Three essays on bubbles in agricultural futures markets

The central theme of this dissertation is to test and understand bubbles in agricultural futures markets. Building on current bubble literature, the dissertation applies a recent-developed forward and backward recursive bubble testing procedure that is capable of date-stamping the exact origination and termination dates of multiple bubbles to various commodity futures markets under various specifications. To account for conditional heteroskedasticity, inferences of the bubble testing procedure is derived based on the recursive wild bootstrap. Though organized under the same framework and examined using the same procedure, the three essays of this dissertation focus on different aspects of speculative influences on recent commodity price booms, with each providing a separate explanation for the booms. All our analyses indicate that speculators may not have been the culprit of recent heightened commodity price volatility as many market analysts have argued. Specifically, in the first essay, we focus on an in-depth analysis of historical patterns of bubbles in a variety of agricultural markets including grains, softs, and livestock products. We find that in general, bubbles are rather rare events and did not become more common in the most recent episode, as compared to the 1971-1976 episode. In addition, bubbles tend to be very short-lived and of small magnitude. In the second essay, we focus on analyzing the market impact of financial index investment in grain futures markets during explosive and non-explosive periods of price behavior. We find that on balance, the effect of index investment on grain futures spikes is very limited, even during periods when prices are explosive and may have been of central concern to policy-makers. In the third essay, we go beyond date-stamping bubble periods and attempt to relate the timelines of bubble occurrence with various fundamental and non-fundamental factors. Using a logit model that accounts for bias due to events rarity, we find that in the presence of bubbles, inventory and exchange rate tend to be lower, while aggregate global demand, crude oil prices, and U.S. export sales tend to be higher. For speculative activities, we find that the market may have benefited from the added liquidity brought in by speculators, which apparently reduces the probability of bubble occurrence