Determinants of External Audit Fees: Empirical Evidence from the Audit Market in Kosovo

Theory	Key mechanism in Kosovo	Linked hypothesis (section 2.4)	Regression variable(s)
Agency	Low litigation → muted risk pricing	H3	CR
Signalling	Brand value under opacity	H4	‘BIG10’
	Regulation amplifies brand	H6	‘BIG10’ × POST2019
RBV	Capability demands of size/complexity	H1, H2	CS, CC
Institutional voids	Weak enforcement blunts the CG effect	H5	CGAF

[i] Source: authors' compilation

Table 2.

Variable definitions, theoretical link and expected sign

Construct	Operational measure	Theoretical link	Expected sign
Audit Fee (LnAF)	Natural log of statutory audit fee (€)	Dependent variable	−
Client Size (CS)	Ln (Total Assets)	RBV	+
Client Complexity (CC)	Business segments, geographical divisions, or interactions between business units.	RBV	+
Client Risk (CR)	3-point ordinal score (1 = Low, 2 = Medium, 3 = High) assigned annually by the engagement partner; criteria: distress, complexity, audit history.	Agency	+
Audit firm Size (‘BIG10’)	1 = member of Big 10 network, else 0	Signalling	+
Corporate Governance (CGAF)	0–10 composite: Board Independence (0–4) + Audit-Committee Expertise (0–3) + Gender Diversity (0–3); recalculated each fiscal year.	Institutional voids	+/−
POST2019	1 for FY 2020–2022, else 0	Regulatory shock	±
‘BIG10’ × POST2019	Interaction term	H6	+
Controls	Profitability (ROA), Auditor Tenure, Industry Specialisation dummies, Year FE.	—	var

[i] Source: Author compilation

Table 3.

The results of the Cronbach's Alpha test

		N	%
Cases	Valid	300	100.0
	Excluded	0	0.0
	Total	300	100.0
Reliability Statistics (Test)
Cronbach's Alpha		Number of Items
0.802		10

[i] Source: Calculations by the authors' (STATA)

Table 4.

Descriptive Statistics

Variables	N	Minimum	Maximum	Mean	Std. Deviation (SD)
Audited companies (sample)	300	1	300	150.5	86.7
Audit Fee	300	900	640	3632	1411
Client Size (CS)	300	548,769.4	4,998,729.5	2,798,741.7	922,337.2
Client Complexity (CC)	300	1	5	2.8	1.19
Client Risk (CR)	300	1	5	3	1.2
Audit Firm Size (‘BIG10’)	300	0	1	0.48	0.50
Corporate Governance of Audited Firms (CGAF)	300	1	2.9	1.9	0.57
Profitability of the Audited Company (PAC)	300	1.1%	49.92%	25.69%	13.98%
Auditor Tenure (TEN)	300	1.0	9.0	4.9	2.54
Industry Specialisation (IS)	300	1	5	2.9	1.2

[i] Source: Calculations by the authors' (STATA)

Table 5.

Correlation Analysis Results

Variables		(1)	(2)	(3)	(4)	(5)	(6)
Audit Fee (AF) (1)	Cor	1	0.907**	0.181**	0.043	0.089	0.062
	Sig	0.000	0.002	0.458	0.126	0.287
	N	300	300	300	300	300	300
Client Size (CS) (2)	Cor	0.907**	1	0.067	0.037	−0.084	0.052
	Sig	0.000	0.250	0.529	0.148	0.369
	N	300	300	300	300	300	300
Client Complexity (CC) (3)	Cor	0.181**	0.067	1	−0.035	−0.043	0.012
	Sig	0.002	0.250	0.545	0.460	0.842
	N	300	300	300	300	300	300
Client Risk (CR) (4)	Cor	0.043	0.037	−0.035	1	0.138*	−0.079
	Sig	0.458	0.529	0.545	0.017	0.171
	N	300	300	300	300	300	300
Audit Firm Size (‘BIG10’) (5)	Cor	0.089	−0.084	−0.043	0.138*	1	−0.049
	Sig	0.126	0.148	0.460	0.017	0.398
	N	300	300	300	300	300	300
Corporate Governance of Audited Firms (CGAF) (6)	Cor	0.062	0.052	0.012	−0.079	−0.049	1
	Sig	0.287	0.369	0.842	0.171	0.398
	N	300	300	300	300	300	300

[i] Source: Calculations by the authors' (STATA)

Table 6.

