Workload stress and conservatism: An audit perspective

Figures & data

Table 1. Descriptive statistics

Panel A quantitative variables
				Quantiles
Variable	N	Mean	S.D.	Min	0.25	Mdn	0.75	Max
CLIENT NUMBER	7,091	3.58	5.55	1.00	1.00	2.00	4.00	87.00
WLSIZE	7,091	0.00	1.00	−0.47	−0.47	−0.28	0.08	15.04
TENURE	7,091	11.28	8.93	1.00	4.00	8.00	16.00	38.00
AUDFEES	7,091	13.63	1.55	9.80	12.59	13.79	14.71	16.81
ACCRUAL	7,091	−0.50	2.23	−14.56	−0.19	−0.02	0.06	3.36
COMPLEX	7,091	6.06	2.71	0.04	4.20	6.25	8.01	11.27
LNMV	7,091	6.24	2.51	0.63	4.36	6.42	8.05	11.29
ROA	7,091	−0.43	1.96	−15.65	−0.17	0.03	0.09	0.39
LEV	7,091	0.62	2.66	−9.86	0.00	0.31	0.97	13.21
BM	7,091	0.36	1.11	−6.42	0.14	0.34	0.63	4.43
Panel B dichotomous variables
		0		1
		N	(%)	N	(%)
MAO	7,091	5,540	78.13	1,551	21.87
GCO	7,091	6,101	86.04	990	13.96
BIG4	7,091	2,462	34.72	4,629	65.28
SPECIALIZATION	7,091	6,036	85.12	1,055	14.88
Panel C audit tenure categories
		N	(%)
Short		1,245	17.56
Medium		2,304	32.49
Long		3,542	49.95
Total	7,091

This table shows the distribution of the categorical variables. Panel A presents the distribution of the dichotomous variables and Panel B displays the distribution of the audit tenure categories.

Table 2. Correlation analysis

Variables	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)	(9)	(10)	(11)	(12)	(13)	(14)
(1) MAO	1.000
(2) GCO	0.711*	1.000
(3) NCLIENT	0.225*	0.335*	1.000
(4) WLSIZE	0.233*	0.326*	0.754*	1.000
(5) BIG4	−0.275*	−0.396*	−0.343*	−0.294*	1.000
(6) DUM_TENURE	−0.102*	−0.131*	−0.112*	−0.088*	0.209*	1.000
(7) SPECIALIZATION	0.009	0.021	0.041*	−0.011	0.099*	−0.053*	1.000
(8) ACCRUAL	−0.059*	−0.063*	−0.010	−0.001	0.025*	−0.002	−0.080*	1.000
(9) COMPLEX	−0.361*	−0.553*	−0.431*	−0.355*	0.652*	0.259*	−0.049*	−0.005	1.000
(10) AUDFEES	−0.316*	−0.498*	−0.450*	−0.377*	0.662*	0.225*	−0.014	0.004	0.898*	1.000
(11) LNMV	−0.348*	−0.518*	−0.380*	−0.309*	0.639*	0.237*	0.003	−0.003	0.903*	0.836*	1.000
(12) ROA	−0.323*	−0.456*	−0.298*	−0.314*	0.271*	0.108*	−0.021	0.078*	0.435*	0.402*	0.331*	1.000
(13) LEV	−0.056*	−0.104*	−0.070*	−0.059*	0.118*	0.032*	−0.018	−0.012	0.186*	0.161*	0.144*	0.081*	1.000
(14) BM	−0.206*	−0.300*	−0.119*	−0.125*	0.116*	0.018	−0.026*	0.055*	0.225*	0.126*	0.133*	0.245*	0.145*	1.000

This table displays the Pearson correlation of all variables in this study. The sample is non-financial public companies listed on Compustat for the years 2013 to 2017. *** p < 0.01, ** p < 0.05, * p < 0.1.

