WORKING PAPER SERIES
CEEAplA WP No. 06/2006
Computable General Equilibrium Models: A
Literature Review
António Gomes de Menezes
Mário Fortuna
Francisco Silva
José Cabral Vieira
May 2006
Universidade dos Açores
Universidade da Madeira
Computable General Equilibrium Models:
A Literature Review
António Gomes de Menezes
Universidade dos Açores (DEG)
e CEEAplA
Mário Fortuna
Universidade dos Açores (DEG)
e CEEAplA
Francisco Silva
Universidade dos Açores (DEG)
e CEEAplA
José Cabral Vieira
Universidade dos Açores (DEG)
e CEEAplA
Working Paper n.º 06/2006
Maio de 2006
CEEAplA Working Paper n.º 06/2006
Maio de 2006
RESUMO/ABSTRACT
Computable General Equilibrium Models: A Literature Review
Applied general equilibrium models have become popular tools used on
ongoing economic policy debates. In this paper we discuss at length the most
proeminent features of applied general equilibrium models in a comprehensive
and non-technical way, thus accessible to the reader interested in economic
policy but with no prior formal exposure to economic modeling. We rationalize
the increasing political demand for such models as policy analysis tools. We
argue that applied general equilibrium models are best equipped to model
regional economies.
Keywords: Regional Economic Modeling; CGE Models.
JEL Codes: R11; C68.
António Gomes de Menezes
Departamento de Economia e Gestão
Universidade dos Açores
Rua da Mãe de Deus, 58
9501-801 Ponta Delgada
Mário Fortuna
Departamento de Economia e Gestão
Universidade dos Açores
Rua da Mãe de Deus, 58
9501-801 Ponta Delgada
Francisco Silva
Departamento de Economia e Gestão
Universidade dos Açores
Rua da Mãe de Deus, 58
9501-801 Ponta Delgada
José Cabral Vieira
Departamento de Economia e Gestão
Universidade dos Açores
Rua da Mãe de Deus, 58
9501-801 Ponta Delgada
Computable General Equilibrium Models: A
Literature Review
António Gomes de Menezes
Mário Fortuna
Francisco Silva
José Cabral Veiray
University of the Azores and CEEAplA
April 2006
Abstract
Applied general equilibrium models have become popular tools used on ongoing
economic policy debates. In this paper we discuss at length the most proeminent
features of applied general equilibrium models in a comprehensive and non-technical
way, thus accessible to the reader interested in economic policy but with no prior
formal exposure to economic modeling. We rationalize the increasing political demand
for such models as policy analysis tools. We argue that applied general equilibrium
models are best equipped to model regional economies.
Keywords: Regional Economic Modeling; CGE Models.
JEL Codes: R11; C68.
This paper was written within the development of a project to develop “An Instrument of Economic
Policy Analysis for the Azores”, …nanced by several institutions, namely, the US Department of Agriculture,
the Luso-American Development Foundation and the Regional Government of the Azores. The project was
managed by CEEAplA, an FCT supported center of the Universities of the Azores and Madeira.
y
We thank Ali Bayar for his extensive contribution, especially to Section 4. All authors with CEEAplA
and Department of Economics and Management, University of the Azores. Rua da Mãe de Deus. 9501-801
Ponta Delgada, Portugal. E-mail: menezesa@notes.uac.pt
1
1
Introduction
Policy makers and professionals alike are commonly interested in learning the direct and
indirect e¤ects of speci…c policy measures on economic outcomes. The ubiquitous existence
of rich data sets, both at a macro and at a micro level, combined with abundant computing
power, explains the increasing political demand for quantitative assessments of the economic
impacts of actual or eventual policy choices.
Since the beginning of the 1980s Computable General Equilibrium (CGE) models have
become increasingly popular tools to analyze the consequences of policy choices and the
allocation of resources in developing as well as in developed economies. In fact, CGE models
are used nowadays not only in universities and research institutions, but also by governments
worldwide in policy formulation and debate.
In this paper, we provide an introduction to the essence of CGE models at an economically
intuitive level, and, in the process, we rationalize the growing popularity of CGE models as
policy analysis tools. In order to achieve this, we set ourselves out to motivate answers to
the following questions:
1. Why do we need economic models?
2. Why do we need CGE models?
3. What are the building blocks of CGE models?
4. What are the weaknesses and the strengths of CGE models?
5. How do CGE models compare with other competing economic models?
6. How have CGE models been applied?
7. Where are CGE models headed?
As this paper aims a general audience interested in economic policy analysis, but not
necessarily with prior formal exposure to economic modeling, the paper employs an intuitive, non-formal, approach to answer the questions above. The paper provides, thus, a
comprehensive non-technical introduction to CGE models. However, we do refer to economic objects and concepts such as utility functions or neoclassical theory with little or no
2
attached explanations whatsoever, so some basic economic understanding is assumed. We
note that our primary interest lies in regional models. Hence, we focus on keys aspects of
regional modeling and, therefore, answer to the questions above bearing in mind our end
goal: to discuss the salient features of an applied regional general equilibrium model. We
shall use the terms applied general equilibrium models and CGE models interchangeably.
As there are many excellent surveys on CGE models, we borrow extensively from existing
work and capitalize, thus, on the existing literature.
The paper is organized as follows. Section 2 discusses why we need economic models.
Section 3 lays out basic desiderata of what constitutes a good economic model. Section
4 discusses the most prominent features of CGE models, including its building blocks and
applications. Section 5 reviews some well-known criticisms of CGE models. Section 6 points
out research directions and concludes.
2
The Need for Economic Models
First and foremost, it is instructive to motivate the need for economic models. Economic
models are, in their essence, abstract representations of reality, build to promote our understanding of the di¤erent economic interrelationships at play in the real world.
Since policy choices a¤ect economic outcomes and, concomitantly, our general well being,
there is the need to anticipate the e¤ects of such policy choices. As with other social sciences,
experimentation is not, usually, a choice. Trial and error learning processes are just inviable
since policy changes involve a¤ecting human experiences. Hence, economic models act as
laboratories where one learns the expected direct and indirect economic e¤ects of speci…c
policy choices. There is, thus, an obvious need for economic models.
Iqbal and Siddiqui (2001) motivate the need for economic models citing Rust (1997),
who, in turn, states the need of an economic model as follows: "To have a complete understanding [of economic systems], we need to be able to calculate detailed implications and
predictions of these abstract theories and determine whether the predictions of these models
are consistent with what we observe in the real world. So we can not pretend to have a
complete understanding of real economies until we can show that the detailed implications
of our theories provide su¢ciently accurate representations of the real world that we would
take our models seriously for forecasting and policy analysis."
3
In the same line of reasoning, Iqbal and Siddiqui (2001) cite Shoven and Whalley (1992)
who stated "...the virtue of using applied general equilibrium models is that, once constructed, they yield a facile tool for analyzing a wide range of possible policy changes. Such
analysis generates results that either yield an initial null hypothesis, or challenge the prevailing view. It may be that subsequently the conclusions from the model are rejected as
inappropriate; the assumptions may be considered unrealistic; errors may be unearthed, or
other factors may undermine con…dence in the results. But there will be a situation in
which the modeler and those involved in the policy decision process will have gained new
perspectives as a result of using the model." These new, formally educated, perspectives on
the functioning of the economy, that follow from the modeling process, are invaluable to the
policymaker.
