REVISED AUGUST 27, 2009
IN THE UNITED STATES COURT OF APPEALS
FOR THE FIFTH CIRCUIT
United States Court of Appeals
Fifth Circuit
FILED
No. 07-30443 August 10, 2009
Charles R. Fulbruge III
JOHN THOMPSON Clerk
Plaintiff-Appellee
v.
HARRY F. CONNICK, in his official capacity as District Attorney;
ERIC DUBELIER, in his official capacity as Assistant District Attorney;
JAMES WILLIAMS, in his official capacity as Assistant District Attorney;
EDDIE JORDAN, in his official capacity as District Attorney;
ORLEANS PARISH DISTRICT ATTORNEY’S OFFICE
Defendants-Appellants
Appeals from the United States District Court
for the Eastern District of Louisiana
Before JONES, Chief Judge, and KING, JOLLY, DAVIS, SMITH, WIENER,
BARKSDALE, GARZA, BENAVIDES, STEWART, CLEMENT, PRADO, OWEN,
ELROD, SOUTHWICK, and HAYNES, Circuit Judges.1
PER CURIAM:
By reason of an equally divided en banc court, the decision of the district
court is AFFIRMED. The panel opinion was vacated by the grant of rehearing
en banc.
1
Judge Dennis is recused and did not participate in this decision.
No. 07-30443
JONES, Chief Judge, would reverse for additional reasons:
I concur in Judge Clement's fine opinion and would also highlight the
troubling tension between this unprecedented multimillion dollar judgment
against a major metropolitan District Attorney's office and the policies that
underlie the shield of absolute prosecutorial immunity. The Supreme Court
ought to address whether holding governmental entities liable for Section 1983
violations is consistent with absolute prosecutorial immunity from such
violations. Imbler v. Pachtman, 424 U.S. 409, 96 S. Ct. 984 (1976).
The Supreme Court recently issued a unanimous opinion affording
absolute immunity from personal Section 1983 liability to Los Angeles County's
chief prosecutors for failure to train or supervise their staff, or failure to
establish appropriate systems in regard to the advocacy function of their office.
Van de Kamp v. Goldstein, ___ U.S. ___, 129 S. Ct. 855 (2009). Much as in this
case, a plaintiff had been freed from custody after he discovered that important
evidence had been withheld during his prosecution. The Court made a number
of observations that are prescient of the circumstances leading to liability in this
case. These bear repeating or paraphrasing with my editorial analogies to
prosecutor’s offices.
No. 07-30443
1. The 'public trust' in the prosecutor's office would suffer were he to have
in mind his own potential liability when making prosecutorial decisions. Van de
Kamp, 129 S. Ct. at 860 (quoting Imbler). Likewise, public confidence will erode
if the public believe a prosecutorial office is motivated by the impulse to cover
itself when challenged by difficult prosecutions.
2. The frequency with which criminal defendants bring suits creates real
fear about public perception as well as the independence of judgment exercised
by prosecutors under the constant threat of lawsuits. Van de Kamp, id.
3. Such suits, whether against the prosecutor – or the office –, “ 'often
would require a virtual retrial of the criminal offense in a new forum, and the
resolution of some technical issues by the lay jury.' ” Van de Kamp, id. (quoting
Imbler). See footnote 41 of Judge Clement's opinion.
4. A prosecutor “ ‘inevitably makes many decisions that could engender
colorable claims of constitutional deprivation.’ ” Van de Kamp, id. (quoting
Imbler). See Judge Clement's opinion at text adjoining footnote 53.
5. Defending against such claims, “ ‘often years after they were made,
could impose unique and intolerable burdens upon a prosecutor [or office]
responsible annually for hundreds of indictments and trials.’ ” Van de Kamp, id.
3
No. 07-30443
(quoting Imbler). A crucial witness here had died, and other prosecutors could
not recall this case as distinct from the hundreds or thousands they had handled.
