Panel A of Table 3 reports the mean (median) accuracy and dispersion of analysts’
forecasts conditional on the auditor choice. The sample sizes (n) are reported in the first
row. As reported in column (1), the mean (median) forecast accuracy is 0.0937 (0.0137)
when the auditor is one of the non-Big 5, whereas it is 0.0224 (0.0024) when the auditor
is one of the Big 5 (i.e., column (2)). The difference is negative and significant at 1 percent
level in both t- and Wilcoxon z-tests (t 19.45 and z 20.80), suggesting that forecast
accuracy is higher for the firms audited by a Big 5 auditor. Similarly, the forecast dispersion
is lower for the firms audited by a Big 5 auditor as evidenced by the significant difference
reported in the bottom two rows of Panel A of Table 3 (t 15.04 and z 18.41).