3
Model Design
Having established the need for economic models, we are, then, left with the following
question: What is a good economic model? We answer this question bearing in mind that
our main goal is to analyze economic policy, and not necessarily, say, short-term forecasting.
This distinction is of the essence since analyzing economic policy implies that we should be
able to uncover causal relationships, from economic policy instruments to economic outcomes,
in an economically intuitive manner. However, if one is interested in forecasting per se, there
are statistical models that may deliver acceptable forecasts, in a statistical sense, with no say
whatsoever on the underlying economic mechanisms at play. To answer to the question of
what is a good economic model we borrow extensively from Devarajan and Robinson (2002)
who claim that for an economic model to be useful for policy analysis it should exhibit the
following desirable characteristics:
1. Policy Relevance. The model should explicitly link values of policy instruments to
economic outcomes of interest to policymakers.
2. Transparency. The links between policy instruments and economic outcomes should
be traceable and easy to explain.
3. Timeliness. The model should be based on recent, relevant data, if it is to be used
on ongoing policy debates.
4
4. Estimation and Validation. The model parameters must be validated for the domain of application of the model.
5. Diversity of Approaches. The model results should be validated by the results from
other models or approaches, whenever possible.
Policy Relevance and Transparency These two criteria argue bluntly for using
structural models. In structural models, some of the endogenous variables (that is, the
variables explained within the model) are expressed as functions of some other endogenous
variables. In contrast, in reduced-form models all endogenous variables are expressed as
functions of exogenous variables (exogenous variables are variables explained outside the
model, and, thus, taken as given by the modeler) and model parameters (that is, parameters
that describe economic behavior which are also taken as given by the modeler) only, but not
as a function of other endogenous variables. Hence, reduced-form models typically do not
explicitly model the links between policy instruments and economic outcomes or, at best,
they do it in a way that it is di¢cult, if not impossible, to identify the underlying structural
relations. Therefore, it is quite di¢cult to trace out the links between policy instruments
and economic outcomes when using reduced-form models. In short, reduced-form models
are "black-boxes" whose results are hard to explain. As it will become clear in the following
sections, one of the most prominent features of applied general equilibrium models is that
they are the opposite of what one would regard as a "black box".
Policy analysis ought to go hand in hand with policy agenda and public debate. Therefore,
policy analysis must focus on measures with which people relate to at a very familiar level,
such as, say, regional employment and growth. Hence, policy analysis typically should not
be con…ned to esoteric measures such as compensating variation or equivalent variation.
The transparency criteria argues for a very stylized model, since this modeling strategy
facilitates the tracing out of causal economic links between the policy instruments and the
economic outcomes of interest. However, policy relevance usually speaks for more institutional and sectoral detail, since policymakers are naturally interested in identifying at a quite
disaggregated level the winners and losers of a given policy package. Thus, policy relevance
promotes the employment of large, complex models. However, by construction, CGE models
explicitly model economic behavior at the individual level. Therefore, it is only natural with
5
CGE models to have detail, complexity and intuition simultaneously. There is no obvious
need to trade-o¤ detail for ease of interpretation in CGE modeling.
Timeliness The model should be based on recent, relevant data, if it is to be used
on ongoing policy choices. Quite naturally, for a model to be credible, it must employ upto-date data. If one uses historical data to draw lessons, then one must also be capable of
demonstrating that the structural form of the economy has not substantially changed, for
the lessons drawn to be of any value.
Estimation and Validation The issue of validation of a policy model also argues,
such as the transparency and the policy relevance criteria, for a structural model. The
domain of applicability of a reduced-form econometric model must be contained within the
historical range of the data used to estimate the model. The domain of applicability of a
structural model depends on the applicability of the structural relations and on the stability
of its parameters in the period of analysis.
In model validation, there is a trade-o¤ between using a structural model, which requires
estimation of a large number of structural parameters, and a reduced-form model with far
fewer parameters.
Structural models are highly non-linear, unlike most of its predecessors which were often
linear and easy to handle. However, and quite fortunately, the recent advances in computing
methods have turned this feature of non-linearity almost a non-bidding constraint.
Structural models, in general, and CGE models, in particular, are based on data such as
input-output tables which are not usually continuously available, but rather only for a few
periods, with long gaps. In contrary, reduced-form models are usually much less demanding
on the number of parameters that must be estimated.
However, some reduced-form (econometric) models require time-series data which, sometimes, are available for only a short period or not available at all. This is not necessarily
the case of CGE models, where one needs to be con…dent about the validity of the available
data, which may be available for one or a few recent years.
Diversity of Approaches An ultimate goal of constructing an economic model may
be policy guidance. However, before leaping from policy simulations to policy implemen-
6
tations one must feel con…dent with one’s model. In order to build up such con…dence, it
is convenient to assess our model predictions by confronting them with other approaches,
whenever possible. Hence, it is often the case that researchers are interested in confronting
results from di¤erent models.1
In summation, it follows from this scrutiny that, aside for those who care for short-term
forecasting per se, if our interest is in economic policy guidance, the economic model should
be a structural model. Given the appropriate data, CGE models are natural candidates to
ful…ll the desiderata above.
4
CGE Models
4.1
Prominent Features
In this section, we highlight prominent features of CGE models, bearing in mind that we are
mainly interested in modeling regional economies.
Microeconomic Principles As stated in Rege (2003), general equilibrium models
possess a rigorous structure derived from sound microeconomic optimization theory. This
feature leads to results with a clear structural or theoretical interpretation.
General Equilibrium Nature Households and markets are modeled in an extensive
way, and not in a partial setting. This is important because it builds complex interactions
between the di¤erent economic agents into the model. These interactions must be considered
as indirect e¤ects usually work in important and, sometimes, counter-intuitive ways. Such
general equilibrium nature naturally evokes the underpinnings of Walrasian general equilibrium (see Varian 1992). As lucidly exposed by Wing (2004) Walrasian general equilibrium
has in its cornerstone the following accounting rules:
Conservation of product: "... re‡ects the physical principle of material balance that the
quantity of a factor with which households are endowed, or of a commodity that is produced by …rms, must be completely absorbed by the …rms or households (respectively)
in the rest of the economy."
1
West (1995) provides an extensive account of a comparison between di¤erent models at the regional
level.
7
Conservation of value: "...re‡ects the accounting principle of budgetary balance that
for each activity in the economy the value of expenditures must be balanced by the
value of incomes."
These accounting rules are usually depicted in the form of the popular circular ‡ows.
Detail and Disaggregation Considerable detail is devoted to individual behavior.
Agents - households, …rms, governments, .... - and markets - commodities, factors of production, ... - are modeled in an explicit way with as much detail as desired. CGE models
are well known for treating with great detail issues such as, say, tax structure, production
structures, and so on. Acknowledging detail is a requirement for policy guidance, which
cannot make ado with aggregation and simpli…cation. As a result, CGE models are best
equipped to model regional economies in the sense that any good regional model should
consider regional speci…cities.
4.2
Building Blocks and Essence
The main building blocks are the agents and the markets that one needs to specify in the
model in order to capture all the key causal chains in the economy. The agents’ set comprises,
usually, households, …rms, governments and institutions, whereas the markets’s set entails
factors of production, commodities, and, less often, …nancial assets.
