PUBLISHED

UNITED STATES COURT OF APPEALS
                 FOR THE FOURTH CIRCUIT


DEBRA GARIETY; HORST O. BISCHOFF,         
as Trustee of Bischoff Family Trust;
PAMELA HANYZEWKI; JOHN J. CLINE;
THOMAS ALLEN, Individually and on
Behalf of all Others Similarly
Situated; THOMAS J. SHANNON, JR.,
as Trustee, Natural Parent and
Guardian of; SMV HOLDING
COMPANY, PLL; VINCENT PAUL;
CHARLES THORNTON, Individually and
as Trustee, SEP; FRED L. MILLNER;
WARREN H. HYDE; CARYL HYDE;
TEEN RESPONSE, INCORPORATED;
ELIZABETH B. SPONSELLER; MICHAEL
J. SPONSELLER; TERRY OVERHOLSER;
and LAWRENCE CORMAN, on behalf               No. 03-1629
of themselves and all others
similarly situated,
                  Plaintiffs-Appellees,
                 and
CRYSTAL N. SHANNON; CITY
NATIONAL BANK, a banking
corporation; COAST PARTNERS
FINANCIAL CORPORATION, formerly
known as Coast Partners Securities,
Incorporated, a California
Corporation,
                           Plaintiffs,
                 and
                                          
2                GARIETY v. GRANT THORNTON, LLP


FEDERAL DEPOSIT INSURANCE               
CORPORATION,
                Intervenor-Plaintiff,
                 v.
GRANT THORNTON, LLP,
             Defendant-Appellant,
                and
HERMAN & CORMANY; MICHAEL
GRAHAM; BILLY JEAN CHERRY; TERRY
LEE CHURCH; ESTATE OF J. KNOX
MCCONNELL, Philip C. Petty,
Administrator; LOUIS J. PAIS;
MICHAEL F. GIBSON; ANDREW L.
RAGO; JULIAN G. BUDNICK; GARY
ELLIS; J&J CONSTRUCTION COMPANY;
HERMIE CHURCH; DIVERSIFIED CAPITAL
MARKETS; MICHAEL PATTERSON; E.E.        
POWELL & COMPANY, INCORPORATED;
JOHN DOES 1-100; ROBERT WAGNER;
QUANTUM CAPITAL CORPORATION;
FERRIS, BAKER, WATTS,
INCORPORATED; SCOTT &
STRINGFELLOW; NANCY VORONO;
GUNNALLEN FINANCIAL INCORPORATED;
JACK INGOLD; REGIS SECURITIES
CORPORATION, a/k/a Quantum
Securities Corporation; NANCY
VARGO; TOM DOOLEY; GARVIN
TANKERSLEY; ELLEN TURPIN; LORA
MCKINNEY; MICHAEL PATTERSON,
INCORPORATED; KUTAK ROCK;
ALAN D. STRASSER, an individual;
MICHAEL D. LAMBERT, an individual;
VIDA HEADRICK; EVELYN HERRON;
                                        
                 GARIETY v. GRANT THORNTON, LLP   3


JEANIE WIMMER; VIRGINIA BURKS;         
DEBRA BAILEY; SUSAN DALTON;
ROBBIN WHITE; CONSTANCE EVANS,
                       Defendants,
                and
MELISSA QUIZENBEURY; HOG PEN
ENTERPRISES, INCORPORATED; JLT
ENTERPRISES; KEYSTONE HARDWARE;
C&H RANCH; MARBIL,
INCORPORATED; HC & TC TRUST;
DOLORES HUGHES; BILL PACK;
WENDY PACK; TAMMY FISHER; LARRY
FISHER; HERMAN FISHER; JUDY KAHL;
ANDREW T. RAPOFF; DONNA MARIE
RAPOFF; MICHAEL A. RAPOFF; DANNY
                                       
RAPOFF; NANCY E. KELANON; DAVID
RAPOFF; ANDREW J. RAPOFF; MICHAEL
J. CHERRY; ROSLYN A. CHERRY; LEAH
M. CHERRY; RACHEL L. CHERRY;
DANIEL T. HALSEY; SUNRISE
AUTOMOTIVE GROUP, INCORPORATED;
NANCY E. KELAHAN; TERRY L. ROSE,
             Intervenors-Defendants,
                and
H. LYNDEN GRAHAM, JR.,
                            Trustee,
                 v.
                                       
4                 GARIETY v. GRANT THORNTON, LLP



ADVANTA MORTGAGE CORPORATION             
USA; CELINK, INCORPORATED,
formerly known as Compu-link
Service, Incorporated,
             Third Party Defendants,
                 and
HARGRAVE MILITARY ACADEMY, a
non-profit corporation; WAYNESBURG
COLLEGE, a non-profit corporation;
MICHAEL CHERRY; TIMMY CLINE;
VICKIE CLINE; THE OFFICE OF THE          
COMPTROLLER OF THE CURRENCY;
FIRST COMMERCE OF AMERICA,
                  Parties in Interest.


UNITED BANK, INCORPORATED,
formerly known as United National
Bank; TERRY WITT; GREGORY WITT;
DAVID WITT; SCOTT WITT; GARY
WITT; DENNIS WITT,
                         Movants.
                                         
             Appeal from the United States District Court
      for the Southern District of West Virginia, at Charleston.
                David A. Faber, Chief District Judge.
     (CA-99-992-2; CA-99-1115-2; CA-00-81-1; CA-00-655-1;
                     CA-99-862-2; CA-02-344-1)

                       Argued: February 27, 2004

                        Decided: May 12, 2004

       Before NIEMEYER and SHEDD, Circuit Judges, and
               HAMILTON, Senior Circuit Judge.
                   GARIETY v. GRANT THORNTON, LLP                       5
Affirmed in part, vacated in part and remanded by published opinion.
Judge Niemeyer wrote the opinion, in which Judge Shedd and Senior
Judge Hamilton joined.


