1. Jamal & Sunder, Cerfication Society Regulation, Competition and Independence in a Certification Society: Financial Reports vs. Baseball Cards Karim Jamal and Shyam Sunder Research Institute for Economics and Business Administration, Kobe University, Kobe, JAPAN January 25, 2008

2. Jamal & Sunder, Cerfication Society Summary Economic consequences of mandatory audit remain controversial Cross market analyses can yield limited but useful insights Exploration of audit certification in the broader context of certification activities in the economy Extent and nature of certification services –Ubiquity of certification and conflicts of interest –Fineness of certification reports (2-100) Field study of unregulated market for certification of baseball cards –Twenty three Certification firms –Market values Cross Sellers and strict grading service yields higher returns to its buyers (no race to the bottom) Implications for auditing: How important are independence and absence of conflict of interest in certification?

3. Jamal & Sunder, Cerfication Society Regulatory Attempts to Improve Auditing Seven decades of regulation under the SEC – regulation and audit failures Major changes in regulation in 2002 (SOX) to promote auditor independence (and Quality?) Controversy over whether these reforms improve auditing (Kinney et al., 2004)

4. Jamal & Sunder, Cerfication Society Extent of Certification Services In 1996, US Government documented 93,000 national standards developed by 80 government and 604 private agencies (Jamal and Sunder 2006) National and international standards cover virtually all aspects of the economy (Jamal and Sunder 2006) To what extent are standards accompanied by availability of services to certify compliance?

5. Jamal & Sunder, Cerfication Society The Audit Society “Audit explosion” in society (Power 1994): education, healthcare, government supported activities Demand for audit fueled by political demands for accountability and control (Power 1999) U.S. audit firms attempted to expand and re-label their “assurance” services into other sectors of the economy in the 1990s (Elliott 1998) Retail customer as the new client for services Audit firms failed in attempt to expand into e-commerce assurance market: could not compete with the business models of BBB Online and TRUSTe (Jamal, Maier and Sunder 2003) Attempt to empirically assess the extent of availability of certification services in private sector

6. Jamal & Sunder, Cerfication Society Sample Selection Selected a sample of 817 items sold online and offline during June 12-July 25, 2004 –400 items from eBay.com (online) –358 items from BLS Producer Price Index –59 items from BLS Consumer Price Index

7. Jamal & Sunder, Cerfication Society Availability of Certification Services by Type Expert opinion about compliance with formal written standards (91%) Expert opinion in absence of formal standards (7%) Ratings given by lay people (no standards) 1% Popularity/activity meters (best-seller lists for books, music, films, TV, etc.) 0.5% Total: 99.5% have some kind of certificate All four kinds of certification available for 40%

8. Jamal & Sunder, Cerfication Society Table 1: Certification Services for Products Sold Online and Offline in the US PANEL AEBay (N=400) PPI/CPI (N=417) Total (N= 817) % of Total (817) Expert Opinions Based on Written Standards 34439974391% Expert Opinions Without Written Standards 4514597% Lay People Ratings 5381% Meter 3030.5% No Certification / Rating 3140.5% Panel B: Detailed Breakdown of Certification Activities All Certification Available (Expert Opinions Based on Written Standards, Expert Opinions Without Written Standards, Lay People Ratings, and Meters) 22110132240% Expert Opinions Based on Written Standards only 411011414% Expert Opinions Based on Written Standards, Lay People Ratings and Meter 33579011% Expert Opinions Based on Written Standards and Expert Opinions Without Written Standards 47348110% Expert Opinions Based on Written Standards, Expert Opinions Without Written Standards, and Lay People Ratings 1734516% Expert Opinions Based on Written Standards, and Meter 828364% Expert Opinions Based on Written Standards, Expert Opinions Without Written Standards, and Meter 1115263% Expert Opinions Based on Written Standards, and Lay People Ratings 320233% All other combinations 5618749%

9. Jamal & Sunder, Cerfication Society Extent of Certification Results support for Power’s characterization of ours as a “audit society” AICPA failed in its attempt to fill what they thought were empty spaces in the market for assurance services Power (1994) studies the public sector, and attributed “audit explosion” to political forces But certification is also ubiquitous for private goods Demand for certification for private goods likely to be driven by broader economic forces

10. Jamal & Sunder, Cerfication Society Data Obtained from One Official Partner of eBay (XYZ) 128 experts Offer opinions on 1,850 separate items $9.95 fee for basic service $29.95 for enhanced service Experts do not follow any written standards Opinions do not reference any written standards Provides descriptions of experts (accreditation by professional institutes, education, related business, relevant experience): two examples