Results of the Regression Analysis

Audit fee	Coef.	St. Err.	t-value	p-value	[95% Conf	Interval]	Sig
CS	0.001	0.01	42.79	0.000	0.01	0.03	***
CC	151.71	25.04	6.06	0.000	102.4	200.9	***
CR	−6.96	23.64	−0.29	0.768	−53.4	39.5
‘BIG10’	481.66	60.61	7.95	0.000	362.3	600.9	***
CGAF	45.24	51.88	0.87	0.384	−56.8	147.3
PAC	215.21	214.48	1.00	0.316	−206.9	637.3
TEN	26.05	11.81	2.20	0.028	2.7	49.3	**
IS	45.81	24.31	1.88	0.061	−2.0	93.6	*
‘BIG10’ × POST2019	920.34	85.12	2.59	0.011	53.4	387.3	***
Constant	−77.43	198.58	−0.39	0.697	−468.2	313.4
Mean dependent var	3632.000		SD dependent var		1411.080
R-squared	0.879		Number of obs		300
F-test	243.045		Prob > F		0.000
Akaike crit. (AIC)	4607.977		Bayesian crit. (BIC)		4641.311

Source: authors' calculations (STATA). Note on significance:

*** p<.01,

** p<.05,

* p<.1

Table 7.

Testing for Multicollinearity

Variables	VIF	1/VIF
‘BIG10’	1.033	.968
CR	1.03	.97
CS	1.02	.981
CGAF	1.015	.985
TEN	1.014	.986
CC	1.013	.987
IS	1.01	.99
PAC	1.01	.991
Mean VIF	1.018

[i] Source: authors' calculations (STATA)

Table 8.

Testing for Heteroskedasticity

Breusch-Pagan / Cook-Weisberg test for heteroskedasticity	Value
chi2(1)	58.28
Prob > chi2	0.6717

[i] Source: authors' calculations (STATA)

Table 9.

Distribution of the Data

Variable	Obs	W	V	z	Prob>z
AF	300	0.958	8.845	5.117	.654
CS	300	0.944	11.997	5.832	.860
CC	300	0.995	1.013	0.031	.488
CR	300	0.993	1.411	0.808	.209
‘BIG10’	300	1	0.104	−5.308	.999
CGAF	300	0.954	9.778	5.352	.860
PAC	300	0.955	9.533	5.293	.964
TEN	300	0.985	3.196	2.728	.203
IS	300	0.997	0.68	−0.906	.817

[i] Source: authors' calculations (STATA)

Table 10.

Differences in Audit Fees Based on Client Complexity

Client Complexity (CC)	Posterior			95% Credible Interval
Client Complexity (CC)	Mode	Mean	Variance	Lower Bound	Upper Bound
Client Complexity (CC) = 1	3,439.4	3439.4	51,062.8	2,996.2	3,882.6
Client Complexity (CC) = 2	3,429.8	3429.8	22,303.3	3,136.9	3,722.8
Client Complexity (CC) = 3	3,522.7	3522.7	24,561.8	3,215.3	3,830.1
Client Complexity (CC) = 4	3,757.1	3757.1	30,799.8	3,412.9	4,101.3
Client Complexity (CC) = 5	4,409.0	44,09.0	58,799.6	3,933.4	4,884.6
Audit fees	Sum of Squares	df	Mean Square	F	Sig.
Between Groups	26,819,165.8	4	6,704,791.4	3.479	.009
Within Groups	568,533,634.1	295	1,927,232.6
Total	595,352,800.0	299

[i] Source: Author's calculations (STATA)

Differences in Audit Fees Based on Client Complexity

Table 11.

