Dealing with Heterogeneous Preferences Concerned with Lake Urmia Restoration Using Multilevel Latent Class Model

اقتصاد و توسعه کشاورزی

Article 2, Volume 30, Issue 4, December 2017, Pages 285-296   

Document Type: Research Article

DOI: 10.22067/jead2.v30i4.52878

Authors

B. Hayati¹; M. Salehnia^{* 1}; M. Molaei²

¹University of Tabriz

²University of Urmia

Abstract

Introduction: In recent years, most environmental evaluation studies have managed the heterogeneity of preferences among individuals. Differing sensitivities are the basis for targeted communication programs and promotions. As consumer preferences and sensitivities become more diverse, it becomes less and less efficient to consider the society in the aggregate. In this research, we will show how multilevel latent class approach is ideal for these problems as it is possible to produce group -level parameter estimates. Investment in the ecological health of the Great Lakes basin is pivotal to its long-term economic success. Lake Urmia in the northwestern corner of Iran is one of the largest permanent hyper-saline lakes in the world and the largest lake in the Middle East. The lake’s surface area has been estimated as large as 6100 km2 but since 1995 it has generally been declining and was estimated from satellite data to be only 926 km2 in 2014. Worldwide Experiences indicate that sustainability of wetlands depends primarily on the extent to which local communities are active in their management. Local communities should, therefore, be fully engaged in the conservation and management of the Lake and its satellite wetlands. Hence, we will conduct a choice experiment for investigating public preferences on the nonmarket economic benefits of Lake Urmia restoration.
Materials and Methods: Cluster analysis is the evolution of aggregate estimation. Clustering algorithms find groups of individuals with similar tastes among the whole sample. The preferences of the individuals are estimated in a semi-individual way by assuming that the respondent utility is equal to the cluster utility, allowing for heterogeneities across segments of respondents, but not within the cluster. The fundamental assumption of standard LC models is that observations are independent. However, this assumption is often violated for example the analysis of populations that are hierarchically structured. The multilevel model can be used to define LC models for nested data. Four attributes consist of animal habitat, climate regulation and prevention from salt storms, aesthetic and ecotourism, and education and research were considered. The required data have been collected from 13 districts located in the northwest of Iran and Exogenous stratified random sampling applied as sampling strategy.
Results and Discussion: The estimated model identified three lower-level groups of individuals, each with their own structure of preferences. The first class is a representation of the average individual with no highly stressed motivations and show greater variability in the responses. For class 2, the most important attribute is the climate regulation and prevention of salt storms and the cost is the least important attribute for them. It seems that the second class has an environmental concerns than others. The third class often chooses those options with no gain in environmental levels even if there is deterioration in environmental conditions. Generally, this class shows opposite opinions. The estimation process detected two grand classes in which the covariates were the factor scores from the factor analysis. Grand class 2 is a little bigger than grand class 1, and most people are in lower-level class 2, the group with the highest proportion of high well-educated young people. In this paper, a novel feature named scale-adjusting was considered and two scale classes were determined. Scale class 1 refers to lower error respondents which more than half of the people belongs to. Whereas, scale class 2 is included of less certain or higher error respondents. To summarize, people from the closer districts to Lake Urmia were classified into one specific grand class with considerable homogeneity.
Conclusion: Human society is complex and this complexity and its manifestations impact the preferences of individuals in potentially many ways that currently are not well-understood, but may be associated with preference heterogeneity. Thus, it seems logical to suggest that not only socioeconomic characteristics influence consumer choices, but geophysical characteristics of their surrounding environment also shape their attitudes, behaviors, and beliefs. In particular, results suggest that it may be possible to incorporate many additional individual different measures in models to capture observable heterogeneity in systematic (deterministic) utility components instead of leaving them, exclusively, to random components. Future work should be directed toward better understanding of the complex interplay of environmental choices and market choices. The results also suggest that we can explain a large part of this preference heterogeneity in the systematic (deterministic) component of utility, which in turn may help to manage a geographical area with greater citizen participation and acceptance.

Keywords

Choice Experiment; Heterogenous preferences; Lake Urmia; Multilevel latent class model; Willingness to pay

References

1- Asparouhov T., and Muthen B.O. 2008. Multilevel mixture models: in Hancock G.R. and Samuelsen K.M. (eds). Advances in latent variable mixture models, IAP.

2- Bartholomew D.J. and Knott M. 1999. Latent Variable Models and Factor Analysis, London: Arnold.

