The Perception Of Auditors Independence Accounting Essay

Published:

4.1 Introduction

The aim of this dissertation is to investigate whether the perception of auditors' independence is being compromised by providing the auditors' firms their clients with a high level of non-audit services. Otherwise, this study examines evidence in the UK about whether auditors providing more non-auditors, as their independence would be affected.

The current chapter presents the empirical results of this research study. This chapter is organised according to following: the next section 4.2 presents descriptive statistics results, whilst the third, forth, and fifth sections 4.3, 4.4 and 4.4 present the key regression models' finings of testing the main three research hypothesis. The sixth section 4.6 presents the some statistics results of the sensitivity tests. The last section will presents some discussions regards to the main findings in this research.

4.2 Descriptive Statistics

Table 1 (panel A) presents descriptive statistics. The mean of audit fee paid by the sample firms are £ 2,563,643 in 2005, £ 4,158,636 in 2006, £4,376,788 in 2007 and £5,149,828. The mean of non-audit fees paid by audit clients to their auditors are £3,527,600 in 2005, £2,293,610 in 2006, £2,063,765 in 2007 and £2,665,149 in 2009.Overall, it can be seen from the table that the mean audit fees paid by the sample companies have doubled during the period of five years from £ 2,563,643 in 2005 to reach £5,149,828. Whereas the mean non-audit fees have fluctuated during the study period. The ratio of non-audit fee to total fees is presented in Panel B. In addition, Panel A is also providing descriptive statistics for a variety of control variables. The mean of the audit's client's total assets is £61.308 billion in 2005, £67.831 billion in 2006, £83.438 billion in 2007 and £91.242 billion in 2009, as it is obviously that the mean of total asset has increased by 50% during the period. This might due to the global increasing of the competition in the worldwide market. Furthermore, in compassing to recent studies concluded in US, Australia and New Zealand, the UK companies examined here are slightly similar. They purchase lower level of non-audit services than those firms in recent studies using US data. Nevertheless, the levels are similar and comparable to those in New Zealand studies such as Hay et al., (2006) and Li et al,. (2003).

Lady using a tablet
Lady using a tablet

Professional

Essay Writers

Lady Using Tablet

Get your grade
or your money back

using our Essay Writing Service!

Essay Writing Service

Panel B of Table 1 shows the ratio of non-audit fees to total audit and non-audit fees for each year of the study period including the years 2005, 2006, 2007 and 2009. It excludes 7 observations for which there were either zero audit fees or zero non-audit fees. In order to have a consistent base comparison, it is also includes only the firms for which the data are available for the all four years. The mean ratio of non-audit fees to total audit and non-audit fees rapidly decreases by almost 50% from 55.% in 2005 to 29.77% in 2006 and then to reach 25.97% in 2007, but it has slightly increased to almost 30% in 2009. The decrease trend of non-audit services provided by the independence auditors during the years 2005 to 2007 is consistent with prior findings in New Zealand (Wong et al., 2002; Hay et al., 2006) and in other countries. In addition, a Mann-Whitney test is applied in order to compare the ratio of non-audit fees to total audit and non-audit fees for 2005 against 2006, 2006 against 2007, and 2007 against 2009. The consequences provide evidence that even thought there are observation decreased they are not significant.

Table 1: Descriptive Statistics

Panel A: Continues variables

Variable

Year

Mean

Median

Std.dev

Min

Max

N

Audit Fees(£000)

2005

2,563.643

1,625.000

3,013.808

0.000

18,303.000

70

2006

4,158.636

2,200.000

6,116.310

206.000

40,679.000

77

2007

4,376.788

2,291.000

6,103.646

200.000

31,140.000

85

2009

5,149.828

2,737.000

7,319.941

216.000

38,600.000

93

Non-audit fees(£000)

2005

3,527.600

2,150.000

4,820.547

0.000

33,333.000

70

2006

2,293.610

1,500.000

2,592.104

0.000

12,000.000

77

2007

2,063.765

1,200.000

2,353.686

0.000

12,701.000

85

2009

2,665.194

1,600.000

3,792.545

0.000

28,121.000

93

Total assets(£000)

