Edward Morey: Abstracts of Recent and Online Papers

Using choice experiments and latent-class modeling to investigate and estimate how academic economists value and trade off the attributes of academic positions, November, 2012

Edward Morey and Jennifer Thacher

Abstract: We investigate how economists would choose among academic positions as a function of the levels of attributes such as department rank, their teaching load, and their salary. We identified all 1735 faculty in the top fifty economics departments and asked each five "Would you prefer to work in Department A or Department B," questions." A latent-class choice model was estimated, indentifying four classes; only average cites/year influence choice.What is preferred and the rates at which attributes substitute for attributes, and for money, vary substantially across the four classes. The largest two classes made choices consistent with homo economicus.

From travel-cost models to moral philosophy, September 25, 2009

Edward Morey

Abstract: The intent of applied welfare economics is to distinguish good from bad, so it is a branch of moral philosphy/ethics. Said another way, welfare economists are in the business of determining whether a policy will increase or decrease happiness. For thirty-odd years I have done environmental valuation: trying to estimate in dollars or Euros individuals' willingness-to-pay for changes in the environment, estimating how and why willingness-to-pay might vary across individuals, and using these estimates to guide policy decisions. The foundations of this practice are the following: Economists believe, or assume, an individual has one, and only one, stable ordering of states of the world, believe the individual knows their ordering and will choose the highest-ranked availabe state, and believe that achieving a higher ranked state is preferred to a lower-ranked state in the sense that the individual is better off in the higher-ranked state. This essay reviews what I have learned - maybe I have it wrong - about these economic beliefs by scratching some surfaces in philosophy, ethics, psychology, neuroscience and evoluntionary biology. What are the implications for the way we, as economists, determine whether an action is good or bad?

Can your personality explain where and with whom you recreate? A latent-class site-choice model informed by estimates from a mixed-mode LC cluster model with latent-personality traits, Ecological Economics 138 (2017) 223-237.

Edward Morey and Mara Thiene

keywords: latent traits, personality, sensation seeking, extroversion, competitiveness, mixed-model latent-class cluster model, ordinal, nominal and cardinal indicators, latent-class choice model

additional tables not in the paper

Abstract: We test and find that personality traits interact with site characteristics and the ability of a potential companion to determine where, and with whom you recreate. 4605 mountain bikers chose between multiple pairs of hypothetical mountain-bike rides, and, in addition, answered Likert-scale questions on sensationseeking,competitiveness and extroversion. For each personality trait, a mixed-mode latent-class cluster model was estimated, accounting for that fact that the indicators can have ordinal, cardinal or nominal meaning. Most LC models ignore these distinctions. Our model also allows the scores on questions to be correlated, even after conditioning on class (typically assumed away). Then, a latent-class choice model of trail attributes and companion’s ability was estimated using the choice-pair data, with the estimated latent personality-traits as covariates. Three choice classes are identified and the odds of being in each varies by personality: estimated choice probabilities and WTP estimates vary significantly and substantially by class and personality type.

On the adequacy of scope test results, Ecological Economics 130 (Oct 2016): 356-60. http://dx.doi.org/10.1016/j.ecolecon.2016.05.022

D.J. Chapman, R.C. Bishop, W.M. Hanemann, B.J. Kanninen, J.A. Krosnick, E.R. Morey and R. Tourangeau

Abstract: Desrvousges et al. (2012) investigate criteria for judging the adequacy of scope test differences in contingent valuation studies. They focus particular attention on our study (Chapman et al. 2009), arguing that, while it demonstrated a statistically significant scope effect, the effect is too small. Unfortunately, DMT misinterpreted Chapman et al., an error that makes DMT's criticisms of our study invalid.

It's not where you do it, it's who you do it with? Journal of Choice Modeling 5(2012) 176-191

Edward Morey and David Kritzberg

Abstract: Individuals often recreate with others, but models of recreation-site choice (which ski area, climbing route, golf course, or bike trail) ignore the social aspect—a trait they share with most choice models. Site-choice models seek to explain site choice as a function of only the cost of visiting each site, the physical characteristics of the sites, income, and other characteristics of the individual. They ignore the influence of others on site choice. We find, using choice experiments, that having a companion and the companion's relative ability are critical determinants of site choice—what social psychology would predict. One will often choose a site less preferred in terms of its costs and characteristics if one has a companion of one's ability at the lesser site but not at the better site. Companions of comparable ability are preferred over companions that are better or worse. And, importantly, how one values the physical characteristics of sites depends on whether one has a companion. The magnitudes of our estimated companion effects suggest recreation-demand models that ignore them, all do, omit a critical endogenous variable. An implication is that observed trip patterns can be generated by social-interaction game playing (“where I go depends on where you go and …”), not utility maximization in isolation. This paper does not model the game; it estimates a recreator's utility/reaction function with companion effects, showing the importance of the social component.

Using Attitudinal Data to Identify Latent Classes that Vary in Their Preference for Landscape Preservation, May 2008.

Edward Morey, Mara Thiene, Maria De Salvo and Giovanni Signorello

Ecological Economics, Vol.68 (1-2), 536-546, December 2008.

