In the

United States Court of Appeals
               For the Seventh Circuit

No. 08-1590

S TEVEN J. T HOROGOOD , individually and on behalf
    of all others similarly situated,

                                                    Plaintiff-Appellee,
                                  v.

S EARS, R OEBUCK AND C OMPANY,
                                               Defendant-Appellant.


             Appeal from the United States District Court
        for the Northern District of Illinois, Eastern Division.
           No. 06 C 1999—Harry D. Leinenweber, Judge.



  A RGUED S EPTEMBER 15, 2008—D ECIDED O CTOBER 28, 2008




 Before P OSNER, K ANNE, and E VANS, Circuit Judges.
  P OSNER, Circuit Judge. The plaintiff, a Tennessean,
bought a Kenmore-brand clothes dryer from Sears
Roebuck (Kenmore is a Sears brand name). The words
“stainless steel” were imprinted on the dryer, and point
of sale advertising explained that this meant that the
drum in which the clothes are dried inside the dryer was
made of stainless steel. The plaintiff says he thought it
meant that the drum was made entirely of stainless steel.
2                                                 No. 08-1590

Part of the front of the drum, a part the user would see
only if he craned his head inside the drum, is made of a
ceramic-coated “mild” steel, which is not stainless steel
because it doesn’t contain chromium; stainless steel is a
steel alloy that is at least 11.5 percent chromium. The
plaintiff alleges that the mild-steel part of the drum
rusted and stained the clothes that he dried in his dryer.
  He filed this class action suit in federal district court
on behalf of himself and the other purchasers, scattered
across 28 states plus the District of Columbia, of the half
million or so Kenmore dryers advertised as containing
stainless steel drums. He claims that the sale of a dryer
so advertised is deceptive unless the drum is made
entirely of stainless steel, since if it is not it may rust and
cause rust stains on the clothes in the dryer. His individual
claim is that the representation that the dryer contained
a stainless steel drum violated the Tennessee Consumer
Protection Act, Tenn. Code. Ann. §§ 47-18-101 et seq. The
Act provides in pertinent part that “any person who
suffers an ascertainable loss of money or property, real,
personal, or mixed, or any other article, commodity, or
thing of value wherever situated, as a result of the use or
employment by another person of an unfair or deceptive
act or practice declared to be unlawful by this part, may
bring an action individually to recover actual damages.”

The members of the class that the
plaintiff represents are alleged to have similar claims
under similarly worded state consumer protection
statutes in their own states. Although some members of
the huge class are citizens of the states of which Sears is
a corporate citizen (New York and Illinois), so that diver-
sity of citizenship is not complete, the suit properly
No. 08-1590                                               3

invoked federal jurisdiction under the Class Action
Fairness Act, 28 U.S.C. §§ 1332(d), 1453, 1711-1715, since
the amount in controversy exceeds $5 million. The
district court certified the class, and we have accepted
the defendant’s appeal from the class certification. Fed. R.
Civ. P. 23(f).
   The class action is an ingenious device for economizing
on the expense of litigation and enabling small claims to
be litigated. The two points are closely related. If every
small claim had to be litigated separately, the vindica-
tion of small claims would be rare. The fixed costs of
litigation make it impossible to litigate a $50 claim
(our guess—there is no evidence—of what the average
claim of a member of the plaintiff’s class in this case
might be worth) at a cost that would not exceed the
value of the claim by many times. But the class action
device has its downside, or rather downsides. There is
first of all a much greater conflict of interest between
the members of the class and the class lawyers than
there is between an individual client and his lawyer.
The class members are interested in relief for the class
but the lawyers are interested in their fees, and the class
members’ stakes in the litigation are too small to
motivate them to supervise the lawyers in an effort to
make sure that the lawyers will act in their best interests.
Saylor v. Lindsley, 

, 900-01 (2d Cir. 1972)
(Friendly, J.); see also Susan P. Koniak & George M. Cohen,
“Under Cloak of Settlement,” 

. Rev. 1051, 1053-
57 (1996) (describing the class action as “lawyer self-
dealing on a grand scale,” 

Jonathan R. Macey
& Geoffrey P. Miller, “The Plaintiff’s Attorney’s Role in
4                                                No. 08-1590

