Developing new methods for improving prediction performance of both macro time series and effects of monetary and fiscal policies by combining multiple DSGE models

2017 Fiscal Year Final Research Report

• Project/Area Number: 15K03439
• Research Category: Grant-in-Aid for Scientific Research (C)
• Allocation Type: Multi-year Fund
• Section: 一般
• Research Field: Economic policy
• Research Institution: Tokyo Metropolitan University
• Principal Investigator: Iiboshi Horokuni 首都大学東京, 社会科学研究科, 教授 (90381441)
• Project Period (FY): 2015-04-01 – 2018-03-31
• Keywords: ニューケインジアンモデル / 非線形動学モデル / ベイズ統計学 / DSGEモデル / マルコフ連鎖モンテカルロ法 / 予測分布 / モデル結合

Outline of Final Research Achievements

Using Bayesian statistics, I examine and develop new methods of improving forecast performance of macro time series by combining multiple prediction densities derived from individual macroeconomic models, compared with the density from a single macro model. In addition, by expanding data from only output and inflation to term structure of interest rates, I verify that not only combining of multiple models but also combining of multiple data sets are useful to improve it. And I also estimate Japanese economy under deflation and zero lower bound after 1999, using nonlinear new Keynesian model.

Free Research Field

経済政策