Depressive disorders often co-occur with personality disorders. The extent to which depressive disorders influence treatment outcome in personality disorders remains unclear. The aim of this study was to determine the impact of co-morbid depression on recovery from personality disorders and improvements in psychosocial functioning. This study drew data from a randomized-controlled trial in which patients (N = 320) with cluster-c (92%), paranoid, histrionic and/or narcissistic personality disorders received schema-therapy, treatment-as-usual, or clarification-oriented psychotherapy. Recovery from personality disorders at three-year follow-up and improvements in psychosocial functioning over a course of three years was predicted by the diagnostic status of depressive disorders at baseline using mixed model regression analyses. Based on the number of axis-I and axis-II disorders, personality disorder severity and global symptomatic distress and functioning a baseline severity index was computed and included in subsequent analyses to test the specificity of baseline depression in predicting outcomes. Patients with co-occurring depression reported higher baseline severity compared to patients without co-occurring depression. Depression at baseline was associated with lower recovery rates at three-year follow-up (p = 0.01) but this effect disappeared after controlling for baseline severity. Patients with depression at baseline reported higher psychosocial impairments throughout treatment (p < 0.001). Depression at baseline did not moderate treatment effects except for one psychosocial outcome measure. In conclusion, depression is associated with lower recovery rates from personality disorders but this effect disappears when general severity is taken into account. Patients with primarily cluster-c personality disorders and co-occurring depression might benefit from additional depression treatment in terms of improved psychosocial functioning.