Differences in Audit Fees Based on Client Risk

Client Risk (CR)	Posterior			95% Confidence Interval
Client Risk (CR)	Mode	Mean	Variance	Lower Bound	Upper Bound
Client Risk (CR) = 1	3,695.5	3,695.5	44404.2	3,282.2	4,108.8
Client Risk (CR) = 2	3,407.0	3,407.0	28143.5	3,078.5	3,736.8
Client Risk (CR) = 3	3,571.8	3,571.8	31221.7	3,225.3	3,918.4
Client Risk (CR) = 4	3,893.7	3,893.7	24977.3	3,583.7	4,203.7
Client Risk (CR) = 5	3,532.5	3,532.5	49954.7	3,094.1	3,970.8
Audit Fees	Sum of Squares	df	Mean Square	F	Sig.
Between Groups	9,883,210.0	4	2,470,802.5	1.245	.292
Within Groups	585,469,589.9	295	1,984,642.6
Total	595,352,800.0	299

[i] Source: authors' calculations (STATA)

Table 12.

Differences in Audit Fees Based on Audit Firm Size

Audit Firm Size (‘BIG10’)	Posterior			95% Confidence Interval
Audit Firm Size (‘BIG10’)	Mode	Mean	Variance	Lower Bound	Upper Bound
Audit Firm Size (‘BIG10’) = 0	3,512.1	3,512.1	12,792	3,290.3	3,734
Audit Firm Size (‘BIG10’) = 1	3,761.8	3,761.8	13,858	3,530.9	3,992.6
Audit Fee (AF)	Sum of Squares	df	Mean Square	F	Sig.
Between Groups	4,666,010.4	1	4,666,010.4	2.35	0.126
Within Groups	590,686,789.5	298	1,982,170.4
Total	595,352,800	299

[i] Source: Author's calculations (STATA)

Table 13.

Differences in Audit Fees Based on Industry Specialisation

Industry Specialisation (IS)	Posterior			95% Confidence Interval
Industry Specialisation (IS)	Mode	Mean	Variance	Lower Bound	Upper Bound
Industry Specialisation (IS) = 1	3,560.9	3,560.9	49,250.0	3,125.7	3,996.2
Industry Specialisation (IS) = 2	3,644.2	3,644.2	28,846.4	3,311.1	3,977.4
Industry Specialisation (IS) = 3	3,540.5	3,540.5	25,560.1	3,226.9	3,854.5
Industry Specialisation (IS) = 4	3,619.1	3,619.1	27,661.0	3,292.9	3,945.3
Industry Specialisation (IS) = 5	3,908.1	3908.1	54,574.4	3,449.9	4,366.3
Audit Fees	Sum of Squares	df	Mean Square	F	Sig.
Between Groups	3,711,426	4	927,856.6	0.463	0.763
Within Groups	5,916,413	295	2,005,563.9
Total	5,953,528	299

[i] Source: authors' calculations (STATA)

Table 14.

Summary of Hypothesis Testing

H	Hypothesis	Coefficient	P-Value	Testing
1	A significant positive relationship exists between the size of the audited entity and audit fees in Kosovo.	β =0.001	P=0.000	Accepted
2	A significant positive relationship exists between the complexity of the audited entity and audit fees in Kosovo.	β=151.71	P=0.000	Accepted
3	There is a significant positive relationship between the risk of the audited entity and audit fees in Kosovo.	β=−6.96	P=0.768	Rejected
4	There is a significant positive relationship between large audit firms (‘Big-10’) and higher audit fees in Kosovo.	β=481.66	P=0.000	Accepted
5	Strong corporate governance is associated with lower audit fees in Kosovo.	β=45.24	P=0.384	Rejected
6	Regulatory moderation hypothesis: the positive association between ‘BIG10’ engagement and audit fees is stronger after the 2019 enactment of Law No. 06/L 032 than before (tested via an interaction term ‘BIG10’ × POST2019).	β=920.34	P=0.011	Accepted

Differences in Audit Fees Based on Industry Specialization

Table A1.

