Grantee Research Project Results
2004 Progress Report: Measuring Economics Benefits for Amenity Consequences of Land Cover Changes
EPA Grant Number: R829508Title: Measuring Economics Benefits for Amenity Consequences of Land Cover Changes
Investigators: Smith, V. Kerry , Palmquist, Raymond B. , Phaneuf, Daniel J.
Institution: North Carolina State University
EPA Project Officer: Hahn, Intaek
Project Period: October 1, 2001 through September 30, 2004
Project Period Covered by this Report: October 1, 2003 through September 30, 2004
Project Amount: $299,855
RFA: Decision-Making and Valuation for Environmental Policy (2001) RFA Text | Recipients Lists
Research Category: Environmental Justice
Objective:
The objectives of this research project are to: (1) describe household choices in a framework that recognizes the assumptions and informational requirements of revealed preference methodologies (e.g., hedonic, random utility, and other models used to value changes in site specific amenities); and (2) incorporate a more explicit account of how economic activities impact environmental resources in these models. The specific focus of our research involves the quality of watersheds and related resources.
Our approach is intended to compare three methods (hedonic property value, random utility, and locational equilibrium models) applied to a common area (Wake County, North Carolina). This comparison was viewed as a first step toward developing an integrated model. We proposed to use several existing databases, augment them with locationally delineated information on soil characteristics and water quality, and collect new data on a sample of households living in Wake County. The first database to be used for this analysis involves housing sales for Wake County. The second database involves measures of water quality likely to be impacted by land uses, relying on U.S. Environmental Protection Agency (EPA), U.S. Geological Survey, and local North Carolina Department of Water Quality measurements as well as a detailed evaluation of watersheds commissioned by Wake County. We discovered in the process of developing these data that the consulting firm CH2M Hill had undertaken a detailed evaluation of the Wake County watersheds. We were able to assemble all the data developed for their evaluation as well as the details of their methods. We also planned and implemented a mail survey of county homeowners to elicit information about their recreational activities and reasons for selecting their homes, as well as conventional demographic and economic characteristics. The last data set to be considered in our analysis is the North Carolina component of the 2000 National Recreation Survey.
Progress Summary:
Our activities during Year 3 of the project focused on work associated with the conceptual and empirical modeling of the interrelationships between hedonic, random utility, and locational equilibrium models for estimating the economic value of environmental amenities related to watershed resources. This research is divided into three sets of activities: (1) cleaning of the stated choice, recreational use, and demographic and economic data from the household survey; (2) development of conceptual analyses of the relationship between amenities and the various choice margins available to recover preference information; and (3) empirical analyses and preparation of manuscripts.
Survey Data Cleaning
Dr. Smith managed the survey and data entry activities. Cleaning required developing an inventory of all response codes falling outside the anticipated domain (e.g., reporting ranges rather than numbers, text answers, answers that displayed inconsistencies with questions, etc.). Cross checks for entry errors also were developed as part of data cleaning. The process involved a systematic review of all nonstandard responses and outliers by Dr. Smith working with a graduate student (Brian Stynes) supported by the project. Each outlier was rechecked against the original questionnaire. The end product of the process was a set of recodes documented in STATA do-files that provides all the transformations to the original data file to convert nonstandard responses and correct data entry errors. These do-files were reviewed by Drs. Palmquist and Phaneuf and then archived.
A second component of the data analysis required three steps. Many respondents misinterpreted the maps and site codes distributed with the mail surveys. A separate map and site legend was distributed for short outings, for 1-day trips, and for 2-day trips. Many respondents did not follow directions and used the same map for all three types of trips. In addition, respondents marked map sites that were not identified in either legend.
The first step in the process required a complete check of the site identification to evaluate which legend was used correctly and, if errors were made, to correctly identify the actual site selected for use. Dr. Smith worked with a graduate student (Brian Stynes) to review every returned survey and developed a protocol for corrections. These changes were documented and archived in STATA Do-files. This process assures the original survey remains intact and the decisions required because of inconsistencies in respondent answers are documented.
The second step in the process required geocoding all sites identified in the source materials as well as the new sites identified by respondents. This process is essential for specifying the choice set of recreation available to each household and for estimating the distance and travel time for each potential choice alternative. Two graduate students (Jaren Pope and Brian Stynes) developed a complete inventory of 231 sites, including both the sites identified in the survey materials and the new sites reported in the survey responses.
