BUSINESS CYCLES, TREND ELIMINATION, AND THE HP FILTER

• 成果类型: Article
• 署名作者: Phillips, Peter C. B.; Jin, Sainan
• 署名单位: Yale University; University of Auckland; University of Southampton; Singapore Management University
• 刊物名称: INTERNATIONAL ECONOMIC REVIEW
• ISSN/ISSBN: 0020-6598
• DOI: 10.1111/iere.12494
• 发表日期: 2021
• 页码: 469-520
• 关键词: hodrick-prescott filter; time-series; unit-root; robust; asymptotics; regression; frequency; priors

摘要

Trend elimination and business cycle estimation are analyzed by finite sample and asymptotic methods. An overview history is provided, operator theory is developed, limit theory as the sample size n ->infinity is derived, and filtered series properties are studied relative to smoothing parameter (lambda) behavior. Simulations reveal that limit theory with lambda=O(n4) delivers excellent approximations to the HP filter for common sample sizes but fails to remove stochastic trends, contrary to standard thinking in macroeconomics and thereby explaining spurious cycle effects of the HP filter. The findings are related to the long run effects of the global financial crisis.