Robustness of CGE Simulation Results in the Context of Structural Changes – The Case of Poland
Robustness of CGE Simulation Results in the Context of Structural Changes – The Case of Poland

Summary/Abstract: It is common to address the problem of uncertainty in computable general equilibrium modeling by sensitivity analysis. The relevant studies of the effects of parameter uncertainty usually focus on various elasticity parameters. In this paper we undertake sensitivity analysis with respect to the parameters derived from calibration to a benchmark data set, describing the structure of the economy. We use a time series of benchmark databases for the years 1996-2005 for Poland to sequentially calibrate a static CGE model, and examine the dispersion of endogenous variables’ responses in three distinct simulation experiments.We find a part – though not the most – of the results to be significantly sensitive to the choice of calibration database (including ambiguities about the direction of response). The dispersion of the results and its sources clearly depend on the shock in question. Uncertainty is also quite diverse between variables. It is thus recommended that a thorough parametric sensitivity analysis be a conventional part of a simulation study. Also, the reliability of results would likely benefit even from simple, trend-based updates of the benchmark data, as the responses of endogenous variables exhibit systematic changes, observed when the model is calibrated to the data for consecutive years.

• Details
• Contents