Abstract: The likelihood of significant heterogeneity in preferences for landscape preservation should be accounted for when designing WTP questions, estimating WTP, and formulating resulting policy recommendations. Herein, heterogeneity in preferences for landscape preservation is investigated in the context of a latent-class model under the assumption of the existence of some finite number of preference classes/groups. The number of classes is estimated, so few restrictions are placed on the form of the heterogeneity. One estimates the probability that individual i belongs to class c where these probabilities are a function of observable characteristics of the individual (covariates); this is much more flexible than assuming, for example, that all farmers have the same preferences. This paper aims to identify preference classes for landscape preservation in the IBLEO, a rural and beautiful part of Sicily. Estimation of classes is performed using only attitudinal data consisting of answers to Likert-scale questions about the importance of preservation and why the respondent thinks preservation is, or is not, important. Summarizing the results, estimation indicates four distinct preference classes. The classes vary in the level of importance attached to preservation and the motivation for preservation (e.g. use vs. non-use motivations), and include one group that has little interest in preservation.

A parsimonious, latent-class methodology for predicting behavioral heterogeneity in terms of life-constraint heterogeneity,

Edward Morey and Mara Thiene

Ecological Economics 74 (2012) 130-144

Abstract: Our conjecture is that for many recreational activities a significant amount of the variation in the sites visited can be explained, and predicted, by variation in life constraints such as kids, BMI (body-mass index), fitness, skill, and health. The objective is to develop a parsimonious method for identifying behavioral heterogeneity caused by life-constraint heterogeneity and separating it from that caused by preference heterogeneity. We estimate, for two different recreational activities, with two independent data sets, how much behavioral heterogeneity can be attributed to life-constraint heterogeneity. We develop and estimate a stacked latent-class approach to life constraints, assuming individuals have many correlated life constraints. First, at the bottom of the stack, a latent-class life-constraint model is specified and estimated; the life-constraint class becomes a covariate in a behavioral latent-class model of participation and site selection. We find, with both simple statistics and behavioral models, that life-constraint classes explain a significant amount of the observed behavioral heterogeneity. Prediction is a critical reason to distinguish the influence of current constraints from the influence of current preferences: it is easy to directly observe life-constraint levels. Stacked latent-class models have many potential applications, besides ours.

Keywords: latent-class models, life constraints, preference heterogeneity, constraint heterogeneity, behavioral heterogeneity, choice, behavior, fitness, BMI, children, skill, exercise, health, disease, recreation, hiking, climbing, mountain biking

A joint latent-class model: combining Likert-scale preference statements with choice data to harvest preference heterogneity

Bill Breffle, Edward Morey, Jennifer Thacher

Environmental and Resource Economics Vol. 50(1). 83-110, 2011

Abstract: In addition to choice questions (revealed and stated choices), preference surveys typically include other questions that provide information about preferences.This preference-statement data includes questions on the importance of different attributes of a good or the extent of agreement with a particular statement. The intent of this paper is to model and jointly estimate preference heterogeneity using stated-preference choice data and preference-statement data. The starting point for this analysis is the belief that the individual has preferences, and both his/her choices and preference statements are manifestations of those preferences. Our modeling contribution is linking the choice data and preference-statement data in a latent-class framework. Estimation is straightforward using the E-M algorithm, even though our model has hundreds of preference parameters. Our estimates demonstrate that: (1) within a preference class, the importance anglers associate with different Green Bay site characteristics is in accordance with their responses to the preference statements; (2) estimated across-class utility parameters for fishing Green Bay are affected by the preference-statement data; (3) estimated across-class preference-statement response probabilities are affected by the inclusion of the choice data; and (4) both data sets influence the number of classes and the probability of belonging to a class as a function of the individual’s type.

keywords: Latent class, E-M algorithm, choice data, preference statements, Likert-scale, preferences, heterogeneity.

data and programs available at http://www.colorado.edu/economics/morey/dataset.html


Patient Preferences for Depression Treatment Programs and Willingness To Pay for Treatment: Heterogeneity and Anhedonia, December 2006

Edward Morey, Jennifer Thacher, and W. Edward Craighead

Journal of Mental Health Policy and Economics, Vol. 10 (2), 87-99, 2007.


Background: Current estimates of the societal costs of depression do not include estimates of how much individuals diagnosed with Major Depressive Disorder (MDD) would be willing to pay to eliminate their depression or how much they would have to be paid in order to accept continued depression. Choice experiment data and discrete-choice random-utility models provide a useful method for valuing changes in mental health and mental-health treatment programs.

Aims of the study: (1) To demonstrate how choice questions and discretechoice random-utility models can be used to estimate preferences over treatment programs for depression and willingness-to-pay (WTP) to eliminate depression. (2) To model and estimate the magnitude of the anhedonia impact of depression: consumption provides less utility when one is depressed. (3) To model heterogeneity in preferences for treatment programs for depression. (4) To derive preliminary estimates of WTP and willingness-to-accept (WTA) for eliminating and reducing depression, both with, and without side effects. Methods: The data are from a choice experiment survey of 104 individuals diagnosed with a new episode of MDD. Individuals indicated their preferred treatment from options that varied in effectiveness, hours of psychotherapy per month, use of anti-depressants, money costs, and side effects (weight gain, little or no interest in sex, inability to orgasm). Choices over treatment alternatives, including no treatment, were modeled using a discrete-choice random utility model. Preference parameters were estimated using maximum likelihood


Results and Discussion: Estimated WTP to eliminate MDD is large but side effects can substantially reduce WTP. Preferences over treatment programs, and WTP, vary as a function of the individual’s age, gender, income category, body-mass-index, and family composition. Some depressed individuals seeking treatment have a high estimated probability of choosing no treatment. Depression has both a direct and indirect effect on utility. The indirect effect on utility (the anhedonia effect), where the utility from consumption varies with emotional state, causes a divergence between WTP and WTA. The results may only be generalizable to those who are referred to or directly seek treatment at a mental-health clinic and should be replicated with a larger sample.