Agents’ behavior follows strictly from microeconomic principles, with explicit modeling
of objective functions, control variables, expectations rules and functional forms.
Markets operate according to institutional features and regulatory constraints. Market
structure is modeled following Industrial Organization models of both perfect and imperfect
competition.
The environment is either static, partial-dynamic (recursive) or fully dynamic, in the
sense that one may have long or in…nitely lived agents who care about the future and whose
problems are related over time. This is particularly important for matters such as capital
accumulation. Rege (2003) provides an interesting discussion of how CGE models treat time.
The general equilibrium nature of these models is their most distinguishing feature and,
concomitantly, explains their increasing popularity, over partial equilibrium models. General
equilibrium models explicitly model all key aspects of a given economy, in contrast with
8
partial equilibrium models which take as given important aspects of the economy. In fact,
if a policy instrument is used to achieve a change in an economic target variable, other
economic variables than the target variable will be a¤ected. The resulting …nal economic
outcomes may well di¤er substantially from the intention of policy makers and the direct
e¤ects predicted by partial equilibrium models.
A CGE model may be de…ned as the fundamental macroeconomic general equilibrium
links among incomes of various groups, the pattern of demand, the balance of payments
and a multisector production structure.2 In addition, a CGE model incorporates a set of
behavioral equations describing the optimizing economic behavior of the agents identi…ed in
the model and the technological, endowment and institutional constraints that these agents
face.
Devarajan and Robinson (2002) distinguish between stylized and applied CGE models.
Stylized models stay as close as possible to theory in order to isolate the empirical importance
of a link that theory suggests as potentially important. As the name indicates, stylized
models are not constructed with the intention of being realistic, since they are designed to
address a particular causal mechanism.
CGE models are aggregate representations of the economy and are based on the ‡ow
equilibrium in product and factor markets in real as well as in nominal terms. Opposite to
input-output models, both quantities and relative prices are endogenous, while consumption
is no longer exogenous but linked to income. The general equilibrium approach, in contrast
with partial equilibrium approaches which analyze the di¤erent sectors separately under
ceteris paribus assumptions, intends to model all links within the economy that represent
a transaction of money or goods. The analysis is usually based on comparative numerical
static analysis.
2
The model is in general equilibrium because a set of prices exists such that all excess demands for
commodities and services, in nominal and in real terms, are zero. That is, the model is in general equilibrium
when the Walras’ Law is valid and the equilibrium price vector is not equal to the zero price vector. In
mathematical terms: p:z(p) = 0 and z(p ) = 0, with z equal to the excess demand function, p equal to the
price-vector and p equal to the equilibrium price vector. See Varian (1992) for an extensive discussion on
Walras’ Law and General Equilibrium Theory, rooted on the seminal work by Debreu (1954, 1959).
9
4.3
Steps in Applied CGE Modeling
In this section we present a number of steps usually undertaken in applied CGE modeling
exercises, following Bayar (2005). We …nd this exercise instructive as it highlights the tasks
at hand and, concomitantly, it guides us in our discussion of the weaknesses and strengths
of CGE models below.
1. Specify dimensions of the model and choose key causal relations. In this step, one
answers the following questions: How many di¤erent types of agents to consider?
Which goods and factors are produced and employed in production? How many regions
and countries are there? Which markets are active? As stated above, the model should
feature all key causal relations in the economy.
2. Choose functional forms. To make the model operational one needs to go from general
algebraic representations or formulae to speci…c functional forms. In this step, it is
important to skillfully balance analytical convenience and adherence to reality.
3. Construct consistent data set. The data must go hand in hand with theory. For
instance, one should specify the set of factors according to the data available.
4. Calibration and econometric estimation. The key causal relations captured by the
model must be traced back to real data. The model should be able to replicate real
case scenarios.
5. Counter-factual experiments, scenario and impact analysis. This step amounts to
putting the model to work.
6. Regular updating. As new data become available, one should update all the previous
steps. This is important because one must prove that there are no structural breaks
or, to be more general, the model is being applied within its domain of applicability:
where the causal relations and their functional and numerical structures are valid (on
this, see below the section on criticisms).
4.4
Applications of CGE Models
In this section we discuss the main applications of applied general equilibrium models. After
a …rst look at the literature from an historical perspective, we then discuss how the litera10
ture has evolved according to key aspects of applied general equilibrium modeling, namely,
production structures, consumer behavior, exports and imports, regional governments, imperfect competition and increasing returns to scale, and, …nally, dynamics.
Applications: A Brief Literature Survey The range of issues on which CGE models
have been applied is quite wide and growing. The spectrum of applications of CGE models
includes international trade, public …nance, agriculture, transportation, welfare, environment
and income distribution. There are many excellent surveys of the literature. For the sake of
preserving space, we focus here only on the seminal works and regional modeling works.3
CGE models are, in their essence, the modern version of Walras’ model of the competitive
economy. Johansen (1960) is usually referred to as the …rst main attempt to use a large CGE
model to study a real economy. Shoven and Whalley (1972), Whalley (1975, 1977), Shoven
(1976), and Miller and Spencer (1977) are found among the earliest followers of the seminal
work by Johansen.
As discussed in Wing (2004), CGE models have been used in areas as diverse as …scal,
social policy and development planning (e.g., Pereira 1993, Perry et al 2001, Gunning and
Keyzer 1995, Ballard et al 1985, Ballard and Goulder 1985, Bandara 1991, Bayar 1993,
Dewatripont et al 1991, Goulder and Summers 1989), international trade and factor mobility
(e.g., Martins and Winters 1996, Harrison et al 1997, Bovenberg and Goulder 1989), and
increasingly, environmental regulation (e.g., Jorgenson and Wilcoxen 1990, Weyant 1999,
Bovenberg and Goulder 1996, Goulder 1992, Prost and Van Regemorter 1990). In fact,
several authors have recently applied CGE models to study the implications of introducing
tax or subsidy distortions that aim to reduce dioxide carbon emissions (see Wing 2004 for
an excellent, easy to read introduction to this subject).
In the 1980s, many applied general equilibrium models were built for the United States,
Canada and the United Kingdom and for several developing countries. However, widespread
applications of regional CGE models are more recent.
One of the main reasons to this late development of regional applied general equilibrium models is the poverty of the data available at the regional level. Indeed, applied
general equilibrium models require a considerable number of extremely detailed data: sectoral production, intersectoral exchanges, commercial ‡ows, distribution of income, private
3
We thank Ali Bayar for his extensive contribution to this section.
11
consumption, investments, public expenditures and taxes, just to list a few pieces of data
usually required by modern applied general equilibrium models. These data are sometimes
di¢cult to gather even at a national level, but the problem becomes even worse at a regional
level.
Moreover, other complications emerge when one moves from the national level to the
regional level, such as, for example, the greater openness of the regional economy to the
external world. Being a more open economy has an important in‡uence on interregional
mobility of labor and capital but also on di¤erentiation of the regional products with respect
to the other areas and on interregional commercial ‡ows. These di¤erences with the national
models accentuate the di¢culties in transposing the national results into the regional models
(for example, estimates of certain parameters of elasticity) and thus suggest divergences of
structure between national and regional general equilibrium models.
