Do recessions occur concurrently across countries? A multinomial logistic approach

Arbeitspapier

We develop a novel multinomial logistic model to detect and forecast concurrent recessions across multi-countries. The key advantage of our proposed framework is that we can detect recessions across countries using the additional informational content from the cross-country panel feature of the data. Furthermore, in a simulation study, we show that our proposed model accurately captures the true underlying probabilities. Finally, we apply our proposed framework to a US and UK empirical application. In terms of recession forecastability, the multinomial logistic model with both countries' interest rate spread and the weekly US NFCI as the set of exogenous predictors was the best performing model. For the counterfactual analysis, we found that a previous US recession will increase the probability of a recession occurring jointly in the US and the UK. However, a tightening of the US NFCI and a negative interest rate spread in both countries only increases the probability of a recession exclusively in the US and UK, respectively.

Object Details

• Language: Englisch

References and Relationships

• Bibliographic citation: Series: Working Paper ; No. 11/2022

Classification and Topics

• Classification: Wirtschaft ; Business Fluctuations; Cycles ; Prices, Business Fluctuations, and Cycles: Forecasting and Simulation: Models and Applications ; Single Equation Models; Single Variables: Time-Series Models; Dynamic Quantile Regressions; Dynamic Treatment Effect Models; Diffusion Processes ; Single Equation Models; Single Variables: Discrete Regression and Qualitative Choice Models; Discrete Regressors; Proportions; Probabilities
• Subject: Recession prediction ; multinomial logistic ; cross-country ; mixed frequency ; Bayesian estimation

Contributors, Places and Time

• Event: Geistige Schöpfung
• (who): Poon, Aubrey ; Zhu, Dan
• Event: Veröffentlichung
• (who): Örebro University School of Business
• (where): Örebro
• (when): 2022

Further information

• Handle: http://hdl.handle.net/10419/265681
• Last update: 10.03.2025, 11:42 AM CET