databases and compare them with the beneficial use values database. In the following, we briefly highlight the differences and similarities between the three economic values databases, then describe areas of further work that can be used to improve the beneficial use values database. Other Databases with Economic Values for Water-related Amenities Currently, there is only one searchable on-line database of economic values, ENVALUE. The New South Wales Environment Protection Authority (NSW EPA) originally developed ENVALUE in Australia as an on-line collection of environmental valuation studies in 1995. Its explicit goal was to “assist decision makers in government and industry as well as academics, consultants and environmental groups, to incorporate environmental values into cost-benefit analyses, environmental impact statements, project appraisals and overall valuation of changes in environmental quality.” (http://www.epa.nsw.gov.au/envalue). It was last updated in May of 1998. In addition to the ENVALUE on-line database of economic values, a second database contains economic values for water indirectly as a source of sportfishing benefits. This database, the U.S. Fish and Wildlife Service (FWS) sportfishing values database (Boyle, et al., 1998), contains values for recreational sportfishing derived exclusively from travel cost and contingent valuation studies. The 109 studies in the database were conducted between 1975 and 1996. The values contained in the database apply exclusively to domestic fishing trips. Like ENVALUE, this database was compiled primarily to assist policy-makers improve policy analysis. In certain ways, the ENVALUE and FWS databases are similar to the beneficial use values database described in this report. Each of the databases provides information on environmental valuation studies, presenting detailed descriptions of economic values and their interpretations. As one would expect, both the ENVALUE and FWS databases share common information categories with the beneficial use values database, such as valuation method used, a description of what the value estimate represents, the units of the value estimate, document descriptions, and bibliographic information. However, there are significant differences between all three databases that are important to note. In particular, each database differs in the scope of the literature covered, the intended audience and focus of the database, and specific content. Scope of Coverage 71 The database constructed by the New South Wales Environment Protection Authority, the ENVALUE database, is the broadest in scope. In addition to some economic values for water, water-related amenities, and water quality, this database contains economic values for air and land quality, health risks, and other non-water related amenities. Studies from multiple countries, not just Australian studies, are represented in the database. It also currently contains the largest number of studies. For instance, there are 88 studies valuing water quality and 127 valuing natural areas (including wetlands, rivers and lakes, and beaches) in ENVALUE. Unlike the FWS and BUVD databases, it includes purely conceptual studies that report no estimated values. The U.S. FWS database is the most specific in scope, encompassing only those studies employing TCM and CVM to value fishing opportunities. This database is limited to domestic studies, and includes 109 studies valuing sportfishing trips, primarily studies of the northeastern and western United States, with a small number conducted in other regions in the continental United States and Hawaii. The content of the database was constrained by a publication date restriction: studies conducted before 1975 were not included in the database, “because non-market valuation studies were not very common in the 1960s and early 1970s” (Boyle, et al., 1998, page 2-1). In terms of scope of coverage, the BUVD falls somewhere between the ENVALUE and FWS databases. It contains fewer studies related to uses for water and water quality than ENVALUE, but more than the FWS database. Also, it is not limited to strictly domestic studies like the FWS database is. In addition, like the ENVALUE database, the beneficial use values database includes studies from the 1960s and 1970s. However, unlike ENVALUE the BUVD does not include conceptual studies, instead focusing on studies that include economic values for water-based amenities. Focus and Intended Audience Another major difference between the databases is the extent to which users are encouraged to employ the database contents for benefits transfers. ENVALUE is the only database that explicitly states that its contents are intended to be used for benefit transfers.6 All values in ENVALUE have been converted into 1997 Australian dollars in an effort to make the values directly comparable between other studies in the database, thus facilitating benefit transfers. Information on studies and values contained within ENVALUE is tailored specifically for use in benefits transfers. Moreover, the database itself “provides guidance on transferring estimates to other sites” (http://www.epa.nsw.gov.au/envalue/). Each document containing economic values in ENVALUE is linked to a page that enumerates potential problems with using the values from the particular study in a benefit transfer. In contrast, the values in the FWS and BUVD are reported as they were presented in the original documents, though they differ in the extent information is presented in a way compatible with benefit transfers. The FWS database is less explicit as a tool for 6 In fact, the homepage states: “It is expected that the ENVALUE database will assist decision makers in government and industry as well as academics, consultants and environmental groups, to incorporate environmental values into cost-benefit analyses, environmental impact statements, project appraisals and overall valuation of changes in environmental quality.” This statement is followed by a definition of benefit transfers. 72 benefit transfers than ENVALUE, although it is structured in a way that makes it most appropriate as a foundation for benefit transfer. Its goal is to provide “information on numerous studies from the large body of economic literature of sport fishing values” to improve the “efficacy of, and consistency in, [FWS] analyses involving the economic evaluation of sport fishing opportunities” (Boyle, et al., 1998, 1-1). The information contained in the database is coded in a manner that is conducive to benefit transfer. Unlike either ENVALUE or the BUVD, the FWS database contains an extensive set of dummy variables and standardized categories to describe details about the values, methodologies, and research assumptions made for each study. This “standardization” of information loses information relative to the open-ended comment fields used in ENVALUE and BUVD, but offers a convenient way for researchers to adjust transferred benefits for differences in conditions at the site being studied relative to those in the database. It also identifies “most-preferred” values where there are multiple estimates for a particular amenity. . Because of these elements, the FWS database is perhaps the most benefit –transfer-friendly database. However, the database assumes a fairly high level of familiarity with theoretical and empirical issues pertaining to TCM and CVM, and is the least accessible to the lay public. The current plan for the BUVD front-end (on-line interface) is to present the information as objectively as possible, as an information tool and starting point for inquiries on economic values for beneficial uses. Its organization and content reflect this. The principal value of the BUVD, as currently configured, is as an educational tool, not as a benefit transfer tool. Although it provides a roadmap to values for various beneficial uses, users should look up the original studies to gather the appropriate information for a benefit transfer. Specific Content The databases also differ substantively with respect to specific content. Although each contains information related to both a study and the values it generates, the types of values and studies included in each database differ, as does the specificity of information provided in each. In general, ENVALUE covers the widest range of amenities and studies, and contains information comparable to the BUVD with respect to methodology and relatively more details about the sample data than does the BUVD. The ENVALUE database also evaluates the performance of each study for benefit transfer by including evaluation criteria for each specific method. For example, for studies employing CVM, the database reports whether the report accounted for biases encountered, estimation problems, and socioeconomic differences between the sample and population. On the other hand, the FWS database contains the smallest number of studies of the three databases, yet reports the most detailed information. Studies that provide values for several recreation activities in combination with fishing are excluded, as well as studies that only report aggregate (i.e., population) value estimates for sport fishing activities. Additionally, values that are determined by the author(s) of each study as being “unreliable” are not in the database. According to Boyle, et al. (1998), unreliable values are those that were included in a study for comparison with a proposed model or those that were deemed inappropriate or not useful by the authors of the study. This 73 contrasts with the BUVD and ENVALUE, both of which include all values reported in studies. In the way of more detailed information, the FWS sportfishing database includes information about baseline site quality conditions, site quality changes, data collection, and econometric estimation above and beyond that contained in the other databases. For instance, variables used in estimation, and whether each is statistically significant, are reported along with details of the econometric methods used in estimation of demand or WTP functions. In addition, variability of welfare estimates are reported for every value that has an estimated variance or standard error. Both ENVALUE and the FWS databases contain at least marginally more information per document than the BUVD. This difference is primarily a reflection of the intended audience and focus. Since both the ENVALUE and FWS databases are to be used primarily by analysts as a tool for policy analysis, more technical information is needed to be able to manipulate values for use in these applications. In contrast, the BUVD has at its core the goal of providing information about the range and types of economic values for beneficial uses to policy analysts and others interested in water issues, irrespective of technical familiarity. Thus, although the BUVD provides some technical information, it contains far less than the other two databases and provides resources (such as a glossary) that allow users to demystify some of the technical jargon used. Further Work The beneficial use values database (BUVD) contains