11. Jamal & Sunder, Cerfication Society Self-Assessment of XYZ’s Experts 128 experts as of July 12, 2004 39% had formal accreditation from a professional body 25% formal relevant educational credentials 79% ran a business in the products for which they provided certification Only 9% reported themselves to be hobbyists

12. Jamal & Sunder, Cerfication Society

13. Implications for Auditing Possibility of conflict of interest (from trade in the relevant good or service) appears to be pervasive (norm) in certification services in general (79%); it is not confined to auditing Unacceptability of advisory services to audit clients in a regulated market (Francis 2004) contrasts sharply with this data from unregulated domains

14. Jamal & Sunder, Cerfication Society Fineness of Ratings/Certification Auditor reports are essentially pass/fail (with some additional qualification) Auditor gains a broad understanding of the quality of internal controls, governance, accounting policies, estimates, and disclosure, and future prospects of business The coarse pass/fail grading does not attempt to convey this additional information to the client Would a “finer” report be more informative to users?

15. Jamal & Sunder, Cerfication Society Available Model Dubey and Geanakoplos (2005): In general, an optimal grading scheme should have an intermediate level of fineness –Coarse reports have less information and do not motivate the reporting agent to exert more effort –Fine reports magnify the consequences of measurement errors –Intermediate grading scale may be optimal Grading scheme should create a small elite grade Absolute grading schemes dominate relative

16. Jamal & Sunder, Cerfication Society Data on the Fineness of Government Standards Visited websites of 80 departments of the federal government that set standards (Toth 1996) Were able to access standards of 64 (80%) of these agencies 53/64 (83%) set minimum requirements (pass/fail only) 11/64 (17%) use finer grading (e.g., AAA beef) DG: Most federal agencies use absolute standards, but do not set elite grades or use intermediate fineness scales (FDA is an exception)

17. Jamal & Sunder, Cerfication Society Greater Variability in Fineness of Scales in Private Standards Pass/fail (e.g., UL) Multiple seals (TRUSTe for e-commerce privacy 10-100 point rating scales (baseball cards, Consumer Reports on cars) Often characterized by competition among rival standards and scales Rating scales as a dimension of competition among services (Jin, Cato and List 2004)

18. Jamal & Sunder, Cerfication Society Market for Baseball Card Certification Services Jin, Cato and List (2004) –Dominant player is PSA: 1991, 10-point scale with 1 point increments, not graded on curve, 10% cards get 10 (no super elite grade) –BGS: 1999, 10-point scale with ½ point inc., numerical and qualitative labels (mint), sub-grades for corners, centering, surface, edges; 0.1% cards get 10, publishes grade distribution on-line; three brands of service (BVG, BGS, and BCCG) –SGC: 1999, smaller market share, 100-point scale (confusion, conversion table to 10-point scale) –20 other smaller players (Table 3), largely consistent with DG predictions (no curve, intermediate fineness, super elite)

19. Jamal & Sunder, Cerfication Society Table 3: Sports Card Grading Services (Summer 2004) NameYear Foun ded Scale Catego ries Gradin g Cost Guar antee Cro ss- sell er Feature 1. Accugrade (ASA) 198813$5-15No Invented 10-point scale + First online rating agency 2. Professional Sports Authentication (PSA) 199110$9-50Yes Membership Fee ($99 – get 6 cards graded free) + Largest Market Share 3. KSA 199614$12- 19 No Canadian 4. American Authentication (AAI) 199610$12- 22 No 5 card minimum per order 5. Finest Grading (FGS) 199714$5-25NoYes Value Pricing Card Shows 6. Map Industries 199814$8No Free Shipping 7. Beckett Grading Service (BGS) –Vintage Grading (BVG) –Collectors Club (BCCG) 199919$8-25 $9-26 $5 Yes Price guide + provide sub- grades for centering, corners, edges, and surface 8. Sportscard Guaranty (SGC) 199918$7- 50NoYes Started with 100- point scale 9. The Final Authority (TFA) 199919$5-13YesNo Value Pricing 10. Collection Monster (CM) 199919$10NoYes Government Grading

20. Jamal & Sunder, Cerfication Society Comparison with the Market for Audit Services Only binary pass/fail reports, little further detail permitted If the purpose of SOX is to convey information about internal controls, will switch from pass/fail to graded report on IC, governance, quality of accounting estimates and accounting policy choices give better value to investors? The current system suppresses the detailed knowledge gained by the auditor; more consistent with govt. not private markets

21. Jamal & Sunder, Cerfication Society Value of Certification in Sports Card Market Prospective sellers of sports cards can hire a certification agency to grade and certify their merchandise before selling Independence and audit quality are difficult to observe in the audit market Baseball card market includes pure and multi-service certifiers, allows a cleaner measure of “audit quality” in grading strictness, observation of an unregulated market and dimensions of competition (e.g., value pricing, computer grading, multiple experts grading the same card, letting customers choose their own grade) Anxiety about race-to-the-bottom and independence in accounting (Dye and Sunder 2001)