Class Action and Derivative Litigation,” 58 U. Chi. L. Rev.
1, 22-26 (1991).
  The defendants in class actions are interested in mini-
mizing the sum of the damages they pay the class and
the fees they pay the class counsel, and so they are willing
to trade small damages for high attorneys’ fees, especially
since, as Judge Friendly put it, “a juicy bird in the hand
is worth more than the vision of a much larger one in
the bush, attainable only after years of effort not
currently compensated and possibly a mirage.” Alleghany
Corp. v. Kirby, 

, 347 (2d Cir. 1964); see also
Bruce L. Hay, “Asymmetric Rewards: Why Class Actions
(May) Settle for Too Little,” 48 Hastings L.J. 479, 485-89
(1997); Bruce L. Hay & David Rosenberg, “ ‘Sweetheart’
and ‘Blackmail’ Settlements in Class Actions,” 75 Notre
Dame L. Rev. 1377, 1389-92 (2000). The result of these
incentives is to forge a community of interest between class
counsel, who control the plaintiff’s side of the case, and the
defendants. (For a notable example, see Reynolds v. Benefi-
cial National Bank, 

(7th Cir. 2002); see also
Mars Steel Corp. v. Continental Illinois National Bank & Trust
Co., 

, 681-82 (7th Cir. 1987).) The judge who
presides over the class action and must approve any
settlement is charged with responsibility for preventing
the class lawyers from selling out the class, but it is a
responsibility difficult to discharge when the judge con-
fronts a phalanx of colluding counsel.
  A further problem with the class action is the enhanced
risk of costly error. “When enormous consequences turn
on the correct resolution of a complex factual question, the
No. 08-1590                                                  5

risk of error in having it decided once and for all by one
trier of fact rather than letting a consensus emerge from
several trials may be undue.” Mejdrech v. Met-Coil
Systems Corp., 

, 912 (7th Cir. 2003); see also
Castano v. American Tobacco Co., 

, 746 (5th Cir.
1996); Lance P. McMillian, “The Nuisance Settlement
‘Problem,’ ” 31 Am. J. Trial Advoc. 221, 252-53 (2007); Jeffrey
W. Stempel, “Class Actions and Limited Vision,” 83 Wash.
U. L.Q. 1127, 1213-14 (2005). Suppose a company is sued
in a number of different cases for selling a defective
product. It wins some of the cases and loses some, so
that the aggregate outcome is a fair reflection of the
uncertainty of the plaintiffs’ claims. But when the
central issue in a case is given class treatment and so
resolved by a single trier of fact, a trial becomes a roll of
the dice; a single throw will determine the outcome of
a large number of separate claims—there is no averaging
of divergent responses from a number of triers of fact
having different abilities, priors, and biases.
  The risk is asymmetric when the number of claims
aggregated in the class action is so great that an adverse
verdict would push the defendant into bankruptcy, for
then the defendant will be under great pressure to
settle even if the merits of the case are slight. In re Rhone-
Poulenc Rorer, Inc., 

, 1298-99 (7th Cir. 1995);
Hay, “ ‘Sweetheart’ and ‘Blackmail’ Settlements in Class

; Barry F. McNeil & Beth L.
Fancsal, “Mass Torts and Class Actions: Facing Increased
Scrutiny,” 

, 489-90 (1996). It is true that in
principle and often in practice, shareholders whose
shares in a particular company are part of a diversified
6                                                 No. 08-1590

portfolio of securities are indifferent to the fortunes of a
particular stock in their portfolio. But corporate manag-
ers—the shareholders’ imperfect agents—are not indif-
ferent to bankruptcy and so they are unwilling to bet
their company on the outcome of a trial. This, however, is
not such a case, as the aggregate claims are well within
Sears Roebuck’s ability to pay.
   There is still another downside to the class action, and
it is well illustrated by this case. It is the tendency, when
the claims in a federal class action are based on state
law, to undermine federalism. In re Bridgestone/Firestone,
Inc., 

, 1020-21 (7th Cir. 2002); In re Rhone-
Poulenc Rorer, 

, 51 F.3d at 1300-02; Elizabeth M. v.
Montenez, 

, 788 (8th Cir. 2006). Our plaintiff
wants to litigate in a single federal district court half a
million claims wrested from the control of the courts of
the 29 jurisdictions in which those claims arose and the
laws of which govern the claimants’ entitlement to and
scope of relief. The instructions to the jury on the law it
is to apply will be an amalgam of the consumer protec-
tion laws of the 29 jurisdictions, and procedural rules
by which particular jurisdictions expand or contract relief
will be ignored. The Tennessee Consumer Protection Act,
for example, does not authorize class actions. Walker v.
Sunrise Pontiac-GMC Truck, Inc., 