2005

61,308.730

6,348.400

180,124.400

4.268

923,671.000

70

2006

67,831.090

8,358.200

188,350.400

0.000

996,023.000

77

2007

83,439.090

8,586.300

275,706.500

0.000

1,897,575.000

85

2009

91,242.910

8,938.000

287,905.100

927.000

1,689,447.000

93

INVREC

2005

0.1781

0.1418

0.1666

0.000

0.6300

70

2006

0.1767

0.1417

0.1812

0.000

0.9847

77

2007

0.1653

0.1328

0.1852

0.000

Lady using a tablet
Lady using a tablet

Comprehensive

Writing Services

Lady Using Tablet

Plagiarism-free
Always on Time

Marked to Standard

Order Now

0.9767

84

2009

0.1544

0.1194

0.1456

0.000

0.9108

93

ROA

2005

8.4944

7.5030

6.6276

0.000

26.6050

69

2006

10.8486

8.8015

8.9398

0.000

35.4670

74

2007

10.2573

8.0195

9.7838

0.000

55.7530

84

2009

6.8079

6.1260

5.4677

0.000

27.0460

83

TD/TA

2005

0.2418

0.2321

0.1751

0.000

0.7956

70

2006

0.2400

0.2420

0.1601

0.000

0.7414

76

2007

0.2591

0.2532

0.1638

0.000

0.6632

84

2009

0.2389

0.2262

0.1649

0.000

0.6296

93

CA/CL

2005

1.9063

0.8575

6.3815

0.000

50.0000

70

2006

1.0966

0.9049

1.3591

0.000

8.6480

77

2007

0.9705

0.8644

1.1699

0.000

7.9394

83

2009

1.1350

1.0037

1.2171

0.000

7.4822

92

FEEDEP

2005

1.0000

1.0000

0.0000

1.0000

1.0000

70

2006

0.8716

0.9091

0.1438

0.2000

1.0000

77

2007

0.8719

0.9219

0.1380

0.3967

1.0000

85

2009

0.9107

0.9438

0.1011

0.5150

1.0000

93

Notes: INVREC: the ratio of the sum of inventories and accounting receivable to total assets; ROA: Ratio of return on assets; TD/TA: Ratio of financial leverage (total debt to total assets); CA/CL: the current ratio (total current assets to total current liabilities); FEEDEP: Free dependence measured as the Auditee's audit and non-audit payment divided by the total audit and non-audit fees of the external auditor.

Panel B: The ratio of Non-audit Fees to Total audit and non-audit fees( for firms with non-zero audit fees or non-audit fees and with data for 4 years)

2005

2006

2007

2009

NAF/(AF + NAF)

NAF/(AF + NAF)

NAF/(AF + NAF)

NAF/(AF + NAF)

N

69

76

89

90

Mean

0.5971

0.3353

0.3131

0.3191

Median

0.5556

0.2977

0.2597

0.3043

Std. Dev.

0.1775

0.1802

0.1785

0.159

Minimum

0.1064

0.0435

0.0339

0.0821

Maximum

0.9031

0.8463

0.8333

0.7918

Table 1 (Continued)

Note: NAF/(AF+ NAF) : The ratio of non-audit fees to the total of audit fee and non-audit fees.

4.3 Key findings of testing the first hypothesis: Audit Fees and Non-audit Fees

The consequences of tests investigating the relationship between audit and non-audit fees in 2005, 2006, 2007 and 2009 [1] (excluding the year 2008 due the unavailable data of audit and non-audit fees as earlier reported) as reported in Table (2). In Panel A, the main findings of OLS regression are reported. The key point of the all models' results show that all models have adjusted R- squared of greater than 63%. Consistent with prior research studies such as Basioudis et al., (2008), It has been found that in this study that there is a insignificant positive relationship between non-audit fees and audit fees, which gives indicates that the first hypothesis which assumes that non-audit fees have a significant positive relation with audit fees is supported.

