Scholarlink

<p>We propose a&#xa0;novel method to address the end-point problem in filtering economic time series. The main idea is to replace filtered quarterly observations at the end of the sample with static forecasts from a&#xa0;MIDAS regression using higher-frequency time series. This method has the potential to improve the stability of output gap estimates or other cyclical series, as we confirm by empirical analysis on selected central and eastern European (CEE) countries and the United States. Stability may still, however, be violated by structural breaks in business cycles or an excessive amount of short-term noise. While MIDAS regressions can offer an improvement over the HP filter approach in estimating output gaps, it is crucial to consider the major role of country-specific circumstances.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

MIDAS Regression: A New Horse in the Race of Macroeconomic Time Series Filtering

  • 期刊论文

  • 发表日期：

    2025年8月6日

查看全文
  • Michal Benčík

摘要

We propose a novel method to address the end-point problem in filtering economic time series. The main idea is to replace filtered quarterly observations at the end of the sample with static forecasts from a MIDAS regression using higher-frequency time series. This method has the potential to improve the stability of output gap estimates or other cyclical series, as we confirm by empirical analysis on selected central and eastern European (CEE) countries and the United States. Stability may still, however, be violated by structural breaks in business cycles or an excessive amount of short-term noise. While MIDAS regressions can offer an improvement over the HP filter approach in estimating output gaps, it is crucial to consider the major role of country-specific circumstances.