Elsevier

Journal of Policy Modeling

Volume 38, Issue 3, May–June 2016, Pages 436-447
Journal of Policy Modeling

Econometric general equilibrium modeling

https://doi.org/10.1016/j.jpolmod.2016.02.004Get rights and content

Abstract

The point of departure for the study of the impact of energy and environmental policies is the neo-classical theory of economic growth formulated by Cass (1965) and Koopmans (1967). The long-run properties of economic growth models are independent of energy and environmental policies. However, these policies affect capital accumulation and rates of productivity growth that determine the intermediate-run trends that are important for policy evaluation.

Heterogeneity of different energy producers and consumers is critical for the implementation of energy and environmental policies. To capture this heterogeneity it is necessary to distinguish among commodities, industries, and households. Econometric methods are essential for summarizing information on different industries and consumer groups in a form suitable for general equilibrium modeling.

In this paper we consider the application of econometric general equilibrium modeling to the U.S., the economy that has been studied most intensively. The framework for our analysis is provided by the Intertemporal General Equilibrium Model (IGEM) introduced by Jorgenson and Wilcoxen (1990). The new version of the IGEM presented in this paper is employed for the evaluation of proposed legislation on climate policy by the U.S. Environmental Protection Agency (2011).

Introduction

Economic growth is a critical determinant of U.S. demand for energy. Emissions from the combustion of fossil fuels are an important source of U.S. requirements for pollution abatement. An essential first step in modeling the impact of energy and environmental policies is to analyze the growth of the U.S. economy. The appropriate point of departure for modeling U.S. economic growth is the neoclassical theory of economic growth, originated by Solow, 1956, Solow, 2005. This theory has been developed in the form appropriate for modeling the interrelationships among energy, the environment, and U.S. economic growth by Cass (1965) and Koopmans (1967)1.

Maler (1974) and Uzawa (1975) have presented neo-classical theories of economic growth with pollution abatement. A recent survey by Brock and Taylor (2005) summarizes the extensive literature on this topic. Solow, 1974a, Solow, 1974b has provided a theory of economic growth that includes an exhaustible resource. The classic textbook treatment of this topic remains that of Dasgupta and Heal (1979), who also give a detailed survey of the literature. In this paper we focus on pollution abatement, since the U.S. economy is relatively open to trade in natural resources, exporting coal and importing oil and natural gas.

In the neoclassical theory of economic growth wage rates grow at the same rate as productivity in the long run, while rates of return depend on productivity growth and the parameters that describe saving behavior. These long-run properties of economic growth are independent of energy and environmental policies. The neoclassical theory of economic growth also provides a framework for analyzing intermediate-run growth trends. These trends reflect the same determinants as long-run trends, but also depend on energy and environmental policies through their effects of capital accumulation and rates of productivity growth. In this context the “intermediate-run” refers to the time needed for the capital-output ratio to converge to a long-run stationary value. This often requires decades, so that the impact of energy and environmental policies on intermediate-run trends is critical for policy evaluation.

The slowdown of the U.S. economy during the 1970s and 1980s and the acceleration of growth during the 1990s and 2000s are striking examples of changes in intermediate-run trends. Two events associated with the slowdown – the advent of more restrictive environmental policies and the increase in world petroleum prices – have led to a focus on the interactions of energy supplies and prices, environmental quality and its cost, and the sources of economic growth. Similarly, Jorgenson (2009a) has demonstrated that the rapid development of information technology is the key to more rapid growth in the 1990s and 2000s.

Nordhaus, 2008, Nordhaus, 2010 has applied the Cass–Koopmans theory of economic growth to the analysis of energy and environmental policies in his important studies of climate policy for the world economy. The necessarily schematic modeling of technology limits consideration of issues that are very important in implementation of energy and environmental policies at the national level, such as the heterogeneity of different energy producers and different consumers. To capture this heterogeneity we distinguish among commodities, industries, and households. We employ an econometric approach to summarize information on different industries and different consumer groups in a form suitable for general equilibrium modeling. We next consider the application of the econometric approach to the U.S. economy.

The framework for our econometric analysis of the impact of energy and environmental policies is provided by the Intertemporal General Equilibrium Model (IGEM) introduced by Jorgenson and Wilcoxen (1990) and summarized below2. The organizing mechanism of this model is an intertemporal price system balancing demand and supply for products and factors of production. The intertemporal price system links the prices of assets in every time period to the discounted value of future capital services. This forward-looking feature is essential in dealing with the critique of macro-econometric models by Lucas (1976).

Forward-looking behavior of producers and consumers is combined with backward linkages among investment, capital stock, and capital services in modeling the dynamics of economic growth. These mechanisms are also featured in the Cass–Koopmans neoclassical model of economic growth. The alternative time paths for economic growth depend on energy and environmental policies through the impact of these policies on intermediate-run trends.