many values of beneficial uses for water spanning the majority of State Water Resource Control Board beneficial uses. This report has described the structure and contents of the database as well as the process by which economic values and studies were collected, screened, and analyzed for information relevant to the database. Difficulties encountered were described and discussed as they relate to both the content of the database and the process of procuring information for the database. In addition to developing a web-based version of the database to make the database accessible more broadly, several areas need further work. In particular, several beneficial uses are under-represented in the database. The absence or underrepresentation of values for several beneficial uses is a result of time constraints. With additional time and research effort, it is possible that studies may be found in the gray literature that represent beneficial uses, such as SAL or WARM, not found in the original literature review. Of course, it is possible that values do not currently exist in the literature for amenities of these types. Still, the pool of economic values for beneficial uses is constantly expanding as researchers continue to develop new valuation methods and estimation techniques and apply them to water resources and their uses. This continual expansion of the literature suggests the need for periodic updates to the BUVD. This will be an important activity for keeping the database current, for correcting errors that may be found, and for finding economic values for previously under-represented categories. As the literature grows, 74 researchers may well generate estimates of beneficial uses that currently do not have many (or any) economic values in the database increases. In sum, the BUVD as presented in this report provides a useful framework within which economic values for beneficial uses of water can be collected and cataloged. The particular organization used in the database is especially practical for project evaluations in California that measure the effects on beneficial use values and for understanding the variety of values associated with beneficial uses of water. Although the database contains only those beneficial use values available in the “readily-available literature” and thus only represents a partial account of the entire literature, the majority of SWRCB beneficial use values are represented. As suggested above, additional efforts should be made to update the database to cover a larger portion of the literature, especially the gray literature, where the majority of the under-represented beneficial use values are likely to be found. Well-trained analysts with expertise in the area of non-market valuation should conduct these updates. 75 References Boyle, Kevin, Richard Bishop, Jim Caudill, John Charbonneau, Doug Larson, Marla A. Markowski, Robert E. Unsworth, and Robert W. Paterson. “A Database of Sport Fishing Values.” Report to Economics Division, Fish and Wildlife Service, U.S. Department of the Interior, October 1998. Downing, Mark and Teofilo Ozuna, Jr. "Testing the Reliability of the Benefit Function Transfer Approach." Journal of Environmental Economics and Management, vol. 30, pp. 316-322, 1996. Economic Analysis, Inc. "Measuring Damages to Coastal and Marine Natural Resources: Concepts and Data Relevent for CERCLA Type A Damage Assessments." Report to CERCLA 301 Project, United States Department of Interior, 1986. Freeman, A. Myrick III. The Measurement of Environmental and Resource Values: Theory and Methods. Washington, D.C.: Resources for the Future, 1993. Hardner, Jared J. “Measuring the Value of Potable Water in Partially Monetized Rural Economies.” Water Resources Bulletin, 32(6):1361-1366, 1996. Kalman, Orit, Jay R. Lund, Daniel K. Lew, and Douglas M. Larson. “Benefit-Cost Analysis of Stormwater Quality Improvements.” Environmental Management, 26(6): 615-628, 2000. Kirchhoff, Stephanie, Bonnie G. Colby, and Jeffrey T. LaFrance. "Evaluating the Performance of Benefit Transfer: An Empirical Inquiry" Journal of Environmental Economics and Management, vol. 33, pp. 7593, 1997. Loomis, John B. "The Evolution of a More Rigorous Approach to Benefit Transfer: Benefit Function Transfer" Water Resources Research, vol. 28(3), pp. 701-705, 1992. "Natural Resource Damage Assessments", Federal Register, Vol. 51, No.148, 1986. 27674-27753. (43 CFR Part 11) Sorg, Cindy F. and John B. Loomis. “Empirical Estimates of Amenity Forest Values: A Comparative Review.” General Technical Report, RM-107, Rocky Mountain Forest and Range Experiment Station, Forest Service, USDA, Fort Collins CO, 1984. Walsh, Richard G., Donn M. Johnson, and John R. McKean. “Issues in Nonmarket Valuation and Policy Application: A Retrospective Glance.” Western Journal of Agricultural Economics, vol.14(1), pp. 178188, 1989. 76 Appendix A Economic Valuation of the Beneficial Uses of Water Valuation Methods 77 Economic Valuation of the Beneficial Uses of Water: Valuation Methods The California State Water Resources Control Board (SWRCB) designates 26 beneficial use categories for water that, taken together, represent the uses for water deemed to be valuable to residents of the state. These beneficial use categories cast a wide net, encompassing consumptive and nonconsumptive uses, instream uses, and ecological functions of water. As a result, commercial, recreational, and environmental uses for water are recognized as valuable. Due to the heterogeneity of these uses, it should not be surprising that there is an equally diverse set of methods used by economists, engineers, and other analysts to value these beneficial uses of water. Table A.1 summarizes the principal methods used to value types of beneficial uses in the database. The principal aim of this appendix is to introduce the reader to the methods used by researchers to value beneficial uses of water. The beneficial use values reported in the database range from theoretically rigorous measures to rough indicators of economic value. For extensive discussions of the economic value of water, the reader is pointed to Young and Gray (1972) and Gibbons (1986). Table A.1. Valuation Methods Used in Studies for Each Searchable Beneficial Use Category. Water Quality Groundwater Aquaculture Commercial Agriculture Navigation X X X X X X X X X X X X X Recreation Wetland/ Floodplain Habitat/ Ecosystem Municipal Industrial Searchable Beneficial Use Category Averting Behavior Approaches Contingent Valuation Methods Conjoint Analysis Damage Function Approaches Hedonic Methods Market Valuation Approaches Opportunity Cost Methods Optimization Models Other Methods Replacement Cost Methods Simulation Models X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X Travel Cost Methods An “x” denotes valuation method used to estimate beneficial use value. 78 Power Measuring the Economic Value of Water: Valuation Methods This section discusses the methods by which beneficial uses for water are measured. They are grouped by the general valuation method categories used in the database, which are: • • • • • • • • • • • • market valuation, simulation models, contingent valuation methods (CVM), conjoint analysis, travel cost methods (TCM), averting behavior approaches, hedonic methods, damage function approaches, replacement cost methods, opportunity cost methods, optimization models, and other methods. As Figure A.1 shows, the majority of values were generated by means of CVM and TCM. Conjoint analysis is not represented in Figure A.1 because no studies in the database used this method. Gibbons (1986) and Young and Gray (1972) discuss some, but not all, of the valuation methods used in the database to value water uses and thus provide a useful alternate overview of several of the methods discussed here. In developing the database, an initial set of valuation methods that would likely be encountered in the literature was developed. This was modified during the search process to reflect what was actually found in the literature. The one valuation method we expected to find, but did not, was conjoint analysis. It was retained in the database on the expectation that future searches will identify useful studies of this type. The “Other Methods” category includes valuation methods that were not easily categorized. This residual category includes valuation methods that were not well represented in the database. 79 Figure A.1: Distribution of Values by Valuation Method 700 641 600 650 500 Number of Studies 400 300 200 181 141 100 35 0 Contingent Valuation Methods Travel Cost Methods Hedonic Methods Averting Behavior Approaches 82 21 Damage Function Approach 82 99 93 9 Replacement Cost Method Opportunity Cost Method Optimization Models Market Valuation Simulation Models Other Methods Valuation Method Contingent Valuation Methods Contingent valuation methods involve directly asking people questions in the form of a survey or interview in order to elicit their WTP. They are a fairly flexible set of methods that can be used to value non-market goods and quality changes. Good sources of detailed discussion are Mitchell and Carson (1989), Freeman (1993), and Cummings, Brookshire, and Schulze (1986). Differences in CVM approaches arise from the framework or question formats used to elicit responses. In general, there are two general types of CVM approaches— direct methods and indirect methods. Direct methods involve asking individuals to directly state their WTP, either through an open-ended question or by choosing a value out of a proffered set of values. Open-ended, payment cards, and bidding game CVM formats fall into this category of CVM approaches. These approaches do not require any analysis to get a measure of WTP (Freeman, 1993). Alternatively, indirect methods involve asking individuals questions to reveal information about their preferences for the amenity in question that can be used to model and consequently to measure their WTP. Researchers employing indirect CVM approaches are required to construct models that estimate WTP using the data obtained from asking these questions. 80 Referendum, or discrete-choice, approaches fall into this group of CVM approaches. Below is a brief explanation of each of these types of contingent valuation methods. • Direct CVM Approaches Perhaps the most straightforward manner of obtaining an estimate of an individual’s WTP for an amenity is simply to ask him or her for it. Typically, respondents are asked to simply state their maximum WTP for a specified environmental quality improvement or other amenity, which is referred to as the open-ended approach. A primary difficulty is that individuals may have difficulty providing an estimate of their WTP for something that may not be familiar to them without a reference to base their answer upon. "Respondents often find it difficult to pick a value out of the air...[and] as a consequence,...