مزایای تعادل عمومی برای بهبود محیط زیست: کاهش ازن پیش بینی شده در تحلیل آتی EPA برای حوزه هوایی لس آنجلس

This research demonstrates how locational equilibrium models can be used for benefit measurement with the detail required to match EPA's benefit analysis for the first Prospective Analysis. Using the projected changes in ozone concentrations for 2000 and 2010 together with the Sieg et al. (Int. Econ. Rev., forthcoming) estimates for household preferences for housing, education, and air quality, this paper measures general equilibrium willingness to pay for the policy scenarios developed for the Prospective study as they relate to households in the Los Angeles area. Benefits are evaluated taking account (at the household level) of initial air quality conditions, relocation based on changes in ozone, and price changes. The framework generalizes the partial equilibrium/general equilibrium comparisons available with conventional computable general equilibrium and property capitalization models. Estimated general equilibrium gains from the policy range from $33 to $2400 annually at a household level (in 1990 dollars).

Research using revealed preference methods in environmental economics has generally sought to measure the benefits associated with small changes in environmental quality. For example, hedonic property value methods offer measures of the incremental willingness to pay for small changes in site specific amenities. By contrast, the tasks that must be addressed by policy analysts require a framework capable of measuring the benefits from large, often spatially diverse, changes in amenities. In our previous research, summarized in [29], we demonstrated that locational equilibrium models offer a new framework for evaluating these types of policies. In that paper, we developed the basic general equilibrium framework, discussed estimation of the model's structural parameters, and illustrated how it could be used with actual changes in air quality conditions between 1990 and 1995. In this paper, we expand this line of research by considering EPA's policy alternatives, developed for the evaluation of the 1990 Clean Air Act Amendments and reported in the first Prospective Analysis. The main purpose of this paper is to demonstrate that locational equilibrium models of the type considered in [29] can be used for benefit measurement at the level of detail required for realistic policy assessments. In particular, this paper uses the same projected spatial variation in ozone concentrations as was developed for the Agency's benefit analysis for the LA Air Basin as part of the Prospective study to compute partial and general equilibrium benefit measures. Our analysis indicates that the estimated annual general equilibrium benefits in 2000 and 2010 associated with the ozone improvements due to continuing the policies mandated under the 1990 Clear Air Act Amendments will be dramatically different by income group and location within the South Coast Air Quality Management District. The gains range from $33 to about $2400 per household (in 1990 dollars). These differences arise from variations in air quality conditions, income, and the effects of general equilibrium price adjustment. This paper builds on a large literature in environmental economics considering the relationship between partial and general equilibrium welfare measures.1 Models for comparing partial (PE) and general equilibrium (GE) effects can take a variety of forms. Most of the empirical measures have used static computable general equilibrium (CGE) models that focus on consistent treatment of product and factor markets with boundary conditions that utilize both the zero profit condition for firms and (in static models) zero savings for consumers. For the most part, they have evaluated general equilibrium price effects. Examples of this research have found that large changes in environmental regulations [17], or in climate attributes influencing production [19], can result in appreciable price changes outside the sector(s) directly affected. Hicksian measures of consumer surplus indicated marked differences between the partial and the general equilibrium welfare measures. A second set of models considering PE/GE comparisons has been framed in terms of the capitalization of exogenous policies into land (or property) values.2 In this case, the analyses focus on: (a) the conditions when rents capitalize the effects of exogenous, location specific, changes in amenities or disamenities; and (b) the relationship between these rent changes and Hicksian welfare measures. To our knowledge there has been no effort, until recently, to use these models to develop numerical comparisons of PE/GE welfare measures.3 Most summaries of this second line of research focus on Starrett [32] and Scotchmer [26] and [27] as providing the most complete analyses of these two questions. However, both make restrictive assumptions that are relaxed in our analysis. For example, to derive the fundamental capitalization relationship for his “internal capitalization” model, Starrett assumes the equivalent of weak complementarity and additive separable preferences in the numeraire good. Scotchmer's [27] short run analysis also assumes income effects are negligible, noting in her evaluation of the importance of these assumptions that: “income effects will be troublesome when income classes are segregated in space…” (p. 72). Her long run analysis adopts a form comparable to Starrett that is transferable in the numeraire, so heterogeneity in the effects of income on demand for land (or housing) is not important. As a result, her long run corrections to short run benefit measures focus on the changes in population density induced by an exogenous amenity change. As noted above, this paper follows a different approach, which is based on a class of hierarchical locational equilibrium models first estimated by Epple and Sieg [12]. In Sieg et al. [29], we have shown how to adapt the locational equilibrium framework to estimate partial and general equilibrium benefits of large changes in spatially delineated local public goods and amenities. The approach used in this paper generalizes the PE/GE comparisons available with conventional CGE and property capitalization models in three ways. First, heterogeneity in household income effects is integral to our analysis of the effects of large-scale policy changes and is used, along with unobserved heterogeneity in household tastes for public goods, to characterize the locational equilibrium. Second, the framework allows for spatial delineation in household adjustments to policy changes at a scale consistent with the requirements of modern environmental policy analyses.4 Households can adjust to changes in a spatially delineated environmental public good by moving. As a result, the price an individual pays for housing, his selected level of the public good, and the amount of housing demanded are likely to be different from what would have been realized with the baseline community. Thus, the benefit measures reflect endogenous changes in both prices and the environmental resources. Third, and finally, there is a consistent accounting of the distributional effects of these large-scale policies, tracking how heterogeneous (in income and tastes for public goods) households gain (or lose) as a result of these policies. 5 Section two reviews the methodological framework used in this paper. Section three discusses how some of the policy scenarios used in EPA's Prospective Analysis can be evaluated within this framework. Section four summarizes the main findings of our study. In contrast to methods used by the EPA in the Prospective Analysis, our analysis relies exclusively on revealed preference arguments. This feature raises the question of whether the results of this study can be compared to EPA's. We discuss this question and related issues in section five. Section six offers some concluding remarks.