Table 2, panel C, results indicate that third-to-fourth-quarter ETR reductions
(ETR4_ETR3) were greater in post-SOX years than in pre-SOX years; mean
ETR4_ETR3 is 0.44 percent before the passage of SOX and 0.67 percent after
its enactment. A t-test (p 0.04) indicates that this difference is significant; a Wilcoxon
rank-sum test (p 0.01) also supports this conclusion. We find that tax fees
paid to auditors declined from pre- to post-SOX periods; the mean of Tax_Fees is
0.21 percent of SG&A before SOX and 0.20 percent of SG&A after SOX, but a
t-test indicates that this decrease is insignificant (p 0.57).10 However, sensitivity
analyses indicate that this result depends on the classification of 2002 as either a
pre- or post-SOX year. In untabulated results, we find that, when 2002 is treated as
a post-SOX year or excluded from the sample, Tax_Fees decreases significantly
from pre- to post-SOX periods. This result is consistent with Omer et al. 2006 and
Maydew and Shackelford 2006, who find decreasing tax fees paid to auditors from
pre- to post-SOX periods. Thus, we find some evidence that this legislation
prompted decreases in tax fees paid to auditors, depending on how we treat year
2002 observations in our sample.