Table 3. The effect of audit workload stress on audit conservatism

VARIABLES	(1) MAO	(2) MAO	(3) GCO	(4) GCO
Panel A probit analysis
NCLIENT	0.13***		0.20***
	(0.04)		(0.05)
WLSIZE		0.11***		0.11***
		(0.02)		(0.03)
AUDFEES	0.12***	0.12***	0.07	0.07
	(0.03)	(0.03)	(0.05)	(0.05)
DUM_TENURE	−0.05*	−0.05	−0.13***	−0.13***
	(0.03)	(0.03)	(0.04)	(0.04)
ACCRUAL	−0.03***	−0.03***	−0.04**	−0.04**
	(0.01)	(0.01)	(0.02)	(0.02)
COMPLEX	−0.05*	−0.05*	−0.21***	−0.21***
	(0.03)	(0.03)	(0.04)	(0.04)
LNMV	−0.14***	−0.14***	−0.32***	−0.32***
	(0.02)	(0.02)	(0.03)	(0.03)
ROA	−0.14***	−0.14***	−0.15***	−0.15***
	(0.02)	(0.02)	(0.03)	(0.03)
LEV	0.01*	0.01*	0.01	0.01
	(0.01)	(0.01)	(0.01)	(0.01)
BM	−0.13***	−0.13***	−0.16***	−0.16***
	(0.02)	(0.02)	(0.03)	(0.03)
Constant	−1.76***	−1.62***	−0.24	0.07
	(0.42)	(0.41)	(0.64)	(0.62)
Pseudo R2	0.19	0.20	0.50	0.50
Observations	7,083	7,083	7,079	7,079
Year and auditor fixed effects	Yes	Yes	Yes	Yes
Panel B marginal effect
NCLIENT	3.0%		2.3%
WLSIZE		2.5%		1.2%
AUDFEES	2.8%	2.9%	0.7%	0.8%
DUM_TENURE	−1.1%	−1.1%	−1.4%	−1.4%
ACCRUAL	−0.7%	−0.8%	−0.4%	−0.4%
COMPLEX	−1.1%	−1.2%	−2.3%	−2.4%
LNMV	−3.4%	−3.4%	−3.6%	−3.6%
ROA	−3.4%	−3.4%	−1.6%	−1.6%
LEV	0.3%	0.3%	0.2%	0.2%
BM	−3.0%	−3.0%	−1.8%	−1.8%

This table shows the results of probit analysis of equation (1)(1) $\begin{array}{rl}& Prob{\left(MAO/GCO\right)}_{i,t}={\beta }_0+{\beta }_{1}NCLIEN{T}_{i,t}+{\beta }_{2}AUDFEE{S}_{i,t}+{\beta }_{3}DUM\mathrm{\_}TENUR{E}_{i,t}+{\beta }_{4}ACCRUA{L}_{i,t}\\ & +{\beta }_{5}COMPLE{X}_{i,t}+{\beta }_{6}LNM{V}_{i,t}+{\beta }_{7}RO{A}_{i,t}+{\beta }_{8}LE{V}_{i,t}+{\beta }_{9}B{M}_{i,t}+{\gamma }_t+{\delta }_j+{\epsilon }_{i,t}\end{array}$(1) . Panel A shows the probit analysis results while Panel B reports the marginal effect. Robust standard errors are in parentheses. ***, **, and * show significance at 1%, 5%, and 10%, respectively.

Table 4. The effect of audit workload stress on the likelihood of Type I/II error

VARIABLES	(1) ERROR1	(2) ERROR1	(3) ERROR2	(4) ERROR2
Panel A probit analysis
NCLIENT	0.21***		−0.10
	(0.05)		(0.12)
WLSIZE		0.11***		−0.23
		(0.03)		(0.15)
AUDFEES	0.07	0.07	−0.03	−0.04
	(0.05)	(0.05)	(0.15)	(0.15)
DUM_TENURE	−0.12***	−0.13***	−0.04	−0.04
	(0.04)	(0.04)	(0.12)	(0.12)
ACCRUAL	−0.04**	−0.04**	0.05	0.05
	(0.02)	(0.01)	(0.03)	(0.04)
COMPLEX	−0.22***	−0.23***	0.13	0.13
	(0.04)	(0.04)	(0.09)	(0.09)
LNMV	−0.31***	−0.31***	−0.09	−0.08
	(0.03)	(0.03)	(0.07)	(0.08)
ROA	−0.14***	−0.14***	0.97**	0.96**
	(0.03)	(0.03)	(0.45)	(0.45)
LEV	0.01	0.01	−0.03	−0.03
	(0.01)	(0.01)	(0.03)	(0.03)
BM	−0.16***	−0.16***	−0.12***	−0.13**
	(0.03)	(0.03)	(0.05)	(0.05)
Constant	−0.31	0.00	−1.86	−1.95
	(0.64)	(0.62)	(1.70)	(1.69)
Pseudo R2	0.50	0.50	0.11	0.11
Observations	7,079	7,079	6,116	6,116
Year and auditor fixed effects	Yes	Yes	Yes	Yes
Panel B marginal effect
NCLIENT	2.3%		−0.1%
WLSIZE		1.2%		−0.2%
AUDFEES	0.8%	0.8%	0.0%	0.0%
DUM_TENURE	−1.4%	−1.4%	0.0%	0.0%
ACCRUAL	−0.4%	−0.4%	0.0%	0.0%
COMPLEX	−2.5%	−2.5%	0.1%	0.1%
LNMV	−3.4%	−3.4%	−0.1%	−0.1%
ROA	−1.6%	−1.6%	0.8%	0.8%
LEV	0.1%	0.1%	0.0%	0.0%
BM	−1.8%	−1.8%	−0.1%	−0.1%