However, In terms of understanding this positive relation between the non-audit and audit fees, one of the explanations for this relationship is that "problem firms" as audit's clients may require a high level of the assuring services (both auditing and non-auditing services). Furthermore, in order to test that potential explanation, it is included the variables of ROA, the current ratio (ca/cl), as well as the ration of problem firms in the model of non-audit fees.

Table 2

Models of Audit Fees, Non-audit Fees and Controlling Variables

 Panel A: OLS Regression

 

 

 

 

 

 

 

 

Ln(AF) = a + b1*Ln(NAF)+ b2*Ln(TA) + b3* INVREC + b4* ROA

+ b5* TD/TA+ b6* CA/CL+ b7* BIG4+ b8* OPINION+ e

2005

2006

2007

2009

Independent Variables

Coefficient

p-value

Coefficient

p-value

Coefficient

p-value

Coefficient

p-value

Intercept

1.942**

0.003

2.032**

0.021

1.372

0.123

2.385**

0.002

Ln(NAF)

0.466**

0.000

0.534**

0.000

0.327**

0.000

0.223**

0.000

Ln(TA)

0.195**

0.000

0.237**

0.001

0.433

0.000

0.446**

0.000

INVREC

1.21

0.084

0.513

0.280

0.091

0.862

0.455

0.407

ROA

0.109

0.503

0.011

0.313

0.034**

0.001

Lady using a tablet
Lady using a tablet

This Essay is

a Student's Work

Lady Using Tablet

This essay has been submitted by a student. This is not an example of the work written by our professional essay writers.

Examples of our work

0.041**

0.010

TD/TA

-0.241

0.638

-0.990

0.062

-1.099

0.062

0.214

0.651

CA/CL

-0.001

0.962

-0.173**

0.033

-0.047

0.588

-0.28

0.687

BIG4

0.012

0.963

-0.130

0.715

-0.253

0.504

-0.728

0.088

OPINION

-0.151

0.409

-0.117

0.479

0.137

0.448

0.175

0.254

Adjusted R-sequared

68.40%

68.13%

63.26%

69.18%

F-Statistic

62.995**

0.000

60.78%

0.000

81.379**

0.000

71.841**

0

Numbe of observations

 

69

72

83

82

Panel B: Two-Stage Least Squares Regression

Ln(NÂF)= a + b1*Ln(TA) + b2* INVREC + b3* ROA + b4*TD/TA+ b6*CA/CL

+ b6* BIG4 + b7*OPINION + e

Ln(AF) = a + b1*Ln(NÂF) + b2*Ln(TA) + b3* INVREC + b4* ROA + b5*TD/TA

+ b7*CA/CL + b8*OPINION + e

2005

2006

2007

2009

Independent Variables

Coefficient

p-value

Coefficient

p-value

Coefficient

p-value

Coefficient

p-value

Intercept

1.802

0.576

1.408

0.367

3.709

0535

4.583

0.310

Ln(NÂF)

0.495

0.417

0.695

0.119

-0.362

0.817

-0.679

0.602

Ln(TA)

0.187

0.304

0.171

0.375

0.670

0.216

0.810

0.136

INVREC

1.131

0.566

0.604

0.267

0.836

0.661

1.843

0.424

ROA

0.012

0.689

0.010

0.398

0.034**

0.024

0.063

0.224

CA/CL

-0.001

0.989

-0.205

0.072

-0.239

0.588

-0.217

0.553

TD/TA

-0.223

0.707

-1.053

0.068

-0.897

0.391

-0.347

0.801

OPINION

-0.134

0.744

-0.105

0.540

0.059

0.856

0.243

0.537

Adjusted R-squared

0.683

0.663

0.140

0.678

F-Statistic

62.940**

0.000

59.23**

0.000

18.09**

0.000

-98.164**

0.005

Numbe of observations

 

69

72

83

82

Notes:

*, **, *** denote significant at the 10%, 5% and 1% levels; respectively ( two-tailed test).