Implications: The WTP estimates suggest that depression imposes a high cost on society beyond the cost of treatment and the cost of lost output. Willingness-to-pay should be included in any benefit-cost analysis of whether additional societal resources should be allocated to the treatment of depression. Side effects from anti-depressants also impose a large cost on society. Estimates such as the ones reported here could provide a mechanism for better matching treatment programs to the patient and thus potentially minimizing non-adherence. The WTP estimates suggest that the pharmaceutical industry could earn significant revenues by making anti-depressants more effective, reducing their side effects, or both.

Key words: Choice questions, discrete-choice random-utility models, depression, treatment preferences, WTP, WTA, income effects, anhedonia JEL classification: I190, Q510

Calculating, with income effects, the compensating variation for a state change, February 2007

Edward R. Morey, and Kathleen Greer Rossmann
Environmental and Resource Economics, Vol. 39(2), 83-90, February, 2008.

Abstract: One can easily obtain exact closed-form solutions for the compensating variation (and equivalent variation) in the presence of income effects when the policy being evaluated can be described as a change in the state of the world and one is willing to assume the policy change does not change the individual's epsilon draw. Alternatively, if one assumes the policy changes the epsilon draw, the expectation of the compensating variation is a complicated integral, typically without a closed-form. The assumption that the policy does not affect one's epsilon draw is common, and often reasonable, but little discussed.

Valuing a change in a fishing site without collecting charactristics data on all fishing sites: a complete but minimal model March 05

Edward Morey and B. Breffle

American Journal of Agricultural Economics, Vol. 88 (1): 150-161, 2006

Abstract: Resource economists are often asked to value a proposed change at one, and only one, recreational site; the model we develop and estimate is applicable for those cases. The application is valuing the elimination of fish-consumption advisories on a large bay on Lake Michigan. The model is a minimal but complete: complete in that the choice set is not restricted, minimal in that only two conditional-indirect utility functions are estimated. It is utility-theoretic and one does not have collect characteristic data on all of the other fishing sites in the region. Data includes the number of trips each individual currently takes to Green Bay, answers to "would you prefer to fish Green Bay under conditions A or B?" and how often each angler says they would fish Green Bay under different sets of conditions.

Using Angler Characteristics and Attitudinal Data to Identify Environmental Preference Classes: A Latent-Class Model, May 2006

Edward Morey, Jennifer Thacher, and B. Breffle

Environmental and Resource Economics, Vol 34(1), 91-115, May 2006

Abstract: A latent-class model of environmental preference groups is developed and estimated with only the answers to a set of attitudinal questions. Economists do not typically use this type of data in estimation. Group membership is latent/unobserved. The intent is to identify and characterize heterogeneity in the preferences for environmental amenities in terms of a small number of preference groups. The application is to preferences over the fishing characteristics of Green Bay. Anglers answered a number of attitudinal questions, including the importance of boat fees, species catch rates, and fish consumption advisories on site choice. The results suggest that Green Bay anglers separate into a small number of distinct classes with varying preferences and willingness to pay for a PCB-free Green Bay. The probability that an angler belongs to each class is estimated as function of observable characteristics of the individual. Estimation is with the E-M (expectation-maximization) algorithm, a technique new to environmental economics that can be used to do maximum-likelihood estimation with incomplete information. As explained, a latent-class model estimated with attitudinal data can be melded with a latent-class choice model.

See also A Discussion of “Using Angler Characteristics and Attitudinal Data to Identify Environmental Preference Classes: A Latent-Class Model”, Provencher and Moore, Environmental and Resource Economics, Vol 34(1), 117-124, May 2006

and Combining attitudinal and choice data to improve estimates of preferences and preference heterogeneity: a FIML, discrete-choice, latent-class model, Dec. 2005, Breffle, Edward Morey, Jennifer Thacher

Using Patient Characteristics and Attitudinal Data to Identify Depression Treatment Preference Groups: A Latent-Class Model

Jennifer Thacher, Edward Morey and Edward Craighead

Anxiety and Depression, Vol. 21(2): 47-54, 2005

Abstract: A latent-class model is used to identify and characterize groups of patients who share similar attitudes towards treating depression. The results predict the probability of preference group membership on the basis of observable characteristics and answers to attitudinal questions. Understanding the types of preference groups that exist and a patient's probability of membership in each of the groups can help clinicians tailor the treatment to the patient and may increase patient adherence. One hundred four depressed patients completed a survey on attitudes towards treatment of Major Depressive Disorder. Analysis shows that treatment preferences vary among depressed patients. Three classes are identified that differ in their sensitivity to treatment costs and side effects. One class cares primarily about treatment effectiveness; side effects and the cost of treatment have little impact on this class's treatment decisions. Another class is highly sensitive to cost and side e®ects. A third class is somewhat sensitive to cost and side effects. Younger and male patients are more likely to be sensitive to treatment costs and side effects.

