Recently developed techniques are combined for modeling mutually correlated crop yields and prices that exhibit heteroscedasticity and autocorrelation, respectively, and follow non-normal probability density functions (pdf's). The importance rigorously modeling these pdf's for financial risk analysis is illustrated through a case study of tropical agroforestry systems for coffee production.