Global identification of linearized DSGE models

• 成果类型: Article
• 署名作者: Kociecki, Andrzej; Kolasa, Marcin
• 署名单位: Narodowy Bank Polski; Warsaw School of Economics
• 刊物名称: QUANTITATIVE ECONOMICS
• ISSN/ISSBN: 1759-7323
• DOI: 10.3982/QE530
• 发表日期: 2018
• 页码: 1243-1263
• 关键词: Global identification; DSGE models; state-space representation

摘要

This paper introduces a computational framework to analyze global identification of linearized DSGE models. A formal identification condition is established that relies on the restrictions linking the observationally equivalent state space representations and on the inherent constraints imposed by the model solution on the deep parameters. This condition is next used to develop an algorithm that checks global identification by searching for observationally equivalent model parametrizations. The algorithm is efficient as the identification conditions it employs shrink considerably the space of candidate deep parameter points and the model does not need to be solved at each of these points. The working of the algorithm is demonstrated with two examples.