[this method] tends to produce an unacceptably large number of nonresponses or protest zero responses to the WTP questions" (Mitchell and Carson,1989,97). Because of this difficulty, researchers often opt for closed-ended question formats that define a range of WTP or bid levels, as in the payment card approach, or present the individual with a specific amount, as in the bidding game approach or referendum approach. An alternative to the open-ended direct question approach is the bidding game format. In this approach, each respondent is asked whether or not they would be willing to pay a specified amount. Then depending upon the answer to this question, the amount is raised (if "yes") or lowered (if "no") and the individual is asked whether they would now be willing to pay the new amount. This process is continued until each respondent's maximum WTP is found (e.g. If respondent initially answered "yes" to $B, then a "no" to $(B+1), then this indicates that $B is the individual's max WTP). Since this approach yields a dataset comprised of individual WTP amounts, welfare estimation proceeds along the same lines as the open-ended data. With the payment card CVM format, respondents are shown a card with a range of payment values from which to choose. They are then asked to choose between the values given or state their own. If they choose a value on the payment card their maximum WTP can either be interpreted as the value chosen or as some amount bounded below by the amount chosen and bounded above by the next higher value. Mitchell and Carson (1989) originally devised this method to "maintain the properties of the direct question approach while increasing the response rates for the WTP questions by providing respondents with a visual aid which contains a large array of potential WTP amounts..." (p.100). The authors note, though, that the method is vulnerable to biases "associated with the ranges used on the cards" (p.100). The appeal of these direct CVM approaches is in their simplicity. Since individuals in the sample provide their maximum WTP, welfare theoretic measures of value are immediately available. Typically, the sample mean or median WTP is reported and assumed to represent the population’s WTP for the amenity. When discrepancies between the socioeconomic characteristics of the sample and population are evident, the observed WTP for each individual in the sample is regressed on demographics (e.g., age, sex, and income) to estimate a WTP function that reveals how characteristics affect WTP. 81 Then, an estimate of the population’s WTP can be calculated by substituting the population’s characteristics into the estimated WTP function. • Indirect CVM Approaches In contrast to direct CVM approaches, estimates of WTP cannot be directly calculated when employing indirect CVM approaches. The most common type of indirect CVM approach is the referendum approach, sometimes referred to as the dichotomous choice question format (Cameron, 1988; Hanemann, 1984). This question format presents each respondent with a question of the form, "Would you vote for a program that will permanently increase water quality by Q% if it would increase your taxes by $b this year?" where b would vary for different random subsamples. Respondents are thus presented with a package consisting of better water quality and less money. They are presumed to say "yes" if they like this combination better than their current situation, where they have lower water quality and more money, and “no” if they do not. A “yes” response indicates their WTP for the proposed water quality is greater than $b, while a “no” indicates it is less. The first versions of the referendum approach asked individuals a single WTP question. A variation on the above referendum approach adds a follow-up question to the questioning format (Hanemann, Loomis, and Kanninen, 1991). This double-bounded dichotomous choice CVM question format starts by asking respondents whether they would pay an initial bid level (b) for a change in quality or the level of an amenity as described above. As in the single-bounded approach, the respondent answers “yes” or “no” to indicate whether their WTP was greater than or less than the initial bid. However, depending upon whether they answered “yes” or “no” to the initial question, respondents are then asked the same question again but with a new bid amount (higher if they had answered “yes” initially, and lower if not). This increases the precision of the WTP estimate. For a description, see Hanemann, Loomis, and Kanninen (1991). Conjoint Analysis Like contingent valuation methods, conjoint analysis is a survey-based method wherein respondents are asked to rank or rate alternatives embodying different levels of attributes. Common attributes that vary between alternatives include environmental quality levels, quantity of amenities, and price associated with the alternative. By identifying individual consumers’ preference ratings or rankings for alternatives, analysts can decompose and estimate the structure of the individual’s preferences. This conjoint analysis approach was popularized in the marketing literature (Green and Srinivasan, 1990) and has been applied in the economics literature to value hunting trips (MacKenzie, 1993) and Atlantic salmon fishing (Roe, Boyle, and Teisl, 1996). When contingent ranking questions are asked, respondents provide an ordinal ranking of the alternatives that are presented. Thus, if Z alternatives are given, rankings from 1 to Z will be returned to the analyst. Typically, one of the Z alternatives represents the status quo. These rankings suggest an ordinal preference ordering for the individual’s utility. Suppose that there are three alternatives, A, B, and C. If the individual ranks the 82 alternatives from best to worst as B, A, and C, the corresponding utility ranking must be V(B) > V(A) > V(C), where V(⋅) is the individual’s indirect utility function. The contingent ranking reveals the structure of an individual’s preferences with respect to the alternatives, and of the attributes associated with them. To analyze this data, a functional form is chosen for V(⋅) and ordered probit or logit estimation is typically used. Once the parameters of the indirect utility function are estimated, welfare measures can be calculated directly. An important feature of contingent ranking data is that it does not reveal anything about the intensity of the preferences. To obtain information about the intensity of preferences, analysts can opt to ask individuals to rate the alternatives on a predetermined scale instead of ranking them. The resulting contingent rating data are often assumed to convey information about the intensity of their preferences, through a transformation function, φ(), that transforms ratings into utility levels. Under fairly restrictive conditions it is possible for researchers to interpret the ratings as utility indices and base welfare measurement directly on analysis of the ratings (Roe, Boyle, and Teisl, 1996). Analysis of data from contingent rating surveys differs from the contingent ranking case. Standard analyses of conjoint analysis ratings data involve regressing ratings for each alternative i (ri) on the attributes of the alternative to estimate the function ri = ri(pi, qi, m), where pi and qi represent vectors of prices and quality attributes associated with the amenities in the alternative i, respectively, and m is the individual’s income. Economic welfare measures of quality changes can then obtained by determining the change in income necessary to make the rating for the status quo (r0) equal to the rating for the alternative associated with the change (r1). Thus, the willingness to pay associated with a change in quality from q0 to q1 satisfies the following equality: r1(p1, q1, m-WTP) = r0(p0, q0, m). Travel Cost Models In the late 1940's, Harold Hotelling wrote a letter to the Park Service in which he laid the groundwork and argued "for an empirical method which would allow the calculation of the value of the service flow, rather than simply counting the use" of recreational sites (Bockstael et. al., 1991, 229). Since then, methods based on the model he laid out, called travel cost methods, have been used to value recreation-related benefits associated with specific recreation resources. The travel cost model is based on the premise that "even though most public recreation sites have zero (or nominal) entry fees, recreationists nonetheless pay an implicit 'price' for a site's services when they visit it" (Smith, Desvousges, and Fisher, 1986, 281). On a very basic level, this amounts to utilizing expenditure data for visitors to a natural resource to extrapolate the associated value of the resource. Typically, travel cost models are used to estimate the in-situ value of recreational sites, such as lakes and beaches, or the services derived at these sites, such as recreation, but also have been used to value water quality improvements (Smith, Desvousges, and Fisher, 1986). Since travel cost methods depend upon observations of individual recreation behavior, these methods 83 measure use values only (Ward and Loomis, 1986). Three principal types of travel cost models are briefly discussed: • • • zonal travel cost models, individual travel cost demand models, and random utility-based travel cost models. The zonal travel cost model, sometimes called the traditional travel cost model because it is the original model suggested by Hotelling, uses aggregate visitor information to construct a demand schedule for a recreation site. This demand function relates the number of trips per capita to the travel costs (viewed as a proxy for the price of trips), substitute prices (travel costs to alternate sites), and socioeconomic and taste characteristics of visitors. Given visitation and demographic data, researchers implement the zonal model in four steps. First, zones are created around the recreation site to be valued. To do this, visitors are identified as coming from selected “origin zones” at varying distances from the site. The average travel cost of traveling to the site is calculated for each individual in the zones by summing both the monetary (gas, food, and lodging) and non-monetary (time) costs involved in the travel experience from their origin in the zone. Second, the relationship between per-capita visitation, which is calculated by dividing the number of visitors by the total population of each zone from that particular zone, and trip costs is estimated econometrically. Third, predicted use from each origin is summed to calculate the total use at a zero price. Finally, given an assumption that individuals respond to an increase in entry fees the same as they would to an increase in travel costs, a demand function is generated for the site using the estimated relationship between the ratio of visits to population and travel costs with hypothetical price increases in fixed increments. Using this estimated recreation site demand function, the consumer surplus can be calculated as a measure of the value of the recreation site or the recreation activities that take place there. One of the primary difficulties with the zonal approach is the inability to link the demand functions and resulting welfare measures with the economic theory of individual