Objective: The research on the association between the working alliance and therapist competence/ adherence and outcome from cognitive behavioral therapy (CBT) is limited and characterized by inconclusive fi ndings. This study investigates the working alliance and competence/adherence as pre- dictors of outcome of CBT for social anxiety disorder (SAD) and panic disorder (PD). Method: Eighty-two clinically referred patients (58.5% female; age: M ¼ 33.6 years, SD ¼ 10.3) with PD ( n ¼ 31) or SAD ( n ¼ 51) were treated with 12 sessions of manualized CBT by 22 clinicians with limited CBT experience in a randomized controlled effectiveness trial. Independent assessors rated the CBT competence/adherence of the therapists using a revised version of the Cognitive Therapy Adherence and Competence Scale, and the patients rated the quality of the working alliance using the Working Alliance Inventory-short form in therapy sessions 3 and 8. The outcome was assessed by independent assessors as well as by patients self-report. A total of 20.7% of the patients (27.5% SAD, 9.7% PD) dropped out during treatment. The association between the alliance, competence/adherence, outcome and dropout was investigated using multiple regression analyses. Results: Higher therapist' competence/adherence early in the therapy was associated with a better outcome among PD patients, lower competence/adherence was associated with dropout among SAD patients. Higher rating of the alliance late in the therapy was associated with a better outcome, whereas lower alliance rating late in the therapy was associated with dropout. Conclusion: The fi ndings indicate that the therapist competence/adherence and the working alliance have independent contributions to the outcome from CBT for anxiety disorders, but in different phases of the treatment.