Among the …rst regional general equilibrium models we …nd Norrie and Percy (1983), who
analyzed Canadian regions, Kimbell and Harrison (1984), who studied California, and Liew
(1984), who focused on Australian regions. This research was also enriched and stimulated
by other contributions, for example, Hertel (1985), Hertel and Mount (1985), who studied
the State of New York, and Jones, Whalley and Wigle (1985), who looked at Canadian
regions, as well. Concomitantly with the recent development of this modeling approach at
the regional level one …nds a very fast proliferation of applied regional studies, mainly in the
United States, Canada and in Australia and a di¤usion of this approach to European regions
(Italy, Germany and United Kingdom) and even to certain developing countries (Brazil and
Malaysia). For the United States, we can cite Berck, Robinson and Goldman (1991), Waters
et al (1997), Berck et al (1996), Buckley (1992), Despotakis and Fisher (1998), Ho¤man
et al (1996), Kilkenny (1993a, 1998), Koh et al (1993), Kraybill et al (1992), Li and Rose
(1995), Morgan et al (1989), Morgan, Mutti and Rickman (1996), Rickman (1992), Schreiner
et al (1999). For Canada we can quote Jones and Whalley (1989, 1990), Whalley and Trela
(1986), Lemelin et al (1993), Gazel (1996), Gazel et al (1995), Wigle (1992); for Australia,
we may cite, among others, Liew (1984) and Peter et al (1996). Conrad and Schroder
(1993) and Hirte (1998) built models for German regions. Harrigan et al (1991; 1992; 1996),
initially, McGregor, Swales and Yin (1995, 1996), thereafter, built a model for Scotland and
D’Antonio et al (1988) built a model for Italian regions.
Harrigan and McGregor (1989), Ko (1985), Ko and Hewings (1986), Watanuki (1996),
12
Ando et al (1997) and Haddad (1999) focused on developing countries (Malaysia, Korea,
China, Indonesia and Brazil). Nowadays, applied general equilibrium models have become
ubiquitous, in the sense that they have been routinely employed in regional studies devoted
to both developed and developing regions.
As a corollary of the vigorous research devoted to applied regional general equilibrium
models, we …nd several literature review essays. Among these we highlight the work by
Lemelin (1994) and Partridge and Rickman (1998), as widely cited articles.
Production Structures In general, regional general equilibrium models assume a demand for intermediate goods based on a Leontief technology (i.e. the assumption of complementarity between intermediate inputs, based on the input/output matrix). Regional and
imported inputs are, usually, supposed to be imperfect substitutes. The substitution between regional and imported inputs is generally speci…ed by means of a function THESE (or
Constant Elasticity of Substitution) in reference to the Armington assumption (Armington,
1969)4 and often employed by regional general equilibrium models.
Concerning factors of production, substitution between factors of production is often
ensured by Cobb-Douglas functions (CD) or THESE functions speci…ed with constant returns
to scale. Very often, an overlapping structure (“nested”) on several levels is adopted to show
substitution between di¤erent factors of production (for example, energy is often modelled
like a factor of production with multiple components such as electricity, oil, natural gas, . . . )
and to overcome the limitations of the Cobb-Douglas functions and THESE functions on a
level which impose, for the …rst, unit elasticities of substitution, and for the second the same
elasticity of substitution between each pair of factors.
In the vast majority of regional models, production functions consider both intermediate
goods and factors of production. Labor and capital are universally speci…ed in regional
general equilibrium models as factors of production. Several studies also consider as a factor
of production natural resources, such as agricultural surfaces or energy production for better
taking account of speci…cities of an area and/or a sector. The following table highlights how
some regional studies have dealt with production structures.
Table 1: Regional Studies and Production Functions
4
The well known Armington Assumption suggests the use of a function THESE to represent imperfect
substitutability between domestic and imported goods.
13
Regional Studies
Regions
Production Functions
Berck, Robinson and Goldman (1991) San Joaquin Valley USA
CD w/ multi-level IO
Berck et al (1996)
California USA
THESE w/ multi-level IO
Buckley (1992)
3 regions - USA
CD w/ multi-level IO
Condrad and Schroder (1993)
Baden-Wurtemberg (Germany) THESE w/ multi-level IO
Despotakis and Fisher (1988)
California USA
Generalized Leontief
Gazel (1996)
4 regions - USA and Canada
THESE w/ multi-level IO
Gazel, Hewings and Sonis (1995)
4 regions - USA and Canada
THESE w/ multi-level IO
Haddad (1999)
3 regions - Brazil
THESE w/ multi-level IO
Harrigan and McGregor (1989)
2 regions - Malaysia
THESE w/ multi-level IO
Harrigan et al (1991; 1992)
Scotland and UK
THESE w/ multi-level IO
Hertel and Mount (1985)
New York State USA
Multi-level translog
Hirte (1998)
10 regions - Germany
THESE w/ multi-level IO
Ho¤man et al (1996)
California USA
THESE/IO/CD
Jones and Walley (1989, 1990)
6 regions - Canada
THESE w/ multi-level IO
Jones, Walley and Wigle (1985)
2 regions - Canada
THESE w/ multi-level IO
Kilkenny (1998)
USA
CD/IO/THESE
Kimbell and Harrison (1984)
California USA
THESE
Koh, Schreiner and Shin (1993)
Oklahoma USA
CD w/ multi-level IO
Li and Rose (1995)
Pennsylvania USA
Generalized Leontief
Liew (1984)
6 regions - Australia
THESE w/ multi-level IO
McGregor, Swales and Yin (1995)
Scotland and UK
THESE w/ multi-level IO
McGregor, Swales and Yin (1996)
Scotland and UK
THESE w/ multi-level IO
Morgan, Mutti and Partridge (1989)
6 regions - USA
Di¤erential linear with CRS
Morgan, Mutti and Rickman (1996)
6 regions - USA
THESE w/ multi-level IO
Peter et al (1996)
8 regions - Australia
THESE w/ multi-level IO
Rickman (1992)
USA (some regions)
THESE w/ multi-level IO
Scheiner et al (1999)
Oklahoma USA
CD/THESE/IO
Walley and Trela (1986)
6 regions - Canada
THESE w/ multi-level IO
Wigle (1992)
6 regions - Canada and USA
THESE
The degree of factor mobility plays an important part in the speci…cation of a regional
14
applied general equilibrium model and di¤ers from one model to another. In general, factor
mobility increases with the span of the period of analysis, so it is quite common to …nd
…xed factor supplies, at least in the short term. With imperfect factor mobility, productivity
wedges across regions may persist over time and, concomitantly, factor returns may also not
be equal across regions as factor returns are linked in some sense to factor productivities.
To specify the “closing” of the regional labor market, the majority of the models use
the neoclassical assumption:5 wages are endogenous and ‡exible so wages adjust in order
to equate labor demand and labor supply; nevertheless, more and more regional models
incorporate Keynesian features: wages (prices) are …xed at predetermined exogenous levels
and there may be factor underutilization (unemployment). See Ginsburgh et al 1986 for an
example of a general equilibrium model with wage rigidities. The following table shows how
di¤erent models have dealt with factor mobility and labor market closure.
Table 2: Regional Studies, Factor Mobility and Labor Market Closure
5
Intutively, prices are ‡exible and adjust in order to clear the market in the sense that demand equals
suply and, thus, there is no factor underutilization.
