A new robustness analysis for climate policy evaluations

• Verfasst von: Hermeling, Claudia [VerfasserIn]
• Verfasst von: Löschel, Andreas [VerfasserIn]
• Verfasst von: Mennel, Tim [VerfasserIn]
• Titel: A new robustness analysis for climate policy evaluations
• Titelzusatz: a CGE application for the EU 2020 targets
• Verf.angabe: Claudia Hermeling, Andreas Löschel, Tim Mennel
• E-Jahr: 2013
• Jahr: April 2013
• Umfang: 9 S.
• Fussnoten: Available online 4 October 2012 ; Teil des Sonderhefts: Special section: Long run transitions to sustainable economic structures in the European Union and beyond ; Gesehen am 22.12.2021
• Titel Quelle: Enthalten in: Energy policy
• Ort Quelle: Amsterdam [u.a.] : Elsevier Science, 1973
• Jahr Quelle: 2013
• Band/Heft Quelle: 55(2013), Seite 27-35
• ISSN Quelle: 1873-6777
• DOI: doi:10.1016/j.enpol.2012.08.007
• Datenträger: Online-Ressource
• Sprache: eng
• Sach-SW: CGE models
• Sach-SW: Climate policy
• Sach-SW: Sensitivity analysis
• K10plus-PPN: 1783493143

Abstract

This paper introduces a new method for stochastic sensitivity analysis for computable general equilibrium (CGE) model based on Gauss Quadrature and applies it to check the robustness of a large-scale climate policy evaluation. The revised version of the Gauss-quadrature approach to sensitivity analysis reduces computations considerably vis-à-vis the commonly applied Monte-Carlo methods; this allows for a stochastic sensitivity analysis also for large scale models and multi-dimensional changes of parameters. In the application, an impact assessment of EU2020 climate policy, we focus on sectoral elasticities that are part of the basic parameters of the model and have been recently determined by econometric estimation, alongside with standard errors. The impact assessment is based on the large scale CGE model PACE. We show the applicability of the Gauss-quadrature approach and confirm the robustness of the impact assessment with the PACE model. The variance of the central model outcomes is smaller than their mean by order four to eight, depending on the aggregation level (i.e. aggregate variables such as GDP show a smaller variance than sectoral output).