The last step in the process required estimating the distance and travel time to each site. To assure maximum flexibility in composing models to describe recreation choices, we structured a separate database. These data include the distance and travel time for every respondent to every site. This implies estimating these variables for every household regardless of whether they actually used all sites. These data serve as a key resource in implementing random utility models to describe how the description of individual behavior is sensitive to the description of the choice set characterizing substitution alternatives.
Constructing these estimates for local outings required significantly more detailed checking than is normally required with the use of available route, distance, and time estimation software because these programs are designed for use with longer trips. The travel time and distance between each survey respondent’s house and each recreation site were calculated using the PCMiler software. PCMiler calculates distances between latitude-longitude points along a road network, and then estimates travel times using speed-limit information. It has been commonly used in travel cost models to calculate travel times and distances, however, it is designed for the trucking industry. Thus, one of PCMiler’s drawbacks is that the road network it uses to calculate the times and distances is composed of roads that are accessible to trucks. The error in estimating travel times and distances using a network of major roads accessible to trucks is likely to be largest for the sites used for local outings.
To decrease the measurement error that might be introduced with PCMiler in these cases, an alternative strategy was developed using ArcView. Using a comprehensive road network, including minor roads developed by “Tigerline” (Census 2000 TIGER/Line Data is provided by the U.S. Bureau of the Census) for Wake County, travel times and distances from survey households to local recreation sites were calculated within ArcView. One exception was the calculation of times and distances to Jordan Lake, a popular local recreation destination located just outside Wake County. By calculating the travel time and distance to the county line, and by adding this time and distance to the PCMiler estimate from the county line to the site, a more accurate time and distance was generated to this recreation site. To determine if the other ArcView estimates were more accurate than the PCMiler estimates for the local recreation site, a sample of 20 households and 10 recreation sites were compared to estimates produced by MapQuest, an online service that also uses major and minor roads in their calculations. Based on the Sum of Squared Errors, it appears that the ArcView estimates were more accurate than the PCMiler estimates for the local recreation sites. Thus, we replaced the PCMiler times and distances with the ArcView times and distances for local recreation sites.
Development of Conceptual and Computational Analyses for Research
Conceptual . Developing integrated models of the role of environmental amenities related to watersheds in economic choice models requires considering how these nonmarket services contribute to individual preferences and the resulting implications for what can be learned from consumers’ choices.
The first set of our research involved the use of alternative preference restrictions. Drs. Palmquist and Smith, in separate research, reconsidered the treatment of weak complementarity in revealed preference models. These activities were undertaken as extensions to research already underway. The result was extensive revision to three initial papers and the preparation of two new papers. Two of these five papers have been accepted for publication and three are being revised in response to revise and resubmit decisions by the journals.
In addition, Drs. Phaneuf and Smith completed extensive revisions for their North Holland Handbook series chapter for the Handbook in Environmental Economics, Volume II, on recreation demand models. This work was completed as part of the required conceptual work associated with developing recreation models to evaluate the amenity services of watershed resources. The chapter is now final and it is expected that the volume will appear in 2005.
Finally, two new sets of conceptual models were developed during the year. The first of these analyses was motivated by insights derived from the focus groups. It arises from the recognition that conventional models for describing the opportunity costs of time are unlikely to be relevant to short (i.e., 3-4 hours or less), local recreational outings. As a result, Dr. Palmquist led the process of reconsidering how a household production framework could be used to recover estimates of the short run or constrained value of time for these objectives. The model was successfully implemented using our survey responses on actual time allocations together with a stated choice question offering flexible services for household activities. A paper describing the model (Palmquist, Phaneuf, and Smith, 2004) was presented at the annual meeting of the American Agricultural Economics Association in August 2004, and the Heartland Environmental Economics Conference at Iowa State in September 2004. This paper is being revised for submission to a journal in early 2005.
The second new conceptual analysis was proposed by Dr. Phaneuf and implemented by Drs. Smith and Palmquist. This framework suggests a new interpretation of choice models as a source of indices of recreation alternatives that allow the complexity of spatially delineated and diverse recreation alternatives to be described as a consistent economic index—the expected value of the maximum utility that can be derived from a choice set available to homeowners because of the selection of a specific neighborhood.