                              COUNSEL

ARGUED: Stanley Julius Parzen, MAYER, BROWN, ROWE &
MAW, L.L.P., Chicago, Illinois, for Appellant. Sigmund S. Wissner-
Gross, HELLER, HOROWITZ & FEIT, P.C., New York, New York,
for Appellees. ON BRIEF: John H. Tinney, Kimberley R. Fields,
THE TINNEY LAW FIRM, Charleston, West Virginia; James C.
Schroeder, Daniel G. Hildebrand, MAYER, BROWN, ROWE &
MAW, L.L.P., Chicago, Illinois; Mark W. Ryan, Andrew J. Morris,
Craig Isenberg, MAYER, BROWN, ROWE & MAW, L.L.P., Wash-
ington, D.C., for Appellant. Joshua I. Barrett, DITRAPINO, BAR-
RETT & DIPIERO, P.L.L.C., Charleston, West Virginia, for
Appellees.


                               OPINION

NIEMEYER, Circuit Judge:

   Grant Thornton LLP, a national accounting firm, appeals the dis-
trict court’s interlocutory order certifying this action as a class action
to adjudicate the plaintiffs’ securities-fraud claims against Grant
Thornton in connection with the 1999 demise of the First National
Bank of Keystone in Keystone, West Virginia. Grant Thornton con-
tends principally that the district court, in finding that the predomi-
nance requirement of Federal Rule of Civil Procedure 23(b)(3) was
satisfied, erred in accepting at face value the plaintiffs’ allegations
that the reliance element of their fraud claims could be presumed
under a "fraud-on-the-market" theory without determining whether
the plaintiffs had any basis for making the allegations. The district
court, concerned about making findings that would overlap with and
therefore prejudice issues on the merits, concluded that the plaintiffs’
assertions alone were "enough at the certification stage" to justify cer-
tifying a class action under Rule 23(b)(3).
6                 GARIETY v. GRANT THORNTON, LLP
   As more fully discussed below, we conclude that the district
court’s reliance on mere assertions did not fulfill the requirements that
the district court take a "close look" at relevant matters, conduct a
"rigorous analysis," and make findings in determining whether the
plaintiffs have demonstrated that the requirements of Rule 23(b)(3)
have been satisfied. We conclude, however, that the district court did
not abuse its discretion, as argued by Grant Thornton, in its appoint-
ment of two of the plaintiffs as lead class representatives. Accord-
ingly, we affirm in part and vacate in part the district court’s
certification order, and we remand this case for further proceedings.

                                    I

   Until the early 1990s, the First National Bank of Keystone
("Keystone") was a relatively small community bank in McDowell
County, West Virginia. In 1992, Keystone embarked on a growth
strategy through which it became a niche lender focusing on subprime
mortgage loans (i.e., loans extended to higher risk borrowers), and in
late 1993, it began buying Federal Housing Authority home improve-
ment loans, pooling them, and selling shares in them to investors —
a process called loan securitization. To pursue its loan securitization
business, Keystone entered into financing relationships with other
banks, paying higher than normal interest rates. In 1997, Keystone
began to securitize its own high loan-to-value loans made to highly
leveraged borrowers with little or no collateral. During these years,
Keystone made its highly risky securitization business its principal
business. From 1992, when Keystone had assets of $107 million, to
1999, Keystone’s business grew almost tenfold. In 1995, Keystone
was reported to be one of the most profitable community banks in the
nation, and by 1999, it reported assets of $1.1 billion. Keystone was
listed No. 1 in American Banker’s June 1999 list of "the 75 most prof-
itable large community banks," with a "whopping" 7.24% return on
average assets in 1998.

   From 1992 until 1999, the Office of the Comptroller of the Cur-
rency ("OCC") examined Keystone’s books annually, but the exami-
nations proved unsatisfactory to OCC because of mutual distrust and
the bank’s resistance to examiners’ findings. The examinations
repeatedly uncovered unsafe and unsound banking practices and regu-
latory violations, yet OCC enforcement actions proved largely inef-
                  GARIETY v. GRANT THORNTON, LLP                      7
fective. In May 1998, pursuant to an agreement with the OCC to hire
an outside auditor, Keystone hired the accounting firm of Grant
Thornton LLP. Grant Thornton issued an audit report on April 19,
1999 (the "Audit Report"), which revealed no problem in Keystone’s
statement of assets. The July 1999 edition of Walker’s Manual of
Unlisted Stocks included a one-page write-up of Keystone that identi-
fied Grant Thornton as Keystone’s auditor and that reported summary
figures taken from Grant Thornton’s Audit Report. During the period
that Grant Thornton was Keystone’s accountant, Grant Thornton also
participated to some extent in the preparation of Keystone’s "Call
Reports," which Keystone filed quarterly with the Federal Deposit
Insurance Corporation ("FDIC"). These Call Reports were available
to the public on the FDIC website.

   While OCC’s examinations of Keystone generally focused on the
credit risk associated with subprime mortgage loan securitizations, it
was not until August 1999 that OCC independently verified that Key-
stone was unable to substantiate $515 million in loan assets, constitut-
ing almost one-half of its assets. On September 1, 1999, the OCC
announced that Keystone was insolvent, and it closed the bank. A few
months later, the FDIC’s Bank Insurance Fund determined that the
Keystone failure would cost the Fund between $750 and $850 mil-
lion, making the loss one of the largest in history. A subsequent inves-
tigation by the Office of Inspector General determined that Keystone
had been suffering heavy losses early in its growth period and that by
late 1996 Keystone had become insolvent. Keystone concealed its
financial condition by continuing to record loans as assets even after
they had been sold to investors as part of a securitized loan pool. The
Office of Inspector General concluded that "[a]lleged fraudulent
accounting practices, uncooperative bank management and reported
high profitability may have all served to mask the bank’s true finan-
cial condition from OCC examiners."