In disaggregating the economic impacts of U.S. energy and environmental policies, we preserve the key features of more highly aggregated intertemporal general equilibrium models like those of Nordhaus. One important dimension for disaggregation is to distinguish among industries and commodities in order to measure policy impacts for narrower segments of the U.S. economy. This makes it possible to model differences among industries in responses to changes in energy prices and the imposition of pollution controls and the policy impacts on markets for different fuels.

A second avenue for disaggregation is to distinguish among households by level of wealth and demographic characteristics. This makes it possible to model differences in responses to price changes and environmental controls. It is also essential for analyzing the distributional effects of energy and environmental policies, as in Jorgenson, Slesnick, and Wilcoxen (1992) and Jorgenson, Goettle, Ho, Slesnick, and Wilcoxen (2011). We begin our discussion of econometric intertemporal general equilibrium modeling by outlining the methodology.

At the outset of our discussion it is necessary to recognize that the predominant tradition in general equilibrium modeling does not employ econometric methods. This tradition originated with the seminal work of Leontief (1951), beginning with the implementation of the static input-output model. Leontief (1953) gave a further impetus to the development of general equilibrium modeling by introducing a dynamic input-output model. This model can be regarded as an important progenitor of the intertemporal general equilibrium model described below. Empirical work associated with input-output analysis is based on determining the parameters that describe technology and preferences from a single inter-industry transactions table.

The usefulness of the “fixed coefficients” assumption that underlies input-output analysis is hardly subject to dispute. By linearizing technology and preferences Leontief solved at one stroke the two fundamental problems that arise in practical implementation of general equilibrium models. First, the resulting general equilibrium model can be solved as a system of linear equations with constant coefficients. Second, the “input-output coefficients’ can be estimated from a single data point. The data required are now available for all countries that have implemented the United Nations, Commission of the European Communities, International Monetary Fund, Organisation for Economic Co-operation and Development, and World Bank (2009) 2008 System of National Accounts (2008 SNA).

An input-output approach to modeling environmental policy was introduced by Kneese, Ayres, and d’Arge (1970). Their work was particularly notable for introducing a “materials balance” implied by conservation of mass for all economic activities. Materials balances bring out the fact that material not embodied in final products must result in emissions of pollutants. These emissions accumulate as solid waste or enter the atmosphere or hydrosphere and reduce air or water quality. The assumption that pollutants are generated in fixed proportions to output is a natural complement to the fixed-coefficients assumptions of Leontief's input-output models in implementing the materials balance approach.

The obvious objection to the fixed-coefficients approach to modeling energy and environmental policies is that the purpose of these policies is to change the input-output coefficients. For example, the purpose of many environmental regulations is to induce producers and consumers to substitute less polluting inputs for more polluting ones. A prime example is the substitution of low-sulfur coal for high-sulfur coal by electric utilities to comply with regulations on sulfur-dioxide emissions. Another example is the dramatic shift from leaded to unleaded motor fuels in order to clean up motor vehicle emissions.

Johansen (1960) and Johansen (1974) provided the first successful implementation of an empirical general equilibrium model without the fixed-coefficients assumption of input-output analysis. Johansen retained Leontief's fixed coefficients assumption in determining demands for intermediate goods, including energy. However, he employed linear-logarithmic or Cobb–Douglas production functions in modeling the substitution between capital and labor services and technical change.

Johansen also replaced Leontief's fixed coefficients assumption for household behavior by a system of demand functions originated by Frisch (1959). Finally, he developed a method for solving the resulting nonlinear general equilibrium model for growth rates of sectoral output levels and prices and implemented this model for Norway, using data from the Norwegian national accounts. Johansen's multi-sectoral growth (MSG) model of Norway is another important progenitor for the Intertemporal General Equilibrium Model described below3.

Linear logarithmic production functions have the obvious advantage that the capital and labor input coefficients respond to price changes. Furthermore, the relative shares of these inputs in the value of output are fixed, so that the unknown parameters can be estimated from a single data point. In describing producer behavior Johansen employed econometric modeling only in estimating constant rates of productivity growth. Similarly, the unknown parameters of the demand system proposed by Frisch can be determined from a single point, except for a single parameter estimated econometrically.

All the essential features of Johansen's approach have been preserved in the computable general equilibrium models now employed in every area of applied economics. Dixon and Parmenter (1996) and Dixon (2016) have surveyed the extensive literature on Johansen-type models. The unknown parameters describing technology and preferences in these models are determined by “calibration” to a single data point. Data from a single inter-industry transactions table are supplemented by a small number of parameters estimated econometrically.

An important advantage of the Johansen approach, like input-output analysis, is the capacity to absorb the enormous amounts of detail available for a single data point. Dixon and Parmenter describe a model of Australia with 120 industries, 56 regions, 280 occupations, and several hundred family types. The current data base for the GTAP model of global trade constructed by Hertel (1999) and collaborators includes data on trade and production for 113 regions of the world and 57 commodity groups for the year 20044.

