Early warning of bubbles in the agricultural commodity market : evidence from LPPLS confidence indicators

This study leverages the Log-Periodic Power Law Singularity (LPPLS) confidence indicator to effectively identify bubbles in agricultural commodity markets. We analyze five major grain price indices reported by the International Grains Council (IGC) from January 2000 to April 2023, successfully identifying several notable bubble periods. These include a positive bubble in 2004 driven by decreased food production, a substantial positive bubble during the 2008 global financial crisis, a negative bubble in 2016, and positive bubbles triggered by the COVID-19 pandemic and the Russo-Ukrainian conflict since 2020. As the critical point is approached, the LPPLS confidence indicator exhibits strong early warning capabilities. To verify the model's robustness, we employ the Bai-Perron test for multiple structural breaks, detecting five such breaks in each agricultural commodity price series. LPPLS indicators provide strong early-warning signals prior to these break dates. Finally, we investigate the predictability of price bubbles in the agricultural commodity market. Using a Markov regime-switching model, our findings confirm that the geopolitical risk index possesses significant predictive power for bubble formation.