Ln(NAF): Natural logarithm of non-audit fees; Ln(TA): Natural logarithm of total assets; INVREC: the ratio of the sum of inventories and accounting receivable to total assets; ROA: Ratio of return on assets; CA/CL: the current ratio (total current assets to total current liabilities); TD/TA: Ratio of financial leverage (total debt to total assets); OPINION: Companies having received a qualified or modified report (1), others (0).

The beta coefficients for the variables ROA, current ratio (CA/CL) and the financial leverage ratio (TD/TA) are not significant in any of the four years tested. This results cannot support the potential explanation that the positive relationship between non-audit and audit fees caused by the problem firms. Nevertheless, companies may familiar other problems that require a high level of the assurance services (audit and non-audit services).

Consistent with recent studies ( Antle el al., 2002; whisenant et al., 2003; Li et al., 2003; Hay et al., 2006) it is also examined the proposition that audit and non-audit fees are jointly determined. In the case that it occurs, therefore the significant positive coefficient of the variable of the Natural logarithm of non-audit fees Ln(NAF) in this study as well as in earlier studies may simply be due to misspecification of the model curried out. Panel B of Table 2 shows the key points of the models' results of two-stage least squares regression. However, in order to have sufficient instrumental variables to be used in estimating the new variable of Ln(NÂF), it was required to eliminate one explanatory variable from the audit fees' model, it is found that the most appropriate choice being the variable of Auditor ( BIG4), which appears to be more closely related to the variable of non-audit fees Ln(NAF) but not the variable of audit fees Ln(AF). The main findings are reported in Panel B of Table 2. In addition, as reported by whisenant et al., 2003; Li et al., 2003 and Hay et al., 2006, that non-audit fees don't have any significant relation with audit fees in the case of currying out two-stage least squares, suggesting that they are jointly determined.

As more ambitious test of this issue, it has been further investigated by adding the new variable of the ratio of non-audit fees to audit fees ( NAF/AF), in order to test whether knowledge spillovers or lower level of independence occur in the case when comparing the level non-audit fees NAF to be higher than in the audit fees AF. According the main findings of OLS regression model curried out, it has been found that the coefficient on this variable was negative and significant, which was an intriguing suggestion of knowledge spillovers or lower level of independence- on the other hand, the coefficient when two-stage least squares regression model was uses is not significant, which it can be concluded that the variables are jointly determined.

4.4 Key findings of testing the second hypothesis: Audit Fees and Audit Opinions

The key findings of the logistic regression model which testing the relationship between the non-audit fees and audit opinion as it reported in Table 4. It has found that the beta coefficient of the variable of interest, the ratio of non-audit fees to audit fees NAF/AF is negative although it is not significant in all years. The coefficient of the variable of natural logarithm of total assets is negative and marginally significant in 2005 which gives us an indicator that the larger companies are less likely to receive qualified or modified audit reports. Furthermore, it can be also seen from the table 4 that the variable of financial leverage TD/TA is significant and positively to the audit opinion over all years of the study period. This however concludes that audit's clients with higher leverage are more likely to receive audit qualification or modification. The coefficient of the variable of the ratio of the sum of inventories and accounting receivable to total assets (INVREC) is significant in the year of 2005. This could be conclude that this appears to be due to that the most of audit's clients that received qualified or modified audit reports are working in the financial and investment sector, as the property sector, the media and communications sector, or the leisure and tourism sectors that have relatively less inventories and accounting receivable.

In addition, it is also included in the model of audit opinion a measure variable of the dependence of the audit companies' earnings on the individual client (FEEDEP). This however based on the work of Firth (2002, p.676), who reports that:

"…..impairment of auditor independence may be more likely if the client audit fee and consultancy fee repr4esents a large proportion of the auditing firm's total fee income."