Section snippets

Econometric general equilibrium modeling

The obvious disadvantage of the calibration approach is the highly restrictive assumptions on technology and preferences required to make calibration feasible. Almost all general equilibrium models retain the fixed-coefficients assumption of Leontief and Johansen for modeling the demand for intermediate goods. However, this assumption is directly contradicted by massive empirical evidence of price-induced energy conservation in response to higher world energy prices beginning in 1973.

British

Econometric modeling of technology and preferences

As in the descriptions of technology by Leontief and Johansen, production in the econometric approach of Jin and Jorgenson is characterized by constant returns to scale in each sector. As a consequence, commodity prices can be expressed as functions of factor prices, using the non-substitution theorem of Samuelson (1951). The non-substitution theorem permits a substantial reduction in the dimensionality of the space of prices determined by the model. This greatly facilitates the solution of the

Conclusion

We conclude that econometric general equilibrium modeling is a very important addition to economic methodologies for evaluating energy and environmental policies. The traditional approach originated by Johansen is based on calibration of the models of household and producer behavior to data for a single data point. This is a very useful simplification, but imposes highly restrictive assumptions on technology and preferences, such as the fixed coefficients assumption for intermediate goods

References (74)

  • R.J. Barro et al.

    Economic growth

    (2004)
  • E.R. Berndt et al.

    Production structure

  • British Petroleum

    BP energy outlook 2030

    (January, 2011)
  • M. Browning et al.

    Micro data and general equilibrium models

  • D. Cass

    Optimum growth in an aggregative model of capital accumulation

    Review of Economic Studies

    (1965)
  • L.R. Christensen et al.

    Transcendental logarithmic production frontiers

    Review of Economics and Statistics

    (1973)
  • P.S. Dasgupta et al.

    Economic theory and exhaustible resources

    (1979)
  • C. Dawkins et al.

    Calibration

  • A.S. Deaton et al.

    An almost ideal demand system

    American Economic Review

    (1980)
  • P.B. Dixon et al.

    Computable general equilibrium modeling for policy analysis and forecasting

  • P.B. Dixon

    The present state of CGE/AGE/Johansen-type modeling

    Journal of Policy Modeling

    (2016)
  • R. Frisch

    A complete scheme for computing all direct and cross demand elasticities in a model with many sectors

    Econometrica

    (1959)
  • W.M. Gorman

    Community preference fields

    Econometrica

    (1953)
  • T.W. Hertel

    Global trade analysis: Modeling and applications

    (1999)
  • E. Holmøy

    The development and use of CGE models in Norway

    Journal of Policy Modeling

    (2016)
  • E.A. Hudson et al.

    U.S. Energy Policy and U.S. Economic Growth, 1975–2000

    The Bell Journal of Economics and Management Science

    (1974)
  • L. Johansen

    A multi-sectoral model of economic growth

    (1960)
  • L. Johansen

    A multi-sectoral study of economic growth

    (1974)
  • D.W. Jorgenson et al.

    Relative prices and technical change

  • D.W. Jorgenson et al.

    Industry-level productivity measurement and the 2008 system of national accounts

    Review of Income and Wealth

    (2013)
  • D.W. Jorgenson et al.

    Environmental regulation and U.S. economic growth

    The Rand Journal of Economics

    (1990)
  • D.W. Jorgenson et al.

    Productivity and U.S. economic growth

    (1987)
  • D.W. Jorgenson et al.

    The transcendental logarithmic model of aggregate consumer behavior

  • D.W. Jorgenson et al.

    Carbon taxes and economic welfare

    Brookings Papers on Economic Activity: Microeconomics

    (1992)
  • D.W. Jorgenson et al.

    Information technology and the American growth resurgence

    (2005)
  • D.W. Jorgenson et al.

    Analyzing environmental policies with IGEM, an intertemporal general equilibrium model of U.S. Growth and the Environment

    (2009)
  • Cited by (6)

    • A hybrid stochastic model based Bayesian approach for long term energy demand managements

      2020, Energy Strategy Reviews
      Citation Excerpt :

      Another framework is bacterial foraging optimization algorithm which is a nature inspired global optimization algorithm that has been applied to all demand and supply sectors of energy systems [33,36]. Also, particle swarm optimization algorithm [37–39], price-based models [40], grey box algorithm [41–43], econometrics approach [49–52], statistical energy model [53–55] and semi-parametric algorithm [56,57] possess such a methodological approach. Traditional energy forecasting models which are mentioned in literature review have not considered uncertainties in energy systems.

    • Conditions accommodating a dominant stakeholder in the design of renewable air conditioning systems for tourism complexes

      2019, Energy
      Citation Excerpt :

      Therefore, the formulations for compromise solutions are linked to equality assumptions, which represent an idealized condition. This equality assumption has been used to address general economic problems [45] and specifically, problems associated with the operation of energy systems [46]. However, it is important to consider the dominance relationships in conflict resolutions.

    View full text