consumers. Nevertheless, due to the fact that this model has minimal data requirements, it became a popular method for valuing recreation for several decades following its conception. For a more detailed exposition of this method, see Durden and Shogren (1988). The other two travel cost models require the use of more detailed, disaggregated data on recreation behavior, costs, and individual characteristics. At the same time, these methods have a firmer grounding in economic theory than the macro-level analysis implied by the zonal model and are in fact derivable from standard consumer theory (Bockstael and McConnell, 1983). Individual travel cost demand models use observations of individual recreation behavior over a season or year to estimate the relationship between the number of trips taken by the individual to a site and the costs of accessing the site for recreation. Using individual-level data on the number of trips taken, travel costs, and socioeconomic information, such as income, a demand function is estimated by assuming that travel costs are an appropriate proxy for price. The estimated demand function can be used to approximate a representative individual’s WTP by calculating consumer’s surplus 84 welfare measures or be integrated back to yield quasi-exact welfare measures of WTP like compensating variation. When quality is included as an argument of demand, the value of quality changes can be calculated. Early applications used simple functional forms and relatively simple econometric techniques (e.g. ordinary least squares). The recent literature has begun to address important issues which were not dealt with in earlier applications such as the integer nature of the quantity variable (trips taken) (Hellerstein and Mendelsohn, 1993), the presence of substitute recreation sites (Rosenthal, 1987), and the appropriate manner in which to incorporate the value of time (Bockstael, Strand, and Hanemann, 1987; Smith, Desvousges, and McGivney, 1983), to name a few. Some of the problems associated with the individual travel cost method, and travel cost methods more generally, are enumerated in reviews of the method, including Randall (1994) and Ward and Loomis (1986). Random utility models (RUM) address both the decision to participate in recreation, by taking a trip, and where to go if a trip is taken. The same principles that motivate the discrete choice CVM applications discussed above also appear in the RUM of recreation demand. The individual’s choice is based on the utility they receive from the different alternatives, which include staying home or visiting one of a number of recreation sites. For example, suppose a person is observed to choose to visit site A instead of staying home or visiting sites B and C. Visiting each site requires the individual to incur travel costs pi to gain access to the ith site (i = A, B, or C) and enjoy the quality of the site qi. The indirect utility function related to visitation of site i is then Vi(pi, qi, M; ε), where M is the individual’s income and does not vary across sites, and ε is a random error. If A is chosen, this implies VA ≥ VB and VA ≥ VC. We could equivalently write this in terms of probabilities. The statement Pr(choosing site A) = Pr(VA ≥ VB and VA ≥ VC) states that the probability of choosing site A is equal to the probability that the utility associated with A is greater than or equal to the utilities associated with the other choices. For a further discussion of this manner of modeling choice, see Ben-Akiva and Lerman (1985). Once a functional form for Vi is chosen, and an assumption is made regarding the distribution of ε, the parameters of Vi can be estimated. 7 The resulting estimated per-occasion indirect utility function can be used to generate per-choice occasion welfare measures, such as compensating variation, for price or quality changes (Small and Rosen, 1981). The names associated with RUM based travel cost models include multinomial logit, multinomial probit, nested logit, and repeated nested logit models of recreation demand. Difficulties associated with the RUM travel cost approach include the fact that it generates per-occasion or per trip based welfare measures instead of per season welfare estimates, which are often desired in policy analysis. That is, RUM-based models answer the question of where to go on a given choice occasion, but do not answer the question of how often to go. Therefore, instead of being concerned with the frequency of trips taken overall or to individual sites, it focuses on modeling or predicting the choice between 7 Most applications to date have assumed Vi is linear. Also, when ε follows a standard normal distribution, probit estimation techniques are used. If it instead follows a type I extreme value distribution, a logit model is used. 85 sites. Another significant issue is defining the set of relevant alternatives faced by the individual whose choices are being modeled. Researchers have developed two principal ways of addressing the “trip frequency” problem. The first is to divide the season into discrete choice occasions and thereby treat the trip frequency and site selection choices as a series of discrete choices. This “repeated” model approach has been used by Feenberg and Mills (1980), Caulkins, Bishop, and Bouwes (1986), and Morey, Rowe, and Watson (1993). A difficulty with this approach is determining the number of choice occasions an individual has, because one of the options is to decide to stay home, in which case no trip is observed. The second approach is to model trip frequency separately from the site selection choice problem (Creel and Loomis, 1992). In this approach, the trip frequency model and site selection model are linked by a variable, commonly called the inclusive value, which is a measure of the expected utility of a trip. The estimation of these “linked models” proceeds in two stages. In the first stage, the site choice problem is modeled to estimate the per-occasion indirect utility function. From this estimated equation, the inclusive value is calculated. In the second stage, a trip frequency demand function is estimated that includes the calculated inclusive value, income, and other demand shifters as right-hand side variables. The travel cost and quality to the site are not included as independent variables because the inclusive value measure summarizes the influence of these variables on site choice and therefore represents an index of sorts for these variables (Creel and Loomis, 1992). Bockstael, Hanemann, and Kling (1987), Creel and Loomis (1992), and Feather, Hellerstein, and Tomasi (1995) are a few examples of studies that have utilized this approach to account for trip frequency and estimate per-season welfare values in RUM travel cost models. A drawback of this approach is that the two separate models that are used to model behavior, the site choice model and the trip frequency model, are not typically consistent with a well-behaved set of individual preferences. A variation on the RUM travel cost model that does use a utility-consistent approach is the Kuhn-Tucker model of recreation demand. This model applies the work by Wales and Woodland (1983), who proposed an estimation method for data where some goods are not consumed, to the case of recreation. This model takes advantage of the inequality constraints that restrict an individual's choices to estimate parameters. It represents a generalization of the RUM approach that allows for both the number of trips made and the set of sites chosen to be simultaneously modeled and estimated. Readers interested in details of the method are referred to Herriges, Kling, and Phaneuf (1999). A variant of the travel cost method presented above combines elements of hedonic theory with a travel cost model. First proposed by Brown and Mendelsohn (1980), the hedonic travel cost (HTC) method combines individual travel cost models with hedonic price theory to derive demand functions for recreation site characteristics. The hedonic travel cost method postulates that the travel costs individuals accrue when choosing where to recreate are directly a function of the level of site characteristics, such as water availability and quality. This method is implemented in two steps: First, individual travel costs are regressed on site characteristics to get a hedonic price function relating the “price” of a recreation trip to characteristics of the site being visited. The second step involves estimating demand functions for the site characteristics (e.g., water level) based on the derived marginal cost of the characteristic (first partial derivative of the hedonic price function with respect to the characteristic), which acts as the implicit 86 price of the characteristic. From the estimated characteristic demand functions, the value of the characteristic is approximated by the calculated CS. Difficulties with the hedonic travel cost model can include negative implicit prices for quality characteristics (Bockstael, Hanemann, and Kling, 1987; Smith and Kaoru, 1987). Bockstael, McConnell, and Strand (1987) suggest that this may be a consequence of a missing relationship between quality characteristics and travel costs. The problem may be one of identifying the separate relationships for demanders and suppliers of the quality characteristics of recreation sites, given there is no market in which to observe information about marginal values Averting Behavior Approaches In contrast to replacement cost methods, which value an amenity by the cost required to replace or repair it once it is lost, averting behavior approaches infer the value people place on an amenity by what they spend to prevent its removal or degradation. For instance, the value of safe drinking water can be approximated by summing up the money expenditures households spend on water filters and bottled water, and the time costs associated with boiling water and other averting activities. The sum of the expenditures associated with the averting activities can be interpreted as a lower bound on the individual’s WTP. In practice, it is difficult to isolate the portion of individuals’ expenditures that are attributable to averting activities. In the example above, the household could have a preference for bottled water instead of tap water, which complicates the determination of how much of the expenditure on bottled water is attributable to water quality. In this case, it is not appropriate to count all bottled water purchases as averting expenditures. Instead, the amount of increased bottled water expenditures due to water quality problems must be determined. Often, this is done using survey methods. Another difficulty with the averting expenditures method of valuing an amenity is the assumption that current expenditures made with certainty are worth incurring for a future uncertainty. That is, the method assumes that the individual has perfect foresight of the consequences of not taking preventative measures and has subsequently determined their maximum willingness-to-pay to avoid these consequences and accept a mitigated level of loss associated with the level achieved after the preventative activities have been undertaken (Winpenny, 1991). See Winpenny (1991) for a more detailed discussion of averting behavior methods. Hedonic Methods Hedonic methods exploit the idea that information on a public good, such as water quality, is contained in the prices and consumption levels of private goods. They assume the existence of a continuous function relating the price of a good to its attributes, called the hedonic price function. A principal way that hedonic methods appear in the database is to value water quality through analysis of property values. For example, the hedonic property value model assumes that consumers inherently recognize that the price they pay for a house reflects the level of environmental quality in the area, among its other characteristics. 