Price, inventories, and volatility in the global wheat market

Kyösti Pietola, Xing Liu, Miguel Robles
ifpri discussion paper

The study estimates a conditional mean model for international wheat prices and inventories. Endogenous price volatility and exogenous shocks in the price and inventory series are controlled for in estimation. Redressing the empirical linkage between volatility, prices, and inventories is important because volatility increases returns to inventories, which in turn may imply prices. The problem is also important from the regulator perspective, because publicly funded inventory programs have been traditional measures in stabilizing prices and improving food security by providing a buffer against adverse yield shocks and stock-outs. The structural model underlying the estimating equations is based on a dynamic inventory optimization problem. The data suggest that the price of both wheat and wheat inventories is nonstationary and that they are significantly linked to each other in the short run but do not exhibit a stationary long-run equilibrium relationship. Price volatility is an important determinant in the short-run conditional mean processes for both the price and inventories. The pairwise causal relationships have only one direction each. Inventories imply price volatility, price volatility implies price, and price implies inventories, but not vice versa. The parameter estimates suggest that when inventories decrease, price volatility increases. Thus, low inventories have likely been among the necessary conditions, but have not been a sufficient condition by themselves, for the price surge observed in 2008. The price and inventory movements have a significant negative relationship in the very short run, but it is leveled off over time. A decreasing price implies either inventory build-ups or postponement of inventory withdrawals. Overall, the current and past inventory and price movements are not very valuable in predicting the future price movements, and it is likely that the inventory information announced each month is already in the prices.