15
Regional Studies
Inputs
Closure6
Berck, Robinson and Goldman (1991) K, L, M, R, water; K, L …xed Endogenous wages
Berck et al (1996)
K, L, M; L ‡exible
Endogenous wages
Buckley (1992)
K, L, M; K, L …xed
Endogenous wages
Condrad and Schroder (1993)
K, L, M, E; K, L …xed
Exogenous wages
Despotakis and Fisher (1998)
K, L, M, E; K, L …xed
Endogenous wages
Gazel (1996)
K, L; K …xed, L free
Endogenous wages
Gazel, Hewings and Sonis (1995)
K, L; K, L …xed
Endogenous wages
Haddad (1999)
K, L, M; K, L …xed
Wage di¤erentials
Harrigan and McGregor (1989)
K, L, M; K …xed, L free
Several cases
Harrigan et al (1991; 1992)
K, L, M; K …xed, L free
Several cases
Hertel and Mount (1985)
K, L, M, E; K …xed, L free
Exogenous wage bill
Ho¤man et al (1996)
K, L, M; Several cases
Several cases
Jones and Walley (1989, 1990)
K, L, M, R; K, L free
Endogenous wages
Kimbell and Harrison (1984)
K, L, M; K free, L …xed
Endogenous wages
Koh, Schreiner and Shin (1993)
K, L, M, R; L free
Several cases
Li and Rose (1995)
K, L, M, E; K, L …xed
Several cases
Liew (1984)
K, L, M, R; L …xed
Endogenous wages
McGregor, Swales and Yin (1995)
K, L, M; Several cases
Several cases
McGregor, Swales and Yin (1996)
K, L, M; Several cases
Several cases
Morgan, Mutti and Partridge (1989)
K, L, R; K, L free
Endogenous wages
Morgan, Mutti and Rickman (1996)
K, L, R; K, L free
Endogenous wages
Peter et al (1996)
K, L, M; K …xed, L free
Several cases
Rickman (1992)
K, L, R; K …xed, L free
Several cases
Scheiner et al (1999)
K, L, R, M; K …xed, L free
Endogenous wages
Walley and Trela (1986)
K, L, R; K, L free
Endogenous wages
Abbreviations: K: capital; L: labor; M: intermediate goods; E: energy; R: natural
resources (surfaces).
6
Endogenous wages means that wages are ‡exible and calculated in order to clear the labour market en-
suring, thus, that there does not exist involuntary unemployment in equilibrium. When wages are exogenous
then involuntary unemployment may arise in equilibrium.
16
Consumer Behavior Cobb-Douglas (CD) and THESE functions are generally used
to represent consumer utility but the homotheticity of these functions restricts consumption goods income elasticities to one. In addition, CD functions imply zero cross-price
elasticities. To avoid these elasticity restrictions, several studies employ the more general,
non-homothetic, “linear system of the expenditures” (Linear Expenditure System or LSE).
Another method generally used to allow for various elasticities of substitution between the
various sets of consumer goods is to use a nested structure of consumption. Import and
domestically produced consumption goods are generally perceived as imperfect substitutes,
following the well known Armington assumption.
Exports and Imports With regard to exports, the functions THESE of imports act as
the source of request for exports in multi-regional models. In the mono-regional models, the
goods produced in the area are often supposed to be delivered either to the local markets
or to the external markets, the choice being made with the same assumption as for the
imported goods, according to a constant elasticity of substitution, namely “constant elasticity
of transformation” (Constant Elasticity of Transformation or CET). Nevertheless, certain
models simply assume a constant elasticity of exports with respect to prices.
With regard to imports, regional general equilibrium models typically follow the Armington assumption to take account of imperfect substitution between the regional domestic
goods and the imported goods.
Regional Governments In some regional general equilibrium models the regional
government is lumped together with the federal government and treated as exogenous. However, some regional general equilibrium models do treat the regional government as a separate
entity from the federal government, but take both of them as exogenous nonetheless.
The most complex treatments of the regional government are found without surprise
in the studies which focus on regional budget policies. In these studies, the expenditure
of the regional government is related to regional household incomes and, in some of them,
the regional government is regarded as a sector of production which requires inputs. Some
models allow for interactions between regional and federal taxes.
17
Imperfect Competition and Increasing Returns to Scale Most applied regional
general equilibrium models assume, for the sake of convenience, perfectly competitive markets, with …rms acting as price takers in both factor and commodities markets, under constant
returns to scale production structures. Given the importance of imperfect competition in
certain sectors, several authors moved away from the perfect competition paradigm and integrated imperfect competition aspects in their regional models, following the industrial organization (IO) and game theory literatures. For instance, Hertel (1985) and Hertel and Mount
(1985) consider oligopolistic behavior in some sectors; Kilkenny (1993b, 1998) and Whalley
and Trela (1986) model certain manufacturing sectors following monopolistic competition
IO models; Mello and Tarr (1992) use duality theory to incorporate imperfect competition
into some goods markets, namely, cars, iron and steel; …nally, Tembo, Vargas and Schreiner
(1999) consider the factor market for wood as a monopsony. Vargas et al 1999 present a
regional CGE model where with monopsony markets in the forest products industry in the
US state of Oklahoma. Some authors also deviated from the traditional perfect competition
with constant returns to scale benchmark models by assuming increasing returns to scale in
their production structures and imperfect competition (see Harris 1984).
Dynamics The simplest applied general equilibrium models are static in the sense
that time is omitted from the analysis. Intuitively, all agents solve a one-shot problem.
Clearly, this modeling strategy is inadequate to study a vast array of economic problems
such as investment or capital accumulation. Savings and investment decisions, to name a
few, are some of the economic problems which must be dealt with in a dynamic setting,
that is, where time is explicitly treated in the analysis. Dynamic models are either recursive
or fully-dynamic, with the latter being more complicated than the former. In recursive
settings, the agents problems are typically related pair-wise over time, meaning that today’s
problem a¤ects tomorrow’s problem only. In fully-dynamic settings, all periods are related
to one another simultaneously. Time dynamics are important if one is interested in the
transition path from one equilibrium state of the economy to another. Comparative static
analysis cares only about comparing one equilibrium against some other equilibrium. This
is possible to analyze within a static framework. However, transitional dynamics, or how
the economy actually moves from one equilibrium to another may be quite important to
have a quantitative feeling of the welfare implications of a given policy package. Who gains
18
and who loses from a given policy change? The answer to this question may depend on the
time frame. Hence, welfare analyses are much more interesting when undertaken in a fully
dynamic setting. See Rege (2003, Ch. 2) for an interesting discussion on this topic.
Applications: A Tentative Classi…cation The regional general equilibrium models
listed here were used in many …elds. Any given classi…cation scheme of such applications
will, inevitably, be somewhat arbitrary. We can classify them in four important topics:
1. Topic 1: To study the direct and indirect economic e¤ects of exogenous changes in
…nal demands (say, government demand);
2. Topic 2: To evaluate trade or federal budget policies;
3. Topic 3: To evaluate regional budget policies;
4. Topic 4: To evaluate certain speci…c policies, such as agricultural, transportation and
environmental.
The following table lists some applications classi…ed according to the above structure.
