Computable general equilibrium
Computable general equilibrium (CGE) models are a class of economic model that use actual economic data to estimate how an economy might react to changes in policy, technology or other external factors. The word « equilibrium » means that values taken by endogenous variables in the model allow the resolution of all equations. CGE models are descended from the input-output models pioneered by Wassily Leontief, but assign a more important role to prices. Thus, where Leontief assumed that, say, a fixed amount of labour was required to produce a ton of iron, a CGE model would normally allow wage levels to (negatively) affect labour demands. One of the main interests of CGE models is their dynamic characteristic enabling to make projections up to 100 years.
A CGE model consists of (a) equations describing model variables and (b) a database (usually very detailed) consistent with the model equations. Data required are the following :
- Tables of transaction values, showing, for example, the value of coal used by the iron industry (read the section above “regional accounting by matrix Input-output”). Usually the database is presented as an input-output table or as a social accounting matrix (SAM). In either case, it covers the whole economy of a country (or even the whole world), and distinguishes a number of sectors, commodities, primary factors and perhaps types of household. I-O tables can be taken from Eurostat (consulted in January 2007), from the national statistic of the studied country or from the governmental (or regional) economic department. For more details on SAM, read Pyatt and Round (1985).
- Elasticities: dimensionless parameters that capture behavioural response to policy scenarios. For example, export demand elasticities specify by how much export volumes might fall if export prices went up (e.g. due to a tax on green house gas emitted by merchandise transportation). Data on elasticities are usually taken from literature survey (Böhringer, 2004).
Nowadays, CGE models are made of thousands of equations. They can give simulations till 100 years time horizon and have regional, national and international spatial dimensions. Hecq (2006a) categorized those models by economical mechanisms, and identified four categories described below. They can be divided into two families : "bottom-up" et "top-down".
Bottom-up models are built on a detailed representation of productive system (supply side) and demand (demand side). Based on data on cost and effectiveness of technologies as well as on basic resources utilization such as energy, they calculate minimum cost strategies. Their weakness stems from the translation of feedback in term of macroeconomic equilibrium.
Example of bottom-up model : Technological optimization models (MARKAL, DNE21+, GMM, MESSAGE,…). They are based on technological data for each sector of a country. They model energetic demand taking into account technical constrains. They allow us to highlight optimal technological options (cost-effectiveness) in order to achieve environmental targets at diverse horizons (e.g. CO2 emissions).
Top-down models describe the economic system in a global way through aggregates and their interrelations in the frame of a general equilibrium built on the base of microeconomic theory.
Example of top-down model : Macroeconometric models (NEMESIS, E3MG, HERMES, etc.). They are numerous and are more developed and accurate than the previous models mentioned above. However, they are neo-Keynesians. Therefore, they operate according to the demand in the economy, which is not always in equilibrium with the supply (structural unemployment is possible to take into account in those models). In addition, production functions are stemming from econometric techniques based on historical series that affect their structure.
Practical use of CGE models
Though models are still subject to discussions, they find numerous applications (simulation, prevision, research) among others evaluation of environmental policies impacts such as taxes on CO2/energy and polluting emissions trading schemes. CGE are relevant when willing to evaluate the economic implications of policy intervention on resource allocation and incomes of agents (for example, a tax on energy and green house gas emissions might affect fuel prices, the consumer price index, and hence perhaps wages and employment). For instance a relevant question to be answered by CGE would be “what is the optimal tax policy to maximize economic performance given minimum constraints on the level of environmental quality or distributional concerns.”
Since CGE are partly based on I-O tables, what is written above in section “regional accounting by matrix Input-output” applies also to CGE. See there for additional information concerning relevancy of CGE.
Limits of the method
CGE models are complex to implement and their results are highly dependents on key economic parameters on which remain uncertainties. In addition, those models are expensive and time consuming (it takes months to years to build a CGE model).
In addition, the model equations tend to be built upon an underlying theory (i.e. traditional economics, as defined by Maréchal and Lazaric, 2007), often assuming cost-minimizing behaviour by producers, average-cost pricing, and household demands based on optimizing behaviour. However, most CGE models conform only loosely to the theoretical general equilibrium paradigm. For example, they may allow for:
1. non-market clearing (general equilibrium is not reached : supply is not equal to demand), especially for labour (unemployment) or for commodities (inventories) 2. imperfect competition (e.g. monopoly pricing) 3. demands not influenced by price (e.g. government demands) 4. a range of taxes 5. externalities, such as pollution
However, even with these corrections to match more the economic reality, Maréchal and Lazaric (2007) estimate that the use of CGE models is highly disputable because of their inadequate account of the real behaviour of economic agents. Indeed, CGE models are built upon a traditional economics that has been strongly questioned by scholars from different fields. For instance, the Homo Oeconomicus paradigm is completely at odds with empirical evidence contained in studies showing that economic decisions are partly guided by feelings and thus emotionally coloured (providing human beings with "intelligent emotions and emotional intelligence"). An individual is not able to make optimal decisions in order to reach its goals. He can only make satisfactory decisions because he is limited by its capacity, its habits and unconscious reflexes; its values and concepts of the goal to reach (which can even be different from the goal decided by the enterprise); and by its knowledge and the imperfect information he has access to. Not grasping this, is a major drawback of CGE models.
Lastly, CGE, similarly to I-O tables, are not able to take into account issues with small impacts because of their too high aggregation level.
- EUROSTAT, consulted in January 2007. ESA 95 Input-Output tables. Available on Internet : http://epp.eurostat.ec.europa.eu/portal/page?_pageid=2474,54156821,2474_54764840&_dad=portal&_schema=PORTAL#IOT)
- Pyatt and Round, 1985. Social Accounting Matrices: A Basis for Planning. The World Bank.
- Böhringer C., 2004. Sustainability impact assessment : the use of computable general equilibrium models. Economie internationale 2004/3, n° 99, pp. 9-26.
- Hecq W., 2006a. Aspects économiques de l’environnement. Fascicule 4. Economie de l’environnement. Université Libre de Bruxelles, 12ème édition, P.U.B.