(Tenn.
2008).
  Sears argues that the Tennessee rule precludes the
maintenance of the present case as a class action. That is
wrong. The procedure in diversity suits is governed by
federal law. What is true is that some procedural rules
No. 08-1590                                                   7

are intended to implement substantive policy, and such
rules do control in diversity cases. The clearest example
is the parol evidence rule of contract law. Another is the
contract doctrine of “mend the hold,” which limits the
right of the defendant in a breach of contract suit to
change his defense in the course of litigation and is thus
a facet of the doctrine of good-faith performance of con-
tracts. Harbor Ins. Co. v. Continental Bank Corp., 

,
364-65 (7th Cir. 1990). We gave another example in S. A.
Healey Co. v. Milwaukee Metropolitan Sewerage District, 

, 310 (7th Cir. 1995): “Suppose a state (as many
states have done) establishes a compulsory arbitration
mechanism in medical malpractice cases in order to cut
down on litigation and reduce malpractice insurance
premiums. The state’s goals are substantive—designed
to shape conduct outside the courtroom and not just
improve the accuracy or lower the cost of the judicial
process—though the means are procedural. The goals
would be thwarted if parties having access to a federal
district court under the diversity jurisdiction could
thumb their noses at the compulsory procedure.” In
contrast, the holding of the Walker decision that consumer
protection suits can’t be maintained under Tennessee
law as class actions was a “plain meaning” statutory
interpretation and did not suggest that the class action
had been precluded in consumer protection suits in
order to advance a substantive policy concerning con-
sumer protection.
  Still, Sears is on to something. Even though the plaintiff
bases his claim, and that of any other Tennesseans who
happen to be members of the class, on Tennessee law, he
8                                                 No. 08-1590

and they are seeking a breadth of relief that Tennessee
does not offer them in its courts. Maybe that is a defect of
Tennessee law. But the purpose of the diversity juris-
diction is to protect out-of-state residents against state
judicial bias in favor of residents; it is not to expand the
relief obtainable under state law.
  The concerns that we have expressed (which are ampli-
fied, and supported by copious references, in our opinion
in In re Rhone-Poulenc, 

, and in such later opin-
ions, apart from those cited already, as Szabo v. Bridgeport
Machines, Inc., 

(7th Cir. 2001); Blair v. Equifax
Check Services, Inc., 

, 834 (7th Cir. 1999); Parker
v. Time Warner Entertainment Co., L.P., 

, 22 (2d
Cir. 2003), and Newton v. Merrill Lynch, Pierce, Fenner &
Smith, Inc., 

, 165-68 (3d Cir. 2001)) suggest
caution in class certification generally. And this case turns
out to be a notably weak candidate for class treatment.
Apart from the usual negatives, there are no positives:
not only do common issues of law or fact not predominate
over the issues particular to each purchase and purchaser
of a “stainless steel” Kenmore dryer, as Rule 23(b)(3) of
the Federal Rules of Civil Procedure requires, but there
are no common issues of law or fact, so there would be
no economies from class action treatment.
  The plaintiff claims to believe that when a dryer is
labeled or advertised as having a stainless steel drum,
this implies, without more, that the drum is 100 percent
stainless steel because otherwise it might rust and cause
rust stains in the clothes dried in the dryer. Do the other
500,000 members of the class believe this? Does anyone
No. 08-1590                                                 9

believe this besides Mr. Thorogood? It is not as if Sears
advertised the dryers as eliminating a problem of rust
stains by having a stainless steel drum. There is no sug-
gestion of that. It is not as if rust stains were a common
concern of owners of clothes dryers. There is no sug-
gestion of that either, and it certainly is not common
knowledge. (At argument the plaintiff’s lawyer, skeptical
that men ever operate clothes dryers—oddly, since his
client does—asked us to ask our wives whether they are
concerned about rust stains in their dryers. None is.)
Stainless steel appliances are common even when no
issue of rust is presented. A porcelain sink does not rust,
but many people prefer a stainless steel sink, partly
because it does not stain, partly because when polished
it looks better (some people think) than porcelain, but
not because they think a sink made of “mild” steel coated
with ceramic would cause rust stains on their dishes;
ceramic doesn’t rust. It is true that the drum is inside
the dryer, and you just see its shiny surface when you
open the door; but the same is true of dishwashers,
many of which are stainless steel too.
  Stainless steel clothes dryers are not advertised as
preventing rust stains on clothes. The only reference to
rust in Sears’s marketing that the plaintiff refers to or that
we have found is “Stainless Steel Drum resists rust and
won’t chip, peel or snag clothes.” The only thing poten-
tially deceptive about this claim is that a ceramic coating
on non-stainless steel is unlikely to rust, chip, peel, or
snag clothes either. But that is not Mr. Thorogood’s
complaint. His concerns are idiosyncratic. A further
indication of this is that to rally fellow victims of the
10                                                No. 08-1590