  Using the expected expenditure formula to calculate the exact expected compensating variation in GEV random utility models with income effects
Anders Karlstrom and Edward Morey, May 04.
Abstract:An exact formula for the expected compensating variation is derived for logit and nested-logit models with income e ects. Intuition, examples, and an application are provided. The appendix contains a formal proof. The formula is applied to estimate the E[cv]s salmon anglers in Maine would associate with changes in catch rates at Maine and Canadian Rivers.
See our web page for using the expected expenditure formula

A simple method of incorporating income effects into Logit and Nested-Logit models: theory and application
Edward R. Morey, Vijaya R. Sharma, and Anders Karlstrom
American Journal of Agricultural Economics 85(1), 250-255, February 2003.

Abstract: Substantive income effects are incorporated in a logit or nested-logit model by assuming utility is a piece-wise linear spline function of residual income. Specific
income data is not required, only income by category. Expected compensating variation is easily and accurately approximated by the difference between expected
maximum utility in the proposed and initial state, multiplied by the inverse of the individual’s initial marginal utility of money. This approximation is almost exact
because while any policy can, in theory, cause an individual to jump income categories, for most policies this probability will be very small.

Estimating the Benefits and Costs to Mountain Bikers of Changes in Trail Characteristics, Access Fees, and Site Closures: Choice Experiments and Benefits Transfer:
Edward R.Morey, Terry Buchanan, and Donald M. Waldman
Journal of Environmental Management 64(4), 411-422, 2002
Abstract: Mountain biking is very popular. Consequences are trail degradation and conflicts with hikers and other users. Resource managers often respond by closing trails to mountain biking, making mountain bikers worse off but improving the welfare of other users. The impact of such changes on mountain bikers is evaluated by developing and estimating a model that predicts the effects of trail characteristics, access fees, and characteristics of the individual on trail selection. The model is used to estimate each individual's per-ride consumer's surplus associated with different policies. They vary significantly as a function of the individual's gender, budget, and interest in mountain biking. Estimation is with stated preference data, specifically choice experiments. Hypothetical mountain bike trails were created and each surveyed biker was asked to make five pair-wise choices. A benefit-transfer simulation is used to show how the model and parameter estimates can be transferred to estimate the benefits and costs to mountain bikers in a specific area.

Estimating Malaria Patients' Household Compensating Variations for Health Care Proposals in Nepal:
Social Science and Medicine 57 (2003) 155-165.
Edward R.Morey, Vijaya Sharma, and Anne Mills
Abstract: A logit model is used to estimate provider choice from six types by malaria patients in rural Nepal. Patient characteristics that influence choice include travel costs, income category, household size, gender, and severity of malaria. Income effects are introduced by assuming the marginal utility of money is a step function of expenditures on the numeraire. This method of incorporating income effects is ideally suited for situations when exact income data is not available. Significant provider characteristics include wait time for treatment and wait time for laboratory results. Household willingness to pay is estimated for increasing the number of providers and for providing more sites with blood testing capabilities. Willingness to pay estimates vary significantly across households and allow one to assess how much different
households would benefit or lose under different government proposals.

Using Stated-Preference Questions to Investigate Variations in Willingness to Pay for Preserving Marble Monuments: Classic Heterogeneity, Random Parameters, and Mixture Models
Edward R. Morey, and Kathleen Greer Rossmann
Journal of Cultural Economics, Vol. 27 (3/4), 2-5-229, November 2003. Winner of the Werner Pommerehne prize for best paper in the JCE
Abstract: This paper investigates heterogeneity in the preferences/WTP (willingness to pay) to preserve marble monuments in Washington, D.C. This is done in the context of three different discrete-choice random-utility models. The main focus is to estimate a mixture model of choices over preservation programs. This model captures the best features of random-parameters models and models that assume preference parameters are deterministic functions of observable characteristics of the individual. The mixture model, and it alone, predicts that increased preservation is a bad for a significant proportion of young, non-Caucasians. That some proportion of the population might consider preservation a bad is a contingency that should be planned for in efforts to value cultural resources. Data and computer code are available at http://www.colorado.edu/economics/morey/dataset.html .

Estimating Recreational Trout Fishing Damages in Montana's Clark Fork River Basin: Summary of a Natural Resource Damage Assessment

Edward R. Morey, W.S. Breffle, R.D. Rowe and D. Waldman

Journal of Environmental Management 66(2), 159-170. Reprinted in The New Economics Of Outdoor Recreation (N. Hanley, Douglass Shaw and R. Wright, Eds.), Edward Elgar Publishing Ltd. 2003.

Abstract:This paper summarizes a natural resource damage assessment for the State of Montana. Mining wastes have caused signiÆcant reductions in trout stocks in a 145-mile stretch of Montana's Silver Bow Creek and Clark Fork River. To estimate economic damages from decreases in catch rates, we develop and estimate an individual-based utility-theoretic model of where and how often an angler will fish as a function of travel costs, catch rates, and other influential characteristics of the sites and individuals. The model includes resident and nonresident anglers who currently fish in Montana, and allows them to have different preferences. Demand parameters and expected catch rates are simultaneously estimated. The value of time is endogenously estimated as a proportion of the wage rate. Catch rates are linked to trout stocks through a stock-catch function. Collection of the angler data involved a three-step process: anglers were intercepted at 26 study sites, a subsample of anglers was selected to reflect the population trip-taking proportions to the study sites, and these anglers received follow-up surveys through the Æshing season. Avidity weights are used to correct for the higher level of avidity inherent in intercept samples.


Combining Stated-Choice Questions with Observed Behavior to Value NRDA Compensable Damages: Greenbay, PCBs and Fish Consumption Advisories

William Breffle, Edward R. Morey, Robert D. Rowe, and Donald M. Waldman

The Handbook of Contingent Valuation (J. Kahn and A. Alberini, Eds.), Edward Elgar Publishing Ltd. ISBN13 978 1 84064 208 7. March 2006.