22. Jamal & Sunder, Cerfication Society Data for Assessing the Value of Certification Services 321,045 cards traded on eBay during August 19- September 3, 2004 Partitioned cards by decades of origin, single- rookie, graded/ungraded (Table 4) Randomly selected 1,000 rookie and 1,000 singles cards from graded cards Market shares of six major firms (Table 5) in the sample of graded cards (PSA 78 in singles, 39 rookie; BGS 6 singles, 34 rookie) Tables 5A, B

23. Jamal & Sunder, Cerfication Society Table 4: Baseball Card Certification Services (Baseball Cards Traded on eBay, N=321,045) Singles (272,399 cards) Rookies (48,646 cards) Issue Date Grade d UnGrade d Total%Grade d Grade d UnGrade d Tota l %Grade d Pre1930s1,9352,9334,86840%---- 1930s1,3761,6893,06545%---- 1940s 4171,4481,86522%202182388% 1950s6,16022,86929,02921%596066659% 1960s7,29128,00735,29821%16175091118% 1970s6,85814,17121,02933%3031,2211,52 4 20% 1980s3,6928,53312,22530%5,4706,37811,8 48 46% 1990s2,69325,16327,85610%3,6699,57813,2 47 28% 2000s1,356135,808137,16 4 1%2,60817,60520,2 13 13% Total31,778240,621272,39 9 11%12,29036,35648,6 46 25%

24. Jamal & Sunder, Cerfication Society Assessing Strictness of Grading Jin, Cato and List (2005) gave the same 212 cards to PSA, BGS and SGC online and to three offline dealers Average scores: 8.5 for BGS and two dealers; 8.7 for PSA and one dealer, and 8.9 for SGC (BGS tighter cut-offs, precise); Table 5C Tables 5A and B: empirical frequencies in our sample (self- selection bias): BVG’s premium service stands out (GAI, SGC) Becket’s value brand (BCCG) is lower priced, also has relaxed standards GEM gives the top grade of 10 to 73% percent of the singles cards In rookie Table 5B: roughly similar results Is the market able to adjust itself for differences in grading?

25. Jamal & Sunder, Cerfication Society

26. Table 5A: Frequency of Grades Given by 3rd Party Certification Services (For a Sample of 1,000 Graded Single Baseball Cards) Grade 3 rd Party Grader Market Share % 109.59.08.58.07.57.06.05.0<5<5 TotalAverage Score PSA78.169 114902830129613356781 7.5 BGS 3.2 011 107 10 0102 32 8.5 BCCG 1.8 9 0 80 00 0100 18 9.3 BVG 1.1 0 0 20 00 0324 11 5.1 GEM 1.1 8 0 00 30 0000 11 9.5 GEM-E 0.3 1 2 00 00 0000 3 9.7 SGC 5.3 2 0 32 74 68129 53 6.1 GAI 4.6 0 1 33 123 8547 46 6.6 PGS 1.5 0 0 62 60 010015 8.3 Other 3.013 3 40 22 311130 8.7 Total100102181851431491468152791000

27. Jamal & Sunder, Cerfication Society Table 5A: Frequency of Grades Given by 3rd Party Certification Services (For a Sample of 1,000 Graded Rookie Baseball Cards) Grade 3 rd Party Grader Market Share % 109.59.08.58.07.57.06.05.0< 5TotalAverage Score PSA 36.8105 0159 080 017421368 8.9 BGS 22.7 185 73371314 1111227 8.9 BCCG 10.4 87 0 16 0 0 0 0100104 9.8 BVG 0.9 0 0 0 2 0 1 0123 9 5.3 GEM 11.6116 0 0 0 0 0 0000 10.0 GEM-E 4.2 42 0 0 0 0 0 0000 10.0 SGC 2.9 4 0 13 6 4 1 100029 8.8 GAI 2.3 2 9 8 2 1 0 010023 9.1 PGS 1.6 16 0 0 0 0 0 0000 10.0 Other 6.6 33 2 208 2 0 001066 9.4 Total100%406962895510016198651000

28. Jamal & Sunder, Cerfication Society Pricing and Value of Baseball Card Certification Services PSA charged a $99 membership fee (dropped recently) Table 3: Price varies ($2-50), extra for faster turnaround For each graded card in the sample, we located an un- graded card matched by player, maker, year of issue, single-rookie For unrated cards, we gathered price estimates from Beckett Baseball Card Monthly Guide (August 2004, Issue #234 online) For rated cards, we recorded player, year, maker, grader, grade, buyer reputation, seller reputation, number of bids, and selling price Table 6: Gross and net returns on baseball card grading services provided by six major service providers