  The plaintiffs, who were purchasers of Keystone stock during the
period after Grant Thornton issued its Audit Report on April 19,
1999, and before the OCC closed the bank on September 1, 1999,
commenced this action to recover damages suffered by reason of the
bank’s failure. The plaintiffs proposed to represent a class of all pur-
chasers of Keystone stock during the April-to-September period.
Their amended complaint, which alleged four counts of federal secur-
8                  GARIETY v. GRANT THORNTON, LLP
ities fraud and seven counts of fraud under state law, named as defen-
dants Grant Thornton, other accountants, former Keystone directors
and officers, broker-dealers, and other Keystone stockholders who
allegedly dumped Keystone stock based on inside information.

   Plaintiffs settled their claims with a number of the defendants,
including Keystone’s former auditors, Keystone’s outside directors,
and certain of the broker defendants. To implement the settlement
agreement, the district court certified a class and, after a final fairness
hearing conducted on June 23, 2003, approved the class settlement as
to the settling defendants. No one objected to the settlement, and it
has now become final. Grant Thornton, however, was not a party to
the settlement and has continued to defend this action.

   In the complaint’s counts against Grant Thornton, the plaintiffs
alleged violations of § 10(b) of the Securities Exchange Act and Rule
10b-5 promulgated thereunder, as well as five state common-law
counts for fraud, constructive fraud, negligent misrepresentation, aid-
ing and abetting tortious conduct, and civil conspiracy. The plaintiffs
alleged that Grant Thornton "participated in falsifying certain material
financial information regarding Keystone" and issued the Audit
Report "knowing that at least $515 million of Keystone’s assets . . .
did not exist, or at a minimum, recklessly representing that such
assets did exist." They also alleged that Keystone insiders illegally
dumped Keystone stock at inflated prices with the participation of
certain stockbrokers who, when pitching Keystone, relied on, and in
some cases provided class members with, the Audit Report, the write-
up in Walker’s Manual, the Call Reports, and the American Banker
ratings. The plaintiffs purport to represent a class consisting of 150
or more persons who purchased stock between April 19, 1999 (the
date of the Audit Report) and September 1, 1999 (the date when the
OCC closed Keystone). During the class period, these class members
engaged in approximately 244 trades of Keystone stock involving, in
the aggregate, 130,500 shares.

   In their motion to certify the action as a class action under Federal
Rule of Civil Procedure 23(b)(3), the plaintiffs recognized that the
element of reliance in their fraud claims "may require individualized
determinations." To satisfy commonality and the predominance of
common issues, however, they relied on a "presumption of reliance"
                  GARIETY v. GRANT THORNTON, LLP                       9
that can be made under a "fraud-on-the-market theory" when shares
are purchased in an efficient market. Pointing to the allegations of
their complaint, they maintained that they "adequately alleged that the
market for Keystone stock did in fact exhibit the five hallmarks of an
efficient market," as necessary for a fraud-on-the-market theory. Con-
sequently, they claimed that they were entitled to a presumption of
reliance, which would permit the district court to find fulfilled the
requirement of Rule 23(b)(3) that common issues predominate over
issues involving only individual class members.

   By an order dated March 28, 2003, the district court ruled that
plaintiffs could maintain this action as a class action under Rule
23(b)(3), and it appointed Horst Bischoff and Debra Gariety as lead
representative plaintiffs. In its order, the court acknowledged that
"plaintiffs were in fact exposed to differing combinations of omis-
sions and misrepresentations, including some oral, making individual
reliance a live issue." To conclude that common issues predominated
over individualized ones, as required by Rule 23(b)(3), the court
relied on a presumption of reliance based on a fraud-on-the-market
theory, which requires that the plaintiffs have purchased in an effi-
cient market for Keystone stock. In response to Grant Thornton’s
argument that the plaintiffs were not entitled to the presumption
because Keystone stock was not in fact traded on an efficient market,
the court stated:

    This court declines to make a detailed examination here of
    the efficiency of the market in which Keystone shares were
    traded because it would require the court to make judgments
    and conclusions regarding the extent and nature of the fraud
    alleged by the plaintiffs. Such determinations, in the absence
    of the procedural safeguards available at trial or the legal
    standard of sufficiency for summary judgment, could theo-
    retically prejudice a party. . . . The court finds the fact that
    the plaintiffs have asserted that the Keystone market was
    efficient is enough at the certification stage to find the mar-
    ket efficient.

In response to Grant Thornton’s additional argument that Rule
23(b)(3) could not be satisfied because plaintiffs’ claims involved the
application of laws from several different States, the district court
10                GARIETY v. GRANT THORNTON, LLP
acknowledged that the laws of at least six States would apply but con-
cluded, with little further explanation, that the variation in those laws
was "minimal."

   Pursuant to Federal Rule of Civil Procedure 23(f), Grant Thornton
requested permission to appeal the district court’s interlocutory certi-
fication order, and on May 23, 2003, we granted that request. This
appeal followed.

