A quantile autoregression analysis of price volatility in agricultural markets

This paper investigates the dynamics of agricultural price volatility based on a quantile autoregression (QAR) model. The QAR model provides a flexible representation of the distribution of price and its dynamics. The approach is applied to U.S. wheat and corn markets over the period of 1980–2017. This period is of significant interest as it covers important changes in agricultural policy and increased reliance on markets. The price analysis is conducted conditional on stocks held in the previous period. We show how increasing previous stocks shift the price distribution to the left and decreases the odds of facing price spikes (by shifting down the upper tail of the price distribution). Our analysis also examines the effects of changing public stocks on prices. For both wheat and corn, this reflects changing agricultural policy, contrasting the 1980s (when public stocks were relatively high) with the post‐2005 period (when public stocks became zero). We document how higher public stock ratio during the previous period did not lower the odds of facing price spikes. Applied to the wheat and corn markets, we also uncover evidence of local dynamic instability in the upper tail of the price distribution, suggesting that price instability becomes more pronounced when previous stocks are low.