87 Using regression models, estimates are made regarding the value of an environmental amenity based on price differentials of properties in areas with differing levels of environmental quality, controlling for all other housing characteristics. Hedonic models of this type have been used to value benefits from amenity improvements (Bartik, 1988) and public projects (Kanemoto, 1988). A good discussion of these methods and their problems can be found in Palmquist (1991) and Freeman (1993). An empirical difficulty with this approach is finding the variation in water quality levels across properties necessary for the estimation of the water quality parameter in the hedonic price regression. Even when this proves difficult, it still may be possible to calculate the value of water quality through this approach. If one has data on property in an area with high quality water and an area with low quality water, and can estimate hedonic price functions which contain all other characteristics important to the home buyer, estimates of the value of the water quality difference can be found by comparing the prices predicted by the hedonic price functions. If all attributes important to the individual are held constant in the hedonic price regression, the difference between hedonic prices, holding all non-water quality characteristics equal, must in principle be attributable to water quality. This is a strong assumption to make in practice. Replacement Cost Method Replacement cost methods use the monetary cost of replacing or restoring an amenity as a measure of the value of the amenity being valued, and are the basis for natural resource damage assessments that have been conducted in response to oil spills and other environmental incidents. The approach is most often applied to habitats, such as wetlands, whose value is approximated by the “replacement value”, or the amount that would be required to recover wetland resources which have been lost or to restore the wetland to its former productivity. Therefore, the value of the flood control functions of a wetland may be approximated by either the cost of building flood control projects that can provide the same level of flood control benefits as the wetland or by the cost of restoring the wetland to its original flood control productivity. In using the replacement cost approach, one must in principle incorporate cost estimates for replacing or repairing all the lost services associated with the removal or loss of the amenity. This can be a very difficult task for multi-faceted resources like wetlands. Wetlands provide many services, including flood control benefits, recreation benefits, habitat benefits, water supply benefits, and water quality benefits. If the replacement wetland or restored wetland does not provide the same types and levels of services as the original wetland, the costs will underestimate the wetland’s value. Moreover, unique services may be irrevocably lost when a wetland is damaged and are therefore difficult to value with restoration or replacement costs, such as in the case of the loss of wetlands that support rare or endangered species. Replacement cost methods implicitly assume that the replacement or restoration costs are indicators of willingness-to-pay. A problem with this assumption is that it mixes the demand side (WTP) of the “market” for water quality with the supply side (replacement cost). Also, though individuals collectively may be willing to repair or restore some amenities, it is also possible that they would not be willing to pay the amount necessary to fully restore or replace a damaged or lost amenity. As a result, 88 replacement or restoration costs cannot be considered accurate measures of economic value. For a further discussion of replacement cost methods, see Winpenny (1991) or McNeely (1988). Winpenny (1991) includes a further discussion of difficulties surrounding the use of replacement cost methods. Damage Function Approach The damage function approach is a common way to value welfare changes associated with pollution effects. As used in studies in the database, these approaches seek to determine a "dose-response" relationship between an environmental quality change and some physical effect on market goods. Market values (such as prices or repair costs) for the estimated response (physical effect) are then used to determine a monetary value for the overall dose (environmental quality change). A typical application of the damage function model contains three steps. First, the relationship between dose (environmental pollution measure) and damage (measure of physical effects of pollution) is estimated. Second, the estimated function is used to generate an estimate of the physical damages caused by the observed change in pollutants. The third step involves multiplying the amount of estimated physical damage by the unit value(s) of the damaged good(s). What constitutes the unit value depends upon the nature of the damaged good. For instance, if the pollutant was salt in irrigation water and the goods damaged by the salinity were crops in a particular region, the market prices of the damaged crops could be used as the unit values. In cases where market values do not exist for the relevant damaged factor, values are drawn from studies that have used indirect methods to derive dollar values. Freeman (1993) calls this damage function approach a “naïve” valuation method because it assumes individual consumers and/or firms will not react to the exogenous change in environmental quality by changing their behavior, either by reducing consumption of the damaged good(s), consuming substitute goods, or changing production. Furthermore, pollution or other changes in environmental quality that alter a physical quantity of a good may result in an increase in the prices for the affected goods (say, as a result of scarcity induced by damage) which also has welfare consequences. In other words, the approach assumes the absence of a market response to environmental change and therefore does not account for behavioral responses to the changes being valued. Still, as pointed out by Hanley and Spash (1993), the method is appealing because it can be applied quickly and inexpensively since it has only modest informational requirements. Still, economists warn that the values resulting from the approach should only be viewed as a first approximation to the economic value of pollution damages. Market Valuation When price and quantity information is available, economists may be able to estimate demand, supply, or production relationships and use market values to derive measures of value. From estimated demand and supply functions, estimates of consumer and producer surpluses (or others described above) can be calculated. Most often, this 89 type of information is available only when water is used for drinking supply and is only sufficient to estimate the demand for water by municipal users (Gibbons, 1986). When water is used as an input for industrial purposes, other methods are typically necessary to identify value. In these cases, economists sometimes estimate production functions using data on input use, capacity, and output. By knowing the relationship between inputs and outputs, the marginal productivity of each input, including water, can be calculated. Then, the marginal value of water (its per-unit value on the margin) can be calculated by multiplying the marginal product of water by the output price to yield the value marginal product, which must be equal across inputs for production to be efficient and input markets to clear. See Chapter 9 of Freeman (1993) for a mathematical treatment of these types of analyses. In certain special circumstances, willingness-to-pay can be observed directly. This occurs when individuals make payments to special funds or charities that will be directly used to improve, recover, or preserve a specific amenity (see Foster, Bateman, and Hurley, 1998). In effect, this is what contingent valuation methods seek to emulate by the construction of hypothetical markets. While appealing in principle, there can be several difficulties with applying this approach in practice. First, the range of amenities that can be valued by observing actual donations is limited to those amenities for which there exist charitable organizations or interest groups, and even if they do exist, procuring the required data may be problematic. Second, if the fund is not truly dedicated to a single purpose (such as protecting wildlife in a specific region), it may be extremely difficult to isolate what portion of contributions really go to, and therefore represent willingness to pay for, the amenity of interest. Third, such funds are typically based on voluntary contributions, and there may be incentives for “free-riding” behavior, so that fund contributions do not fully represent willingness to pay. These potential problems point to the need to carefully examine what the individual making the donation is told about how the money will be used. Simulation Models Several studies in the database used simulation models to aid in the valuation of amenities. They include bioeconomic fishery models, crop yield models, and biological growth models. These models are usually used to determine the biological or physical response to pollution or other physical event, and the resulting effect on human behavior, such as increased fishing effort in response to changes in a population of fish. The results from these models are then incorporated into economic models to obtain WTP measures associated with changes in exogenous factors. Optimization Models Optimization models are often used in economics to value water. The class of optimization models used to value beneficial uses includes mathematical programming and dynamic optimization models. Both types of models find a “best” solution to a problem of constrained choice, which might involve maximizing profit, minimizing cost, or maximizing net economic value (to name just a few). As a description of empirical 90 studies in the database, mathematical programming models are typical static, or oneperiod, models, while dynamic optimization models determine optimal choices each period in a many-period framework, often using dynamic programming methods. • Mathematical Programming Models In the database, mathematical programming methods are often used to determine the value of irrigated water and groundwater used in numerous uses or the value of a water quality change from a representative firm’s or a central planner’s perspective. Programming models are appealing when there is detailed data for a few representative firms, but market data sufficient for econometric estimation of demand, supply, or productivity functions are lacking. In general, mathematical programming is used to model economic problems in which the economic agent (consumers, central planner, or firm) seeks to optimize (maximize or minimize) their welfare (e.g., surplus, costs, profit, or revenues) while facing inequality constraints that restrict their ability to choose certain levels of inputs or outputs. Inequality constraints are of the form g(x, y) ≥ c, where g(⋅) is a function dependent on the variables represented by the vectors x and y, which are chosen by the optimizer, and c is a constant. The general form of a mathematical programming model has the following three elements (Chiang, 1984): 1. An objective function f(x, y) which is to be maximized or minimized by the choice of x and y. This objective function often represents profit, revenues, surplus, costs, and net return. 