6. The Court also said: "We do not see how...differences in the pattern of
liability among a group of prosecutors in a single office [i.e. distinguishing
between the supervisors and the line prosecutors] could alleviate Imbler’s basic
fear, namely, that the threat of damages liability would affect the way in which
prosecutors carried out their basic court-related tasks.” Van de Kamp, 129 S. Ct.
at 862. Moreover, “. . . ‘it is the interest in protecting the proper functioning of
the office, rather than the interest in protecting its occupant, that is of primary
importance.’ (internal citation omitted)." Id. Authorizing Section 1983 liability
against the office creates the same stress on the proper functioning of the office.
7. With regard to liability for supervisory actions related to the trial
process, the Court held that "a suit charging that a supervisor made a mistake
directly related to a particular trial, on the one hand, and a suit charging that
a supervisor trained and supervised inadequately, on the other, would seem very
much alike." Van de Kamp, 129 S. Ct. at 863.
8. "It will often prove difficult to draw a line between general office
supervision or training and specific supervision or training related to a
particular case.” Van de Kamp, id. “To permit claims based upon the former is
4
No. 07-30443
inevitably to permit the bringing of claims that include the latter." Id. In this
case, the jury was permitted to infer Section 1983 deliberate indifference and
causation based on a single incident of withheld Brady evidence.
9. "[O]ne cannot easily distinguish, for immunity purposes, between
claims based upon training or supervisory failures related to Giglio [at issue in
Van de Kamp] and similar claims related to other constitutional matters
(obligations under Brady v. Maryland, 373 U.S. 83 (1963), for example). And
that being so, every consideration that Imbler mentions militates in favor of
immunity." Id.
10. If the threat of damages liability for a trial error could lead a trial
prosecutor to take account of that risk when making trial-related decisions, so,
too, could the threat of office liability for the same error affect the decisions of
other prosecutors. Van de Kamp, 129 S. Ct. at 863. So, too, could the office
policies become infected with risk aversion.
11. Because “better training or supervision might prevent most, if not all,
prosecutorial errors at trial, permission to bring such a suit [ in Van de Kamp]
would encourage claims [by other criminal defendants], in effect claiming
damages for (trial-related) training or supervisory failings.” Van de Kamp, id.
Such suits could, "given the complexity of the constitutional issues," "pose
5
No. 07-30443
substantial danger of liability even to the honest prosecutor.” Id. (quoting
Imbler). Indeed, only four convictions of the New Orleans District Attorney's
office were overturned for Brady violations in the decade preceding Thompson's
conviction (Judge Clement's opinion at footnotes 49-50), and none involved lab
reports.
12. Practical anomalies result from the coexistence of absolute
prosecutorial immunity with potential Monell liability of the prosecutor's office.
As the Court observed in Van de Kamp, id., "[s]mall prosecution offices where
supervisors can personally participate in all of the cases would...remain immune
from [damage suits]; but large offices, making use of more general office-wide
supervision and training, would not."
13. "Most important, the ease with which a plaintiff could restyle a
complaint charging a trial failure so that it becomes a complaint charging a
failure of training or supervision would eviscerate Imbler." Van de Kamp, id.
This seems true whether the potential defendant is a supervisor, as in Van de
Kamp, or the governmental office itself, as in this case.
The Court has not specifically excluded municipal Section 1983 liability
for prosecutorial offices, nor has it ruled that they are vulnerable. Still, every
reason advanced in Van de Kamp and Imbler for protecting the independence
6
No. 07-30443
and integrity of prosecutors in trial-related actions and supervision suggests
that holding a government entity liable in their stead for the same violations is
simply untenable. The Court recognized, "as Chief Judge Hand pointed out [in
Imbler], that sometimes such immunity deprives a plaintiff of compensation that
he undoubtedly merits; but the impediments to the fair, efficient functioning of
a prosecutorial office that liability could create lead us to find that Imbler must
apply here." Van de Kamp, 129 S. Ct at 864. Today’s judgment raises issues
that will continue to plague honest prosecutors’ offices.
