Forecasting and Empirical Analysis By Multivariate Time Series Models Using Bayesian Model Averaging

• Project/Area Number: 20K01591
• Research Category: Grant-in-Aid for Scientific Research (C)
• Allocation Type: Multi-year Fund
• Section: 一般
• Review Section: Basic Section 07030:Economic statistics-related
• Research Institution: University of the Ryukyus
• Principal Investigator: Sugita Katsuhiro 琉球大学, 国際地域創造学部, 教授 (50377058)
• Project Period (FY): 2020-04-01 – 2024-03-31
• Project Status: Completed (Fiscal Year 2023)
• Budget Amount: ¥4,420,000 (Direct Cost: ¥3,400,000、Indirect Cost: ¥1,020,000); Fiscal Year 2022: ¥910,000 (Direct Cost: ¥700,000、Indirect Cost: ¥210,000); Fiscal Year 2021: ¥1,300,000 (Direct Cost: ¥1,000,000、Indirect Cost: ¥300,000); Fiscal Year 2020: ¥2,210,000 (Direct Cost: ¥1,700,000、Indirect Cost: ¥510,000)
• Keywords: 時系列モデル / ベイズ法 / ベイジアンモデル平均法 / 多変量モデル / 計量経済学 / 時系列予測 / 時系列分析 / 多変量時系列分析 / 経済時系列モデル / 多変量時系列モデル / ベイジアン / MCMC / 多変量解析

Outline of Research at the Start

ベクトル自己回帰モデル等の多変量時系列モデルは、様々なマクロ経済や金融計量分析、そして予測において有用であるが、モデルの推定パラメータ数が多く、そして不必要なパラメータを多く含んでいるのが問題である。そこで本研究では、ベイジアン・モデル平均法を多変量時系列モデルに応用し潜在的モデルの不確実性を考慮し過剰適合の問題を回避し予測精度の向上を図る。本研究ではBMA法をVARモデル、そして非線形VARモデルや多変量GARCHモデルに応用し、マクロ経済や金融の計量分析ならびに予測に関する研究を行う。

Outline of Final Research Achievements

Multivariate times series models such as Vector autoregressive (VAR) model are often used for analysis and forecasting of macro economic and finantial econometrics. However, these models have many parameters to infer and contain unnecesary parameters in models. This study investigates multivariate time series models using a Bayesian model averagin method to improve forecasting performance, considering uncertainty in models and avoiding the overparameterization problem. This study consider various VAR models and nonlinear VAR models, and multivariate GARCH models using the BMA method for econometric analysis and forecasting for macro economics and financial econometrics.

Academic Significance and Societal Importance of the Research Achievements

研究成果の学術的意義は不必要なパラメータを多く含む多変量時系列モデルに対してできるだけ不必要なパラメータを除去しモデル自身を単純化することによってより信頼性の高い計量モデルを構築し、その結果非線形多変量モデルなどによってより高い予測精度が得られることにある。

Research Products

• [Journal Article] Time Series Forecasting Using a Markov Switching Vector Autoregressive Model with Stochastic Search Variable Selection Method (2022)
  • Author(s): Sugita Katsuhiro
  • Journal Title: Financial Econometrics: Bayesian Analysis, Quantum Uncertainty, and Related Topics, Studies in Systems, Decision and Control 427 Volume: 427 Pages: 147-170
  • DOI: 10.1007/978-3-030-98689-6_10
  • ISBN: 9783030986889, 9783030986896
• [Journal Article] Forecasting with Bayesian vector autoregressive models: comparison of direct and iterated multistep methods (2022)
  • Author(s): Sugita Katsuhiro
  • Journal Title: Asian Journal of Economics and Banking Volume: 6 Issue: 2 Pages: 142-154
  • DOI: 10.1108/ajeb-04-2022-0044
  • Peer Reviewed
• [Presentation] Time Series Forecasting Using a Markov Switching Vector Autoregressive Model with Stochastic Search Variable Selection Methods (2021)
  • Author(s): 杉田勝弘 (Katsuhiro Sugita)
  • Organizer: The Fifth Econometric Conference of Vietnam - ECONVN2022
  • Int'l Joint Research / Invited