These cases present the question whether an insured person
or entity may under Illinois common law maintain a claim
against an insurer for its bad faith conduct in handling a
claim under an insurance policy. We are invited to reconsider
our recent holding in Kelly v. Stratton, 552 F. Supp. 641
(N.D.Ill. 1982), that such a common law cause of action exists.
An Illinois statute, Ill.Rev.Stat. ch. 73, § 767 (1981),
provides for an award of attorney's fees, costs, and a limited
penalty upon a showing that an insurer acted vexatiously and
unreasonably in connection with an insurance claim.*fn1 Some
districts of the Illinois Appellate Court have interpreted the
most recent version of § 767 as precluding any common law
recovery based upon an insurer's bad faith. See Hamilton v.
Safeway Insurance Co., 104 Ill. App.3d 353, 60 Ill.Dec. 97,
432 N.E.2d 996 (1st Dist. 1982); Tobolt v. Allstate Insurance Co.,
75 Ill. App.3d 57, 30 Ill.Dec. 824, 393 N.E.2d 1171 (1st Dist.
1979).*fn2 These cases rejected Ledingham v. Blue Cross Plan,
29 Ill. App. 339, 330 N.E.2d 540 (5th Dist. 1975), rev'd as to
costs, 64 Ill.2d 338, 1 Ill.Dec. 75, 356 N.E.2d 75 (1976),
which held that an independent cause of action existed. Another
court, in Hoffman v. Allstate Insurance Co., 85 Ill. App.3d 631,
40 Ill.Dec. 925, 407 N.E.2d 156 (2d Dist. 1980), while
holding that § 767 barred a common law punitive damages claim
against an insurer, held that the statute did not limit
recovery of common law compensatory damages.
Given the lack of an Illinois Supreme Court ruling on the
effect of § 767 and the divergent appellate holdings, the issue
faced in Kelly presented difficult questions of determination
of state law under Erie R. Co. v. Tompkins, 304 U.S. 64, 58
S.Ct. 817, 82 L.Ed. 1188 (1938). The defendant in No. 82 C 6789
has challenged our application of Erie as well as our
substantive holding in Kelly. In addition, a recent opinion by
our colleague, Judge Milton I. Shadur, provides an extensive
critique of our Erie analysis in Kelly. Commercial Discount
Corp. v. King, 552 F. Supp. 841, 847-52 (N.D.Ill. 1982). These
factors militate in favor of a second look at these
The first question is the effect of the various Illinois
intermediate appellate decisions concerning § 767. In Kelly, we
held that while the appellate court decisions provided "data"
for our determination of state law, they were not controlling.
Kelly, 552 F. Supp. at 644-45. Rather, we held that we were
required to determine how the Illinois Supreme Court would
decide the issue. In doing so, we expressed our disagreement
with Judge Shadur's analysis of the Erie issue. See, e.g.,
Slate Printing Co. v. Metro Envelope Co., 532 F. Supp. 431, 434
(N.D.Ill. 1982); Bonanno v. Potthoff, 527 F. Supp. 561, 563
(N.D.Ill. 1981); Instrumentalist Co. v. Marine Corps League,
509 F. Supp. 323, 329 (N.D.Ill. 1981); National Can Corp. v.
Whittaker Corp., 505 F. Supp. 147, 148-49 n. 2 (N.D.Ill. 1981).
In those cases, as in Commercial Discount, Judge Shadur held
that a court of this district, in determining state law, must
act as a state trial court. Where the various Illinois
appellate districts are in conflict, Judge Shadur further held,
a judge of the Northern District of Illinois must apply the law
of the First District of the Illinois Appellate Court, the
appellate district in which this federal district court sits.
In Commercial Discount, he noted that were a case to be
presented in which the proper Illinois venue was other than the
First District, then the law of the district of proper venue
must be applied. Commercial Discount, 552 F. Supp. at 850.*fn4
It is true, as Judge Shadur suggests, that we must avoid
applying a rule that would permit a litigant to forum shop by
choosing to bring his action in federal court if he found the
law of the relevant state appellate district unfavorable,
hoping to find a more sympathetic ear on the federal bench.
See Bernhardt v. Polygraphic Co. of America, Inc.,
350 U.S. 198, 203-04, 76 S.Ct. 273, 276-77, 100 L.Ed. 199 (1956); Klaxon
Co. v. Stentor Electric Manufacturing Co., 313 U.S. 487, 496,
61 S.Ct. 1020, 1021, 85 L.Ed. 1477 (1941). However, we think
that Commercial Discount begins with an erroneous premise: that
a federal court sitting in diversity jurisdiction "must decide
substantive questions . . . in the same way that a state trial
judge counterpart sitting at the same location would."
Commercial Discount, 552 F. Supp. at 847 (quoting National Can
Corp. v. Whittaker Corp., 505 F. Supp. 147, 148 n. 2 (N.D.Ill.
1981)) (emphasis supplied). The error of this premise affects
the remainder of the Commercial Discount analysis.