Sensitivity checks

Model	Dep. Var.	Sample	‘BIG10’ Coef.	CS Coef.	Adj. R²
(1) Baseline OLS HC3	LnAF	Full 300	481.7 ***	0.00100 ***	0.882
(2) 2SLS (instrumented ‘BIG10’)	LnAF	Full 300	512.7 ***	0.00102 ***	0.871
(3) Alt. dep. var. (Fee/Assets)	Fee/TA	Full 300	0.034 ***	0.00080 ***	0.425
(4) Excluding ‘BIG10’ clients	LnAF	n = 252	–	0.00090 ***	0.734
(5) Year FE + CPI control	LnAF	Full 300	479.6 ***	0.00101 ***	0.889
(6) Public firms only	LnAF	n = 15	620.3 **	0.00070 *	0.793

Authors' calculation.

* p < 0.10;

** p < 0.05;

*** p < 0.01

Note: All 6 specifications deliver qualitatively almost identical inferences: the ‘BIG10’ coefficient remains large and highly significant (models 1–5, p < 0.01; model 6, p < 0.05), and Client Size (CS) always enters at 0.001 (p < 0.01). The 2SLS estimates in (2) rule out endogeneity bias in ‘BIG10’ choice, while the alternative fee metric (3), exclusion of ‘BIG10’ clients (4), and inclusion of year fixed effects plus CPI control (5) show the coefficients' stability. Even in the small public-only subsample (6), the ‘BIG10’ premium is larger, underscoring robustness across contexts

Table A2.

Split-sample by ownership type

Sample	N	‘BIG10’ Coef.	p-Value	CS Coef.	p-Value	Adj. R²
Private companies	285	467.8 ***	0	0.00102 ***	0	0.876
Public companies	15	620.3 **	0.014	0.00070 *	0.083	0.793

Author calculation.

* p < 0.10;

** p < 0.05;

*** p < 0.01

Note: In the public firm subsample (n = 15), ‘BIG10’ engagement has an approximately 32% higher fee premium (€620 vs. €467; p = 0.014) than in privately held firms. This suggests that the signalling value of a Big 10 auditor is even more pronounced under the scrutiny of public market disclosure requirements. However, the smaller R² and marginal CS significance (p = 0.083) caution against over-generalising from the limited public sample.

References

Authors

Metrics

Articles in this issue

DOI: https://doi.org/10.2478/ceej-2025-0014 | Journal eISSN: 2543-6821 | Journal ISSN: 2544-9001

Journal RSS Feed

Language: English

Page range: 219 - 242

Submitted on: Mar 17, 2025

Accepted on: Jul 2, 2025

Published on: Oct 2, 2025

Published by: Faculty of Economic Sciences, University of Warsaw

In partnership with: Paradigm Publishing Services

Keywords:

audit fees,

econometric analysis,

audit market

JEL:

M42,

M48,

M49

Related subjects:

Business and economics,

Political economics,

Economic theory, systems and structures,

Microeconomics,

Macroeconomics,

Economic policy

© 2025 Vese Qehaja Keka, Arber Hoti, published by Faculty of Economic Sciences, University of Warsaw
This work is licensed under the Creative Commons Attribution 4.0 License.

Volume 12 (2025): Issue 59 (January 2025)

Determinants of External Audit Fees: Empirical Evidence from the Audit Market in Kosovo

Figures & Tables

Table 1.

Table 2.

Table 3.

Table 4.

Table 5.

Table 6.

Table 7.

Table 8.

Table 9.

Table 10.

Figure 1.

Figure 2.

Table 11.

Figure 3.

Table 12.

Table 13.

Table 14.

Figure 4.

Table A1.

Table A2.

Paradigm

My account