7
No. 07-30443
E. Grady Jolly, Specially Concurring:
Ordinarily, when an en banc case results in a tie vote, we affirm the
district court judgment without opinion. That is the way I would prefer it today.
However, notwithstanding that there is no majority opinion, and that no opinion
today will bind any court or future party in this circuit, each side has now
written for publication, and judges are joining one or the other of the respective
opinions. I join Judge Clement’s opinion because, as between the two, it shows
the intellectual fortitude of meeting head-on, in a specific workmanlike manner,
the truly difficult legal issues presented by this case.
8
No. 07-30443
EDITH BROWN CLEMENT, Circuit Judge, with whom JONES, Chief Judge,
and JOLLY, SMITH, GARZA, and OWEN, Circuit Judges, join, would reverse
the district court for the following reasons:
We believe it imperative to explain why the result in this case should not
encourage the extension of single incident municipal liability under Monell.
John Thompson, the plaintiff-appellee, was convicted of murder and spent
fourteen years on death row for a crime he did not commit because prosecutors
failed to turn over a lab report in a related case. In this 42 U.S.C. § 1983 case,
he sought compensation for the years spent in prison and on death row. The
jury awarded Thompson $14 million against the Orleans Parish District
Attorney in his official capacity. This appeal asks whether Harry Connick, the
District Attorney, was deliberately indifferent to an obvious need to train his
assistants on their obligations under Brady v. Maryland;1 it further asks
whether a lack in Brady training was the moving force behind Thompson’s
constitutional injury. The district court, and a panel of this court, held that the
evidence was legally sufficient to support both of these claims. The panel
opinion was vacated by our order for en banc rehearing.
Only under the most limited circumstances may a municipality be held
liable for the individual constitutional torts of its employees. Considering the
strict standards of culpability and causation applicable here, we conclude that
the evidence supporting Thompson’s claim was legally inadequate to hold the
District Attorney’s Office liable for this employee failure. Along similar lines, we
also conclude that the jury instructions given on “deliberate indifference” were
plainly erroneous.
1
373 U.S. 83 (1963).
9
No. 07-30443
FACTS
In 1985, a few weeks before his murder trial, John Thompson was tried
and convicted of attempted armed robbery. Because of the attempted armed
robbery conviction, Thompson decided not to testify in his own defense in his
trial for the murder of Raymond T. Liuzza, Jr. Thompson was convicted of
murder and sentenced to death.
Fourteen years later, in 1999, an investigator in Thompson’s habeas
proceedings discovered that prosecutors had failed to turn over a crime lab
report in the attempted armed robbery case. That lab report indicated that the
perpetrator had type B blood. Because Thompson has type O blood, the
attempted armed robbery conviction was vacated. In 2002, the Louisiana Fourth
Circuit Court of Appeals granted post-conviction relief and reversed Thompson’s
murder conviction, holding that the improper attempted armed robbery
conviction had unconstitutionally deprived Thompson of his right to testify in his
own defense at his murder trial.2 Thompson was retried for Liuzza’s murder in
2003 and found not guilty.
After his release, Thompson brought suit alleging various claims against
the District Attorney’s Office, Connick, James Williams, Eric Dubelier, and
Eddie Jordan—the District Attorney in 2003—in their official capacities; and
Connick in his individual capacity (collectively, “Defendants”).3 The only claim
that proceeded to trial was a claim under § 1983 for wrongful suppression of
exculpatory evidence in violation of Brady v. Maryland. Thompson presented
two theories of liability to the jury: (1) that the Brady violation was due to an
unconstitutional official policy of the District Attorney’s Office, and (2) that the
Brady violation was caused by Connick’s deliberate indifference to an obvious
2
State v. Thompson, 825 So. 2d 552, 557–58 (La. Ct. App. 2002).
3
Connick was dismissed in his individual capacity prior to trial.
