An empirical test of spatial competition in the audit market

advertisement
Competitive pressure,
audit quality and industry specialization
Wieteke Numan
(KU Leuven)
Marleen Willekens
(KU Leuven)
University of Oklahoma Price College of Business
Steed School of Accounting Conference
Regulation and the Audit Industry
18 May 2012
Motivation
o Audit market has high level of concentration, a given demand for
audits and high entry barriers → Regulators are concerned
o Former SEC Chairman Christopher Cox on the 2005 AICPA
National Conference:
“… within the accounting profession and within the SEC, we are forced to ask
ourselves: “Is this intense concentration in the market for large public
company auditing good for America?” If you believe, as I do, that genuine
competition is essential to the proper function of any market the answer is
no.”
Motivation
o U.S. Government Accountability Office, Audits of public
companies, 2008:
“Firms with significant market power have to potential to reduce the quality
of their products… however the presence of high market shares does not
necessarily mean anticompetitive behavior is occurring…
… competition in an oligopoly can also be intense and result in a market with
competitive prices, innovation and high-quality-products”
Motivation
o Green paper, European Commission, October 13, 2010:
“ The market appears to be too concentrated in certain segments and deny
clients sufficient choice when deciding on their auditors.
Moreover, being an auditor of large listed companies seems to create a
reputational endorsement”
o UK - House of Lords Inquiry July 2010 / report Economics
Affairs Committee March 2011
“….. explore concerns that market domination by a small number of firms
damages competition and reduces choice in the audit market as well as raising
questions about the quality of audited accounts…”
Intended contribution
o What are the effects of competition on audit quality?
o Is supplier concentration a good measure when
assessing competition in the audit market?
o Which economic theory fits the audit market and
competition within that market?
Key findings
o Concentration per se does not seem to affect audit
quality
o The pressure from a closely competing audit firm
does. Higher pressure is associated with lower quality
o The likelihood of issuing a going concern opinion and
the accruals based quality of earnings are affected by
industry expertise dominance over the closest
competitor.
Competition surely exists even in oligopolies….
The Herfindahl measure of competition
Market A (city-industry)
Auditor Market share
1
60%
2
10%
3
10%
4
10%
5
10%
Herf = 40%
Market B (city-industry)
Auditor Market share
1
30%
2
25%
3
20%
4
15%
5
10%
Herf = 22.5%
Which market is most competitive? And why?
Competitive Pressure?
- (Absolute difference between incumbent office’s market
share and the market share of the office that has the closest
market share to that of the incumbent office)
Market A (city-industry)
Market B (city-industry)
Auditor Mkt - Abs. diff
Auditor Mkt - Abs. diff
1
60%
-50%
1
30%
-5%
2
10%
0%
2
25%
-5%
3
10%
0%
3
20%
-5%
4
10%
0%
4
15%
-5%
5
10%
0%
5
10%
-5%
Herf = 0.40
Herf = 0.225
Problems with the Herfindahl
• The Herfindahl is derived theoretically from a model of
Cournot competition
– Cournot may not be a good description of the
matching process between audit firms and clients
• The conclusion that concentrated markets are less
competitive does not hold when:
– there are search costs (Stiglitz, JPE 1987)
– competition is on price rather than quantity
• Assumes that the level of competition is the same for
all firms within a given market
Concentration and quality
o Previous studies examine the relationship between audit quality and
market concentration, but find mixed results
o Kallapur et al. (2010) and Newton et al. (2011) – positive
association
o Francis et al. (2012) – international study
Two facets of concentration:
o B4 concentration level: positive association
o Within B4 herfindex: negative association
Prior empirical studies
audit quality and differentiation?
o Differentiation through industry specialization (market share)
o Lim and Tan (2008): no main effect
Only conditional on high joint supply of NAS
o Reichelt and Wang (2010): positive main effect
→ Effects of Industry specialization are studied,
but not the effect of competitive pressure?
Prior empirical studies about
competition in the audit market
o Based on theoretical front end in Simunic (1980)
= neo-classical economics:
o Perfect competition vs. monopolistic pricing
o No strategic elements included
o However, is this the appropriate model to study
competition in the audit market?
o Oligopolistic competition
o Product differentiation (horizontal /vertical)
Competition through differentiation
o Oligopolistic competition through differentiation
o The audit market characterized by a large number of
heterogeneous clients that pay unique fees: heterogeneity
of consumer tastes → horizontal differentiation
o Clients are willing to pay higher fees for audits
performed by different type of auditors → horizontal
and vertical product differentiation
Chan et al. (2004): location model
- auditors strategically choose to become ex post
heterogeneous through specialization
- audit firms modeled as Bertrand oligopolists who
simultaneously choose specializations w.r.t. client
characteristics
- empirically tested by merger Price Waterhouse
and Coopers & Lybrand in Australia
Competition and audit pricing
o Numan and Willekens (2012) show that on a city level, audit fees not
only increase in the level of industry expertise, but also in industry
expertise dominance over the closest competitor
→ First aspect of market location through differentiation
= Industry specialization as measure of auditor fit with client
.