stainless steel Kenmores he posted his bad experience on
a web site (“Fight Back.com,” http://fightback.com) that
advertises itself as a “conduit for consumer problem
solving and redress.” No one responded, as the “feed
back” file on the web site invites persons who agree
with a posted complaint to do.
   The evaluation of the class members’ claims will
require individual hearings. Each class member who
wants to pursue relief against Sears will have to testify to
what he understands to be the meaning of a label or
advertisement that identifies a clothes dryer as con-
taining a stainless steel drum. Does he think it means
that the drum is 100 percent stainless steel because other-
wise his clothes might have rust stains, or does he
choose such a dryer because he likes stainless steel for
reasons unrelated to rust stains and is indifferent to
whether a part of the drum not easily seen is made of a
different material? Sears does not advertise its stainless
steel drum as a protection against rust stains on clothes;
it does not even say that the drum itself will not rust—only
that it “resists rust.” Advertisements for clothes dryers
advertise a host of features that might matter to con-
sumers, such as price, size, electrical usage, appearance,
speed, and controls, but not, as far as anyone in this
litigation has suggested except the plaintiff, avoidance
of clothing stains due to rust.
  In granting class certification, the district judge said that
because “Sears marketed its dryers on a class wide
basis . . . reliance can be presumed.” Reliance on what? On
stainless steel preventing rust stains on clothes? Since
No. 08-1590                                             11

rust stains on clothes do not appear to be one of the
hazards of clothes dryers, and since Sears did not
advertise its stainless steel dryers as preventing such
stains, the proposition that the other half million buyers,
apart from Thorogood, shared his understanding of
Sears’s representations and paid a premium to avoid rust
stains is, to put it mildly, implausible, and so would
require individual hearings to verify.
  An additional variable in the class action calculus is
relief. Even if some consumers, like Mr. Thorogood, would
not pay a premium for a stainless steel drum that was
not 100 percent stainless steel, the amount of damages
will vary from consumer to consumer. Some may (though
we are dubious) have experienced rust stains, or be
fearful of experiencing them, and therefore seek as dam-
ages the difference between the resale value of their
stainless steel dryer and what a new dryer would cost.
Some may have bought a Kenmore at a discount and so
ended up paying no more than they would have paid
for a machine with a porcelain drum. And some—because
the stainless steel drum is packaged with other
premium features rather than offered as a separately
priced option—may have incurred no damages at all
because on the whole they prefer their stainless steel
dryer to any other dryer they could buy even if the stain-
less steel feature itself was a neutral or even negative
consideration in their purchasing decision. The plaintiff
is seeking on behalf of himself and the members of the
class actual damages, not statutory damages, which
might not require individual proof.
12                                               No. 08-1590

  The difficulty of determining the relief to which the
individual class members are entitled, though serious, is
not the deal breaker. If it were proved that X thousand
buyers of Kenmores had been deceived, a settlement that
provided each with an amount equal to an estimate of
the average damages they had sustained would be a
sensible and legally permissible alternative to remitting
all the buyers to individual suits each of which would
cost orders of magnitude more to litigate than the
claims would be worth to the plaintiffs. “Aggregate class
proof of monetary relief may . . . be based on sampling
techniques or other reasonable estimates, under accepted
rules of evidence.” 3 Herbert B. Newberg & Alba Conte,
Newberg on Class Actions § 10.3, p. 480 (4th ed. 2002); see
also 

Stewart v. General Motors Corp., 

, 452-53 (7th Cir. 1976); United States v. City of Miami,

, 1299-1300 (11th Cir. 1999); Pettway v. Ameri-
can Cast Iron Pipe Co., 

, 259-63 (5th Cir. 1974).
The deal breaker is the absence of any reason to believe
that there is a single understanding of the significance
of labeling or advertising clothes dryers as containing a
“stainless steel drum.”
  The district court is instructed to decertify the class.
                                                  R EVERSED.




                           10-28-08