Recreational Fishing Damages from Fish Consumption Advisories in the Waters of Green Bay November 1, 1999
William S. Breffle, Edward R. Morey, Robert D. Rowe, Donald M. Waldman, and Sonya M. Wytinck
Abstract: Lower-bound estimates of the compensating variations associated with elimination of Green Bay fish consumption advisories are estimated from a probit-type, random-utility model that combines observed frequency data with stated-choice data and stated-frequency data. Each data type has a separate error variance. The model incorporates correlations across data types. Variations on the basic model include estimation with subsets of the data and the incorporation of random parameters.

This work has benefitted greatly from comments and suggestions by Vic Adamowicz, David Allen, Robert Baumgartner, Rich Bishop, Don Dillman, Pam Rathbun, Paul Ruud, V. Kerry Smith, Roger Tourangeau and Michael Welsh.


Valuing and Preserving Site-Specific Cultural Resources in Italy: Some of the Issues Nov 7, 2001.

Abstract:The paper outlines the challenges and opportunities associated with valuing site-specific cultural resources. The discussion is framed around the issue of valuing cultural resources in Italy, a country with a large and deteriorating stock of such resources, resources that have great use and nonuse value for locals, Italians, and foreigners. All of the standard issues in nonmarket valuation are relevant, including, but not limited to, estimating use and nonuse values, incentive compatibility, describing the scenario, and the informational content of stated verses revealed preference data. Nonmarket valuation can help in determining what should, and should not, be preserved, and, if preserved, how to raise the required funds


  Gaussian Quadrature versus Simulation for the Estimation of Random Parameters: Some Evidence from Stated Preference Recreational Choice Data, August 30, 2004
William S. Breffle, Edward R. Morey, and Donald M. Waldman
forthcoming in Applications of Simulation Methods in Environmental and Resource Economics, A. Alberini and R. Scarpa, eds., Springer, Dordrecht, The Netherlands.
Abstract: In environmental economics, numerical simulation using random draws is the most commonly used method to estimate joint probabilities of individual choices in discrete choice random parameters models. This paper compares simulation to another method of estimation, Gaussian quadrature, on the basis of speed and accuracy. The comparison is done using stated preference data consisting of the answers to choice questions for fishing in Green Bay. Each sampled individual chose between a pair of Green Bay scenarios with different fishing conditions. Quadrature is found to be as accurate as simulation based on random draws, but Gaussian quadrature attains stability in estimated parameters considerably faster.

Joint Estimation of Catch and Other Travel-Cost Parameters: Some Further Thoughts
Edward R. Morey and Donald M. Waldman
Journal of Environmental Economics and Management, Vol. 40, 82-85, 2000
Abstract: Morey and Waldman (see next entry) propose and use an estimator that jointly estimates catch rates and the parameter on those catch rates. It is designed for situations where there is limited catch data for some of the sites in the choice set and takes advantage of the fact that trip patterns reflect variations in catch rates: anglers do not like to fish at sites with few fish. Train, McFadden and Johnson note that this estimator is particularly sensitive to omitted variable bias and suggest an alternative estimator, but their estimator effectively eliminates poor sites from the choice set.

Measurement Error in Recreation Demand Models: The Joint Estimation of Participation, Site Choice and Site Characteristics

Edward Morey and Donald Waldman

Journal of Environmental Economics and Management, Vol. 35, 262-276, 1998 (available online at www.idealibrary.com)

Abstract: This paper develops and demonstrates a solution to the errors-in-variables when there are repeated measurements on the variable measured with error. The application is the demand for recreational fishing sites, where an important explanatory variable differentiating sites is the unobserved expected catch rate. The usual proxy used in empirical work is the average of the observed catch rates for each site. Since observed catch rate is subject to sampling variability, the use of this proxy gives rise to an errors-in-variables problem which causes the estimated parameters on both expected catch and the other explanatory variables to be biased and inconsistent. In particular, the parameter on catch is biased downward, so the literature has underestimated the importance of catch on site choice. We show that consistent and efficient estimates of both expected catch rates and the parameters in the demand functions for the sites can be obtained by simultaneously estimating them in a maximum likelihood framework. Empirical implementation is straightforward. A policy relevant empirical example demonstrates the importance of simultaneously estimating expected catch rates and the influence of expected catch.

see also the above entry

TWO RUMs unCLOAKED: Nested-Logit Models of Site Choice and
Nested-Logit Models of Participation and Site Choice
Chapter 4 in "Valuing the Environment Using Recreation Demand Models,"
(C.L. Kling and H. Herriges, Eds.), Edward Elgar Publishing Ltd. 1999
An earlier version was presented at the W-133 Meetings, Tucson Arizona, February, 1994.

Abstract: Nested logit is increasingly advocated as a tool of recreational demand and benefit estimation. The intent of this short monograph is to lay out, in a simple fashion, the theory behind the nested-logit model of site choice and the nested-logit model of participation and site choice. Rigorous but straightforward derivations of the properties of nested-logit models are provided, including the probability of choosing a particular alternative, likelihood functions, expected maximum utility, and compensating and equivalent variations. Also discussed are the properties of the underlying distribution, estimation, regularity conditions, the interpretation of the scaling parameters, and the relation between those scaling parameters and the Independence of Irrelevant Alternatives ( IIA) assumption(s) embedded in both the nested logit model and its special cases. Examples are used to link the theory to recreational demand and benefit estimation.

