This study suffers two limitations. The first lies in the sample data. Because audit fees may be calculated in a
variety of ways (e.g., they may or may not be inclusive of travel expenses and the fees for interim reports), we
cannot infer whether the data on companies that do not disclose detailed audit fees are consistent with those of
other companies. The second limitation lies in self-selection. Although we perform a sensitivity test using the
two-stage procedure developed by Heckman (1976), we cannot completely rule out the influence of self-selection
owing to the complexity of dealing with such a problem.