We study the performance of affine policies for two-stage adjustable robust optimization problem under a budget of uncertainty set. This important class of uncertainty sets provides the flexibility to adjust the level of conservatism in terms of probabilistic bounds on constraint violations.