3- Breffle W. B., Morey E. R., and Thacher J. A. 2011. A joint latent-class model: combining Likert-scale preference statements with choice data to harvest preference heterogeneity. Environment and Resource Economics, 50: 83–110.

4- Borghi C. 2009. Discrete choice models for marketing, New Methodologies for Optional Features and Bundles. Ms.c thesis. Mathematisch Instituut, Universiteit Leiden.

5- Eimanifar A., and Mohebbi F. 2007. Urmia Lake (Northwest Iran): A brief review. Saline Systems, 3, 5.

6- Farizo B. A., Joyce J., and Solino M. 2014a. Dealing with heterogeneous preferences using multilevel mixed models. Land Economics, 90: 181-198.

7- Farizo B. A., Louviere J. J., and Solino M. 2014b. Mixed integration of individual background, attitudes and tastes for landscape management. Land Use Policy, 38: 477-486.

8- Food and Agriculture Organization of the United Nations (FAO). 2003.Fisheries Management 2. The ecosystem approach to fisheries. FAO technical guidelines for fisheries. Rome.

9- Hanley N., Mourato S., and Wright R. 2001. Choice modeling approaches: A superior alternative for environmental valuation. Journal of Economic Surveys, 15: 435-462.

10- Izadimehr N. 2014. Estimating the value of nonmarket functions of Lake Urmia: Choice experiment approach. Ms.c thesis, Faculty of agriculture, University of Urmia.(In Persian)

11- Johnston R.J. 2007. Choice experiments, site similarity and benefits transfer. Environmental and Resource Economics, 38: 331-351.

12- Khodaverdizadeh M., Khalilian S., Hayati B., and Pishbahar E. 2014. Estimating the value of functions and services of Marakan protected area using choice experiment. Journal of applied economical studies of Iran, 3 (10): 267-290.)in Persian(

13- Kuhfeld W. F. 2010. Marketing research methods in SAS. SAS institute Inc. Cary, NC, USA.

14- Lancaster K. 1966. A new approach to consumer theory. Journal of Political Economics, 74: 217-231.

15- Magidson J., and Vermunt J. K. 2004. Latent class models, in Kaplan D. (ed.). The Sage Handbook of Quantitative Methodology for the Social Sciences, 175-198. Thousand Oakes: Sage Publications.

16- McFadden D. 1974. Conditional Logit Analysis of Qualitative Choice Behavior. In: Zarembka, P. (Ed.), Frontiers in Econometrics. Academic Press, New York.

17- Morey E., Thacher J., and Breffle W. 2006. Using angler characteristics and attitudinal data to identify environmental preference classes: a latent-class model. Environmental and Resource Economics, 34: 91-115.

18- Perman R., Ma Y., and Mc Gilvray J. 1996. Natural Resources and Environmental Economics. Translated byArbab H.R. longman.

19- Rose J.M., and Bliemer M. C.J. 2013. Sample size requirements for stated choice experiments. Transportation, 40: 1021-1041.

20- Salehnia M., Hayati B., Ghahremanzadeh M., and Molaei M. 2014. Estimating the value of improvement in Lake Urmia’senvironmental situation: An application of choice experiment. Journal of Agricultural Economics andDevelopment, 27(4): 267-276. (in Persian)

21- SEDAC .2010. Gridded population of the world: future estimates. Socioeconomic Data and Applications Center (SEDAC); collaboration with CIESIN, UN-FAO, CIAT. Accessed December 14, 2011 at: http://sedac.ciesin.columbia.edu/gpw.

22- Sharzei G., and Javidi N. 2011. Internalization of external costs of electrisity production using choice experiment. Journal of Economical Studies of Energy, 29: 1-29.)In Persian(

23- Train K.E. 2003. Discrete choice methods with simulation. Cambridge University Press.

24- Vermunt J.K. 2003. Multilevel latent class models. Sociological Methodology, 33: 213-239.

25- Vermunt J.K. 2008. Latent class and finite mixture models for multilevel data sets. Statistical Methods in Medical Research, 17: 33-51.

26- Vermunt J. K., and Magidson J. 2005. Technical guide for Latent GOLD 4.0: basic andadvanced. Statistical Innovations Inc.

27- West Azerbaijan Department of Environment. 2014.Lake Urmia in the past and present, consequences of the crisis.Research studies of the Lake Urmia, University of Urmia.

28- Zarghami M. 2011. Effective watershed management; Case study of Urmia Lake, Iran. Lake and Reservoir Management, 27(1): 87-94.