Table 4

Logistic Regression Model of Auditing Opinion, Non-audit Fees and Control Variables

 Panel A: OLS Regression

 

 

 

 

 

 

 

 

Auditor opinion(0,1) = a + b1*(NAF/AF)+ b2*Ln(TA) + b3* INVREC + b4* FEEDEP

+ b5* ROA + b6* TD/TA+ b7* BIG4+ + e

2005

2006

2007

2009

Independent Variables

Coefficient

p-value

Coefficient

p-value

Coefficient

p-value

Coefficient

p-value

NAF/AF

-0.896

0.459

-1.215

0.527

-1.163

0.264

-1.567

0.335

Ln(TA)

0.961*

0.081

0.882

0.489

1.438**

0.047

1.073

0.681

INVREC

0.373

0.604

0.226

0.288

0.940

0.963

3.655

0.134

FEEDEP

0.896

0.563

1.368

0.857

0.514

0.714

0.491

0.780

ROA

0.945

0.244

0.990

0.773

1.022

0.383

1.041

0.436

TD/TA

5.167*

0.055

0.109*

0.068

0.393**

0.054

0.751**

0.048

BIG4

0.667

0.636

0.764

0.787

0.950

0.959

0.957

0.975

Constant

Model Chi-square

4.98

0.545

2.99

0.885

7.30**

0.002

4.72**

0.000

-2 log likelihood

44.979

48.547

53.732

53.419

Numbe of observations

 

69

73

83

83

Notes:

*, **, *** denote significant at the 10%, 5% and 1% levels; respectively (two-tailed test).

NAF/AF: The ratio of non-audit fee to audit fee; Ln(TA): Natural logarithm of total assets; INVREC: the ratio of the sum of inventories and accounting receivable to total assets; FEEDEP: Free dependence measured as the Auditee's audit and non-audit payment divided by the total audit and non-audit fees of the external auditor; ROA: Ratio of return on assets; TD/TA: Ratio of financial leverage (total debt to total assets); BIG4: Big 4 auditors (1), others (0);

Following the work of Firth (2002), Li et al (2003) and Hay et al (2006), the total of the audit fees and non-audit fees for a particular audit client is divided by the sum of the audit and non-audit fees obtained from all of its clients in the top 100 listed companies in the index FTSE-100. The negative coefficient of the variable would point out that when the fee payment accounts of audit client for a high percentage of an auditor's total earnings, then the auditor is less likely to maintain independence and issue audit qualification of modification. Nevertheless, the coefficient of this variable (FEEDEP) in not significant in any of the years covered during the study period. Generally, the insignificant coefficient of the variable of the ratio of non-audit fees to audit fees (NAF/AF) was unsuccessful in order to support the second hypothesis which assumes that the level non-audit fees are negatively related with the audit qualification.

In addition, it has also tested the last stage of the second hypothesis, as it is estimated another OLS regression model of the non-audit fees as stated in the Table number 5. The key findings of this model suggest that there is a high positive significant relationship between audit fees and the level of non-audit fees, in every of the four years of the study period. It is also shown in this model that the qualified or audit opinion have insignificant positive relation to the audit fees, reporting that the auditors' firms are no less likely to give a qualified or modified audit opinion to audit clients from whom they receive a high level of the non-audit services. This concludes to the fact that these findings are also unsuccessful in order to support the second hypothesis.

OLS Regression Models of Non-audit Fees, Opinion and Controlling Variables

The OLS Regression model is:

 

 

 

 

 

 

 

 

Ln(NAF) = a + b1*Ln(TA)+ b2*AF + b3* OPINION + b4* BIG4

+ b5* MINING+ b6* FINANCE+ e

2005

2006

2007

2009

Independent Variables

Coefficient

p-value

Coefficient

p-value

Coefficient

p-value

Coefficient

p-value

Intercept

1.508 ***

0.017

5.237***

0.000

4.494***

0.002

3.037***

0.036

Ln(TA )

0.042

0.425

0.265**

0.003

0.183

0.206

0.198

0.155

Ln(AF)

0.781**

0.000

0.005**

0.020

0.008**

0.013

0.002**

0.007

OPINION

0.087

0.676

-0.125

0.589

-0.209

0.478

0.010

0.976

BIG4

1.303**

0.027

0.748**

0.031

0.551**

0.039

1.954**

0.015

MINING

0.029

0.932

0.355

0.320

-0.517

0.295

-0.986**

0.033

FINANCE

-0.343

0.198

-0.167

0.559

-0.097

0.849

-0.097

0.851

Adjusted R-sequared

71.913

0.531

0.553

0.598

F-Statistic

62.49**

0.000

43.222**

0.000

39.742**

0.000

69.079**

0.000

Numbe of observations

69

74

84

93

Table 5

Notes:

*, **, *** denote significant at the 10%, 5% and 1% levels; respectively (two-tailed test).