Those familiar with the 1994 version of this paper will find a few corrections, many more references, and much more elaboration, particularly with respect to the extreme value distribution, regularity conditions, parameter estimation, and consumer's surplus estimation.

Using Contingent Valuation to Estimate a Neighborhood's Willingness to Pay to Preserve Undeveloped Urban Land :
William S. Breffle, Edward R. Morey, and Tymon S. Lodder
Urban Studies, Vol. 35, No. 4, 715-72, February 1988
Abstract: Contingent valuation (CV) is used to estimate a neighborhood's willingness to pay (WTP) to preserve a 5.5-acre parcel of undeveloped land in Boulder, Colorado, that provides views, open space, and wildlife habitat. Households were surveyed to determine bounds on their WTP for preservation. An interval model is developed to estimate sample WTP as a function of distance, income, and other characteristics. The model accommodates individuals who might be made better off by development and addresses the accumulation of WTP responses at zero. Weighted sample WTP estimates are aggregated to obtain the neighborhood's WTP. This application demonstrates that contingent valuation is a flexible policy tool for land managers and community groups wanting to estimate WTP to preserve undeveloped urban land.

Investigating Preference Heterogeneity in a Repeated Discrete-Choice Recreation Demand Model of Atlantic Salmon Fishing:
William Breffle and Edward R.Morey
Marine Resource Economics, Volume 15, pp. 1-20.
Abstract: Estimating a demand system under the assumption that preferences are homogeneous may lead to biased estimates of parameters for any specific individual and significantly different expected consumer surplus estimates. This paper investigates several different parametric methods to incorporate heterogeneity in the context of a repeated discrete-choice model. The first is the classic method of assuming utility to be a function of individual characteristics. Second, a random parameters method is proposed, where preference parameters have some known distribution. Random parameters logit causes the random components to be correlated across choice occasions and, in a sense, eliminates IIA. Simulation noise is discussed. Finally, methods are proposed to relax the assumption that the unobserved stochastic component of utility is identically distributed across individuals. For example, randomization of the logit scale, which is a new method, allows noise levels to vary across individuals, without the added burden of explaining the source using covariates. The application is to Atlantic salmon fishing, and expected compensating variations and changes in trip patterns are compared across the models for three policy-relevant changes in fishing conditions at the Penobscot River, the best salmon fishing site in Maine.

Two Nested Constant-Elasticity-of-Substitution Models of Recreational Participation and Site Choice:
An "Alternatives" Model and an "Expenditures" Model:
Edward R.Morey, William Breffle, and Pamela Greene
American Journal of Agricultural Economics. Vol 83, Issue 2, pp 414-427, May 2001.
Abstract: Two demand models of recreational participation and site choice are developed: an "alternatives" model and an "expenditures" model. Both assume maximization of utility over the year, so allow for diminishing marginal utility. They do not impose the restrictive assumption that where one goes on a trip is independent of where one plans to go on other occasions. Estimation is with a nested constant-elasticity-of-substitution preference ordering: it is relatively easy to estimate because of global regularity, it allows sites to be complements, and it has the potential to be locally flexible. The application is to Atlantic salmon fishing.

Modeling and estimating WTP for reducing acid deposition injuries to cultural resources: using choice experiments in a group setting to estimate passive-use values: July, 2000
Edward R.Morey, Kathleen Rossmann, Lauraine Chestnut and Shannon Ragland.
Chapter 7 in "Valuing Cultural Heritage," (Stale Narvud and Richard Ready, Eds.), Edward Elgar Publishing Ltd. 2002.
This paper summarizes and extends the work in "Valuing Acid Deposition Injuries to Cultural Resouces" - see below.
Abstract: A step-income effects logit model is used to estimate the WTP for reducing acid deposition injures to the 100 marble monuments in Washington, D.C. Visitation rates and travel costs indicate a significant use value. However, focus group results suggest that the most important reason for preservation is potential use by others, both now and in the future. The reduction in injuries are presented as hypothetical preservation programs that mimic the expected yet uncertain effects of the SO2 reductions within Title IX of the 1990 Clean Air Act Amendments. Individuals were presented with 10 pairs of preservation alternatives and asked to choose the preferred alternative from each pair. While choice experiments have a long history in marketing and transportion research, they have only recently been used to estimate use values for environmental commodities. The surveys were administered in-person to groups of respondents, a method that reduces implementation costs yet exploits many of the benefits of individual in-person surveys. Estimated WTP is significant and varies by gender, income group, age, and ethnicity. The marginal utility of money is found to be a step function of income. The WTP estimates from the choice experiments are compared with those from a CVM question.