                                   II

   The underlying principles for certifying a class action are not con-
troverted. A district court may, in its discretion, order that an action
proceed as a class action only if it finds that the requirements of Fed-
eral Rule of Civil Procedure 23 have been satisfied. Every class action
must satisfy the four requirements of Rule 23(a) — numerosity, typi-
cality, commonality, and adequacy of representation, with "the final
three requirements . . . ‘tend[ing] to merge.’" Broussard v. Meineke
Disc. Muffler Shops, Inc., 

, 337 (4th Cir. 1998) (quoting
Gen. Tel. Co. of the Southwest v. Falcon, 

, 157, n.13
(1982)). In addition, a proposed class must also satisfy the require-
ments of one of the three Rule 23(b) categories. In this case, the plain-
tiffs requested certification of the class under Rule 23(b)(3), which
requires the court to find (1) that "questions of law or fact common
to the members of the class predominate over any questions affecting
only individual members" (the predominance requirement), and (2)
that "a class action is superior to other available methods for the fair
and efficient adjudication of the controversy" (the superiority require-
ment). Rule 23(b)(3)’s predominance requirement is "far more
demanding" than Rule 23(a)’s commonality requirement and "tests
whether proposed classes are sufficiently cohesive to warrant adjudi-
cation by representation." Amchem Prods., Inc. v. Windsor, 

, 623-24 (1997). The plaintiffs who propose to represent the class
bear the burden of demonstrating that the requirements of Rule 23 are
satisfied. Lienhart v. Dryvit Sys., Inc., 

, 146 (4th Cir.
2001).

   The claims on which the plaintiffs seek to proceed against Grant
Thornton as representatives of a class are essentially securities-fraud
claims under federal and state law. To prove a violation of the federal
                  GARIETY v. GRANT THORNTON, LLP                      11
antifraud provisions of § 10(b) of the Securities Exchange Act of
1934, 15 U.S.C. § 78j(b), and Rule 10b-5 promulgated thereunder, 17
C.F.R. § 240.10b-5, "a plaintiff must prove that, in connection with
the purchase or sale of a security, (1) the defendant made a false state-
ment or omission of material fact (2) with scienter (3) upon which the
plaintiff justifiably relied (4) that proximately caused the plaintiff’s
damages." Longman v. Food Lion, Inc., 

, 682 (4th Cir.
1999) (internal quotation marks and citations omitted). The parties
agree that the requirements for state common law fraud and negligent
misrepresentation have analogous elements. Because proof of reliance
is generally individualized to each plaintiff allegedly defrauded, see
Gunnells v. Healthplan Servs., Inc., 

, 434-35 (4th Cir.
2003), fraud and negligent misrepresentation claims are not readily
susceptible to class action treatment, precluding certification of such
actions as a class action. See id.; 

; Andrews
v. AT&T Co., 

, 1025 (11th Cir. 1996) (decertifying class
in part because "the plaintiffs would . . . have to show, on an individ-
ual basis, that they relied on the misrepresentations, suffered injury as
a result, and incurred a demonstrable amount of damages"); Castano
v. Am. Tobacco Co., 

, 745 (5th Cir. 1996) (concluding
that "a fraud class action cannot be certified when individual reliance
will be an issue"). As the Supreme Court explained in Basic, Inc. v.
Levinson, 

, 242 (1988), "[r]equiring proof of individual-
ized reliance from each member of the proposed plaintiff class effec-
tively would have prevented [plaintiffs] from proceeding with a class
action, since individual issues then would have overwhelmed the
common ones." Both the plaintiffs and the district court recognized
the barriers to certifying a class action where proof of individualized
reliance would have to be proved because such individual issues
would then likely "overwhelm" the common ones and preclude the
court from finding satisfaction of the predominance requirement of
Rule 23(b)(3).

   Reliance can, however, be treated as a common issue if it is pre-
sumed under the theory that the defendants defrauded the market —
not the individual plaintiffs — so long as the plaintiffs purchased their
stock in an efficient market. The "fraud-on-the-market" theory was
recognized by the Supreme Court in Basic as a way to litigate
securities-fraud class actions.
12                GARIETY v. GRANT THORNTON, LLP
   In Basic, the Court acknowledged that Rule 10b-5 actions include
a reliance element and recognized that proof of individualized reli-
ance from each class member "effectively would have prevented" the
plaintiff from proceeding with a class action, given the commonality
and predominance requirements of Rules 23(a)(2) and 

. But the Court held that this barrier could be overcome
with a rebuttable presumption of reliance that may be applied through
a fraud-on-the-market theory. 

Quoting the Third Cir-
cuit, Basic described the theory as follows:

     "The fraud on the market theory is based on the hypothesis
     that, in an open and developed securities market, the price
     of a company’s stock is determined by the available material
     information regarding the company and its business. . . .
     Misleading statements will therefore defraud purchasers of
     stock even if the purchasers do not directly rely on the mis-
     statements. . . . The causal connection between the defen-
     dants’ fraud and the plaintiffs’ purchase of stock in such a
     case is no less significant than in the case of direct reliance
     on misrepresentations."

(quoting Peil v. Speiser, 

, 1160-61 (3d
Cir. 1986)). The Basic Court emphasized the difference between
"modern securities markets, literally involving millions of shares
changing hands daily, [and] the face-to-face transactions contem-
plated by early fraud cases. . . ." 

With face-to-face
transactions, the reliance inquiry focuses on the "subjective pricing"
of information before an investor. 

In a modern securities
market, by contrast,

     "the market is interposed between seller and buyer and, ide-
     ally, transmits information to the investor in the processed
     form of a market price. Thus the market is performing a sub-
     stantial part of the valuation process performed by the inves-
     tor in a face-to-face transaction. The market is acting as the
     unpaid agent of the investor, informing him that given all
     the information available to it, the value of the stock is
     worth the market price."

re LTV Sec. Litig., 

, 143 (N.D. Tex.
1980)). Persuaded that "the market price of shares traded on well-
                   GARIETY v. GRANT THORNTON, LLP                       13
developed markets reflects all publicly available information, and,
hence, any material misrepresentations," the Court concluded that
"[a]n investor who buys or sells stock at the price set by the market
does so in reliance on the integrity of that price." 