→ Second aspect of market location through differentiation
= Distance/closeness of closest competitor
o Hence: competitive pressure (market share distance) from closest
competitor has negative (positive) effect on audit fees
Competition and quality?
- Industry experts can deliver higher quality as specialists earn fee
premiums that they can invest further in industry knowledge of the
client. In an oligopolistic competition setting, the pressure from close
(read: similar) competitors puts fee premiums under pressure, which
could lead to lower effort/quality.
- In a competitive market, an auditor is more likely to lose its clients to
competitors. The auditor has stronger incentives to please
management in order to keep the engagement.
+ Competitive pressure may force the auditor to distinguish himself
on other factors than price and results in innovation, for example to
avoid client switching (GAO report 2008)
The extent to which pressure from close competitor
affects audit quality is an open, empirical question
Design Model
13
Audit quality  ao  a1 competive pressure  a2 Industry expertise   ai Controls  
i 3
Degryse and Ongena (2005) and Numan and Willekens (2012):
Audit quality =
P (going concern)
Accruals-based earnings quality
Competitive pressure =
Industry market share ‘closeness’ to
closest competitor
Industry expertise =
Fit between the auditor and client in
terms of industry specialization
Controls =
i.e. Alternative competition
variables such as the Herfindahl Index
and Big4
Industry specialization measures
Riley and Neal (2004):
Market share approach: “How well has an audit firm differentiated
itself from its competitors in terms of market share within a
particular industry”
• Possible to designate specialists to industries that are too small
• Possible to inadequately recognize large, highly competitive
industries where most major accounting firms generate
significant revenues and devote significant resources to
develop industry audit technologies
Portfolio share approach: “Relative distribution of audit services
and related audit fees across the various industries for each audit
firm considered individually”
• Could be driven by size of the industry
Sample and data
o Audit market m: Audit firms serving a 2-digit SIC industry in a
Metropolitan Statistical Area (MSA /City level; Francis et al. 2005)
o Competitive pressure: ‘Closeness’ of closest competitor based on
industry market share
o Industry expertise: Both portfolio share measures and market share
measures (Neal and Riley 2004)
Going concern analysis
o Industry specialists have the expertise and independence to
more often issue a going concern opinion
(Lim and Tan 2008; Li 2009)
o Sample is restricted to financially distressed firms
(Lim and Tan 2008; Li 2009).
o Sample period 2005-2008
o Data sources: Compustat, Audit Analytics
o 4,134 firm-year observations; about 20% received
a going concern opinion
Descriptives Going-concern sample
N
Mean
StdDev
Min
P25
Median
P75
Max
4134
0.204
0.403
0.000
0.000
0.000
0.000
1.000
Competitive pressure
closest competitor
4134
-0.140
0.219
-0.955
-0.167
-0.034
-0.004
0.000
Industry expertise (office)
4134
0.318
0.341
0.004
0.056
0.170
0.467
1.000
Industry expertise (national)
4134
0.165
0.211
0.000
0.056
0.089
0.172
1.000
Industry leader (office)
4134
0.250
0.433
0.000
0.000
0.000
1.000
1.000
Industry leader (national)
4134
0.138
0.345
0.000
0.000
0.000
0.000
1.000
Industry specialist (office)
4134
0.183
0.387
0.000
0.000
0.000
0.000
1.000
Industry specialist (national)
Herfindex_msa
4134
4134
0.118
0.278
0.322
0.074
0.000
0.164
0.000
0.240
0.000
0.261
0.000
0.291
1.000
0.993
Total assets in million $
Size
Current ratio
Leverage
4134
4134
4134
4134
384
3.974
3.274
0.176
1147
2.119
3.723
0.304
0.050
-2.996
0.007
0.000
15
2.689
1.124
0.000
56
4.033
2.035
0.022
201
5.301
3.978
0.253
11550
9.354
29.396
2.745
Leverage change
Roa
Loss_lag1
Cfo_lag1
Sales_turn
MTB
Big4
Ln_tenure
4134
4134
4134
4134
4134
4134
4134
4134
0.023
-0.679
0.753
-0.325
0.885
2.615
0.483
1.669
0.189
1.962
0.432
0.944
0.863
8.803
0.500
0.792
-1.087
-31.676
0.000
-11.383
0.000
-55.072
0.000
0.000
-0.005
-0.567
1.000
-0.308
0.213
0.718
0.000
1.099
0.000
-0.208
1.000
-0.063
0.663
1.758
0.000
1.792
0.034
-0.049
1.000
0.039
1.293
3.842
1.000
2.197
1.333
0.299
1.000
0.420
4.820
80.351
1.000
3.401
Dependent variable
Going concern
Independent variables
Earnings quality analysis
o High quality auditors are more likely to detect questionable
accounting practices and misrepresentations
o Industry leadership / expertise is associated with higher earnings
quality (i.e. Reichelt and Wang 2010)
o Our measure of earnings quality: performance-adjusted
abnormal accruals based on a cross-sectional Jones (1991) model.
