The existing literature on audit opinion shopping provides inconsistent evidence on whether such shopping has any association with abnormal audit fees. In this paper, we hypothesize that firms engage in audit opinion shopping and pay an abnormal audit fee only when their degree of accounting quality is low. To examine the issue, we group firms on the basis of their change in return on assets(ROA), and show that abnormal audit fees improve audit opinions only among firms that engage local auditors and have a low degree of ROA, but report a large increase in ROA, especially when the ROA change is the result of abnormal accruals. We find no association between abnormal audit fees and audit opinion improvement for other firms. The existing literature on audit opinion shopping provides any inconsistent evidence on whether such shopping has any association with abnormal audit fees. When this paper, we hypothesize that firms engage in audit opinion shopping and pay an abnormal audit fee only when their degree of accounting quality is low To examine the issue, we group firms on the basis of their change in return on assets (ROA), and show that abnormal audit fees improve audit opinions only among firms that engage local auditors and have a low degree of ROA, but report a large increase in ROA, especially when the ROA change is the result of abnormal accruals. We find no association between abnormal audit fees and audit opinion improvement for other firms.