The
Court held that "[b]ecause most publicly available information is
reflected in market price, an investor’s reliance on any public material
misrepresentations, therefore, may be presumed for purposes of a
Rule 10b-5 action." 

   Application of the reliance presumption is not, however, automatic
in all federal securities-fraud actions. To gain the benefit of the pre-
sumption, a plaintiff must prove "(1) that the defendant made public
misrepresentations; (2) that the misrepresentations were material; (3)
that the shares were traded on an efficient market"; and (4) that the
plaintiff purchased the shares after the misrepresentations but before
the truth was revealed. 

n.27 (citing Levinson
v. Basic, Inc., 

, 750 (6th Cir. 1986)).

   In this case, Grant Thornton contends that the fraud-on-the-market
theory cannot be applied to create the reliance presumption because
(1) Keystone’s shares were not traded on an efficient market, and (2)
Grant Thornton did not make any public misrepresentations — i.e.,
misrepresentations on which the market could rely. It therefore con-
tends that as a matter of law, the district court erred in applying the
fraud-on-the-market presumption of reliance. Without that presump-
tion, the plaintiffs must prove individualized reliance and therefore
almost certainly cannot meet the requirement of 23(b) that common
questions predominate over questions affecting only individual mem-
bers. Grant Thornton also contends that the district court erred by "re-
fusing to look beyond the pleadings" in determining that the fraud-on-
the-market presumption of reliance applied in this case.

   In its certification order, the district court recognized that "the pre-
dominance requirement can be a sticking point in securities class
action certifications where fraud is alleged because individual issues
of reliance on fraudulent misrepresentations may be such as to under-
cut a showing of predominance." Indeed, the court observed that in
this case "plaintiffs were in fact exposed to differing combinations of
omissions and misrepresentations, including some oral, making indi-
vidual reliance a live issue." To overcome this barrier to finding pre-
14                 GARIETY v. GRANT THORNTON, LLP
dominance under Rule 23(b)(3), the district court applied a
presumption of reliance through the fraud-on-the-market theory. In
determining whether the conditions for application of this theory were
satisfied, the court examined the plaintiffs’ complaint and concluded
that the complaint had adequately alleged the elements of a fraud-on-
the-market theory, thereby justifying the creation of a presumption of
reliance. Explaining its refusal to look beyond the complaint, the
court stated that it was "declin[ing] to make a detailed examination
here of the efficiency of the market in which Keystone shares were
traded because it would require the court to make judgments and con-
clusions regarding the extent and nature of the fraud alleged by the
plaintiffs." Citing Eisen v. Carlisle & Jacquelin, 

, 177-
78 (1974), for the proposition that "no inquiry into the merits of a
case can be held at the certification stage," the district court simply
relied on Keystone’s allegations in the complaint and based on those
allegations found that

     the fact that the plaintiffs have asserted that the Keystone
     market was efficient is enough at the certification stage to
     find the market efficient. . . . The plaintiffs further assert
     that they all purchased Keystone stock during the period
     when the fraud they allege was perpetrated on the market.
     Accordingly, the plaintiffs have demonstrated to the satis-
     faction of the court that they are entitled to a presumption
     of reliance for their fraud claims under the Securities [ ]
     Exchange Act.*

   *While the court stated it was declining to make any detailed examina-
tion of the efficiency of the market in this case, it also stated that "had
it chosen to make an investigation into the efficiency of the market in
question, there [was] evidence in this case that could support a finding
of market efficiency," citing to the single fact that the shares dropped in
value over the period of a couple of days after the OCC announced that
it had closed the bank. Had the court conducted an inquiry into the facts
presented about the market for Keystone stock, however, it would have
been confronted with at least a serious question about whether plaintiffs
in this case could demonstrate that they purchased their shares on a mar-
ket sufficiently efficient to act as a surrogate for reliance. Prior to June
25, 1999, Keystone stock was unlisted, and thereafter it was listed only
in the "Pink Sheets" and the Nasdaq OTC Bulletin Board. Moreover, dur-
                   GARIETY v. GRANT THORNTON, LLP                       15
   We conclude that, by accepting the plaintiffs’ allegations for pur-
poses of certifying a class in this case, the district court failed to com-
ply adequately with the procedural requirements of Rule 23. Rule
23(b)(3) on its face requires the court to "find[ ] that the questions of
law or fact common to the members of the class predominate over any
questions affecting only individual members. . . ." (Emphasis added).
In addition, the Rule lists numerous factors for the court to consider
in making its "findings." If it were appropriate for a court simply to
accept the allegations of a complaint at face value in making class
action findings, every complaint asserting the requirements of Rule
23(a) and (b) would automatically lead to a certification order, frus-
trating the district court’s responsibilities for taking a "close look" at
relevant matters, 

, for conducting a "rigorous
analysis" of such matters, 

, and for making
"findings" that the requirements of Rule 23 have been satisfied, see
Fed. R. Civ. P. 23(b)(3). Moreover, if courts could only consider the
pleadings, then "parties would have wide latitude to inject frivolous
issues to bolster or undermine a finding of predominance." Robert G.
Bone & David S. Evans, Class Certification and the Substantive Mer-
its, 51 Duke L.J. 1251, 1269 (2002).