Ln(NAF): Natural logarithm of non-audit fees; Ln(TA): Natural logarithm of total assets; Ln(AF) : Natural logarithm of audit fees; OPINION: Companies having received a qualified or modified report (1), others (0); BIG4: Big 4 auditors (1), others (0); MINING : Mining and oil industry (1), others (0); FINANCE: Financial institutions (1), others (0).

4.5 Key findings of testing the third hypothesis: Audit Fees and Auditor Change

The main findings of multivariate analysis of the curried out OLS regression model of non-audit fees including audit tenure as reported in Table 6 shown below. The results suggest that the level of non-audit fees have insignificant relation with and the changing of auditors during the study period of the five years (except the year 2008 as previously mentioned). In addition, the results also suggest that the variables exhibiting a significant association with non-audit fees (NAF) are audit fees (AF), size of auditor (Big4 audit firms) as well as the audit client's size (TA). However, the outcomes of the model are not able to support the third hypothesis which assumes that the level of non-audit fees received by the audit firms has a significant positive association with the duration of auditor tenure.

Table 6

OLS Regression Models of Non-audit Fees, Auditor Change and Controlling Variables

The OLS Regression model is:

Ln(NAF) = a + b1*CHANGE + b2* Ln(TA) + b3* Ln(AF)+ b4* OPINION

+ b5* BIG4+ b6* MINING + b7*FINANCE + e

Independent Variables

Coefficient

p-value

Intercept

-4.391**

0.000

CHANGE

-0.139

0.406

Ln(TA )

0.082**

0.008

Ln(AF)

0.732

0.000

OPINION

-0.182

0.149

BIG4

6.105***

0.000

MINING

-0.161

0.411

FINANCE

0.067

0.700

Adjusted R-squared

41.65%

F-Statistic

41.704***

0.000

Number of observations

320

Notes:

*, **, *** denote significant at the 10%, 5% and 1% levels; respectively (two-tailed test).

CHANGE = Audit's clients change auditors during the research period (1), others (0); Ln(TA): Natural logarithm of total assets Ln(AF) : Natural logarithm of audit fees; OPINION: Companies having received a qualified or modified report (1), others (0); BIG4: Big 4 auditors (1), others (0); MINING :Mining and oil industry (1), others (0); FINANCE: Financial institutions (1), others (0).

Furthermore, in order for testing the third hypothesis further, it is run the model of logistic regression of auditor switch model, for the non-audit fees and the control variables across the four years of the study period. However, the main findings shown in Table 7 also suggest that the variable of interest, the ratio of non-audit fees to audit fees (NAF/AF) is insignificant. Of all the control variables, except the variables of the financial leverage (TD/ TA) which has a significant coefficient of ( ρ = 0.017). These results suggest that audit clients with a high proportion of the financial leverage are more likely to change the audit firm. These findings however could not also to support the third hypothesis which predicts that there is a positive association between the level of non-audit fees and the auditor tenure.

Table 7

Logistic Regression Models of Auditor Change, Non-audit Fees and Controlling Variables

The Logistic Regression model is:

CHANGE (0, 1) = a + b1*NAF/AF + b2* Ln(TA) + b3* OPINION + b4* BIG4

+ b5* ROA+ b6* CA/CL + b7*TD/TA + e

Independent Variables

Coefficient

p-value

NAF/AF

0.901

0.199

Ln(TA)

2.197

0.000

OPINION

0.849

0.605

BIG4)

0.091

0.628

ROA

1.057

0.170

CA/CL

1.023

0.706

TD/TA

0.554**

0.017

Constant

1.209

0.891

Model Chi-square

6.085

0.070

-2 log likelihood

129.133

Number of observations

305

Notes:

*, **, *** denote significant at the 10%, 5% and 1% levels; respectively (two-tailed test).

