Long Memory, the 'Taylor Effect' and Intraday Volatility in Commodity Futures Markets

Abstract

This paper investigates long term dependence in commodity futures markets. Using daily futures returns on cocoa, coffee and sugar, we show that FIGARCH models are able to adequately describe both the long and short run characteristics of commodity market volatility. The paper also considers three measures of risk - squared returns, absolute returns and intraday volatility - and finds that they exhibit the long memory property. Intraday volatility shows the strongest auto correlation structure. Moreover, there is evidence of the so-called "Taylor effect". Key Words: long memory, fractional differentiating; ARFIMA; FIGARCH; squared returns; absolute returns; Taylor effect; intraday volatility.