   When Rule 23(c), which originally required certification orders to
be made "as soon as practicable after commencement of [the] action,"
was amended in 2003 to require the court to determine class certifica-
tions "at an early practicable time," the Advisory Committee on Civil
Rules explained the preexisting and longstanding practice that
prompted the change:

     Time may be needed to gather information necessary to
     make the certification decision. Although an evaluation of

ing the entire class period, there were only 244 trades, with an average
of 2.5 trades a day. More importantly, during the relevant trading period,
there were days on which the spread between the bid and the ask price
was nearly 30%. While that volatility was not constant, it existed in vary-
ing degrees at various times during the period of relevant trading. If the
district court were to take such facts into account, it would have been
faced with formidable evidence that the generally accepted criteria for an
efficient market might not have been satisfied during the period from
April to September 1999.
16                 GARIETY v. GRANT THORNTON, LLP
     the probable outcome on the merits is not properly part of
     the certification decision, discovery in aid of the certifica-
     tion decision often includes information required to identify
     the nature of the issues that actually will be presented at
     trial. In this sense, it is appropriate to conduct controlled
     discovery into the "merits," limited to those aspects relevant
     to making the certification decision on an informed basis.

Fed. R. Civ. P. 23 advisory committee’s note to 2003 amendments
(emphasis added).

   The Eisen decision, upon which the district court relied, does not
require a court to accept plaintiffs’ pleadings when assessing whether
a class should be certified. In Eisen, the Supreme Court held that the
district court’s preliminary hearing on the merits of the case — con-
cluding that the plaintiff was "more than likely" to prevail — was
inappropriate for the purpose of determining whether a class action
could be 

. Eisen simply restricts a
court from expanding the Rule 23 certification analysis to include
consideration of whether the proposed class is likely to prevail ulti-
mately on the merits. See 

; 5 Moore’s Federal
Practice ¶ 23.84[2] [a] (3d ed. 2003). As the Supreme Court itself
stated in a post-Eisen case, "sometimes it may be necessary for the
[district] court to probe behind the pleadings before coming to rest on
the certification question." 

; see also Coopers
& Lybrand v. Livesay, 

, 469 (1978) ("[T]he class deter-
mination generally involves considerations that are enmeshed in the
factual and legal issues comprising the plaintiff’s cause of action"
(internal quotation marks omitted)).

    Thus, while an evaluation of the merits to determine the strength
of plaintiffs’ case is not part of a Rule 23 analysis, the factors spelled
out in Rule 23 must be addressed through findings, even if they over-
lap with issues on the merits. Eisen’s prohibition against assessing
plaintiffs’ likelihood of success on the merits as part of a Rule 23 cer-
tification does not mean that consideration of facts necessary to a
Rule 23 determination is foreclosed merely because they are required
to be proved as part of the merits. The analysis under Rule 23 must
focus on the requirements of the rule, and if findings made in connec-
tion with those requirements overlap findings that will have to be
                   GARIETY v. GRANT THORNTON, LLP                      17
made on the merits, such overlap is only coincidental. The findings
made for resolving a class action certification motion serve the court
only in its determination of whether the requirements of Rule 23 have
been demonstrated.

   The district court’s concern that Rule 23 findings might prejudice
later process on the merits need not lead to the conclusion that such
findings cannot be made. The jury or factfinder can be given free
hand to find all of the facts required to render a verdict on the merits,
and if its finding on any fact differs from a finding made in connec-
tion with class action certification, the ultimate factfinder’s finding on
the merits will govern the judgment. A model for this process can be
observed in the context of the preliminary injunction practice. Courts
make factual findings in determining whether a preliminary injunction
should issue, but those findings do not bind the jury adjudging the
merits, and the jury’s findings on the merits govern the judgment to
be entered in the case. See, e.g., Univ. of Tex. v. Camenisch, 

, 395 (1981).

  At bottom, we agree with the conclusion reached by the Seventh
Circuit when it observed:

     The proposition that a district judge must accept all of the
     complaint’s allegations when deciding whether to certify a
     class cannot be found in Rule 23 and has nothing to recom-
     mend it. . . . Before deciding whether to allow a case to pro-
     ceed as a class action . . . a judge should make whatever
     factual and legal inquiries are necessary under Rule 23. . . .
     And if some of the considerations under Rule 23(b)(3) . . .
     overlap the merits . . . then the judge must make a prelimi-
     nary inquiry into the merits.

Szabo v. Bridgeport Machs., Inc., 

, 675-76 (7th Cir.
2001).

   We must not lose sight of the fact that when a district court consid-
ers whether to certify a class action, it performs the public function
of determining whether the representative parties should be allowed
to prosecute the claims of absent class members. Were the court to
defer to the representative parties on this responsibility by merely
18                GARIETY v. GRANT THORNTON, LLP
accepting their assertions, the court would be defaulting on the impor-
tant responsibility conferred on the courts by Rule 23 of carefully
determining the class action issues and supervising the conduct of any
class action certified.

   Because the district court concededly failed to look beyond the
pleadings and conduct a rigorous analysis of whether Keystone’s
shares traded in an efficient market, we must remand the case to per-
mit the district court to conduct the analysis and make the findings
required by Rule 23(b)(3).

                                   III

   When the district court indicated that it was relying only on the
plaintiffs’ assertions to determine whether the predominance require-
ment had been satisfied, the court also said that "had it chosen to
make an investigation into [the] efficiency of the market in question,"
it would have found evidence to support a finding of market effi-
ciency, citing the drop in price of Keystone shares during the days
after the OCC closed the bank. Not only does this hypothetical find-
ing fail to comport with the rigorous analysis required when determin-
ing whether a case may proceed as a class action, but it also overlooks
the substantive requirements that a plaintiff must satisfy to sustain the
fraud-on-the-market theory.

   The fraud-on-the-market theory was recognized as a surrogate for
individualized reliance, creating a presumption of reliance, based on
the proposition that "[a]n investor who buys or sells stock at the price
set by the market does so in reliance on the integrity of that price."

