THE EFFICIENT MARKET HYPOTHESIS AND THE DYNAMIC BEHAVIOR OF SUGAR FUTURE PRICES

Abstract

This paper examines the market efficiency hypothesis by estimating time-varying coefficients using Kalman Filter techniques for an Autoregressive Moving Average model (ARMA) with Generalized Autoregressive Conditiona l Heteroskedasticity (GARCH) errors. The estimation technique utilized was maximum likelihood for sugar future returns. The higher lag order variable coefficients were the ones which showed greater fall in absolute value over time, which may suggest that variables with higher lags may lose weight as markets get more efficient.