A1964
Title: Model averaging: A case study of the petrochemical sector in a large-scale general equilibrium macro model
Authors: Fakhri Hasanov - KAPSARC (Saudi Arabia) [presenting]
Peter Pedroni - Williams College (United States)
Abstract: Petrochemical production is an important source of economic diversification for many oil-exporting economies. Policy decisions related to the petrochemical sectors in such economies are often informed by the forecasts of large-scale general equilibrium macro models. A standard point of interest is the parameterization of the many equations entailed in these models. We make use of such a model, namely the KAPSARC Global Energy Macroeconometric Model (KGEMM), which represents the economic, energy, and environmental nexus for Saudi Arabia in order to study the extent to which model predictions can be improved by blending forecasts based on alternative estimation approaches. In particular, we estimate VAR models for the petrochemical sectors key indicators such as value-added, feedstocks and employment and derive the parameter implications based on the long-run VAR representations. We combine these with existing KGEMM parameterizations, which are estimated predominately using single equation cointegration methods, and compare the performance relative to the unblended version. Finally, we also use the VAR-based projections of some of the key variables, such as domestic energy prices, in order to further enhance the performance of the KGEMM model simulations.
