Do recessions and bear markets occur concurrently across countries? A multinomial logistic approach

Aubrey Poon, Dan Zhu

Research output: Contribution to journalArticleResearchpeer-review

Abstract

We introduce a novel multinomial logistic model for detecting and forecasting concurrent recessions and bear markets across multiple countries. Our framework leverages cross-country panel features and provides additional information for robust analysis. Through a comprehensive simulation study, we demonstrate the computational efficiency and accuracy of our model, even when handling multiple binary indicators. Applying our framework to empirical data from the United States, the UK, and Euro Area, we find that the multinomial logistic model produces superior medium-term forecasting of concurrent recession and bear market events across countries compared to multiple independent single logistic models. Additionally, our counterfactual analysis reveals that specific events, such as a recession and bear market in the United States, along with the tightening of financial conditions and a negative interest rate spread in the United States, increase the probability of concurrent and individual recession and bear market occurrences in the UK and Euro Area.
Original languageEnglish
Number of pages21
JournalJournal of Financial Econometrics
DOIs
Publication statusAccepted/In press - 2024

Keywords

  • recession prediction
  • bear markets
  • multinomial logistic
  • cross-country
  • mixed frequency
  • Bayesian estimation

Cite this