Abstract

This thesis investigates the relationship between speculation in futures markets and expected and unexpected volatility in the spot markets for 21 different commodities. I use the index of adequate speculation, INDADSP, and the index of excess speculation, INDEXSP, developed and estimated by Shanker (2017), to capture the degree of speculation required to meet hedging demand, and the degree of speculation in excess of hedging demand, respectively. For comparison, I also use Working’s (1960) speculative index T, as a measure of speculation. I estimate the expected volatility (EV) and unexpected volatility (UEV) of the spot market using a GARCH model. The empirical results indicate that the GJR-GARCH model with a Student’s t distribution for the error term is the most appropriate model, among the GARCH-family of models, to capture the volatility of 17 of the 21 spot commodity returns. However, the results of feeder cattle indicate the exists of serial correlation of the residuals for all three GARCH model I used, so I drop it and do the further analysis for the rest of 20 commodities and financial contracts. For each commodity, I create time series of matched weekly indices of speculation, expected volatility and unexpected volatility. Next, I investigate the long-run and short-run relationships between volatilities and speculation using an autoregressive distributed lag model. The results indicate that there is a long term relationship between expected and unexpected volatility and the speculative indices, for all commodities, except the Euro, Eurodollar, and U.S. T-bond, and a short term relationship between volatilities and speculation for all commodities. Finally, I apply the Toda-Yamamoto test to investigate the causal relationship between speculation in futures markets and volatility in spot markets. I find that speculation tends to lead expected volatility more than unexpected volatility for the majority of commodities/financial assets. Expected volatility, rather than unexpected volatility, tends to lead speculation for a majority of commodities/financial assets. There is a bidirectional causality between expected volatility and INDADSP, INDEXSP, and T and between unexpected volatility and INDEXSP for several different commodities and financial assets. However, there is no bidirectional causality between unexpected volatility and the speculative indices INDADSP and T for all 20 commodities/financial assets.