10
No. 07-30443
need to train, monitor, or supervise his prosecutors. The jury found that the
Brady violation was not due to an official policy of the District Attorney’s Office,
but was due to a failure to train. The jury awarded Thompson $14 million in
damages. Defendants filed timely motions for judgment as a matter of
law—before and after the verdict—as well as a motion to amend or alter the
judgment and a motion for a new trial.4 The district court denied all of these
motions, and Defendants appealed.
STANDARD OF REVIEW AND APPLICABLE LAW
We review the denial of a motion for judgment as a matter of law de novo.5
Judgment as a matter of law is appropriate if “the court finds that a reasonable
jury would not have a legally sufficient evidentiary basis to find for the party on
that issue.”6 Under this standard, we consider all of the evidence “in the light
and with all reasonable inferences most favorable to the party opposed to the
motion.”7 Substantial evidence—defined as “evidence of such quality and weight
that reasonable and fair-minded men in the exercise of impartial judgment
might reach different conclusions”—must be presented by the non-moving party,
4
Defendants correctly argue that the district court erred in naming Connick, Dubelier,
and Williams in the judgment and abused its discretion by denying their Rule 59(e) motion to
alter or amend the judgment to correct this error. We have previously held that it is proper
to dismiss allegations against municipal officers in their official capacities when the allegations
duplicate claims against the governmental entity itself. See Castro Romero v. Becken, 256 F.3d
349, 355 (5th Cir. 2001) (noting that the official-capacity claims were duplicative of the claims
against the governmental entities); cf. FED. R. CIV. P. 25(d) (providing that when a public
officer is sued in his official capacity and then ceases to hold office, his successor is
automatically substituted as a party). Because it is undisputed that the claims against
Connick, Dubelier, and Williams, who are no longer employed at the District Attorney’s Office,
are duplicative of the claims against the governmental entity, they should not have been
named in the judgment.
5
Mullins v. TestAmerica, Inc., 564 F.3d 386, 403 (5th Cir. 2009).
6
FED. R. CIV. P. 50(a).
7
Boeing Co. v. Shipman, 411 F.2d 365, 374 (5th Cir. 1969) (en banc), overruled on other
grounds by Gautreaux v. Scurlock Marine, Inc., 107 F.3d 331 (5th Cir. 1997).
11
No. 07-30443
and a “mere scintilla” is insufficient.8 “[A] jury’s freedom to draw reasonable
inferences does not extend so far as to allow the jury to draw an inference which
amounts to mere speculation and conjecture.”9 In reviewing the record, we “give
credence to the evidence favoring the nonmovant as well as that evidence
supporting the moving party that is uncontradicted and unimpeached.”10
Because we are reviewing a Monell11 verdict against a government entity,
our evidentiary review must take into account that § 1983 Monell liability is
fundamentally different from an entity’s vicarious liability, predicated on
respondeat superior, for its employees’ ordinary misconduct. Thus, when a
plaintiff seeks to impose § 1983 liability on a municipality for its failure to train
its employees, normal tort standards are replaced with heightened standards of
culpability and causation.12 Liability will only attach if the municipality was
deliberately indifferent to the constitutional rights of citizens—a showing of
negligence or even gross negligence will not suffice.13 Errors of judgment do not
alone prove deliberate indifference, nor is such heightened culpability
established simply by showing that a municipality could have ordered more or
different training or even misjudged whether training was necessary. Similarly,
to satisfy causation, the plaintiff must demonstrate that the failure to train was
the moving force behind the constitutional violation—“but for” causation is not
8
Id.
9
Nichols Constr. Corp. v. Cessna Aircraft Co., 808 F.2d 340, 352 (5th Cir. 1985)
(quotation omitted).
10
Reeves v. Sanderson Plumbing Prods., Inc., 530 U.S. 133, 151 (2000) (quotation
omitted).
11
Monell v. Dep’t of Soc. Servs., 436 U.S. 658 (1978).
12
City of Canton v. Harris, 489 U.S. 378, 391 (1989).
13
Id. at 388 & n.7.
