Testing Rockets and Feathers Patterns with a Multivariate Quantile Approach : An Application to the Soybean Complex

Output markets usually respond to input price changes asymmetrically, with prices rising faster than they fall, known as the rockets and feathers pattern. We expand the vector autoregressive quantile (VARQ) model by incorporating quantile cointegrating relationships to investigate such asymmetric pattern in conditional quantiles, with an application to two soybean end products, soybean meal and oil, that are jointly produced by crushing soybeans. Our study finds mixed evidence of price asymmetries, driven by different market conditions of end products. Specifically, the prices of end products respond more to input price increases rather than decreases when their own market is bullish but the other product’s market is bearish, conforming the rockets and feathers pattern at the extreme deciles of the price distributions