2. A constraint set comprised of i = 1,…,n inequality constraints of the form gi(x, y) ≥ ci that form the basis for restricting the values that x and y can take. 3. Non-negativity restrictions that require the variables x and y to take on nonnegative values. Since constraints in mathematical programming models are inequalities instead of equalities, the classical techniques of calculus typically used to solve optimization problems cannot be used. Instead, algorithms, typically solved with the aid of computers, have been developed that determine the optimal solution to programming models. Mathematical programming models are useful for determining both marginal and non-marginal values for water as an input. Typically, water enters mathematical programming models as an input constraint.8 For example, the value of an incremental unit of water to a profit maximizing firm can be found by relaxing the water constraint by adding a unit of water available for production and calculating the difference between the optimal value (value of the objective function evaluated at the optimal levels) before and after relaxing the constraint. This marginal value of water is also known as the “shadow value” of water. Non-marginal changes can be evaluated in an analogous manner. Similarly, changes in the shadow value of water can be calculated for exogenous changes in output prices, input prices, or constraints. 8 In some models, excess water not used in production can be sold. This case is not discussed here, because the method for calculating values does not change. Only the value of water sold is added to the objective function. 91 Two final points are worth mentioning. First, mathematical programming models should be “calibrated” against a base year or an average over several years so that the normative results from the models do not show wide divergences between previous and current observed levels. In other words, to ensure that the normative model does not yield unrealistic results, it should be calibrated in a way that yields optimal levels corresponding to a levels corresponding to an observed time period. Several calibration methods have been proposed and employed in practice (Howitt, 1995). Second, there are several types of mathematical programming models worth mentioning because of the frequency with which they can be found in the database. When f(⋅) and all gi(⋅) are linear, the model is a linear programming (LP) model. Alternatively, if either f(⋅) or any gi(⋅) is non-linear, the resulting model is a non-linear programming (NLP) model. For an indepth look of linear programming, see Paris (1989) or Chiang (1984). Chiang (1984) also includes a discussion of non-linear programming. Programming models can also be categorized by the time horizon in which the problem is solved. Static programming models determine the optimal level of the “choice variables” (the variables the optimizer chooses in order to optimize the objective function) for a given time period, while dynamic programming models result in the optimal path of the choice variables over multiple time periods. In the case of dynamic programming models, the objective function for the entire time period considered becomes the sum of the individual objective functions for each time period in the time horizon under investigation. • Dynamic Optimization Models As a general class of methods, dynamic optimization models are differentiated from static models by the time horizon of the analysis. Instead of leading to an optimal solution comprised of values that maximize a single objective function for a specific period of time (as static models do), dynamic optimization models yield optimal paths of choice variables over multiple time periods. Thus, dynamic programming models are both dynamic optimization models and mathematical programming models, as defined above. There are two other types of dynamic optimization models, calculus of variations models and optimal control models. These two types of dynamic optimization models are distinct from the usual formulation of dynamic programming models by the focus on continuous time intervals instead of discrete time intervals. 9 Again, we refer the interested reader to Leonard and Long (1992) or Chiang (1992) for detailed discussions of modeling dynamic economic problems that are outside the scope of this appendix. Like programming models, dynamic optimization models are useful for determining both marginal and non-marginal values for water. In fact, welfare measurement of changes in the quality or quantity of water (or other exogenous parameters) can be measured in a manner analogous to the process used for programming models. The value of changes in exogenous parameters is measured by the difference between the optimal value (value of the objective functional evaluated at the optimal paths of the state and control variables) evaluated with and without the change. These 9 Dynamic programming models can be used to model continuous time problems, as shown in Leonard and Long (1992). 92 models are used to measure the value of water allocation schemes, irrigation policies, and water quality projects, among other things. Opportunity Cost Methods The concept of opportunity cost can be employed to value uses for water. Opportunity cost can be defined as the value foregone by not engaging in the next best alternative. Thus, the opportunity cost associated with using an amenity for one use is the value of the amenity used in its next best alternative use. For instance, opportunity cost methods assert that the value of preserving a wetland area can be approximated by its value in its next best use, not as a wetland, but for commercial development. The value of the land for commercial purposes in this example is a proxy for the worth, or value, of the wetland in its present uses. The reasoning underlying this interpretation of opportunity cost as a measure of value lies in the idea of economic efficiency—since the wetland is not being used for commercial purposes, then the wetland must yield at least the amount that would be yielded from this alternative use in its present use. If it did not yield at least as much value in its present use as the next best use, then a rational owner of the resource would switch to the alternative use. See Gupta and Foster (1975) as an example of this approach. A serious limitation of the approach in practice is that there are many market failures which may prevent the equilibration of the marginal values of wetlands across uses. These market failures include the public good nature of wetland services, and the difficulty of assessing their value because of the absence of a market. There is little reason to expect that the value of wetlands as habitat and provider of water quality services is approximated by their value in a radically different use, such as for providing shopping or housing services. A variant of the opportunity cost approach for valuing amenities uses alternative costs to value water specifically in industrial uses where the decision is whether to purchase additional water or recycle the water already in use. Since firms have the option of recirculating existing water or acquiring additional supplies, the value of additional water can be approximated by the recycling costs required to make existing water supplies usable. Advocates of this approach reason that industry would be willing to pay for water up to the cost of producing water of adequate quality through treatment and reuse (Gibbons, 1986). Thus, this alternate cost accrued by the firm, the recycling cost, represents an indicator of a firm’s WTP for water. Other Methods Valuation methods falling into this category represent a variety of approaches to valuing beneficial uses. They were not treated as distinct valuation methods for one of three reasons. First, their occurrence in the database is infrequent. Second, they were difficult to classify because they represented combinations of methods. Valuation methods that incorporated a mixture of techniques and were not dominated by a particular approach fell into this class of methods. And third, the methodology may not be clearly stated. In such cases, we categorize the method as an "other method." 