The logic associated with this proposal was implemented, taking advantage of the spatial structure built into the design of our integrated hedonic, household survey, and water quality database. This allowed estimation of 19 separate random utility models—one characterizing the recreation choice alternatives for each of the Multiple Listing Zones used to delineate how realtors characterize housing neighborhoods in Wake County.
Estimates of these indices then were used to measure an access index and a quality-adjusted index of the expected recreation opportunities for each home. With the spatially linked hedonic database, it was possible to estimate an upper bound for willingness to pay for improvements in watershed amenities as reflected in enhanced recreation opportunities.
The results of the conceptual and empirical analyses are described in a new paper (Smith, Phaneuf, and Palmquist, 2004), which was presented by Dr. Smith at EPA’s Valuation of Ecological Benefits: Improving the Science Behind Policy Decisions meeting held in October 2004. These results also are scheduled to be presented by Dr. Phaneuf at the Allied Social Science Association (an Association of Environmental and Resource Economists session) in January 2005.
This analysis also takes advantage of the Palmquist, et al. (2004) proposed methodology for measuring the opportunity cost of time to compare the hedonic upper bound with more conventional recreation-based measures of the willingness to pay for quality enhancements in recreation sites. This paper probably will be restructured into at least two separate manuscripts. The paper provides a direct implementation of the integration strategy proposed in our initial research outline, and as a result illustrates the value of spatial integration of hedonic, recreation, and water quality data.
Computational. Our activities in developing methods to improve the feasibility of estimating complex mixed discrete/continuous models have been undertaken in parallel with the data cleaning, conceptual analyses, and paper presentation. One aspect of this work is a paper (von Haefen and Phaneuf, 2004) on the opportunities to use Bayesian methods and Gibbs sampling in the estimation of Kuhn-Tucker corner solution models.
The only area where our planned research has been behind relates to the locational equilibrium model. As our 2003 annual report suggested, our estimates for the price indices estimated at the Multiple Listing Service level, as well as measures for the watershed quality indices, appear consistent with the ordering properties associated with the locational equilibrium model. Estimation has proved more challenging.
Our objective has been to develop a MATLAB set of source code for the framework. A new graduate student familiar with MATLAB (Nicolai Kuminoff) has been added to the project team for the spring semester of the 2004-2005 academic year and for the summer to recode the initial programs and work with Dr. Smith to complete the locational equilibrium model. This also is an area for this student’s Ph.D. thesis research. As a result, we expect to be able to complete this final aspect of our proposal activities during the first half of the last year of the project.
Empirical Analysis and Preparation of Manuscripts
With the data cleaning completed, we undertook analyses of both stated choice and revealed choice data from both the survey data and the hedonic housing sales data for five sets of modeling activities:
- Conventional hedonic models for local water quality measures;
- Evaluation of price indices for alternative spatial definitions of local neighborhoods for the locational equilibrium model;
- Joint estimation of revealed time allocation and stated time service purchased for the model of the opportunity cost of time;
- Nineteen (19) random utility models for local recreation outings based on our homeowner survey; and
- Integrated hedonic and random utility models to estimate an upper bound for the willingness to pay for enhanced local recreation opportunities.
Four of these activities, as discussed in the preceding section, contributed directly to research papers from our project. Research from our activities resulted in five published or forthcoming papers, three papers that received revise and resubmit decisions, and three additional completed papers that have been presented at national or specialty workshops. One of these new papers is under review and the other two are being prepared for submission to refereed journals.
Future Activities:
We will complete a paper on the water quality databases and the locational equilibrium modeling. In addition, we have planned a series of papers that describe how the conceptual research and empirical database on watershed services can be used to measure the economic benefits of protecting or enhancing these services. During the last year of the project, we plan to devote the majority of our time to preparing and revising these papers. We anticipate activities in five areas.
Valuing Time
Based on the comments received at the American Agricultural Economics Association session and the Heartland Environmental Economics Conference, we are revising the Palmquist, et al. (2004) paper for submission to a professional journal.
Our survey also includes a conjoint question eliciting time tradeoffs in another format. Preliminary analysis of these results suggests that a random utility model is quite successful in describing these stated choices in a transportation context. The implicit opportunity costs of time are different from what was measured using the household production logic. Of course, the nature of the time usage and discretion in using time savings was quite different. We expect to also prepare a paper describing these distinctions.
