Using Macro Indicators for Consistent CGE Baselines

Many policy targets are set in such a way that their implementation will not take place immediately but rather at some time in the future. In these cases, quantitative policy analysts are confronted with the problem of compiling a baseline scenario that reflects the most likely state of the economy in the year in question. The methods employed to derive the baseline scenarios are heterogeneous and range from the usage of the last observable year to complete and consistent estimation procedures. In the case of computable general equilibrium (CGE) analyses, the Scenar2020 project (European Commission 2006a) is one example of how projections of macroeconomic indicators (exogenous drivers) are used to construct the baseline as a model scenario: Starting from a calibrated version, exogenous variables are modified until macroeconomic projections are met. However, numerous projections refer to economic indicators which are endogenous variables within the CGE framework, such as gross domestic product (GDP), market prices, or produced quantities. To investigate methods that allow integrating projections for endogenous CGE variables is the main topic of this study. A starting point is the work by Arndt et al (2002), where entropy-based (Golan et al 1996) techniques are employed for the estimation of behavioural parameters by fitting a CGE model to time series on endogenous variables. Following this concept, we apply a method to fit a CGE´s parameters and endogenous variables to market- and macroeconomic projections from major research institutes. The original model is calibrated for the year 2000 for selected Member States of the EU, and the estimation is based on and tested against available observations for 2005.