This table shows the results of a probit analysis of equation (1)(1) $\begin{array}{rl}& Prob{\left(MAO/GCO\right)}_{i,t}={\beta }_0+{\beta }_{1}NCLIEN{T}_{i,t}+{\beta }_{2}AUDFEE{S}_{i,t}+{\beta }_{3}DUM\mathrm{\_}TENUR{E}_{i,t}+{\beta }_{4}ACCRUA{L}_{i,t}\\ & +{\beta }_{5}COMPLE{X}_{i,t}+{\beta }_{6}LNM{V}_{i,t}+{\beta }_{7}RO{A}_{i,t}+{\beta }_{8}LE{V}_{i,t}+{\beta }_{9}B{M}_{i,t}+{\gamma }_t+{\delta }_j+{\epsilon }_{i,t}\end{array}$(1) with Error1/Error2 as the dependent variable. Robust standard errors are in parentheses. ***, **, and * show significance at 1%, 5%, and 10%, respectively.

Table 5. The marginal effect of audit workload stress on audit conservatism by controlling client’ risks

VARIABLES	(1) MAO	(2) GCO	(3) MAO	(4) GCO	(5) MAO	(6) GCO	(7) MAO	(8) GCO
Panel A client’s financial distressed
	Distressed	Distressed	Distressed	Distressed	Non-Distressed	Non-Distressed	Non-Distressed	Non-Distressed
NClient	6.5%***	5.7%***			0. 1%	0.4%
	(0.06)	(0.06)			(0.06)	(0.10)
WLSize			3.7%***	3.3%***			1.3%**	0.2%
			(0.03)	(0.04)			(0.03)	(0.04)
Control	Included	Included	Included	Included	Included	Included	Included	Included
Pseudo R2	0.27	0.36	0.27	0.36	0.07	0.45	0.07	0.45
Observations	2,160	2,160	2,160	2,160	4,912	4,903	4,912	4,903
Year & auditor fixed effects	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Panel B client’s fraud probability
	High F-SCORE	High F-SCORE	High F-SCORE	High F-SCORE	Low F-SCORE	Low F-SCORE	Low F-SCORE	Low F-SCORE
NClient	5.5%***	3.9%***			1.6%	1.6%**
	(0.07)	(0.08)			(0.05)	(0.07)
WLSize			2.6%***	1.4%***			2.3%***	1.0%**
			(0.03)	(0.04)			(0.03)	(0.04)
Control	Included	Included	Included	Included	Included	Included	Included	Included
Pseudo R2	0.30	0.55	0.30	0.55	0.16	0.48	0.16	0.49
Observations	1,768	1,766	1,768	1,766	5,313	5,303	5,313	5,303
Year & auditor fixed effects	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes

This table shows the results of a probit analysis of equation (1)(1) $\begin{array}{rl}& Prob{\left(MAO/GCO\right)}_{i,t}={\beta }_0+{\beta }_{1}NCLIEN{T}_{i,t}+{\beta }_{2}AUDFEE{S}_{i,t}+{\beta }_{3}DUM\mathrm{\_}TENUR{E}_{i,t}+{\beta }_{4}ACCRUA{L}_{i,t}\\ & +{\beta }_{5}COMPLE{X}_{i,t}+{\beta }_{6}LNM{V}_{i,t}+{\beta }_{7}RO{A}_{i,t}+{\beta }_{8}LE{V}_{i,t}+{\beta }_{9}B{M}_{i,t}+{\gamma }_t+{\delta }_j+{\epsilon }_{i,t}\end{array}$(1) run in sub-samples. To simplify the interpretation, the marginal effects are reported in lieu of estimated coefficients. Control variables are included but not reported to abbreviate the table. Robust standard errors are in parentheses. ***, **, and * show significance at 1%, 5%, and 10%, respectively.

