AAA07 Aug. 7, Chicago

1

• AAA07 Aug. 7, Chicago
• FMA, Oct.20, Orlando

2
Contact Information

• Mine Aksu or Kemal Saatcioglu
• Faculty of Management
• Sabanci University
• Orhanli, Tuzla
• 34956 Istanbul Turkey
• E-mail ksaatcioglu_at_sabanciuniv.edu
• maksu_at_sabanciuniv.edu
• Web http//people.sabanciuniv.edu/ksaatcioglu
• http//people.sabanciuniv.edu/
  maksu

3
Motivation

• Recent Corporate Gov. (CG) and financial
  reporting scandals
• Govt and institutional intervention to improve
  CG and disclosure (CG principles, IFRS IAS
  adoption, Audit committees)
• Voluntary disclosure efforts by firms (voluntary
  IFRS adoptions)
• CG and transparency disclosure (TD) ratings
• ? Increased importance of independent, hi quality
  audits

4
Motivation

• Prior research on link between auditor and
  reporting quality  
• auditor quality (higher fees, Big-4)
• ? higher reporting quality (transparent full
  disclosure)
• ? lower agency conflict (asymmetric info)
• ? higher firm performance (lower cost of capital)
• (RichardsonWelker, 2001 Mansi et al, 2007)

5
Motivation

• New developments in auditing
• Establishment of Audit Committees in boards
• Auditor switch every 5 years
• Separation of audit and consultancy services.
• New decree of CMB in June 2006 to adopt IAS
• New Turkish Trade Code requires independent
  audits in all firms
• TD as an important component and indicator of CG
  quality
• McKinsey 2002 survey ranked acct. disclosure top
  priority for Turkey
• TD a proxy for asymmetric info between minority
  and concentrated family shareholders.

6
Motivation

• Competition in the audit market
• AIA newsletter-Nov. 2007
• BDO (5th largest audit firm in the UK) research
  found
• ½ investors concerned if a switch to non-big 4
• 1/3 would review their investment decisions
• ¼ were unable to say why (lack of reputation
  cited the most)
• New BDO International CEO Jeremy Newman sees
  massive market opportunity in challenging the
  dominance of the Big-4

7
Objectives

• Investigate the association between auditor
  selection and client firm characteristics in the
  ISE
• Compare 2 sub-periods (1999-2001 and 2003-2004)
  to observe the impact of the financial reporting
  scandals and the regulatory and economic
  improvements in the aftermath
• Introduce and investigate the effect of a new
  proprietary Transparency and Disclosure (TD)
  measure that could alleviate the basic agency
  conflict in Turkey expropriation by concentrated
  family ownership.

8
Audit Quality

• AQ joint probability of detecting and
  reporting material financial statement errors
  (DeAngelo, 1981)
• Differing levels of audit quality (Simunic, 1980)
  demanded by firms based on their characteristics

9
Proxies for Audit Quality

• Auditor Size ( or sales of clients)
• More clients, duplicate tasks, more investment in
  audit technologies ? Larger audit firms are
  better able to and have the incentives to supply
  a higher audit quality (Francis Wilson, 1988
  De Fond Jiambalvo, 1993 Dye, 1993)
• Audit Fee Premium
• A fee premium exists in the big-6 ? an indication
  of product differentiation. The big-8 receive
  audit fees 15 - 20 greater (Francis, 1984
  Francis Simon, 1987)

10
Proxies for Audit Quality

• Brand Name (Reputation)
• firms develop/ maintain brand name reputations
  for quality to secure/protect rents arising from
  their brand name (Klein et al., 1978 Klein and
  Lefler, 1981, Francis and Wilson, 1988)
• The big-8 are considered to be brand-name audit
  firms that perform higher quality audits.
  (Simunic, 1980 Dopuch Simunic 1982, Smunic
  Stein, 1987 Defond, 92)
• ? We use brand name to categorize audit firms
  into 3 groups (Defond, 1992)
• The big-8? The Big-4
• Second-Tier, international or regional firms
• Local Turkish firms

11
Data and Sample

• 1999-2001
• All ISE firms with data (n 569 firm years)
  
• 2003-2004
• 52 firms with TD scores data
• 44 followed by SP/IFCG index
•  52 ISE firms
• 8 largest and most liquid

12
Data Sources

• Firm specific variables from annual reports and
  ISE website
• TD scores 2003-2004 annual reports and company
  websites searched for the inclusion of 106 info
  items in 3 dimensions
• Ownership structure and investor relations (32
  items)
• Financial transparency and info. disclosure (37
  items)
• Board Management structure and processes (37
  items)
• TD Rankings based on SP methodology
• Count Yes, No, N/A answers (yes 1 pt.) as
  a of maximum possible yes answers in each
  category

13
Design and Methodology

• Short panel data
• Pre and post test design Before and after the
  scandals and institutional interventions
• Multivariate Logit Analysis Binary dependent
  variable
• AUTYPE a ß1 FORSHA ß2 PUBSHA ß3
  FINSEC ß4 LEVER ß5 SIZE ß6 ROA ß7
  MRKBOOK ß8 TDS ß9 TDS PUBSHA

14
Variables Used

• AUTYPE 1, if auditor is Big Five 0, otherwise
• FORSHA of shares held by foreign shareholders
• PUBSHA of shares held by public
• FINSEC 1, if the client is in finance ind. 0,
  otherwise
• LEVER1 Leverage (debt/ debt equity)
• ASSET 1 Log of total assets of the client (Size
  proxy)
• ROA1 Net income/total assets
• MRKBOOK market-to-book ratio (growth proxy)
• TDS Firm-specific transparency and disclosure
  scores for 2003 and 2004 created by CG Forum of
  Turkey in collaboration with SP
• TDS PUBSHA Interaction variable for combined
  effect of public share ownership and transparency

15
Results Univariate Comparisons

• Big Audit Firms (Big 4) have higher of total
  clients with Foreign Ownership
• Mean of foreign shareholding is higher in Big4
• Mean of public shareholding higher in non-Big 4
• Higher of finance firms audited by Big 4
• Mean leverage higher in clients audited by
  non-Big 4
• Clients of Big 4 are larger
• Clients of Big 4 have higher ROA

16
Results Multivariate Logit

• 1999-2001
• Positive associations
• Foreign shareholding, finance industry, size
• Negative associations
• Public shares, leverage (not significant)
• 2003-2004
• Positive associations
• Size, Market-to-book, ROA, TD score, TD and
  public share ownership interaction
• Negative associations
• Public shares

17
Contributions

• The first study on auditor selection in ISE. 
• Incorporates the TD score variable as a proxy for
  info. risk to audit selection literature.
• Contributes an area of usefulness of TD scores
  for rating companies.
• The findings provide a benchmark for comparative
  research
• The findings would help audit firms make better
  market targeting decisions
• ? The findings would be of interest to audit
  committees, local/international investors,
  shareholders, creditors, and regulators.

18
Limitations

• Only one country and 5 years
• Sample size small in the 2003-2004 subsample
• Omitted variables ???
• Causality ???
• Ownership, transparency ? Auditor quality
• ?

19
Future Directions

• Repeat the study on ISE-100 for 2003-2004
• Examine the lead lag relationships
• e.g., do TD scores predict auditor choice
• Use a finer partitioning for audit quality
• Investigate changes
• e.g., change in TD or float rate ? auditor switch
• Try more interaction variables
• Compare the results with other countries.