Table 3: Regional Studies - Some Applications
19
Regional Studies
Topic
Berck, Robinson and Goldman (1991) 4
5
Berck et al (1996)
3
Buckley (1992)
4
Condrad and Schroder (1993)
4
Despotakis and Fisher (1998)
1, 4
Gazel (1996)
2
Gazel, Hewings and Sonis (1995)
2
Haddad (1999)
2, 4
Harrigan and McGregor (1989)
1
Harrigan et al (1991; 1992)
1
Hertel and Mount (1985)
4
Ho¤man et al (1996)
1, 2
Jones and Walley (1989, 1990)
1, 2, 3
Kimbell and Harrison (1984)
3
Koh, Schreiner and Shin (1993)
1
Li and Rose (1995)
4
Liew (1984)
2
McGregor, Swales and Yin (1995)
4
McGregor, Swales and Yin (1996)
4
Morgan, Mutti and Partridge (1989)
3
Morgan, Mutti and Rickman (1996)
3
Peter et al (1996)
4
Rickman (1992)
3
Scheiner et al (1999)
4
Walley and Trela (1986)
2, 4
Comments on CGE Models
In this section we discuss weaknesses and strengths of applied general equilibrium models as
well as where are these models headed.
According to Iqbal and Siddiqui (2001), despite the fact that applied general equilibrium
20
models have raised the level of sophistication of the policy debate, applied general equilibrium
models are subject to criticism on many counts. A CGE model embodies three types of
information: analytical, functional and numerical. The analytical structure is the theoretical
support where one identi…es the variables of interest and postulates their causal relations.
The functional structure is the algebraic representation of the analytical structure necessary
to make the model operational. Finally, the numerical structure consists of the coe¢cients
in the equations considered in the model. The criticism is mainly directed towards the
functional and numerical structures of the calibrated or applied CGE model.
Quality of Data: The quality of the model is heavily dependent on the quality of
the data of the chosen benchmark period, say, without lost of generality, a given year.
This benchmark year is perceived as a good representation of the deep or structural
key causal relations in the economy. Any given year will re‡ect a series of anomalous
events which, in turn, undermines generalizations. In this sense, regular updating is
important to asses eventual structural breaks or the importance of erratic events.
Choice of Parameters: In CGE models, some parameters are taken from empirical
studies, some are chosen arbitrarily and some are calibrated, in the sense that they
are picked in a way that forces the model to replicate the data of a chosen benchmark
year. Several authors o¤er the following criticisms to such practices. In order to
model production structures one typically has to quantify elasticities. If one borrows
elasticities from empirical studies devoted to other regions or time periods, one may be
using information from sectors with di¤erent statistical classi…cations or technologies,
which, in addition, may be obsolete. As with any calibration exercise, there is no way
to formally test the validity of the parameters calibrated, since by construction they
guarantee that the model replicates the benchmark data.
Choice of Functional Forms: Most CGE models assume …rst order functional forms
which are easy to handle but embody unwelcome elasticity restrictions. However, and
as several authors show, it is possible to avoid such restrictions by assuming more
‡exible functional forms and, consequently, the models can represent all the relevant
own and cross price elasticities derived from an arbitrary utility or pro…t function,
without imposing prior constraints.
21
Calibration of the Model: Calibrating a CGE model to a given benchmark year
may be problematic as the benchmark year may not accurately represent the natural
state of the economy or it may not provide enough data.
Static CGE Model: Static CGE models can address questions of what happens to
an economy as it moves from one state of exogenous conditions to another, that is,
comparative static analysis. By design, however, comparative static analysis omits the
time path of response of the economy, which may be important for welfare analysis
and, of course, for forecasting. Moreover, in static models some intertemporal problems
such as consumption and saving are modeled in an arbitrary way, without formal
optimization or guidance from theory. Hence, in a setting short of fully dynamic, the
economy may move along a non-optimal path. In order to avoid non-optimal paths,
fully dynamics must be considered and great care should be devoted to the role of
expectations.
Sensitivity of Results: Results are sensitive to inputs such as key elasticities. Hence,
it is instructive to learn how results vary as one allows these key elasticities to vary
according to guidance from empirical econometric studies.
6
Concluding Remarks
Applied CGE models are nowadays the most popular modeling approach to regional phenomena. This is hardly surprising when one realizes that applied CGE models allow for a
great deal of detail and complexity of institutional features and of regional speci…cities and
are theoretically coherent. Since its origins in the 1960s, applied CGE models have come
a long way. However, there is still an ongoing widespread research e¤ort that continuously
brings more realistic models into the forefront practice. The research agenda includes the
following topics and challenges:
Interactions between regions;
Disintegration of the labor market;
Imperfect competition and increasing returns to scale;
22
Dynamic aspects of the economies;
More ‡exible functional forms;
In addition to calibration, to use econometric estimates whenever possible;
Consider …nancial assets;
Consider transportation costs;
Consider household heterogeneity;
Consider alternative macro closure rules.
The current research agenda and the recent successes will, certainly, reinforce the popularity of applied CGE models and the political demand for their use as policy guidance
tools in the near future. However, it is possible that in some regions, poor data make theory
ahead of measurement.
23
References
[1] Ando, A. and T. Shibata. 1997. A multi-Regional Model for China Based on Price and
Quantity Equilibrium, in Manas Chatterji (ed.), Regional Science Perspectives for the
Future. London : MacMillan, pp. 326-344.
[2] Armington, P. 1969. A theory of demand for products distinguished by place of production. IMF Sta¤ Papers 16 : 159-178.
[3] Ballard, C. L., D. Fullerton, J. B. Shoven, and J. Whalley. 1985. A General Equilibrium
Model for Tax Policy Evaluation. Chicago. The University of Chicago Press.
[4] Ballard, C. L. and L. Goulder. 1985. Consumption Taxes, Foresight, and Welfare :
A Computable General Equilibrium Analysis, in : Piggott, J. and J. Whalley, Eds.
New Developments in Applied General Equilibrium Analysis. Cambridge. Cambridge
University Press. Pp. 253-282.
[5] Bandara, J. S. 1991. Computable general equilibrium models for development policy
analysis in LDCs, Journal of Economic Surveys, Vol 5, No 1.
[6] Bayar, A. H. 1993. BELTAX ET GEMTAX : Deux modèles d’équilibre général de
taxation pour la Belgique, Thèse de doctorat, ULB.
[7] Bayar, A. 2005. Regional CGE Modeling. Paper presented at REAL Meeting in Ponta
Delgada.
[8] Berck, P., S. Robinson, and G. Goldman, eds. 1991. The use of computable general
equilibrium models to assess water policies. In The economics and management of water
and drainage in agriculture. Norwell, MA: Kluwer Academic Publishing.
[9] Berck, P., E. Golan, B. Smith, J. Barnhart, and A. Dabalen. 1996. Dynamic revenue
analysis for California. Sacramento, CA: California Department of Finance.
[10] Bovenberg, A. L. and Goulder, L. H. 1989. Promoting Investment under International
Capital Mobility : An Intertemporal General Equilibrium Analysis. Manuscript.
24
[11] Bovenberg, A. L. and Goulder, L. H. 1996. Optimal Environmental Taxation in the
Presence of Other Taxes: General-Equilibrium Analyses. American Economic Review
86(4): 985-1000.
[12] Buckley, P. H. 1992. A transportation-oriented interregional computable general equilibrium model of the United States. The Annals of Regional Science 26: 331-348.
[13] Conrad, K., and M. Schroder. 1993. Choosing environmental policy instruments using
general equilibrium models. Journal of Policy Modeling 15: 521-543.