Combining Responses to Actual and hypothetical Offers to Estimate WTA for Highly-Polluting Clunkers: Ordered Probit with Distinct, Noisy and Biased Bounds: June 1999,
Edward R.Morey and Tymon Lodder
Abstract: Accelerated vehicle retirement programs (AVRP) are receiving increasing attention as a possible means of cost effectively reducing air pollution. Whether AVRPs are efficient depends on the cost of acquiring vehicles as a function of the amount of pollution they produce. This paper estimates an individual's willingness to accept (WTA) an offer price for his or her vehicle in an AVRP. Data was collected from a pilot program in Denver.
Available information includes: the owner's response to the AVRP's actual offer of $1000, and the owner's response(s) to hypothetical offers from the AVRP. The data is a combination of revealed and stated preferences. Our emphasis is on the information content of the responses to the hypothetical offers.
One model is a multiple-interval ordered probit model of WTA that assumes bounds generated by responses to actual offers are distinct and bounds generated by responses to hypothetical offers are noisy. A special case is the conventional all-bounds distinct. A more restrictive case is bounds generated by hypothetical offers are discarded.
The distinct/noisy model statistically dominates all-bounds distinct; both generate similar predictions. In contrast, discarding the bounds generated by the responses to the hypothetical responses generates significantly different predictions. Parameter estimates indicate that an individual's willingness to sell to an AVRP is a function of the characteristics of the vehicle, not the owner. A second model investigates the possibility that hypothetical responses are systematically biased.
While the application here is WTA for clunkers, the model could be used to estimate WTA or WTP (willingness to pay) with any data that provides bounds but where the responses that generate the bounds are thought to vary in informational content.
A Repeated Nested-Logit Model of Atlantic Salmon Fishing with Comparisons to Six Other Travel-Cost Models
Morey. E.R., R. Rowe and M. Watson
American Journal of Agricultural Economics, Vol. 75 (3), August 1993, pp. 578-592.
Abstract: Participation and siet choice for Atlantic salmon fishing are modeled in the context of a repeated three-level nested logt model. Consumer's surplus measures are derived for different levels of species availability in the Penobscot River, the most important salmon river in New England. For comparison, six other travel-cost models are estimated. These include restricted cases of the nested-logit model, a partial demand model, and two single-site demand models. Comparisons across these models indicate the importance of modeling the participation decision, including income effects and of adopting a nested-logit structure rather than a single-level logit structure.  

Searching for a model of multiple-site recreation demand that admits interior and boundary solutions
Morey. E.R., D. Waldman, D. Assane and D. Shaw
American Journal of Agricultural Economics, Vol. 77 (1), February 1995, pp. 129-140.
abstract: For most recreation demand data sets, different individuals visit different subsets of the available sites. Interior solutions (i.e., individuals who visit all of the recreational sites) are not the norm. Bouldary solutions (i.e., individuals who do not participate, or who visit some, but not all of the sites) predominate. We critque eight demand models in terms of their ability to accommodate boundary solutions. Three models are recommended for consideration when there are multiple sites and the data set includes a signficant number of boundary solutions: a repeated-nested logit model, a multinomial share model, and a Kuhn-Tuckeer model.
Competition in Regional Environmental Policies When Plant Locations are Endogenous
Markusen, J., E.R. Morey and N. Olewiler
Journal of Public Economics. Vol 56, February 1995, pp. 55-77.
Abstract: A two-region model is presented in which an imperfectly competitive firm produces a good with increasing returns at the plant level. Production of the good causes local pollution. The firm decides whether to maintain plants in both regions, serve both regions from a single plant or shut down. If the disutility of pollution is high enough, the two regions will compete by increasing their environmental taxes (standards) until the polluting firm is driven from the market. Alternatively, if the disutility from pollution is not as great, the regions will usually compete by undercutting each other's pollution tax rates.
Keyword(s): Environmental policy; Tax competition; Plant location

Environmental Policy when Market Structure and Plant Locations Are Endogenous
James R. Markusen, Edward R. Morey, Nancy D. Olewiler

Journal of Environmental Economics and Management, Vol. 24, No. 1, Jan 1993, pp. 69-86.


Abstract: A two-region, two-firm model is developed in which firms choose the number and the regional locations of their plants. Both firms pollute, and market structure is endogenous to environmental policy. There are increasing returns at the plant level, imperfect competition between the 'home' and the 'foreign' firm, and transport costs between the two markets. At critical levels of environmental policy variables, small policy changes cause large discrete jumps in a region's pollution and welfare as a firm closes or opens a plant, or shifts production to/from a foreign branch plant. The implications for optimal environmental policy differ significantly from those suggested by traditional Pigouvian marginal analysis.


What Is Consumer's Surplus Per Day of Use, When Is It a Constant Independent of the Number of Days of Use, and What Does It Tell Us about Consumers Surplus?
Edward R. Morey
Journal of Environmental Economics and Management, Vol. 26, No. 3, May 1994, pp. 257-270.
Abstract: An individual's consumer's surplus per day of use for a change in the price of recreational site is the price change, so it is a constant, independent of the number of days of use. Consumer's surplus per day of use for a change in a site's characteristics is not, in general, a constant. When a constant compensating variation per day of use exists, it multiplied by the number of days at the site in the original state (proposed state) bounds the compensating variation, CV, from below (above). The average of these two approximations is an almost second-order approximation to the CV. Simulations indicate the approximation biases can be large.


Separability, Partial Demand Systems and Consumer's Surplus Measures
Hanemann, M., and E.R. Morey
Journal of Environmental Economics and Management. Vol. 22, 1992, pp. 241-258.
Abstract: In practice, complete demand systems are not estimated. Rather, either an incomplete demand system is estimated, or separability is invoked and a partial demand system is estimated. This paper considers the relationship between the conventional compensating variation (equivalent variation) and the corresponding welfare measure that can be derived from a partial demand system and the current budget allocation to the separable group. Even assuming the separability assumption invoked is appropriate, these partial measures provide, in general, only a limited amount of information about the compensating variation and no information about the equivalent variation. Great care is therefore needed when using partial welfare measures to evaluate policy.