93 The methods described below include three cost-based valuation techniques (cost savings methods, residual imputation methods, and income multiplier methods), an energy analysis approach, and an agricultural yield comparison approach. • Cost Savings Methods Determining the value of water by the savings incurred by using the water in its current use versus the next best (cheapest) alternative is the primary aim of cost savings methods. This is a fairly common approach for valuing water for use in transportation, and can be applied to other uses of water as well. The method is distinct from opportunity cost methods in that the difference in the costs in its present use and its next alternative use is used as an indicator of value instead of its value in the alternate use. Cost savings methods are the primary means for valuing water as a source of transportation benefits. To value water as a means for transporting commerce, the cost savings resulting from not having to transport the same commerce via an alternative form of transit, typically by train, is a measure of the value of water for this use. This implies the value of water for transporting goods is equal to the difference between the cost of transporting goods by train and the cost of transporting goods by boat. When applied to transportation uses of water, the approach does not adjust for the characteristic of each travel mode that differs most, time costs (Gibbons, 1986). The method has also been applied to value hydropower. In this case, the value of hydropower can be estimated as the difference between the cost of providing power using hydropower methods and the next cheapest power alternative. Since the approach equates cost savings with value, it implicitly assumes demand will be unresponsive to the change in costs (demand is inelastic over the range being analyzed). • Residual Imputation Methods The residual imputation method is a form of a budget analysis technique that seeks to find the maximum return attributable to water or its use by calculating the total returns (profit, surplus, etc.) and subtracting off all non-water related expenses. The residual value is assumed to be the returns to water and represents the maximum amount the firm or farm would be willing to pay for water and still cover input costs (Naeser and Bennett, 1998). Additionally, if water procurement costs are subtracted from the residual value, a net WTP amount is calculated and may then be comparable to values for water in other uses (Gibbons, 1986). This approach is sometimes categorized as a farm crop budget technique in applications to agriculture. A difficulty is that the residual return after subtracting the costs of all measured non-water inputs is the return to water and all unmeasured inputs. • Income Multiplier Method 94 Analyses of local economy impacts often employ the use of income multipliers. Income multipliers measure the circulation of expenditures through an isolated economy, tracing the flow of money through individual sectors of the economy. The total effect of expenditures in one sector of the economy on the total economy is thus measured by multiplying the income multipliers by the expenditures. In very early recreation valuation studies, this was a fairly common way of measuring the value of recreation opportunities (e.g., Weitzman and Haas [1982] use this approach to value recreational trips to fisheries). However, changes in expenditures measure economic impact, not net economic value or WTP (Sorg and Loomis, 1984). They represent a redistribution of economic activity, which are transfers of surplus between regions, between people, and between industries that sum to zero in full employment economies. Changes in expenditures provide some evidence that the activity or amenity for which expenditures were made is valuable but do not provide guidance on the magnitude of this worth. • Energy Analysis The energy analysis technique is a method developed for valuing ecosystems (Farber and Costanza, 1987). The method examines the total biological productivity of an ecosystem to calculate a rough approximation of its value. In its most general application, the total energy captured by ecosystems is used as an estimate of the total potential the ecosystems have to do useful work for the economy. Costanza, Farber, and Maxwell (1989) argue that this approach provides an upper bound on the economic value of the ecosystem’s products because not all products will be used by the economy. One application of this approach converts the gross primary production, which forms the basis for the food chain and thus supports the production of economically valuable products within the ecosystem (e.g., plants and wildlife), into equivalent energy units using a single dollar to energy conversion factor (Costanza, Farber, and Maxwell, 1989). This conversion factor is measured in dollars per unit of energy. Therefore, once the total energy potential of the ecosystem is measured, the total economic value of system for providing biological products for the economy can be estimated by multiplying the energy potential by the conversion factor. In effect, the resulting welfare measure represents the total consumptive value of the ecosystem being valued. This suggests that non-consumptive benefits, such as some recreation values, are not accounted for in energy analysis-derived values. A difficulty with this method is that there is no reason to expect the conversion factor between energy and dollars to be even approximately constant. • Yield comparison method When farm budget information contains the per acre returns from both irrigated and non-irrigated land, it is possible to estimate the value of irrigated water as this observed difference in returns. This procedure forms the basis for the yield comparison method used to value irrigation water in agriculture. The method presumes the difference in net income associated with adding water to the production process is the producer’s 95 WTP for water. Thus, the difference in returns represents the maximum amount the producer would be willing to pay for the ability to use irrigation water. The appeal of this approach lies in its simplicity—it does not require detailed information on crop mixes. However, this strength also underscores a serious weakness—the heterogeneity of land and crops planted, crop values, and other differences between irrigated and non-irrigated land draw the validity of the difference in net returns being characterized as the net WTP for irrigated water into question. For an example of this approach, see Naeser and Bennett (1998). 96 Appendix B Glossary of Terms Used in the Database Abbreviations, Currencies, and Technical Terms 97 Abbreviations used in the database: The list below contains abbreviations used in the database. In addition to this list, American state abbreviations were used (e.g., CA = California, TX = Texas, etc.). Abbreviation BCA BOD BOR Cfs COD COE CRP CS CVM DOI EPA FBACT Ft FWPCA GLS Kg KWH Lb LS ML MLE N N NE NLS OLS PH PHOS Ppb Ppm PV RUM SEK SMSA TCM TURB WFUD WLS WQ WTA WTP UK US USDA USGS Yr Term Benefit-cost analysis Biochemical oxygen demand United States Bureau of Reclamation Cubic feet per second Chemical oxygen demand United States Corps of Engineers Conservation Reserve Program Consumer surplus Contingent valuation method United States Department of Interior United States Environmental Protection Agency Fecal bacteria Foot/feet Federal Water Pollution Control Act (Clean Water Act) Generalized least squares Kilogram Kilowatt hour Pound Least squares Maximum likelihood Maximum likelihood estimation Sample size Nitrogen Northeast Non-linear least squares Ordinary least squares A measure of acidity Phosphorus Parts per billion Parts per million Present value Random utility model Swedish Krona (currency) Standard Metropolitan Statistical Area Travel cost method Turbidity Wildlife and fish user day Weighted least squares Water quality Willingness-to-accept Willingness-to-pay United Kingdom United States United States Department of Agriculture United States Geological Survey Year 98 Currency: The current database includes values measured in several different types of monetary currencies aside from the U.S. dollar, including English pounds, Filipino pesos, Swedish Krona, Australian dollars, and Indian rupees. Technical Terms: The methodological comments typically explain or mention the estimation approaches used in the study to estimate beneficial use values. The following is a short list of the technical terms used and a brief definition of each. Valuation approaches contained in Appendix A are not listed here. Bioeconomic model Censored data A model that relates economic variables and biological factors, such as the effect of harvest rates on a fish population. A type of sample selection problem. This problem is characterized by inclusion of values in a certain range being reported or transformed into a single value. A measure of net willingness-to-pay or value associated with a change in the quantity or quality of an amenity. A survey-based valuation approach whereby individuals reveal their willingness-to-pay for an amenity or change in an amenity through their answers to questions. These models are typically used to characterize an individual’s decision between alternative choices and are grounded in random utility theory. An econometric technique used to characterize models with a discrete dependent variable by discriminating between different values. Field of economics concerned with the integration of statistics, mathematics, and economic theory to estimate economic relationships. This is the ex ante measure of willingness-to-pay for an amenity that is measured by the sum of expected levels of consumer surplus times their probability of occurrence. A function obtained in economic theory that relates the minimum amount of money obtainable given prices, quality, and other shift variables. Anything that anyone wants. All options or alternatives are goods. Goods can be tangible or intangible. A function obtained in economic theory that relates the maximum amount of utility (or satisfaction) obtainable given prices, quality, and other shift variables. A maximum likelihood-based econometric technique commonly used to model discrete choices that is based on the logistic distribution. An equation of the form log(y) = a + b1log(x1) + b2log(x2) +… + bnlog(xn), where a and bi (i = 1,…n) are constants, y is the dependent variable, and x1, x2,…,xn are independent variables. Regression based on the logistic distribution. An equation of the form y = a + b1x1 + b2x2 +… + bnxn, where a and bi (i = 1,…n) are constants, y is the dependent variable, and x1, x2,…,xn are independent variables. A statistical estimation approach that involves finding the parameters of a stochastic equation (an equation containing an error component) by minimizing the sum of the squared errors. Consumer’s surplus Contingent valuation Discrete-choice models Discriminant analysis Econometrics Expected consumer surplus Expenditure function Good Indirect utility function Logit estimation Logarithmic (log) linear functional form Logistic regression Linear functional form Least squares regression 99 Maximum likelihood estimation Optimization Option price Option value Producer surplus Poisson regression Probit estimation Quasi-option value Random utility models Semi-log functional form Tobit estimation Truncated data Utility function An econometric technique for estimating the parameters of a model that relies on maximizing the probability of observing the sample. A mathematical framework for calculating the maximum or minimum of a function. Maximum sum an individual would be willing to pay to preserve the option of using an amenity in the future. The excess value between option price and expected consumer surplus. A measure of a producer’s willingness-to-pay. Geometrically, it is the area above the supply curve and below the price line. An econometric technique based on the Poisson distribution used to estimate models where the dependent variable takes discrete values but is not a categorical variable (used for count data models). A maximum likelihood-based econometric technique commonly used to model discrete choices that is based on the normal distribution. The value an individual places on information that can be acquired at a future time that would allow the person valuing an amenity to make a better decision. These models are used to model the choices an individual makes between alternative goods. These models presume that the individual, when faced with choosing between several goods, will choose the good that yields the greatest utility (i.e., satisfaction or well-being) and models the choice in terms of the probability of choosing one good out of all possible choices. An equation of the form log(y) = a + b1x1 + b2x2 +… + bnxn, where a and bi (i = 1,…n) are constants, y is the dependent variable, and x1, x2,…,xn are independent variables. A maximum likelihood-based econometric technique that accounts for censoring and truncation in the sample data. A type of sample selection problem. Data whose values represent only a portion of the total, such as a dataset that excludes values above (or below) a threshold value. A construct of economic theory that represents the satisfaction, or wellbeing, of an individual. Direct utility functions give the relationship between utility and the quantities of goods and other exogenously determined characteristics of the individual, while indirect utility functions represent the relationship between utility and the prices of goods and exogenous factors. 