Table 6. The moderating effect of auditors’ resources

VARIABLES	(1) MAO	(2) MAO	(3) GCO	(4) GCO	(5) MAO	(6) MAO	(7) GCO	(8) GCO
Panel A audit firm size
	Big4	Big4	Big4	Big4	Non-Big4	Non-Big4	Non-Big4	Non-Big4
NClient	−1.3%		0.0%		5.7%***		5.0%***
	(0.07)		(0.13)		(0.05)		(0.06)
WLSize		−0.9%		0.3%		2.6%***		1.9%***
		(0.10)		(0.18)		(0.03)		(0.03)
Control	Included	Included	Included	Included	Included	Included	Included	Included
Pseudo R2	0.06	0.06	0.36	0.36	0.26	0.26	0.43	0.42
Observations	4,623	4,623	4,623	4,623	2,447	2,447	2,443	2,443
Year & auditor fixed effects	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Panel B audit specialization
	Specialist	Specialist	Specialist	Specialist	Non-Specialist	Non-Specialist	Non-Specialist	Non-Specialist
NClient	0.7%		−2.7%		3.0%***		2.6%***
	(0.12)		(0.15)		(0.04)		(0.06)
WLSize		1.3%		−1.5%*		2.5%***		1.2%***
		(0.09)		(0.07)		(0.02)		(0.03)
Control	Included	Included	Included	Included	Included	Included	Included	Included
Pseudo R2	0.21	0.21	0.49	0.49	0.20	0.20	0.52	0.52
Observations	1,049	1,049	1,049	1,049	6,012	6,012	6,026	6,026
Year & auditor fixed effects	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes

This table shows the results of a probit analysis of equation (1)(1) $\begin{array}{rl}& Prob{\left(MAO/GCO\right)}_{i,t}={\beta }_0+{\beta }_{1}NCLIEN{T}_{i,t}+{\beta }_{2}AUDFEE{S}_{i,t}+{\beta }_{3}DUM\mathrm{\_}TENUR{E}_{i,t}+{\beta }_{4}ACCRUA{L}_{i,t}\\ & +{\beta }_{5}COMPLE{X}_{i,t}+{\beta }_{6}LNM{V}_{i,t}+{\beta }_{7}RO{A}_{i,t}+{\beta }_{8}LE{V}_{i,t}+{\beta }_{9}B{M}_{i,t}+{\gamma }_t+{\delta }_j+{\epsilon }_{i,t}\end{array}$(1) run in sub-samples. To simplify the interpretation, the marginal effects are reported in lieu of estimated coefficients. Control variables are included but not reported to abbreviate the table. Robust standard errors are in parentheses. ***, **, and * show significance at 1%, 5%, and 10%, respectively.

Table

Variable	Definition
Dependent variables
GCO	1 if client i’s audit opinion is a going concern opinion in year t, zero otherwise
MAO	1 if client i’s audit opinion is not unqualified opinion in year t, zero otherwise
Independent variables
NCLIENT	Natural logarithm of audit client number per partner, count based on audit report date.
WLSIZE	Principal component analysis output of the number of audit clients at firm level and the total assets of the firm’s clients.
Control variables
ACCRUAL	Discretionary accruals of the modified Jones model by Dechow et al. (1995)
AUDFEES	Natural logarithm of 1 plus audit fees of audit firm i in year t
BIG4	1 if the audit firm i is Ernst & Young (EY), PricewaterhouseCoopers (PWC), KPMG, or Deloitte Touche Tohmatsu (Deloitte), zero otherwise
BM	Book value divided market value equity of client i in year t.
CLIENT NUMBER	The total client number of audit firm i at year t
COMPLEX	Natural logarithm of total assets of client i in year t.
DISTRESSED	1 if the client i has negative net income and negative operating cash flow at year t, zero otherwise (Non-Distressed)
ERROR1	1 if client i receives a GCO in year t but does not file for bankruptcy in year t + 1, zero otherwise
ERROR2	1 if client i does not receive a GCO in year t and files for bankruptcy in year t + 1, zero otherwise
F_SCORE	Fraud probability score based fraud prediction Model I in Dechow et al. (2011, p. 61)
High F_SCORE	F_SCORE is high when it is higher than the median
Low F_SCORE	F_SCORE is low when it is lower than the median
LEVERAGE	Total debt divided by total equity of client i in year t.
LNMV	Natural logarithm of market value equity of client i in year t.
ROA	Net income divided by total assets of client i in year t.
SPECIALIZATION	1 if the client’s industry is within the auditor’s industry specialization, zero otherwise. To define auditor industry specialization, I used the highest auditor’s industry market share in a given year.
TENURE	The length of audit firm–client relationship in years
DUM_TENURE	1 (short) if audit tenure is less than or equal to three years, 2 (medium) if audit tenure is longer than four years but less than eight years, and 3 (long) if audit tenure is longer than eight years

Dechow, P. M., Sloan, R. G., & Sweeney, A. P. (1995). Detecting earnings management. The Accounting Review, 70(2), 193–225. Retrieved from www.jstor.org/stable/248303.

 Web of Science ®Google Scholar

Dechow, P. M., Ge, W., Larson, C. R., & Sloan, R. G. (2011). Predicting material accounting misstatements. Contemporary Accounting Research, 28(1), 17–82. https://doi.org/10.1111/j.1911-3846.2010.01041.x

 Web of Science ®Google Scholar

Data availability

Data are available from the public sources cited in the text.