Auditor logfee 1000 0392 0682 0735 0015 0213 0051

AUDITOR
LOGFEE 1.000 0.392 ** 0.682 ** 0.735 ** 0.015 0.213 ** –0.051 –0.117 ** –0.030 0.094 **
LOGNAS 0.359 ** 1.000 0.366 ** 0.224 ** –0.040 0.216 ** 0.060 –0.148 ** –0.053 0.157 **
LOGASSETS 0.683
**
0.339
**
1.000 0.443
**
–0.147
**
0.325
**
0.038 –0.211
**
–0.108
**
0.147
**
LOGSUBS 0.734 ** 0.210 ** 0.476 ** 1.000 0.035 0.118 ** –0.271 ** 0.062 0.055 –0.031
INVREC –0.105
**
–0.113
**
–0.382
**
–0.132
**
1.000 –0.007 –0.017 0.002 –0.009 0.018
LEVERAGE 0.374
**
0.234
**
0.379
**
0.346
**
–0.058 1.000 –0.015 –0.124
**
–0.050 0.018
FOREIGN 0.011 0.092 ** 0.111 ** –0.189 ** –0.010 –0.117 ** 1.000 –0.252 ** –0.033 0.117 **
CHINESE –0.156
**
–0.140
**
–0.239
**
0.034 0.215
**
–0.005 –0.254
**
1.000 –0.014 –0.152
**
OPINION –0.017 –0.053 –0.097
**
0.056 0.040 –0.092
**
–0.056 –0.023 1.000 –0.051
AUDITOR 0.095 ** 0.157 ** 0.152 ** –0.019 –0.129 ** –0.007 0.119 ** –0.149 ** –0.051 1.000
Note. Pearson Correlation is at diagonal up and Spearman Correlation is at diagonal down.
** Correlation is significant at the 0.01 level (2-tailed)
* Correlation is significant at the 0.05 level (2-tailed)
Multivariate Regression Analysis
The regression analysis (i.e. the first model) tests Hypothesis H1. The results are
presented in Table 5. The results contain both the OLS regressions for the full
sample and for NAS-purchased companies. Both models are well specified as
evidenced by high F statistics. The R
2
of both equations are in excess of 70% and
they are consistent with prior studies. Further, the results are similar for both
samples. More importantly, the hypothesis variable LOGNAS is significant at 1%
level for both regressions. However, the sign of the coefficient is in the opposite
direction of the theory proposed earlier. There is a significant positive
relationship between audit fees and NAS fees. The explanation to this finding of
the hypothesis variable is provided in the discussion section.
34

The Provision of Non-Audit Services
TABLE 5
REGRESSION RESULTS USING THE FULL SAMPLES (N = 819) AND
NAS FEE INCURRED COMPANIES (N = 512)
N = 819 N = 512
Variables +
Expected
sign
Coefficient t -statistic Coefficient t -statistic
LOGNAS – 0.051 5.284 ** 0.098 5.988 **
LOGASSETS + 0.251 12.407 ** 0.274 14.597 **
LOGSUBS + 0.560 18.606
**
0.533 19.773
**
INVREC + 0.002 10.246
**
0.214 5.104
**
LEVERAGE + –0.002 –0.453 –0.003 –0.511
FOREIGN + 0.208 3.718** 0.231 3.093
**
CHINESE – –0.053 –1.624* –0.078 –1.937
*
OPINION + –0.015 –0.531 –0.047 –1.028
AUDITOR +
0.022 1.214 0.036 1.451
#
Constant +/–
–0.044 –0.454 –0.293 –2.875 **
Adjusted R
2
F ratio
Prob > F (Two-tailed test)
0.720
198.73
0.000
0.739
161.68
0.000
+ See Table 3 for variable definitions
**Significant at 1% (one-tailed)
* Significant at 5% (one-tailed)
# Significant at 10% (one-tailed)
Other explanatory variables that are significant in both regressions are
LOGASSET, LOGSUBS, INVREC, FOREIGN and CHINESE. The variables are
significant at 1% level in the predicted directions. The results are consistent with
previous studies done in Malaysia and elsewhere (see e.g. Ayoib, 2001; Rose,
1999). As expected, company size and complexities are the main determinants of
audit fees. Similarly, foreign investors are likely to demand higher quality audit
and this is reflected in higher audit fee. Unique to the Malaysian audit market,
Chinese controlled companies pay lower audit fees than other companies due to
the Chinese business practice discussed earlier. However, the variable AUDITOR
is not significant for the analysis of all companies and only (weakly) significant
when the sample of NAS-purchased companies is utilised. Hence, the evidence of
brand name premium is not conclusive.