← Some problems with the Markowitz mean-variance model