[14] D’Antonio, M., Colaizzo, R. et Leonello, G. 1988. Mezzogiorno/centre-North : A tworegion model for the italian economy, Journal of policy modelling, 10 (3).
[15] Debreu G. 1954. Valuation equilibrium and Pareto optimum. Proceedings of the National Academy of Sciences 40 : 588-592.
[16] Debreu G. 1959. Théorie de la valeur. Paris : Editions Dunod.
[17] Despotakis, K. A., and A. C. Fisher. 1988. Energy in a regional economy: A computable general equilibrium model for California. Journal of Environmental Economics
and Management 15: 313-330.
[18] Devarajan, S. and S. Robinson. 2002. The in‡uence of computable general equilibrium
models on policy. In: Kehoe, T., Srinivasan, and J. Whalley. Frontiers in Applied General Equilibrium Modeling. Cambridge University Press.
[19] Dewatripont, C., S. Erlich, V. Ginsburgh at D. Van Regemorter. 1991. The E¤ects
on Unemployment of reducing Social Security Contributions : A General Equilibrium
Analysis for Belgium. In : Don, H., T. Van de Klundert, J. Van Sinderen. Applied
General Equilibrium Modelling. Dordrecht. Kluwer Academic Publishers. Pp. 135-153.
[20] Gazel, R. 1996. Free trade agreements and interregional labor migration: The case of
the U.S. and Canada. Annals of Regional Science 30: 373-390.
[21] Gazel, R., G. J. D. Hewings, and M. Sonis. 1995. Trade, sensitivity and feedbacks:
Interregional impacts of the US-Canada Free Trade Agreement. In Recent advances
in spatial equilibrium modeling, eds. J. C. J. M. van den Bergh, P. Nijkamp, and P.
Rietveld. New York: Springer.
25
[22] Ginsburgh, V. and L. Van Der Heyden 1986. General Equilibrium with Wage Rigidities
: An Application to Belgium. Cowles Foundation Paper. No. 633.
[23] Goulder, L. H. 1992. Do the Costs of a carbon Tax Vanish when Interactions with Other
Taxes are Accounted for ? NBER Working Paper. No. 4061.
[24] Goulder, L. H. and L. H. Summers. 1989. Tax Policy, Asset Prices, and Growth. A
General Equilibrium Analysis. Journal of Public Economics. 38(2) : 265-296.
[25] Goulder, L., ed. 2002. Environmental Policy Making in Economies with Prior Tax Distortions. Northampton MA: Edward Elgar.
[26] Gunning, J., and M. Keyzer. 1995. Applied General Equilibrium Models for Policy
Analysis. In J. Berhman and T. N. Srinivasan (eds.) Handbook of Development Economics Vol. III-A, Amsterdam: Elsevier, 2025-2107.
[27] Haddad, E. 1999. Regional Inequality and Structural Changes : Lessons from the Brazilian experience. Aldershot : Ashgate Publishing Limited.
[28] Harrigan, F., and P. G. McGregor. 1989. Neoclassical and Keynesian perspectives on
the regional macro-economy: A computable general equilibrium approach. Journal of
Regional Science 29: 555-573.
[29] Harrigan, F., P. G. McGregor, and J. K. Swales. 1996. The system-wide impact on the
récipient region of a regional labor subsidy. 0xford Economic Papers 48: 105-133.
[30] Harrigan, F., P. G. McGregor, J. K. Swales, and N. Dourmashkin. 1991. The sensitivity of output multipliers to alternative technology and factor market assumptions:
A computable general equilibrium analysis. In Regional input-ouput modeling: New
developments and interpretations, eds. J. H. L. Dewhurst, R. C. Jensen, and G. J. D.
Hewings. Aldershot, UK: Averbury Press.
[31] Harrigan, F., P. G. McGregor, J. K. Swales, and N. Dourmashkin. 1992. Imperfect
competition in regional labour markets: A computable general equilibrium analysis.
Environment and Planning A 24: 1463-1481.
26
[32] Harris, R. 1984. Applied General Equilibrium Analysis of Small Open Economies with
Scale Economies and Imperfect Competition. American Economic Review. 74 : 10161032.
[33] Harrison, G., Rutherford, T., and D. Tarr. 1997. Quantifying the Uruguay Round.
Economic Journal 107: 1405-1430.
[34] Hertel, T. W. 1985. Partial vs. general equilibrium analysis and choice of functional
form: implications for policy modeling. Journal of Policy Modeling 7: 281-303.
[35] Hertel, T. W., and T. D. Mount. 1985. The pricing of natural resources in a regional
economy. Land Economics 61: 229-243.
[36] Hirte, Georg. 1998. Welfare e¤ects of regional income taxes : Results of an interregional
CGE analysis for Germany. The Annals of Regional Science 32 : 201-219.
[37] Ho¤mann, S., S. Robinson, and S. Subramanian. 1996. The role of defense cuts in
the california recession: Computable general equilibrium models and interstate factor
mobility. Journal of Regional Science 36: 571-595.
[38] Iqbal, Z., and Siddiqui, R. 2001. Critical review of literature on computable general equilibrium models. MIMAP Technical Paper Series 9, Pakistan Institute of Development
Economics.
[39] Johansen F. R. 1960. A Multisector Study of Economic Growth. Amsterdam. NorthHolland.
[40] Jones, R., and J. Whalley. 1989. A Canadian regional general equilibrium model and
some applications. Journal of Urban Economics 25: 368-404.
[41] Jones, R., and J. Whalley. 1990. Regional balance sheets of gains and losses from national policies. Regional Science and Urban Economics 20: 421-435.
[42] Jones, R., J. Whalley, and R. Wigle. 1985. Regional impacts of tari¤s in Canada: Preliminary results from a small dimensional numerical general equilibrium model. In New
Developments in applied general equilibrium analysis, eds. J. Piggott and J. Whalley.
New York: Cambridge University Press.
27
[43] Jorgenson, D. W. and P. J. Wilcoxen. 1990. Intertemporal General Equilibrium Modeling of US Environmental Regulation. Journal of Policy Modeling. 12(4) : 715-744.
[44] Kilkenny, M. 1993a. Rural/urban e¤ects of terminating farm subsidies. American Journal of Agricultural economics 75: 968-980.
[45] Kilkenny, M. 1993b. Agricultural liberalization in segmented or integrated markets, with
scale economies. Journal of Economic Integration 8: 201-218.
[46] Kilkenny, M. 1998. Transport costs and rural development. Journal of Regional Science
38: 293-312.
[47] Kimbell, L. J., and G. W. Harrison. 1984. General equilibrium analysis of regional
…scal incidence. In Applied general equilibrium eds. H. Scarf and J. Shoven. New York:
Cambridge University Press.
[48] Ko, S. 1985. A Regional Computable General Equilibrium Model for Korea. Ph.D.
dissertation. University of Illinois at Urbana-Champaign.
[49] Ko, S., and Hewings, G.J.D. 1986. A Regional Computable General Equilibrium Model
for Korea. The Korean Journal of regional Science, vol. 2, December, pp. 45-57.
[50] Koh, Y. K., D. F. Schreiner, and H. Shin. 1993. Comparisons of regional …xed price and
general equilibrium models. Regional Science Perspectives 23: 33-80.