Earnings quality analysis
o Sample period 2005-2008
o Data sources: Compustat, Audit Analytics
o 7,071 firm-year observations and 2,640 unique clients
Descriptives abnormal accruals
sample
N
Mean
StdDev
Min
P25
Median
P75
Max
Dependent variable
Abs_abn_accruals
7071
0.111
0.119
0.000
0.037
0.079
0.144
1.253
Independent variables
Competitive pressure closest competitor
Industry expertise (office)
Industry expertise (national)
Industry leader (office)
Industry leader (national)
Industry specialist (office)
Industry specialist (national)
Herfindex_msa
Total assets in million $
Size
Std_cfo
Cfo
Leverage
Loss
MTB
Litigation
Current ratio
Roa
Sales_turn
Total_accruals_lag
Ln_tenure
Big4
7071
7071
7071
7071
7071
7071
7071
7071
7071
7071
7071
7071
7071
7071
7071
7071
7071
7071
7071
7071
7071
7071
-0.186
0.243
0.119
0.339
0.202
0.242
0.176
0.281
1444
5.423
0.111
-0.009
0.154
0.411
2.860
0.338
3.014
-0.102
1.026
-0.081
1.866
0.673
0.248
0.295
0.172
0.473
0.402
0.429
0.381
0.074
3702
2.048
0.183
0.282
0.210
0.492
4.209
0.473
2.617
0.386
0.740
0.184
0.838
0.469
-0.970
0.004
0.003
0.000
0.000
0.000
0.000
0.164
0.625
-0.470
0.008
-2.927
0.000
0.000
-24.011
0.000
0.082
-5.283
0.000
-2.553
0.000
0.000
-0.246
0.039
0.033
0.000
0.000
0.000
0.000
0.240
53
3.962
0.033
-0.024
0.000
0.000
1.258
0.000
1.401
-0.119
0.504
-0.094
1.386
0.000
-0.070
0.112
0.074
0.000
0.000
0.000
0.000
0.262
218
5.383
0.058
0.066
0.061
0.000
2.143
0.000
2.166
0.026
0.880
-0.046
1.946
1.000
-0.017
0.320
0.113
1.000
0.000
0.000
0.000
0.296
950
6.856
0.116
0.125
0.249
1.000
3.661
1.000
3.645
0.075
1.374
-0.012
2.485
1.000
0.000
1.000
1.000
1.000
1.000
1.000
1.000
0.993
30734
10.333
2.816
0.327
1.487
1.000
37.904
1.000
19.901
0.298
4.253
0.266
3.526
1.000
Dependent variable: P (going concern)
Expertise = Portfolio share
Parameter
Estimate
Prob
Estimate
Prob
Intercept
0.591
0.0805
0.668
0.0502
-0.506
0.0440
Competitive pressure
Industry expertise (office)
0.499
0.0094
0.450
0.0205
Herfindex_msa
-0.909
0.2286
-1.211
0.1169
Other control variables
Dependent variable: P (going concern)
Expertise = Leader
Parameter
Estimate
Prob
Estimate
Prob
Intercept
0.769
0.0198
0.816
0.0140
-0.479
0.0781
Competitive pressure
Industry expertise (office)
0.274
0.0853
0.153
0.3802
Herfindex_msa
-0.501
0.4880
-0.783
0.2925
Other control variables
Dependent variable: Abnormal accruals
Expertise = Portfolio share
Parameter
Estimate
Prob
Estimate
Prob
Intercept
0.135
0.0002
0.131
0.0003
0.022
0.0135
Competitive pressure
Industry expertise (office)
0.008
0.4316
0.012
0.2973
Herfindex_msa
-0.023
0.2783
-0.012
0.66111
Other control variables
Dependent variable: Abnormal accruals
Expertise = Leader
Parameter
Estimate
Prob
Estimate
Prob
Intercept
0.143
0.0000
0.138
0.0000
0.028
0.0007
Competitive pressure
Industry expertise (office)
0.002
0.5404
0.009
0.0632
Herfindex_msa
-0.014
0.5918
0.004
0.8873
Other control variables
Competition and audit quality
• Competitive pressure ∆↑ 10% P(going concern) ↓ from 12.9% to 12.1%
Earnings quality ↓ ∆ 2%
• Similar results Big 4 and non-Big 4 firms
• No effect of general concentration
Conclusions and interpretations
o The higher the pressure from a closely competing
audit firm, the lower the delivered audit quality of the
incumbent audit firm.
o The likelihood of issuing a going concern opinion or
the accruals based quality of earnings are affected by
industry expertise dominance over the closest
competitor.
Some limitations
o Endogeneity
o Measure of ‘fit’
dimensions of auditor differentiation
o Definition of market segments
o Vertical or horizontal differentiation?
Thank you for your comments and attention
Download