Forecasting Japanese inflation with a news-based leading indicator of economic activities

Keiichi Goshima, Hiroshi Ishijima, Mototsugu Shintani*, Hiroki Yamamoto

*Corresponding author for this work

Research output: Contribution to journalArticlepeer-review

Abstract

We construct business cycle indexes based on the daily Japanese newspaper articles and estimate the Phillips curve model to forecast inflation at a daily frequency. We find that the news-based leading indicator, constructed from the topic on future economic conditions, is useful in forecasting the inflation rate in Japan.

Original languageEnglish
Pages (from-to)111-133
Number of pages23
JournalStudies in Nonlinear Dynamics and Econometrics
Volume25
Issue number4
DOIs
Publication statusPublished - 2021 Sep 1
Externally publishedYes

Keywords

  • Phillips curve
  • nowcasting
  • out-of-sample forecasting
  • text classification

ASJC Scopus subject areas

  • Analysis
  • Social Sciences (miscellaneous)
  • Economics and Econometrics

Fingerprint

Dive into the research topics of 'Forecasting Japanese inflation with a news-based leading indicator of economic activities'. Together they form a unique fingerprint.

Cite this