12
No. 07-30443
enough.14 The Supreme Court has repeatedly cautioned that if we neglect these
stringent standards, we risk collapsing the distinction between vicarious liability
and direct liability.15 Heightened standards also guard against the potentially
“endless exercise of second-guessing municipal employee-training programs,” a
task for which federal courts are ill suited.16
A. Heightened Culpability: Deliberate Indifference
Municipal liability attaches “where—and only where—a deliberate choice
to follow a course of action is made from among various alternatives by city
policymakers. Only where a failure to train reflects a ‘deliberate’ or ‘conscious’
choice by a municipality—a ‘policy’ as defined by our prior cases—can a city be
liable for such a failure under § 1983.”17 Because a municipality rarely decides
not to train its employees out of sheer contempt for citizens’ rights, deliberate
indifference must be proven circumstantially. Deliberate indifference “generally
requires” a showing that the policymaker was made aware of the training
deficiencies by “at least a pattern” of similar deprivations.18 Absent a pattern,
in certain unique circumstances, the plaintiff can establish liability based upon
a single violation of constitutional rights. In such a case, a failure to train
constitutes municipal policy only where the need for training was “so obvious”
that a failure to do so would mean that the policymaker was deliberately
14
Bd. of County Comm’rs of Bryan County v. Brown, 520 U.S. 397, 404 (1997)
(hereinafter, Bryan County).
15
Id. at 415; City of Canton, 489 U.S. at 391-92.
16
City of Canton, 489 U.S. at 392.
17
Id. at 389 (internal quotation and citation omitted).
18
Burge v. St. Tammany Parish, 336 F.3d 363, 370 (5th Cir. 2003) (finding no
deliberate indifference as a matter of law with respect to sheriff’s office training on Brady
obligations).
13
No. 07-30443
indifferent to constitutional rights.19 A need for training is considered
sufficiently obvious only when the deprivation of constitutional rights is a
“highly predictable consequence” of the training deficiency.20 “Deliberate
indifference” implies a sense of callousness21 and is “treated, as it is elsewhere
in the law, as tantamount to intent, so that inaction by a policymaker
deliberately indifferent to a substantial risk of harm is equivalent to the
intentional action that [the] setting [of] policy presupposes.”22 The Supreme
Court has given us a hypothetical example of such callousness: giving untrained
police officers firearms and ordering them to arrest fleeing suspects. In such a
situation, the deprivation of constitutional rights, i.e., the use of excessive force,
is a highly predictable consequence of the lack of training, and therefore the
need for such training is sufficiently obvious.23
Our court has considered single violation liability several times, and, with
only one exception in some thirty years since Monell, has “consistently rejected
application of the single incident exception.”24 The sole exception, Brown v.
Bryan County, involved a failure to train a neophyte on the constitutional limits
19
City of Canton, 489 U.S. at 390.
20
Bryan County, 520 U.S. at 409; see also City of Canton, 489 U.S. at 396 (O’Connor,
J., concurring in part and dissenting in part) (finding that single incident liability for a failure
to train only obtains where “[t]he constitutional duty of the individual officer is clear, and it
is equally clear that failure to inform city personnel of that duty will create an extremely high
risk that constitutional violations will ensue” (emphasis added)).
21
See Estelle v. Gamble, 429 U.S. 97, 106 n.14 (1976); Doe v. Taylor Indep. Sch. Dist.,
15 F.3d 443, 453 (5th Cir. 1994).
22
Bryan County, 520 U.S. at 419 (Souter, J., dissenting) (internal footnote omitted).
23
City of Canton, 489 U.S. at 390 n.10.
24
Gabriel v. City of Plano, 202 F.3d 741, 745 (5th Cir. 2000); see also Cozzo v.
Tangipahoa Parish Council—President Government, 279 F.3d 273, 288 (5th Cir. 2002) (“[T]his
court has often rejected application of the single incident exception.”); Snyder v. Trepagnier,
142 F.3d 791, 798 (5th Cir. 1998) (“[P]roof of a single violent incident ordinarily is insufficient
to hold a municipality liable for inadequate training.”).