. The fraud-on-the-market theory "interpret[s]
the reliance requirement to mean reliance on the integrity of the mar-
ket price rather than reliance on the challenged disclosure." Daniel R.
Fischel, Efficient Capital Markets, the Crash, and the Fraud on the
Market Theory, 74 Cornell L. Rev. 907, 908 (1989). In contrast to
face-to-face transactions, where an investor has no reason to rely on
the integrity of an offered price because the price may not reflect
available information, "[t]he central premise of the fraud on the mar-
ket theory is that prices of actively traded securities reflect publicly
available information." 

see also 

(quoting In re LTV Sec. 

) (contrasting face-to-
                   GARIETY v. GRANT THORNTON, LLP                       19
face transactions with modern securities markets). For face-to-face
transactions, the reliance requirement serves

     to ensure that the plaintiffs who would not have acted differ-
     ently if the true information were known cannot recover.
     The requirement guarantees, in other words, that informa-
     tion that does not affect a buyer’s or seller’s view of the
     merits of a transaction cannot form the basis of a cause of
     action. In organized markets, however, the market has
     already performed the function of distinguishing between
     unimportant and important information.

Daniel R. Fischel, Use of Modern Finance Theory in Securities Fraud
Cases Involving Actively Traded Securities, 38 Bus. Law. 1 (1982).
A reasonable investor will rely on the integrity of the market price,
however, only if the market is efficient, because in an efficient mar-
ket, "the market price has integrity[;] . . . it adjusts rapidly to reflect
all new information." Jonathan R. Macey & Geoffrey P. Miller, Good
Finance, Bad Economics: An Analysis of the Fraud-on-the-Market
Theory, 42 Stan. L. Rev. 1059, 1060 (1990). Only in those circum-
stances, therefore, may a court presume reliance and avoid individual-
ized inquiries.

   Although Basic clearly requires that a market be efficient in order
for the fraud-on-the-market presumption of reliance to be 

n.27, the decision offers little guidance for determin-
ing whether a market is efficient. The Court does refer to "modern
securities markets, literally involving millions of shares changing
hands daily," 

but that is obviously a general statement
offered as a contrast to face-to-face transactions and is not meant as
a necessary requirement for finding that a market is efficient. Of more
relevance, but only of limited guidance, are the Court’s references to
an "open and developed" market and an "impersonal, well-developed
market." See 

247.

   Nonetheless, it is recognized that it is "[r]ivalrous competition
among market professionals [that] drives securities prices to their effi-
cient levels." Macey & 

; see also West v. Pru-
dential Sec., Inc., 

, 937 (7th Cir. 2002) ("The theme of
Basic and other fraud-on-the-market decisions is that public informa-
20                GARIETY v. GRANT THORNTON, LLP
tion reaches professional investors, whose evaluations of that infor-
mation and trades quickly influence securities prices"); Eckstein v.
Balcor Film Investors, 

, 1129 (7th Cir. 1993)
("Competition among savvy investors leads to a price that impounds
all available information, even knowledge that is difficult to articu-
late. We call a market ‘efficient’ because the price reflects a consen-
sus about the value of the security being traded . . ."). Thus, "[t]he
more thinly traded the stock, the less well the price reflects the latest
pieces of information." 

.

   As a consequence, to determine whether a security trades on an
efficient market, a court should consider factors such as, among oth-
ers, whether the security is actively traded, the volume of trades, and
the extent to which it is followed by market professionals. See, e.g.,
Cammer v. Bloom, 

, 1285-87 (D. N.J. 1989) (exam-
ining (1) average trading volume, (2) number of securities analysts
following the stock, (3) number of market makers, (4) whether the
company was entitled to file an S-3 Registration Statement, if rele-
vant, and (5) evidence of a cause and effect relationship between
unexpected news and stock-price changes).

   While the district court’s brief allusion to the drop in the price of
Keystone’s shares during the days after the OCC announced that Key-
stone was insolvent reflects the assimilation of market information at
its grossest level, that single piece of information, standing alone,
does not represent adequate evidence that the plaintiffs in this case
purchased their shares of Keystone stock in an efficient market. But
this is not to say that there is no other evidence for the court to con-
sider. There is. Substantial discovery has taken place since the district
court issued its certification order, and on remand, the court is free to
consider matters developed through this discovery, as well as other
matters it considers appropriate for making its Rule 23 findings.

                                   IV

  Also in connection with the issue of whether the district court con-
ducted a sufficiently rigorous analysis of the predominance require-
ment of Rule 23(b)(3), Grant Thornton contends (1) that the district
court did not adequately determine whether Grant Thornton made
public misrepresentations as required to show fraud-on-the-market,
                   GARIETY v. GRANT THORNTON, LLP                      21
and (2) that the district court did not adequately review the state law
that would be applied to resolve the state-law claims, nor did it deter-
mine whether application of those laws would preclude a finding that
the predominance requirement has been satisfied. We address these
two points seriatim.

                                    A

   Grant Thornton properly notes that in order to rely on presumed
reliance flowing from an application of the fraud-on-the-market the-
ory, the plaintiffs must demonstrate, among other things, that the
defendant made a public misrepresentation. See 

n.27. It also notes that such a misrepresentation must be directly
attributable to Grant Thornton and not to some other person, because
liability under § 10(b) of the Securities Exchange Act and Rule 10b-
5 does not include aiding and abetting liability. See Cent. Bank of
Denver v. First Interstate Bank of Denver, 

(1994). In
Central Bank, the Supreme Court observed that aiding and abetting
liability would, if recognized, permit plaintiffs to "circumvent the reli-
ance requirement" because a "defendant could be liable without any
showing that the plaintiff relied upon the aider and abettor’s state-
ments or actions." 

see also Anixter v. Home-Stake Prod.
Co., 

, 1225 (10th Cir. 1996) ("The critical element sepa-
rating primary from aiding and abetting violations is the existence of
a representation, either by statement or omission, made by the defen-
dant, that is relied upon by the plaintiff").