Modeling Recreation Choices
Our analysis to date has focused on one component of the recreation data that we collected: the local outings. We expect to undertake several sets of analyses of these data:
- Evaluating the robustness of our descriptions of the choice of sites for local outings;
- Modeling 1- and 2-day recreation trips for these households and investigating their interrelationships;
- Investigating the role of an area of water quality measures for all three types of recreation; and
- Comparing our findings to the results from the 2000 National Survey on Recreation and the Environment for North Carolina households in urban areas.
We expect that several of these analyses will lead to distinct papers.
Choice Margins: Recreation Demand and the Hedonic Models
Our paper presented at the EPA workshop can be divided into at least two distinctive papers. We also need to reflect the assessment of the roles of water quality measures for the recreation demand models (i.e., the random utility model analyses) in our analyses of the effects of recreation opportunities for housing values.
Once this is complete, we expect to have sufficient research for two papers completed. The first paper will describe the conceptual logic for our index of recreation opportunities. The second paper will illustrate how it can be used in policy applications.
Locational Equilibrium
One important aspect of our proposed research involved comparing travel cost demand, hedonic, and locational equilibrium models’ estimates of the value of improving the quality of urban watersheds. Once the programming effort described above is completed, we expect this analysis will be possible.
Documentation of Databases
We have a documented record for the Wake County water quality records database. We have distributed these data to several other researchers who requested them. We also expect to be able to provide a database that includes the majority of the nonconfidential records for our survey.
Journal Articles on this Report : 6 Displayed | Download in RIS Format
Other project views: | All 20 publications | 10 publications in selected types | All 6 journal articles |
---|
Type | Citation | ||
---|---|---|---|
|
Atasoy M, Palmquist RB, Phaneuf DJ. Estimating the effects of urban residential development on water quality using microdata. Journal of Environmental Management (in review, 2004). |
R829508 (2004) |
not available |
|
Banzhaf HS, Smith VK. Meta analysis in model implementation: choice sets and the valuation of air quality improvements. Journal of Applied Econometrics 2007;22(6):1013-1031. |
R829508 (2004) R828103 aka R826609 (Final) |
Exit Exit |
|
Palmquist RB. Weak complementarity, path independence, and the Intuition of the Willig Condition. Journal of Environmental Economics and Management (in press, 2004). |
R829508 (2004) |
not available |
|
Smith VK. Krutilla’s legacy: twenty-first century challenges for environmental economics. American Journal of Agricultural Economics 2004;86(5):1167-1178. |
R829508 (2004) |
not available |
|
Smith VK, Evans MF, Poulos C, Banzhaf S. Rehabilitating weak substitution. Journal of Environmental Economics and Management (in revision, 2004). |
R829508 (2004) |
not available |
|
Smith VK, Banzhaf HS. Quality adjusted price indexes and the Willig condition. Economics Letters 2007;94(1):43-48. |
R829508 (2003) R829508 (2004) R828103 aka R826609 (Final) |
Exit |
Supplemental Keywords:
geographic information system, GIS, economic benefits, land cover, water quality, random utility model, hedonic model, locational equilibrium model, Wake County, North Carolina, NC, recreation, opportunity cost, choice, environmental amenities, household, freshwater, survey, homeowner, watershed, geomorphological data, economic, social, and behavioral science research program, economics, economics and decision making, economics and business, decision making, behavior model, benefits assessment, cost benefit, ecosystem valuation, efficient household framework, environmental values, household choice, land cover changes, landowner behavior, measuring benefits, model aggregation methods, recreational value, residential property values, valuing environmental quality,, RFA, Scientific Discipline, Economic, Social, & Behavioral Science Research Program, Economics and Business, decision-making, Ecology and Ecosystems, Economics & Decision Making, ecosystem valuation, residential property values, model aggregation methods, economic benefits, landowner behavior, measuring benefits, valuing environmental quality, recreational value, cost benefit, environmental values, household choice, land cover changes, water quality value, behavior model, efficient household frameworkProgress and Final Reports:
Original AbstractThe perspectives, information and conclusions conveyed in research project abstracts, progress reports, final reports, journal abstracts and journal publications convey the viewpoints of the principal investigator and may not represent the views and policies of ORD and EPA. Conclusions drawn by the principal investigators have not been reviewed by the Agency.