The proposition that we must act as state trial judges stems
from a misapprehension of the commands of Erie and its progeny.
Erie requires a federal court to apply the substantive law of
the forum state; we take this to mean that we must apply the
law that ultimately would be applied were the case to be
litigated in the
state courts. While intermediate appellate decisions exert
upon us a high degree of persuasive force, and while they may
be binding upon state trial courts, the law we must apply is
that which the state supreme court would apply.*fn5 In a
given case we may choose to follow an intermediate appellate
ruling, but we may not end our analysis of state law with mere
citation to such rulings where we are persuaded that the state
supreme court would rule otherwise. See generally In re Air
Crash Disaster Near Chicago, Illinois on May 25, 1979,
701 F.2d 1189, 1196-98 (7th Cir. 1983) (Appeal of American Airlines,
The policy behind this principle is Erie's concern with
avoiding forum shopping between state and federal courts.
Applying the law that the state supreme court would follow is
necessary if we are to avoid creating an incentive for such
forum shopping. The "state trial court" approach creates an
incentive for forum shopping in that it requires the federal
courts to give more weight to state intermediate appellate
decisions than they would be given in the state system. This
may be illustrated by three examples.
In a case in which no supreme court decision exists and the
appellate district of proper state venue has not yet taken a
position on an issue, Commercial Discount would require a
federal court to follow the law as declared by the other
appellate districts. Commercial Discount, 552 F. Supp. at 848
(citing People v. Thorpe, 52 Ill. App.3d 576, 579, 10 Ill.Dec.
351, 354, 367 N.E.2d 960, 963 (2d Dist. 1977) and Garcia v.
Hynes & Howes Real Estate, Inc., 29 Ill. App.3d 479, 482,
331 N.E.2d 634, 636 (3d Dist. 1975)). However, the very fact that
the various Illinois appellate districts sometimes conflict on
an issue of law indicates the problem inherent in the state
trial court approach. The appellate districts, it appears, do
not consider each others' decisions binding; rather, they
regard them as persuasive authority only. Thus, if a litigant
filed suit in a state court in the First District and the only
intermediate appellate decision on a pertinent issue was from
the Fourth District, while the trial court presumably would
follow the Fourth District ruling, on appeal the First District
would not necessarily do so, if it found persuasive
reasons to do otherwise. In such a case, if diversity of
citizenship existed, the litigant favored by the Fourth
District rule could file the case in federal court (or remove
it, if there was diversity of citizenship, in the case of a
non-Illinois defendant) and thereby obtain "insurance" that
the favorable rule of law would be applied and upheld on
appeal, were Commercial Discount to apply.
The same is true even where "First District law" exists. The
First District has five divisions. It appears that the
different divisions do not consider themselves bound by the
holdings of other divisions. See, e.g., Bonnano v. Potthoff,
527 F. Supp. 561, 563-64 (N.D.Ill. 1981) (Shadur, J.)
(discussing conflicting holdings of Third and Fifth Divisions
on issue of date of accrual of cause of action); Commercial
Discount, 552 F. Supp. at 843-45 (discussing apparent difference
in decisions in different divisions on question of effect of
failure to give notice of sale of repossessed goods).*fn6
To note an example of recent origin, a controversy existed
within the First District over the question whether a
prosecutor must justify his or her use, within a particular
case, of peremptory challenges to exclude minority group
members from a jury. Compare People v. Gosberry, 109 Ill. App.3d 674,
65 Ill.Dec. 99, 440 N.E.2d 954 (1st Dist., 3d Div.
1982) and People v. Payne, 106 Ill. App.3d 1034, 62 Ill.Dec.
744, 436 N.E.2d 1046 (1st Dist., 3d Div. 1982) with People v.
Newsome, 110 Ill. App.3d 1043, 66 Ill.Dec. 708, 443 N.E.2d 634
(1st Dist., 2d Div. 1982) and People v. Teague, 108 Ill. App.3d 891,
64 Ill.Dec. 401, 439 N.E.2d 1066 (1st Dist., 1st Div.
1982). Because one division does not consider itself bound by
the others' holdings,*fn7 application of the "law" of the
First District where not all divisions have spoken may result
in a litigant obtaining a result that would not obtain were the
case to be litigated in state court.*fn8 Commercial Discount,
therefore, permits a diversity litigant in whose favor the
non-unanimous but not as yet uncontradicted rule runs to obtain
"insurance" by bringing the case in federal court or removing
The third type of forum shopping permitted by Commercial
Discount is somewhat more subtle. The doctrine of that case
requires federal courts to give more weight to state appellate
decisions than the rendering courts themselves would give them.
The "state law" that Erie requires us to follow also includes
the power of a state court to reexamine its earlier holdings
based upon "data" not considered in the earlier decision; Erie,
we think, permits a federal court to exercise the same
authority. As two noted commentators have stated,
Unless a federal court is allowed this much
freedom and flexibility, the Erie doctrine simply
would have substituted one kind of forum-shopping
for another. The lawyer whose case was dependent
on an ancient or shaky state court decision that
might no longer be followed within the state
would have a strong incentive to bring the suit
in or remove it to federal court, hoping that the
could not be impeached under the mechanical
application of existing state precedents that the
Erie doctrine was once thought to require.
