On the http://scikit-learn.org/stable/modules/generated/sklearn.linear_model.MultiTaskElasticNetCV.html#sklearn.linear_model.MultiTaskElasticNetCV, and http://scikit-learn.org/stable/modules/linear_model.html#multi-task-elastic-net, documentation describes use of l1_ratio (or rho) as:

The optimization objective for MultiTaskElasticNet is:

(1 / (2 * n_samples)) * ||Y - XW||^Fro_2
+ alpha * l1_ratio * ||W||_21
+ 0.5 * alpha * (1 - l1_ratio) * ||W||_Fro^2

But the description of the parameter (and the code) agree that it should be:

(1 / (2 * n_samples)) * ||Y - XW||^Fro_2
+ alpha *(1 -  l1_ratio) * ||W||_21
+ 0.5 * alpha * l1_ratio * ||W||_Fro^2