An Overview of Environmental Economic Studies That Follow and Those That Fail

Document Type:Thesis

Subject Area:Economics

Document 1

It discusses main econometric issues such as observation variations in primary studies, method of selecting samples, summary of the data, heterogeneity of primary data, and heteroscedasticity. It evaluates one hundred and thirty meta-analyses in relation to estimation methods, followed by an exclusive summary of the findings. It concludes with recommendations for environmental economic studies that follow Keywords: Meta-Regression-Analysis; Environmental Valuation; data heterogeneity; INTRODUCTION In the recent years, various meta-analysis of Environmental Economic Studies have been conducted. An important question in environmental economic studies can be posed: which are the guidelines for Meta-Regression-Analysis for studies that follow and those that fail? No consensus answer, is yet to be reached. In order, to answer this question a comprehensive meta- regression analysis is needed1. Meta-analysis started to be applied in Economics in the 1980s. Stanley & Jarrell (1989), developed meta-regression analysis, which forms the framework for this study. This study shows that a significant number of studies out of the sampled one hundred and thirty studies in the environmental economics have been done in the last decade, thus meaning meta-analysis in environmental studies is gaining more focus. This paper aims to integrate widely diverse findings and to explain the variation among these empirical studies, and the reasons for variations. It will identify the effects of different econometric specifications, models multiple sources of potential bias and systematic variation, thereby explaining the excess variation ubiquitous in environmental economics research. It involves collecting different primary studies with similar results, such as the value of ecosystem services provided by lakes, water recreation, wetlands valuation, paying for fresh water quality, elasticity of gasoline and valuating scope and type of ecosystem conservation.

Sign up to view the full document!

Date and place of research, methods of econometrics, specific model used, size of sample, method of evaluation, attributes of the primary data form the explanatory variables. Binary dummies are sometimes used to represent regressors, and drawn from the primary studies. Also, regressors are at times represented as attributes of primary investigators. The objectives of this meta-analysis are: firstly, to provide an estimate of the effect-size, which will be calculated using the fixed effect size model. Secondly, to determine the heterogeneity in effect-sizes or the study to study variation, by identifying variables that explain this heterogeneity. Moreover, suggesting for improvement areas in the model technique, study design, and primary data. Thirdly, to offer dependent variable estimates in a sample of preset conditions. Fourthly, to offer result summary of primary study having different values.

Sign up to view the full document!

Earlier studies in environmental economics do not focus extensively on environmental issues, lack technicality, and uncomprehensive. This study focuses on the primary heterogeneity causes; methodological and factual. It identifies effects variations in various primary studies. This study focuses on the application of different methods of estimation for standard errors and variables. It handles heterogeneity in the literature in two main ways. Firstly, through meta-regression. Second, drawing estimates of effect-size from one primary study. Third, applying same modifications to achieve similar effect-sizes. Fourth, some primary studies sharing unobservable characteristic(s) such as same environmental commodity management. Lastly, some primary studies sharing an observable characteristic(s), such as drawing data from similar location, omitting crucial explanatory variable(s), using similar identical functional forms. This study identify, utilization of different estimates from similar primary studies.

Sign up to view the full document!

No Heterogeneity If the study uses similar methodology, each following primary study should provide values with no bias (Hedges & Olkin, 1985) Therefore, i = βi + ei (1) Where i is the primary study estimate. i ( i = 1,. N ), βi is the population value of this estimate, and ei is a sampling-estimation error. Assuming that the sampling error is normally distributed with mean zero and variance σi2 2. Unexplainable Heterogeneity If there is heterogeneity in the data, on basis of a Q-test or on a priori. Partially Explainable Heterogeneity This study shows that the R-square of the adjusted median is 0. most meta-regressions have imperfect explanatory power (Hox & Leeuw, 2003). Therefore, βi = α0 + α1 xil +…. αkxik + ui (4) Assuming, ui is a normally distributed sampling-estimation error with zero mean and variance r2. When equation (1) is substituted in equation (4) it yields equation (5) below.

Sign up to view the full document!

Selection and Coding of 130 Meta-Analyses This study examines useful environmental economics literature, both electronic and printed. Over three hundred studies were evaluated, and one hundred and fifteen of them chosen for examination with regard to method(s) or context. A few studies in the sampled studies offer multiple analyses, therefore the final sample comprised of one hundred and thirty meta-analyses obtained from one hundred and fifteen studies. This study groups each analysis into different topics based on primary studies. Topics analyzed include: value of ecosystem services provided by lakes, water recreation, wetlands valuation, paying for fresh water quality, elasticity of gasoline and valuating scope and type of ecosystem conservation, air pollution, suitable method, biodiversity, growth of the economy and regulation of the environment, pesticides and wastes, aquatic resources values, and wetlands resources.

Sign up to view the full document!

It summarizes forty two primary studies and one hundred and eight two observations, thus a mean of 6. for each study. Treatment of heteroskedasticity Heteroskedasticity treatment is present in numerous studies, although shown in table one forty three studies did not reveal treatment of any kind. Forty two studies were weighted regressions, twenty five used white, and twenty had standard errors. No Applications, Outliers, and Independence Forty-seven studies indicated zero data correlation treatment. Preliminary Meta-Analysis Majority offer a statistics summary, as well as explanatory variables. Two studies indicated their weighted means. Six studies illustrated their data in a graphical manner. Meta-Regressions Estimation Some studies used complex models of regression. Therefore, this study recommends that follow should try out multiple models of regression. Section Summary This study indicates that these issues do not occur frequently, the solution can be different methods of econometrics.

Sign up to view the full document!

Firstly, selecting effect sizes consistently, and using multiple samples solves the issues of heterogeneity. Secondly, application of panel method, and careful selection of estimation methods can be used to solve the issues of dependencies. Third, application of weighted least-squares solve issues of heteroskedasticity. Lastly, following studies need to define their procedures clearly, and adopt data regression diagnostics. The study reveals many studies have failed to follow the correct methodology for conducting a meta-analysis. They do not define the topic statement and problem statements. However, some studies can be said to have completeness in terms of their literature. As shown in table three some studies act as great examples for studies that follow, but they are still work in progress to attain perfection in all areas of concern. To sum it up, many studies that fail show usage of incorrect methods, and these constitutes to policy related issues.

Sign up to view the full document!

Cavlovic, T. A.  K. H.  Baker, R.  Rietveld, Flooding risk and housing values: An economic assessment of environmental hazard, Working paper 07-02, Purdue University, 2007. Hedges, L. V.  Fixed effects models, in H.  Cooper and L.  and H. O.  Stekler, Meta-analysis, Journal of Economic Perspectives 16, 2002, 225-26. Elvik, R.  Can we trust the results of meta-analyses? Transportation Research Record 1909, 2005, 221-29. P. Gawande, K. A meta-analysis of environmental Kuznets curve studies. Agricultural and Resource Economics Review, 29(1), 32-42. Hox, J. Kennedy, P. E. The use (and abuse) of meta-analysis in environmental and natural resource economics: an assessment. Environmental and resource economics, 42(3), 345-377. Study characteristics Studies number Study characteristics Studies numbers Publications 130  Treatments of Heteroskedasticity Articles 89 No treatment reported 43 Chapters 7 Other controls 20 Working papers 20 Explicit weights 42 Other 14 Huber-White se 20 Newey-West se 7 Primary Studies: Reporting: White se 25 No data provided 56 All data provided 26 Used Explicit weights: Some data provided 48 Other 7 Selection criteria provided 54 No.

Sign up to view the full document!

From $10 to earn access

Only on Studyloop

Original template

Downloadable