Semiparametric inference for multivariate fractional processes and its applications

2023 Fiscal Year Final Research Report

• Project/Area Number: 19K01590
• 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: Okayama University
• Principal Investigator: Narukawa Masaki 岡山大学, 社会文化科学学域, 准教授 (30588489)
• Project Period (FY): 2019-04-01 – 2024-03-31
• Keywords: 実数和分過程 / 多変量時系列 / セミパラメトリック推定 / Taper / 共和分

Outline of Final Research Achievements

In this research, we proposed a two-step approach to semiparametrically estimate fractional cointegration with the fractionally integrated parameters in potentially nonstationary multivariate time series by combining the narrow-band least squares method and the multivariate local Whittle method with the efficient tapering incorporated. Furthermore, a Hausman-type test statistic was constructed from the multivariate local Whittle likelihood to detect the existence of fractional cointegration. In addition to deriving the asymptotic properties of the estimators and the test statistic, their finite sample performance was investigated by numerical simulation.

Free Research Field

計量経済学

Academic Significance and Societal Importance of the Research Achievements

多変量時系列解析において実数共和分分析は有用かつ魅力的な手法の一つであるが，強弱双方の実数共和分関係や一般的な変量数を想定した既存研究は少ない上にほとんどが制約的な側面を有している状況下で，単一方程式モデルに基づいているものの計算負荷が比較的抑えられ効率性も保ちつつ安定したセミパラメトリック推測を行えうる実用性の高い頑健なアプローチを提案している本研究はこの分析に新たな進展をもたらすと考えられる．