JOINT MODELING AND SIMULATION OF AUTOCORRELATED NON-NORMAL TIME SERIES: AN APPLICATION TO RISK AND RETURN ANALYSIS

This study presents a technique that can jointly model and simulate the expected values, variances, and covariances of sets of correlated time-series dependent variables that are autocorrelated and non-normal (right or left skewed and/or kurtotic). It illustrates the method by applying it to risk analysis of diversified tropical agroforestry systems.