Multi‐period portfolio selection with investor views based on scenario tree

摘要

How to measure investor views and apply it in multi-period investment is an important problem in portfolio selection. This paper attempts to construct a portfolio selection model with extreme situations and extend it under the multi-period framework. First, we modify a portfolio selection model to fit the extreme cases of 0% or 100% confidence views, then we establish a new programming problem based on optimization approach and figure out the explicit solutions. Second, we extend the model to multi-period form and discretize the results with scenario tree, which solves the multi-period problems. Third, we build an international portfolio with CVaR risk measurement. The numerical tests show that the new multi-period selection model performs better than the others.