   In this case, the plaintiffs contend that Grant Thornton can be held
liable for public misrepresentations from (1) the contents of its Audit
Report; (2) its alleged substantial participation in the preparation of
Keystone’s Call Reports filed with the FDIC; and (3) "the verbatim
recitation of financial information contained in the Audit Report
included in Walker’s Manual." Grant Thornton argues that it could
not be so liable because the Audit Report was not publicly filed and
Grant Thornton was not the author, nor credited as the author, of the
FDIC Call Reports or the summary contained in Walker’s Manual.

   In addressing this issue and reaching the conclusion that the plain-
tiffs fulfilled the requirement in this respect for satisfying the fraud-
on-the-market theory, the district court again relied on mere allega-
22                GARIETY v. GRANT THORNTON, LLP
tions made by the plaintiffs that Grant Thornton "actively falsified
Call Reports sent to the FDIC," that "the falsified reports influenced
public opinion about Keystone," and that Walker’s Manual contained
a summary of the Audit Report. Because the district court must, on
remand, make a finding of whether common issues predominate over
individual issues in the context of reliance and the application of
fraud-on-the-market theory, it should also address more completely
whether Grant Thornton made a public misrepresentation for which
it may be found primarily liable. See Wright v. Ernst & Young LLP,

, 175 (2d Cir. 1998) ("[A] secondary actor cannot incur
primary liability under the [Securities Exchange] Act for a statement
not attributed to that actor at the time of its dissemination"); 

("[F]or an accountant’s misrepresentation to be
actionable as a primary violation, . . . [he] must [himself] make a false
or misleading statement (or omission) that [he] know[s] or should
know will reach potential investors"). But see In re Software Tool-
works, Inc. Sec. Litig., 

, 628 n.3 (9th Cir. 1994) (acknowl-
edging Central Bank and holding that a secondary actor can be
primarily liable under § 10(b) if the actor "played a significant role"
in the preparation of fraudulent statements).

                                   B

   Apart from the fraud-on-the-market theory, Grant Thornton con-
tends that because the law of several States will have to be applied
to resolve state law claims, such individual determinations would pre-
clude a finding of predominance under Rule 23(b)(3).

   The district court allowed that the laws of at least six States are
applicable, and the plaintiffs do not dispute Grant Thornton’s charac-
terization of these laws as including a reliance requirement. More-
over, as Grant Thornton points out, the laws of the six States
identified relate only to the 16 representative parties and not to the
absent class members. The proposed class includes 150 or more per-
sons, and with respect to the absent class members, the applicable
state law has not yet been identified. The plaintiffs have the burden
of showing that common questions of law predominate, and they can-
not meet this burden when the various laws have not been identified
and compared. Moreover, with the potential for individualized reli-
ance determinations having to be made under the law of at least six
                   GARIETY v. GRANT THORNTON, LLP                      23
different States, the district court did not explain why it could find the
predominance requirement satisfied. At least four of the States identi-
fied so far have declined to adopt the fraud-on-the-market presump-
tion of reliance to substitute for finding individualized reliance. See
In re Medimmune, Inc. Sec. Litig., 

, 968 (D. Md.
1995); Kaufman v. i-Stat Corp., 

, 1200-01 (N.J. 2000);
Akerman v. Price Waterhouse, 

, 192 (App. Div.
1998); Kahler v. E. F. Hutton, 

, 145 (Fla. Dist. Ct. App.
1990). And, as we have indicated, individual inquiries into reliance
typically preclude a finding that common issues of fact predominate.
See 

; 

; see
also 

(going so far as to hold that "a fraud
class action cannot be certified when individual reliance will be an
issue").

   Because the plaintiffs bear the burden in this regard of demonstrat-
ing satisfaction of the Rule 23 requirements and the district court is
required to make findings on whether the plaintiffs carried their bur-
den, we also commend this issue to the district court for further devel-
opment on remand.

                                    V

   Finally, Grant Thornton contends that the district court abused its
discretion in appointing Horst Bischoff and Debra Gariety as lead
class representatives under Rule 23(a)(4). With respect to Bischoff,
Grant Thornton contends that he is "incapable of directing this litiga-
tion" because "he testified repeatedly [in his deposition] that any
knowledge that he had about this case was acquired solely from his
attorneys, and that his lawyers make the decisions in this case." As
to Gariety, Grant Thornton argues that because the court denied her
motion to be lead plaintiff pursuant to the Private Litigation Securities
Litigation Reform Act ("PSLRA"), 15 U.S.C. § 77z-1(a)(3)(B)(i), she
should not be allowed to serve as a lead class representative under
Rule 23(a)(4).

   The district court found that "whatever challenges Mr. Bischoff
may face as a class representative," he "has demonstrated that he is
adequately involved in the litigation by reading pleadings and confer-
ring with Ms. Gariety and his counsel regarding strategy." We con-
24                GARIETY v. GRANT THORNTON, LLP
clude that the district court did not abuse its discretion in making this
finding.

   With respect to allowing Gariety to be a representative party even
though she was denied the role as the lead plaintiff pursuant to the
PSLRA, we again conclude that the district court did not abuse its dis-
cretion. Pursuant to the PSLRA, the court simply determined which
plaintiff was the "most capable of adequately representing the inter-
ests of class members." See 

The district court
"felt that Bischoff was simply a better lead plaintiff" and "did not find
that Gariety . . . was a per se inadequate lead plaintiff." In reaching
this conclusion under the PSLRA, the district court did not preclude
a subsequent ruling that Gariety could serve as a lead class represen-
tative under Rule 23(a)(4), and its decision in this regard has not been
shown to constitute an abuse of discretion.

                         AFFIRMED IN PART, VACATED IN PART,
                                             AND REMANDED