In codes the drift capacity of unreinforced masonry (URM) walls is often estimated as a function of the failure mode and the aspect ratio. The empirical relationships are based on results from quasi-static cyclic tests on single URM walls, which were tested simulating either fixed-fixed or cantilever boundary conditions. In real structures, the stiffness and strength of slabs and spandrels define the boundary conditions of the walls and therefore the moment, shear force and axial force imposed on a pier during an earthquake. Depending on the exact configuration of pier, slab and spandrel, the boundary conditions can vary significantly. In order to investigate the influence of these boundary conditions on the force-deformation behaviour of URM walls, six quasi-static cyclic tests were performed. Different boundary conditions were simulated by varying the axial load ratio and the ratio of top and bottom moment applied to the pier. This article presents the test results and discusses the influence of the boundary conditions on the failure mechanism and the drift capacity of the walls. In addition, the results from 64 quasi-static tests on URM walls of different heights and masonry types are evaluated. These tests confirm the influence of the boundary conditions on the drift capacity. Moreover, they show that a strong size effect is present which leads to smaller drift capacities with increasing pier height. For this reason, an empirical drift capacity equation is proposed which accounts for the moment profile, the axial load ratio and the size effect.