Oil characteristics and the U.S. demand for foreign crude by region of origin

Kohli, U. and E.R. Morey
Atlantic Economic Journal, Vol. 18 (3), 1990, pp. 55-67.
no abstract

An introduction to checking, testing and imposing curvature properties: the true function and the estimated function

E.R. Morey
Canadian Journal of Economics, Vol. 19 (2), May 1986, pp. 207-235.
abstract: This paper has two parts. Part I is intented as a guide for checking the local curvature properties of estimated functions. A set of results is consisely presented, results that are now widely diffused and embedded in proofs and other materials that are of only secondary importance to the empirical economist. Part II considers estimation and how to use the curvature properties of the estimated functions to test the null hypothesis that the true function has the desired curvature property. Part II also considers how one imoses a particular curvature property on the estimated function.

The U.S. demand for foreign crude: a translog approach

Kohli, U. and E.R. Morey
The Journal of Energy and Development, Vol. 11 (2), Autumn 1986, pp. 115-133.
no abstract
Characteristics, Consumer Surplus and New Activities: A Proposed Ski Area
Journal of Public Economics, Vol. 26 (2), March 1985, 221-236.

Confuser Surplus
American Economic Review, Vol. 74, No. 1, 163-173, March 1984,
No abstract available

  The Choice of Ski Areas: Estimation of a Generalized CES Preference Ordering with Characteristics
The Review of Economics and Statistics, Vol. 66, No. 4, 584-590, November 1984,
Abstract: A Generalized CES (GENCES) preference ordering is developed and estimated. It incorporates
characteristics of both the individual and the activities. The GENCES is used to explain the share of ski
time and individual allocates to each ski area as a function of site characteristics, skiing ability, and costs.
The stochastic specification limited the shares to the 0-1 simplex. This specification was found to be more
appropriate than the conventional normality assumption. The null hypothesis that preferences are
homothetic and additive is rejected. Characteristics, ability, and costs are important determinants of
demand. The estimated elasticities provide numerous insights into skier behavior.

The Demand for Site-Specific Recreational Activities: A Characteristics Approach

Journal of Environmental Economics and Management, Vol. 8 (No. 4), December 1981, 245-271


Valuing Acid Deposition Injuries to Cultural Resources: May 14, 1997
*Note that this report consists of html documents rather than pdf documents.
Prepared for National Acid Precipitation Assessment Program by Edward Morey and Kathleen Greer Rossmann, University of Colorado at Boulder, and Lauraine Chestnut and Shannon Ragland, Hagler Bailly Inc.
Abstract: This study estimates the WTP to decrease the effects of pollution on an important set of cultural resources in the United States. A discrete-choice random-utility model of preservation alternatives is specified and estimated. From it, WTP is estimated for preserving the 100 most significant marble monuments in Washington, D.C. The three preservation alternatives valued span the likely reduction in injuries to these monuments that will result from the reduction in sulfur dioxide emissions expected to be achieved under Title IV of the 1990 Clean Air Act Amendments.
Prior to this study, little was known about the value of these resources other than that many people visit them. Focus group participants indicated that society values monuments for a variety of aesthetic, cultural, and historical reasons. While direct use values are important, "use by others" and "use by future generations" was typically listed as a more important reason for preservation, indicating significant passive use values and the need for a valuation technique that would include them.
Photos and scripts were used to identify the monuments, indicate their current condition, indicate what their condition would be at different points in the future if nothing was done to reduce the current rate of injury, and indicate their future condition under three different preservation alternatives. While there are no existing treatments for marble that would reduce the impact of acid deposition, those surveyed accepted the existence of such technologies. The accepted existence of such technologies made it possible to isolate the impact of acid deposition on marble monuments from all of the other injuries it generates.
This material was presented in-person to groups of individuals recruited in Boston and Philadelphia. A group approach allowed for the consistent and controlled presentation of extensive information-both verbal and visual. Any questions were answered one on one, so there was no group interaction. During and after the presentation, each individual provided written answers to survey questions. Each individual was asked to choose his or her most preferred option from each of ten pairs of options, where each option included a preservation alternative and a cost. Some pairs included no treatment as one of the options (always costing zero), but many of the pairs required that the individual choose between two preservation alternatives. Such a presentation encouraged individuals to focus on the trade-off between price and quantity of preservation; did not require individuals to formulate a precise WTP for an unfamiliar good; and allowed them to show support for preservation, yet still reject prices above their willingness to pay. The pair-wise data was used to estimate the parameters of a discrete choice RUM. The estimated WTP for reducing the rate of injury is significant, it increases with the degree of preservation, and it is a function of ethnicity, gender and income. Aggregating with weights indicates a substantial WTP on the part of residents of Boston and Philadelphia to preserve monuments in Washington, D.C.
After making the pair-wise choices, respondents were asked to choose, from a list of dollar amounts (including zero), his or her maximum WTP for the most effective preservation alternative. These data were used to estimate an interval model of WTP. The estimated WTP for preservation from the payment card data is essentially the same as from the pair-wise data when WTP is assumed to have a normal distribution, but much larger when WTP is assumed to have a log-normal distribution. While the answers to the payment card question were potentially affected by the prior pair-wise questions, the payment card estimates suggest that the WTP estimates from the pair-wise data were not biased upward by the decision to not include the status quo as one of the options in each pair.

Contact me via email | Table of Contents | Home Page | Top of Page

Last Update: 05-05-2017