動態隨機一般均衡模型，近年來被大量使用在政策分析上，然而對此模型的預測能力的評比卻稍顯缺乏。本文遵循Ghent(2009)的作法，比較三種具有個體基礎模型對台灣資料的估計及預測結果。作者發現，在對國外詮釋不夠周全時，容易低估市場的競爭程度，換句話說，與市場競爭程度相關之係數，容易出現錯誤判斷，例如，價格僵固係數和中間財廠商的加價能力在對國外詮釋不夠周全時，皆出現高估的情況。除了以上發現外，本文也針對上述三種模型的預測能力做評比，其結果皆顯示，動態隨機一般均衡模型相較於貝氏向量自我迴歸模型有較佳的預測能力，另一方面也發現，若能將國外詮釋得更周全，確實能改善模型對台灣總體變數的預測能力。Dynamic stochastic general equilibrium (DSGE) models have become standard tools for policy analysis in recently years. However, their forecasting properties have not been fully explored for Taiwanese economy. In this paper, we follow Ghent(2009) to compare the re-sults of estimation and prediction among three models with micro-foundations for Taiwa-nese data. Our study shows that the degree of market competition is underestimated; sticky price coefficient and the bargaining power of intermediate goods firm are overestimated when we do not consider the foreign sector completely. Moreover, we evaluate the fore-casting power of the three models and BVAR’s. We also find that the prediction power of the DSGE model with an open sector is the best.
