Portfolio Analysis with Factors and Scenarios
展示如何通过添加新变量使情景模型的协方差矩阵对角化，解释新变量的含义，将组合方差分解为情景内和情景间方差，并推广到因子与情景并存的模型。
ABSTRACT Recently there has been a growing interest in the scenario model of covariance as an alternative to the one‐factor or many‐factor models. We show how the covariance matrix resulting from the scenario model can easily be made diagonal by adding new variables linearly related to the amounts invested; note the meanings of these new variables; note how portfolio variance divides itself into “within scenario” and “between scenario” variances; and extend the results to models in which scenarios and factors both appear where factor distributions and effects may or may not be scenario sensitive.