Decomposing aggregate risk into marginal risks under partial information: A top-down method

This paper proposes a method to decompose a square-integrable random variable into any number of marginal random variables under partial information restrictions. An executable algorithm and a concrete example of the capital allocation problem are also provided.