Further developments of exposure-based therapy (EBT) require more knowledge about transfer of treatment to non-trained everyday contexts. However, little is known about transfer effects of EBT. Using a standardized EBT protocol in 275 patients with panic disorder and agoraphobia we investigated the transfer of EBT to a highly standardized context during a Behavioral Avoidance Test (BAT; being entrapped in a small and dark test chamber) and not part of the exposure sessions. Patients of a treatment group underwent the BATs before treatment (t1), after a preparatory treatment phase (t2), and after an agoraphobic exposure phase (t3) and were compared with wait-list control patients, who repeated BAT assessments across the same time period. We found stronger reductions in avoidance behavior, reported fear, and autonomic arousal during the BAT from t1 to t3 in the treatment group patients who were anxious during t1 relative to the anxious but untreated patients. Fear reduction was related to treatment outcome indicating the contribution of transfer effects to successful EBT. Interestingly, reduction varied for different fear response systems suggesting different processes to may be involved in transfer effects. Importantly, final BAT assessment still evoked residual fear in the treatment group as compared to BAT non-anxious control patients, suggesting limited transfer effects – one possible reason for the return of symptoms in new situations.