NAF/AF: The ratio of non-audit fee to audit fee; Ln(TA): Natural logarithm of total assets; OPINION: Companies having received a qualified or modified report (1), others (0); BIG4: Big 4 auditors (1), others (0); ROA: Ratio of return on assets; CA/CL: the current ratio (total current assets to total current liabilities); TD/TA: Ratio of financial leverage (total debt to total assets).

As mentioned earlier, the third investigation of testing the third hypothesis applies a matched pairs method in order to comparing the ratio of non-audit fees (NAF) to audit fees (AF) among audit clients firms that both for those whom changed the audit firm and did not during the study period. As it is reported in Table 7 which contains the key findings of the logistic regression model of auditor change, non-audit fees and control variables, The result however is insignificant as the coefficient of the variable NAF/AF is ( p= 0.199). In addition, this test also reports that there is no significant difference in non-audit fees between both switching as well as non-switching audit clients.

4.6 Discussion

This research study extends the prior research of investigation the issue of the joint provision of the assurance services (non-audit and audit services) into the UK market and empirically testes the effect of more provided non-audit services by auditors on the independence of the audit firms in the UK. The first question which has been asked that examining the association between the audit and non-audit fees in the UK for the four years from 2005 to 2009 (except the year 2008 as it is excluded due to the data unavailability as noted before). The second research question has been examined by the second hypothesis, investigating the effect of non-audit fees on the propensity of auditors in order to issue a qualified or modified audit report. The third hypothesis examined the research question that concerns whether if the stability of audit tenure is affected by the level of non-audit fees.

The key findings of the regression models suggest that the audit fees are positively related with the level of non-audit fees, after controlling for factors that have been found in order to influence audit fees only in the case of using the OLS regression model and not when implying the two-stage least square regression model. However, the main findings of further tests suggest that the view that both audit and non-audit fees are jointly determined is significantly supported, as the client size (measured by total assets) is the strongest determinant of each of them. However, this might be due to a combination of supply and demand of factors. As far as audit services are being concerned by public, large audit clients (measured by total assets) would have a high level of the number of transactions as well as balances in order to be audited; which can provide a higher level of audit risk to audit firm; and in these cases the auditor can gain more level than normal in providing such a high quality of audit services. In the same way, as far as non-audit services are also concerned, the larger audit clients would have a higher level of verity and complex structures; higher level of complex taxation issues; are more likely to rise finance; and indulge in mergers and acquisitions, while the non-audit services they required can be efficiently provided and delivered in these circumstances by the capable large auditors' firm that they deal with.

According the main findings of the logistic regression model with investigate the association between the level of non-audit services (measured by non-audit fees) and audit opinion, the results suggest that the qualified or modified audit reports have insignificant relationship with the level of non-audit fees. Thus, this does not indicate effect or lack of the perception of auditor independence.

Furthermore, in order to test the association between the level of non-audit fees and auditor change, three different regression models are run. The first one is the OLS regression model of non-audit fees, the second method is a logistic regression model, and the last method used is the test of matched pairs. The key findings of the three diffract regression models suggest that the level of non-audit fees provided by auditors have insignificant relation with the change of auditor during the research study period of the 100 largest UK companies. In the model of non-audit fees, there is no significant relation between the dummy variable for auditor change ( CHANGE) and either the variable of the level of non-audit fees ( NAF) or any of the other independence variables, compared with the results of the logistic regression model of Auditor Change, Non-audit Fees and Controlling Variables, which suggests that there is a marginally significant relation between the variable of auditor change and the variable of financial leverage, these results however suggest that audit tenure is less stable for those companies whom experiencing financial distress. The results of match pairs investigation fails to support any evidence that audit clients which switch have a lower level of non-audit fees.

J