Moreover, to give state court decisions more
binding effect than they would have in the state
court system would undermine the ability of the
federal courts to ensure that the outcome of the
litigation be substantially the same as it would
be if tried in a state court and subjected to
that system's appellate process.
19 C. Wright, A. Miller & Cooper, Federal Practice and
Procedure § 4507 at 89-91 (1982) (emphasis supplied).*fn9
Thus, Commercial Discount does permit forum-shopping of a
sort, in that a federal court is required to give state
intermediate appellate precedent more weight than it would
carry in other state appellate tribunals and even in the
rendering panel itself. These are "uncertaint[ies] already
present in state law." Commercial Discount, 552 F. Supp. at 852
(emphasis in original). The view we espoused in Kelly —
application in all cases of the "Supreme Court predictive"
approach — has the disadvantage of being something less than a
bright line, easily applicable rule.*fn10 It is, however,
faithful to Erie and its progeny.*fn11
In addition to its potential for encouraging forum shopping,
Commercial Discount will, at least in some cases, give rise to
a waste of litigants' and courts' resources. Where a state
appellate court has ignored a critically important "datum" of
state law — in Kelly v. Stratton, certain doctrines of
statutory construction — and has reached a result that is
incorrect even as a matter of state law, Commercial Discount
would require us to follow the appellate court ruling and
reach a similarly erroneous result, despite the existence of
persuasive reasons for believing that the state supreme court
would not so hold.*fn12 As Judge Shadur presumably would not
require a federal appellate court to act as a state trial
court (because the two are not "counterparts"), the result is
to require the district court to commit error and leave it to
the court of appeals to correct the error.*fn13 So read,
Commercial Discount elevates form over substance and promotes
the needless expenditure of courts' and litigants'
resources.*fn14 We do not find in Erie or its progeny the
rigidity that Commercial Discount appears to require.
The principles of Illinois stare decisis to which we have
made reference must be recognized by a federal diversity court
if it is to avoid the forum shopping potential recognized by
Wright and Miller in the quoted passage. Again, the central
principle is that we must give appellate court holdings their
due where the supreme court has not spoken, but we must not
give them more than their due. This will require resort to the
"Supreme Court predictive approach," but to do otherwise would
be to ignore the policy of Erie and its progeny. That policy is
the avoidance of forum shopping. When we apply the law that
ultimately would be applied were the case litigated in state
court, we are fully faithful to Erie. By contrast, to act as a
state trial court, following intermediate appellate decisions
that are erroneous as a matter of state law, not only would
violate the policy of Erie, but would also elevate form over
substance, as the court of appeals, assuming the role of its
state counterpart, would apply the correct rule of state
Erie requires us, in all cases, to apply the rule of law that
the state supreme court would follow. Despite our rejection of
the state trial court approach, however, we are not cast adrift
without a rudder. Several general rules exist to guide our
construction of state law.
One such rule is that a federal court should not attempt
"dramatic innovation" in state law. Murphy v. White Hen Pantry
Co., 691 F.2d 350, 355 (7th Cir. 1982). See also Lamb v. Briggs
700 F.2d 1092, 1096 (7th Cir. 1983).*fn16 This appears to be
a corollary of the rule that a federal court may not
"substantially affect the enforcement of the right as given by
the State." Guaranty Trust Co. v. York, 326 U.S. 99, 109, 65
S.Ct. 1464, 1470, 89 L.Ed. 2079 (1945).
Another set of guideposts is set forth in McKenna v. Ortho
Pharmaceutical Corp., 622 F.2d 657 (7th Cir.), cert. denied,
449 U.S. 976, 101 S.Ct. 387, 66 L.Ed.2d 237 (1980):
An accurate forecast of [a state's] law, as it
would be expressed by its highest court, requires
an examination of all relevant sources of that
state's law in order to isolate those factors
that would inform its decision. The primary
course that must be analyzed, of course, is the
decisional law of the [state] [s]upreme [c]ourt.
In the absence of authority directly on point,
decisions by that court in analogous cases
provide useful indications of the court's
probable disposition of a particular question of
law. . . . [R]elevant state precedents must be
scrutinized with an eye toward the broad policies
that informed those adjudications, and to the
doctrinal trends which they evince.
Considered dicta by the state's highest court
may also provide a federal court with reliable
indicia of how the state tribunal might rule on a
particular question. . . . [However,] [a]s
Professor Charles Alan Wright has written, "much
depends on the character of the dictum." Of
somewhat less importance to a prognostication of
what the highest state court will do are
decisions of lower state courts and other federal
courts. Such decisions should be accorded "proper
regard" of course, but not conclusive effect.
Id. at 662 (footnotes omitted).