100 References Bartik, Timothy J. “Measuring the Benefits of Amenity Improvements in Hedonic Price Models.” Land Economics, vol. 64(2), pp. 172-183, 1988. Ben-Akiva, Moshe and Steven R. Lerman. Discrete Choice Analysis: Theory and Application to Travel Demand. Cambridge: MIT Press, 1985. Bockstael, Nancy E., W. Michael Hanemann, and Catherine L. Kling. “Estimating the Value of Water Quality Improvements in a Recreational Demand Framework.” Water Resources Research, vol. 23(5), pp. 951-960, 1987. Bockstael, Nancy E., Ivar E. Strand, and W. Michael Hanemann. “Time and the Recreational Demand Model.” American Journal of Agricultural Economics, vol. , pp. 293-302, 1987. Brown, Gardner, Jr. and Robert Mendelsohn. “The Hedonic Travel Cost Method.” Review of Economics and Statistics, vol. 66, 427-33, 1984. Cameron, Trudy A. “A New Paradigm for Valuing Non-Market Goods Using Referendum Data,” Journal of Environmental Economics and Management, vol. 15, pp. 355-379, 1988. Caulkins, P., Richard Bishop, and Nicholas Bouwes. “The Travel Cost Model for Lake Recreation: A Comparison of Two Methods for Incorporating Site Quality and Substitution Effects.” American Journal of Agricultural Economics, vol. 68, pp. 291-297, 1986. Chiang, Alpha C. Fundamental Methods of Mathematical Economics. New York: McGraw-Hill, 1984. Chiang, Alpha C. Elements of Dynamic Optimization. New York: McGraw-Hill, 1992. Costanza, Robert, Stephen C. Farber, and Judith Maxwell. “Valuation and Management of Wetland Ecosystems.” Ecological Economics, vol. 1, pp. 335-361, 1989. Creel, Michael and John Loomis. “Recreation Value of Water to Wetlands in the San Joaquin Valley: Linked Multinomial Logit and Count Data Trip Frequency Models.” Water Resources Research, vol. 28(10), 2597-2606, 1992. Cummings, Ronald G. and Glenn W. Harrison. "The Measurement and Decomposition of Nonuse Values: A Critical Review." Environmental and Resource Economics, vol. 5, pp. 225-247, 1995. Cummings, Ronald G., David S. Brookshire, and William D. Schulze. Valuing Environmental Goods: An Assessment of the Contingent Valuation Method. Totowa, N.J.: Rowman and Allanheld, 1986. Downing, Mark and Teofilo Ozuna, Jr. "Testing the Reliability of the Benefit Function Transfer Approach." Journal of Environmental Economics and Management, vol. 30, pp. 316-322, 1996. Durden, Garey and Jason F. Shogren. “Valuing Non-Market Recreation Goods: An Evaluative Survey of the Literature on the Travel Cost and Contingent Valuation Methods.” Review of Regional Studies, vol. 18(3), 1-15, 1988. Economic Analysis, Inc. "Measuring Damages to Coastal and Marine Natural Resources: Concepts and Data Relevent for CERCLA Type A Damage Assessments." Report to CERCLA 301 Project, United States Department of Interior, 1986. Feather, Peter, Daniel Hellerstein, and Theodore Tomasi. “A Discrete-Count Model of Recreational Demand.” Journal of Environmental Economics and Management, vol. 29, pp. 214-227, 1995. 101 Feenberg, Daniel and Edwin S. Mills. Measuring the Benefits of Water Pollution Abatement. New York: Academic Press, 1980. Foster, Vivien, Ian J. Bateman, and David Harley. “Real and Hypothetical Willingness to Pay for Environmental Preservation: A Non-Experimental Comparison.” In Environmental Valuation, Economic Policy and Sustainability: Recent Advances in Environmental Economics. Melinda Acutt and Pamela Mason (eds.). Northampton, MA: Edward Elgar, 35-49, 1998. Freeman, A. Myrick III. The Measurement of Environmental and Resource Values: Theory and Methods. Washington, D.C.: Resources for the Future, 1993. Green, Paul E. and V. Srinivasan. “Conjoint Analysis in Marketing: New Developments with Implications for Research and Practice.” Journal of Marketing, October, pp. 3-19, 1990. Greenley, Douglas A., Richard G. Walsh, and Robert A. Young. "Option Value: Empirical Evidence from a Case Study of Recreation and Water Quality." Quarterly Journal of Economics, vol. 96(4), pp. 657-673, 1981. Gupta, Tirath and John Foster “Economic Criteria for Freshwater Wetland Policy in Massachusetts.” American Journal of Agricultural Economics, vol. 57(1), pp. 40-45, 1975. Hanemann, W. Michael. “Welfare Evaluations in Contingent Valuation Experiments with Discrete Responses,” American Journal of Agricultural Economics, vol. 66, pp. 332-341, 1984. Hanemann, W. Michael, John Loomis, and Barbara Kanninen. “Statistical Efficiency of Double-Bounded Dichotomous Choice Contingent Valuation,” American Journal of Agricultural Economics, vol. 73(4), pp. 1255-1263, 1991. Hanley, Nick and Clive L. Spash. Cost-Benefit Analysis and the Environment. Brookfield, Vermont: Edward Elgar Publishing Limited, 1993. Hellerstein, Daniel and Robert Mendelsohn. “A Theoretical Foundation for Count Data Models.” American Journal of Agricultural Economics, vol. 75, pp. 604-611, 1993. Howitt, Richard E. “Positive Mathematical Programming,” American Journal of Agricultural Economics, vol. 77(2), pp. 329-342, 1995. Herriges, Joseph A., Catherine L. Kling, and Daniel J. Phaneuf. “Corner Solution Models of Recreation Demand: A Comparison of Competing Frameworks.” Chapter 6 in Valuing the Environment Using Recreation Demand Models, Herriges, J.A. and C.L. Kling (eds.) Aldershot: Edward Elgar, 1999. Just, Richard E., Darrell L. Hueth, and Andrew Schmitz. Applied Welfare Economics and Public Policy. Englewood Cliffs, New Jersey: Prentice-Hall, 1982. Kanemoto, Yoshitsugu. "Hedonic Prices and Benefits of Public Projects." Econometrica, vol 56(4), pp. 981-989, 1988. Kirchhoff, Stephanie, Bonnie G. Colby, and Jeffrey T. LaFrance. "Evaluating the Performance of Benefit Transfer: An Empirical Inquiry" Journal of Environmental Economics and Management, vol. 33, pp. 7593, 1997. Larson, Douglas M. “Exact Welfare Measurement for Producers Under Uncertainty.” American Journal of Agricultural Economics, vol. 70(3), pp.597-603. 102 Leonard, Daniel and Ngo Van Long. Optimal Control Theory and Static Optimization in Economics. New York: Cambridge University Press, 1992. Loomis, John B. "The Evolution of a More Rigorous Approach to Benefit Transfer: Benefit Function Transfer" Water Resources Research, vol. 28(3), pp. 701-705, 1992. Mackenzie, John. “A Comparison of Contingent Preference Models.” American Journal of Agricultural Economics, vol. 75, pp. 593-603, 1993. Mäler, Karl-Goran. Environmental Economics: A Theoretical Inquiry. Baltimore: Johns Hopkins Press for Resources for the Future, 1974. Mas-Colell, Andreu, Michael D. Whinston, and Jerry R. Green. Microeconomic Theory. New York: Oxford University Press, 1995. McNeely, Jeffrey A. Economics and Biological Diversity: Developing and Using Economic Incentives to Conserve Biological Resources. Gland, Switzerland: International Union for Conservation of Nature and Natural Resources, 1988. Mitchell, Robert C. and Richard T. Carson. Using Surveys to Value Public Goods: The Contingent Valuation Method. Washington, D.C.: Resources for the Future, 1989. Morey, Edward R., Robert D. Rowe, and Michael Watson. “A Repeated Nested-Logit Model of Atlantic Salmon Fishing.” American Journal of Agricultural Economics, vol. 75, pp. 578-592, 1993. Naeser, Robert B. and Lynne L. Bennett. "The Cost of Noncompliance: The Economic Value of Water in the Middle Arkansas River Valley," Natural Resources Journal, vol. 38(3), pp. 445-463, 1998. "Natural Resource Damage Assessments", Federal Register, Vol. 51, No.148, 1986. 27674-27753. (43 CFR Part 11) Palmquist, Raymond B. "Hedonic Methods," in Measuring the Demand for Environmental Quality. John B. Braden and Charles D. Kolstad, eds. Amsterdam: Elsevier Science Publishing, pp. 77-120, 1991. Paris, Quirino. An Economic Interpretation of Linear Programming. Ames, Iowa: Iowa State University Press, 1991. Pendleton, Linwood. “Reconsidering the hedonic vs. RUM debate in the valuation of recreational environmental amenities,” Resource and Energy Economics, vol. 21(2), 167-189, 1999. Randall, Alan. “A Difficulty with the Travel Cost Method.” Land Economics, vol. 70(1), pp. 88-96, 1994. Roe, Brian, Kevin J. Boyle, and Mario F. Teisl. “Using Conjoint Analysis to Derive Estimates of Compensating Variation.” Journal of Environmental Economics and Management, vol. 31, pp. 145-159, 1996. Rosenthal, David H. “The Necessity for Substitute Prices in Recreation Demand Analyses.” American Journal of Agricultural Economics, vol. 69, pp. 828-837, 1987. Small, Kenneth A. and Harvey S. Rosen. “Applied Welfare Economics with Discrete Choice Models.” Econometrica, vol. 49(1), pp. 105-130, 1981. Smith, V. Kerry, William H. Desvousges, and Matthew P. McGivney. “The Opportunity Cost of Travel Time in Recreation Demand Models.” Land Economics, vol. 59(3), pp. 170-189, 1987. 103 Smith, V. Kerry and Yoshiaki Kaoru. “The Hedonic Travel Cost Model: A View from the Trenches.” Land Economics, vol. 63(2), 179-192, 1987. Sorg, Cindy F. and John B. Loomis. “Empirical Estimates of Amenity Forest Values: A Comparative Review.” General Technical Report, RM-107, Rocky Mountain Forest and Range Experiment Station, Forest Service, USDA, Fort Collins CO, 1984. Wales, T. J. and A.D. Woodland. “Estimation of Consumer Demand Systems with Binding NonNegativity Constraints.” Journal of Econometrics, vol. 21, pp. 263-285, 1983. Walsh, Richard G., Donn M. Johnson, and John R. McKean. “Issues in Nonmarket Valuation and Policy Application: A Retrospective Glance.” Western Journal of Agricultural Economics, vol.14(1), pp. 178188, 1989. Ward, Frank A. and John B. Loomis. “The Travel Cost Demand Model as an Environmental Policy Assessment Tool: A Review of Literature.” Western Journal of Agricultural Economics, vol. 11(2), pp. 164-178. Weitzman, A. Stephen and Mark A. Haas. “Socioeconomic Value of the Trout Fishery in Lake Taneycomo, Missouri.” Transactions of the American Fisheries Society, vol. 111, pp. 223-230, 1982. Winpenny, James T. Values for the Environment: A Guide to Economic Appraisal. London: HMSO, 1991. Young, Robert A. and S. Lee Gray. “Economic Value of Water: Concepts and Empirical Estimates. Department of Economics, Colorado State University. Report #NWC-SBS-72-047. Final Report to the National Water Commission, 1972. 104