Quantile connectedness in agri-commodity markets: What has changed over past six decades?

Agri commodities have been investigated in the past to determine their inter-relationships. However, no study has checked their risk spillover/connectedness for six decades using extreme quantiles. Various shocks (positive/negative) often pose challenges to these commodities over the past six decades. Such shocks' impact is usually observed in extreme quantiles or tails. Therefore, we have investigated fourteen agri commodities (namely Coffee, Cocoa, Soyabean, Wheat, Sugar, Orange, Chicken, Beef, Maize, Tea, Coconut Oil, Groundnut Oil, Palm Oil & Rice) from January 1, 1960 to June 1, 2022 (covering 62 years on a monthly basis), deploying Quantile VAR or QVAR as suggested by [1](extended [2,3] calibration). We found that the risk spillover/ connectedness never came down for these Agri commodities. It is always at a higher level (more than 55%) proving that agri commodities remain vulnerable to various shocks throughout. Spillover looks symmetric as both the extreme tails enjoy about 92-93% connectedness levels, whereas the median is below 60%. Rice, Orange Juice, Chicken, Tea and Groundnut Oil were consistent net receivers across such a long-time frame, whereas Palm Oil, Soyabeans, Maize and Wheat were net emitters all through. Further, we found decreasing complexity (network connectedness reduction) with increased quantiles. Since these findings are over such an extended period, policy decisions can be made based on them.

Automated summary

This study investigates the risk spillover and connectedness of agri commodities over the past six decades using extreme quantiles. The findings show that these commodities exhibit consistently high levels of risk spillover and connectedness, indicating their vulnerability to various shocks. Rice, Orange Juice, Chicken, Tea, and Groundnut Oil are consistent net receivers, while Palm Oil, Soyabeans, Maize, and Wheat are net emitters throughout. Additionally, there is a decreasing complexity of network connectedness with increased quantiles.