Objective: This study aims to evaluate the differential predictive values of age, age of onset and duration of illness on paroxetine and cognitive‐behavioural therapy (CBT) outcome in late‐life panic disorder with agoraphobia. Method: Patients 60 years and older with a confirmed diagnosis of panic disorder with agoraphobia (n = 49) were randomly assigned to paroxetine (40 mg/day) treatment, individual CBT or a waiting‐list control condition. Multiple regression analyses were conducted per treatment arm with post‐treatment avoidance behaviour and agoraphobic cognitions as the dependent variables. Results: Higher age at onset and shorter duration of illness were predictors of superior outcomes following CBT, although these variables did not influence the treatment effects of paroxetine. Conclusions: In late‐life agoraphobic panic disorder, chronological age has no impact on treatment modality outcome. In older patients with a late disease onset or shorter duration of illness, CBT is to be preferred over paroxetine, whereas paroxetine might be the treatment of choice for older people with an early onset and short duration of illness.