[51] Kraybill, D. S. 1993. Computable general equilibrium analysis at the regional level in
Microcomputer-based-input-output modeling : Applications to economic development,
eds. Otto, D. M. and Johnson, T. G. Boulder. Co : Westview Press.
[52] Kraybill, D. S., T. G. Johnson, and D. Orden. 1992. Macroeconomic imbalances: A
multi-regional general equilibrium analysis. American Journal of agricultural Economics
74: 726-736.
[53] Lemelin André, P. Frechette, and M. Boisvert. 1993. La spéci…cation d’un modèle
d’équilibre général appliqué pour les régions métropolitaines de Québec et de Montréal. INRS-Urbanisation, Collection Inédits, In-15 et CRAD, Université Laval, Cahier
de Recherche.
28
[54] Lemelin André. 1994. Analyse économique régionale et équilibre général : un modèle appliqué à la région métropolitaine de Montréal. Revue d’Economie Régionale et Urbaine
n 5.
[55] Li, P., and A. Rose. 1995. Global warming policy and the Pennsylvania economy: A
computable general equilibrium analysis. Economic Systems Research 7: 151-171.
[56] Liew, L. H. 1984. "Tops-down" versus "bottoms-up" approaches to regional modeling.
Journal of Policy Modeling 6: 351-367.
[57] Martins, W. and L. Winters (eds.). 1996. The Uruguay Round and the Developing
Economies. New-York: Cambridge University Press.
[58] McGregor, P. G., J. K. Swales, and Y. P. Yin. 1995. Migration equilibria in regional
economies: A multi-period CGE analysis of an improvement in local amenities, In
Recent advances in .Spatial equibrium modeling, eds. J. C. J. M. van den Bergh, P.
Nijkamp, and P. Rietveld. New York: Springer.
[59] McGregor, P. G., J. K. Swales, and Y. P. Yin. 1996. A long-run interpretation of regional
input-output analysis. Journal of Regional Science 36: 479-500.
[60] Melo, J. de, and D. Tarr. 1992. A General Equilibrium Analysis of US Foreign Trade
Policy. Cambridge. MIT Press.
[61] Miller, M. H., and J. E. Spencer. 1977. The static economic e¤ects of the UK joining the
EEC: A general equilibrium approach. The Review of Economic Studies 38: 175-208.
[62] Morgan, W., J. Mutti, and M. Partridge. 1989. A regional general equilibrium model
of the United States: Tax e¤ects on factor movements and regional production. The
Review of Economics and Statistics 71: 626-635.
[63] Morgan, W., J. Mutti, and D. Rickman. 1996. Tax exporting, regional economic growth
and welfare. Journal a Urban Economics 39: 131-159.
[64] Mutti, J., W. Morgan. and M. Partridge. 1989. The incidence of regional taxes in a
general equilibrium framework. Journal of Public Economics 39: 83-107.
29
[65] Norrie, K. H., and M. B. Percy. 1983. Freight rate reform and regional burden: A general
equilibrium analysis of Western freight rate proposals. Canadian Journal of Economics
16 : 325-349.
[66] Partridge, Mark D., and D.S. Rickman. 1998. Regional Computable General Equilibrium
Modeling : A survey and Critical Appraisal. International Regional Science Review 21,
3 : 205-248.
[67] Pereira, A. M. 1993. A Dynamic General Equilibrium Analysis of Corporate Tax Integration. Journal of Policy Modeling. 10(3) : 401-436.
[68] Perry, G., Whalley, J., and G. McMahon, eds. 2001. Fiscal reform and structural change
in developing countries. New York: Palgrave-Macmillan.
[69] Peter Matthew W., M. Horridge, G.A. Meagher, F. Naqvi, and B.R. Parmenter. 1996.
The Theorical Structure of Monash-MRF. Preliminary Working Paper N OP-85, Centre
of Policy Studies, Monash University, Australia.
[70] Proost, S. and D. Van Regemorter. 1990. Economic E¤ects of a Carbon Tax with a
General Equilibrium Illsutration for Belgium. CES Discussion Paper. No. 11.
[71] Rege, S. 2003. Applied General Equilibrium Analysis of India’s Tax and Trade Policy.
Ashgate Pulbishing Limited.
[72] Rickman, D. S. 1992. Estimating the impacts of regional business assistance programs:
Alternative closures in a computable general equilibrium model. Papers in Regional
Science 71: 421-435.
[73] Rust, J. 1997. Dealing with the complexity of economic calculations. Working Paper,
Yale University.
[74] Schreiner, Dean F., D.W. Marcouiller, G. Tembo, and E.E. Vargas. 1999. Computable
General Equilibrium Modeling for Regional Analysis. West Virginia University : The
Web Book of Regional Science.
[75] Shoven, J. B. 1976. The incidence and e¢ciency e¤ects of taxes on income from capital.
Journal of Political Economy 84: 1261-1283.
30
[76] Shoven, J. B., and J. Whalley. 1972. A general equilibrium calculation of the e¤ects of
di¤erential taxation of income from capital in the U.S. Journal of Public Economics 1:
281-321.
[77] Shoven, J. B., and J. Whalley. 1992. Applying General Equilibrium. Cambridge. Cambridge University Press.
[78] Tembo Gelson, E.E. Vargas, and D.F. Schreiner. 1999. Sensitivity of Regional Computable General Equilibrium Models to Exogenous Elasticity Parameters. Paper presented at the 30th Annual Meeting of the Mid-continent Regional Science Association,
June 10-12, Minneapolis.
[79] Vargas, Eliecer E., and D.F. Schreiner. 1999. Modeling Monopsony Market With Regional CGE : The Oklahoma Forest Products Industry Case. Paper presented at the
30th Annual Meeting of the Mid-continent Regional Science Association, June 10-12,
Minneapolis.
[80] Varian, H. 1992. Microeconomic Analysis. New York: Norton.
[81] Watanuki, M. 1996. Regional Development Policy Analysis of Indonesia: Interregional
CGE Modeling Approach. Paper presented at the 43rd North American Meeting of the
Regional Science Association International, Washington DC, November.
[82] Waters, E., D. W. Holland, and B. A. Weber. 1997. Economic impacts of a property
tax limitation: A computable general equilibrium analysis of Oregon’s Measure 5. Land
Economics 73: 72-89.
[83] West, G. R. 1995. Comparison of input-output, input-output + econometric and computable general equilibrium impact models at the regional level. Economic Systems
Research 7: 209-227.
[84] Weyant, J., ed. 1999. The Costs of the Kyoto Protocol: A Multi-Model Evaluation,
Energy Journal special issue.
[85] Whalley, J. 1975. A general equilibrium assessment of the 1973 U.K. Tax Reform. Economica 42: 139-161.
31
[86] Whalley, J. 1977. The United Kingdom tax system, 1968-1970: Some …xed point indications of its economic impact. Econometrica 45: 1837-1858.
[87] Whalley, J., and I. Trela. 1986. Regional aspects of confederation. Bu¤alo: University
of Toronto Press.
[88] Wigle, R. M. 1992. Transportation costs in regional models of foreign trade: An application to Canada-U.S. Trade. Journal of Regional Science 32: 185-207.
[89] Wing, I. 2004. Computable General Equilibrium Models and Their Use in EconomyWide Policy Analysis. MIT Joint Program on the Science and Policy of Global Change,
Technical Note No. 6.
32