When applying one of our previous Bayesian versions of Black-Litterman to the whole S&P 500, we ran into running time and memory allocation problems since we had to generate a random matrix of size 106GB at each iteration in a Gibbs Sampler. This suggested the fact the we needed to reduce the dimension. Because of this and because of the close relationship between the original Black-Litterman and the CAPM (which is a factor analysis model), we decided to introduce factors to the versions we already had. 

​

Sensitivity analyses for the level of confidence that investors have in their own views were performed. Also, the performance was assessed on a test data-set consisting of returns over the month of January 2018 eve when he investor inputs personal views about different sectors. 

​

Complete work in process of publication.

​

​