Title: Leveraging Labor Market Metrics for Recession Identification in Japan

Abstract:

Recession forecasting tools are frequently perceived as being tailored specifically to the United States economy, prompting inquiries into whether labor market-centric methodologies, such as the Sahm and Michez Rules, can effectively identify downturns in other nations. This study addresses that uncertainty by assessing the adaptability of these rules to the Japanese context, specifically through the calibration of smoothing parameters and thresholds to align with local labor market statistics. We developed a comprehensive dataset comprising 95,832 distinct recession indicators that integrate both vacancy and unemployment figures. The resulting classification models demonstrated flawless statistical performance within the training phase (1970–2021), successfully capturing all 11 historical recessions without producing any false alarms. Of the total, 193 classifiers occupied the anticipation-precision frontier. By focusing on the high-precision segment of this group, we identified six models that maintained a standard deviation in detection errors of less than three months. On average, the chosen ensemble of classifiers indicated the start of a recession just 0.06 months after it actually began. These results indicate that labor market rules based on economic slack offer a robust, universal framework for enhancing the accuracy of real-time recession detection globally.

Source: arXiv Generated at: 2026-06-02 00:00:00 UTC