29. Jamal & Sunder, Cerfication Society Table 6: Average Returns to Grading of Rookie Cards (by Grader and Decade of Issue) RaterNumber of Cards Gross Return % Net Return % 2000+ Gr. =10 % 1990’s Gr.=10 % 1980’s Gr.= 10 % 2000+ Gr.=9 % 1990’s Gr.= 9 % 1980’s Gr.=9 % 1970’s Gr.= 9 % BGS 22726446333----- 4486176----- GAI 2319892185----- 55124128----- SGC 29125 4303-----201(27)(65)(1)----- PSA 368238 (2 )2547178(50)(53)(5)683 BCCG 104112(14)(15)(11)(3)(68)(72)(25)20 GEM 11680.6(23)154(38)(24)-----

30. Jamal & Sunder, Cerfication Society Returns to Certification RaterGross ReturnNet Return BGS264%46% GAI198%92% SGC125%4% PSA238%(2%) BCCG112%(14%) GEM80.6%(23%)

31. Jamal & Sunder, Cerfication Society Table 7: Regression Analysis Ri =  1 +  1 Year i +  2 Grade i +  3 GPSA i +  4 GBGS i +  5 GBeckett i +  6 GGAI i +  7 GGEM i +  8 GSGC i + e i Regression (990)df, R 2 = 0.1747, p < 0.001. (Adjusted R 2 = 0.168) CoefficientStd ErrorT Statistic Intercept108.6812.60 8.63 *** Year -0.058 0.006 -8.97 *** Grade 0.72 0.06 12.65 *** GPSA 0.37 0.16 2.22** GBGS 1.12 0.18 6.25*** GECKETT -0.18 0.19 -0.98 GGAI 1.21 0.31 3.87 *** GGEM -0.73 0.18 -3.97*** GSGC 0.48 0.28 1.68* *** p <0.001 ** p < 0.01 * p < 0.05

32. Jamal & Sunder, Cerfication Society Regression Analysis Table 7: Year, grade and identity of grader have a significant effect on net return from the certification service Certification service is more valuable for older cards (selection effect?) The returns are ranked by strictness of grading scales, negative for the lenient graders Easy grade of 10 from GEM does no good for the value of the card No race to the bottom Beckett able to provide different brands of service, and market is able to differentiate among them without getting confused

33. Jamal & Sunder, Cerfication Society Problems of Cross-Market Inference Baseball cards are physical objects of substantive interest; financial reports are symbolic representations of entities which are of substantive interest In both cases, willingness to pay for certifiers’ expertise Buyers rely on knowledge and reputation Certifiers compete for fee revenues Certifiers choose portfolio of services to offer

34. Jamal & Sunder, Cerfication Society Back to Auditing Regulatory objectives of auditing (quality of certification, independence, etc.) In baseball card market 17/23 offer only certification services but six (including all major) players offer related services (pricing guides, dealers, magazines, shows) as well as handle other collectibles Cross-sellers dominate the market and are able to collect premium price for their service, yielding positive net returns to their clients Independence appears to be neither necessary nor sufficient condition for high quality service in this certification market Why?

35. Jamal & Sunder, Cerfication Society Skepticism about Deregulation Will a deregulated market for auditing function properly (race to the bottom, Dye and Sunder)? Industry expert auditors command higher fees and restrain abnormal accruals (Francis et al. 2005, 2006) Recent policy changes (PCAOB) have gone in the other direction Lenient graders generate negative returns to their clients Jamal, Maier and Sunder (2003) on market for privacy seals in e-commerce dominated by higher quality Does cross-selling of services promote or inhibit grade inflation? Antle et al. 1997; Dopuch 7 King 1991, Abdel- Khalik 1990; Simunic 1984; Zhang 2005; Antle & Demski 1991. What are the real reasons for poor enforcement of GAAP by audit firms?

36. Jamal & Sunder, Cerfication Society Concluding Remarks Evidence on ubiquity of certification Certification by government agencies is mostly binary (pass/fail) Private certification has finer scale Field study of baseball card grading market –Dominated by cross sellers (conflict of interest) –Multiple dimensions of competition –No race to the bottom

37. Jamal & Sunder, Cerfication Society Open Questions Will the tough auditors win or lose? Are Independent auditors tough auditors? Precision of grading – Significant? Should auditors issue more nuanced reports? Does it pay to get audited (vs. reputation, warranty and disclosure) Will de-regulated audit market lead to quality differentiation?