14
No. 07-30443
to the use of force.25 The Brown court focused on the decision by the sheriff to
place his nephew, a completely untrained new deputy, “on the street to make
arrests.”26 Within a matter of weeks, the deputy unconstitutionally injured a
citizen he was trying to arrest. This court concluded, based on an extensive list
of factors, that “the jury reasonably could have concluded that Sheriff Moore
made a conscious decision not to train [his nephew], yet still allowed him to
make arrests.”27 There was “unmistakable culpability and clearly connected
causation” sufficient to hold the municipality liable.28
In Brown, the conscious decision was to not train a specific deputy, and the
excessive use of force—which was the “highly predictable consequence” of failing
to train the deputy—occurred soon after the officer went out on the streets.
Several panels of this court, however, have reviewed cases where a decision not
to train was made long before the alleged violation, and found that the lack of
any similar violations indicates that a violation could not be the “highly
predictable consequence” of failing to train.29 This approach reflects common
25
219 F.3d 450 (5th Cir. 2000) (hereinafter, Brown). The facts of Brown demonstrate
that single violation liability applies only in extreme circumstances. The offending officer had
been on the job for only a few weeks and had no education or experience whatsoever in law
enforcement. Moreover, shortly before joining the sheriff’s office, he had been arrested for
several crimes, including assault and battery. In contrast, the assistant district attorneys
involved in Thompson’s armed robbery case had three years of legal schooling, prosecutorial
experience, and no history of past misconduct.
26
Id. at 458.
27
Id. (emphasis added); see id. (listing seven specific factors that put the sheriff on
notice that more training was needed).
28
Id. at 461. Brown said that “the evidence must withstand a vigorous test” for single
violation liability to attach: (1) “it [must] have been obvious to [the policymaker] that the
highly predictable consequence of not training” would be a constitutional violation, and (2)
“this failure to train” must have been “‘the moving force’ that had a specific causal connection
to the constitutional injury.” Id.
29
See Rios v. City of Del Rio, 444 F.3d 417, 427 (5th Cir. 2006) (“Here there is no
allegation of any prior incident in which [a similar violation occurred].”); Johnson v. Deep E.
15
No. 07-30443
sense: if there have been thousands of opportunities for municipal employees to
violate citizens’ constitutional rights, and yet there have been no previous
violations, then the need for training is simply not “so obvious.”30
B. Heightened Causation: Moving Force
To safeguard the boundaries established in Monell, the Supreme Court has
made clear that in addition to a heightened standard of culpability, plaintiffs
must meet a heightened standard of causation in order to hold a municipality
liable under § 1983.31 A § 1983 plaintiff must prove that the municipal entity’s
custom or policy—in this case the failure to train—was the “moving force” that
caused the specific constitutional violation; stated differently, the plaintiff must
establish a “direct causal link” between the municipal policy and the
constitutional injury.32 We have said that the “connection must be more than a
Tex. Reg’l Narcotics Trafficking Task Force, 379 F.3d 293, 310 (5th Cir. 2004) (finding that “no
similar incident had occurred in . . . ten years on the job” and that the evidence showed that
officers generally followed the law); Cozzo, 279 F.3d at 288 (“[T]he elapsing of almost two
decades without any [similar constitutional] complaint being lodged suggests that the
inadequate training was not obvious and obviously likely to result in a constitutional
violation.” (quotation omitted)); Conner v. Travis County, 209 F.3d 794, 797 (5th Cir. 2000)
(“We can reasonably expect—if the need for training in this area was ‘so obvious’ and the
failure to train was ‘so likely to result in the violation of constitutional rights’—that the
[Plaintiffs] would be able to identify other instances of harm arising from the failure to train.
The fact that they did not do so undercuts their deliberate indifference claim.”).
30
Of course, there will be situations in which we would not expect to see a pattern of
violations. One example would be where the Supreme Court develops a new duty. See Walker
v. City of New York, 974 F.2d 293, 300 (2d Cir. 1992) (focusing on the fact that Brady had been
decided only seven years prior to the alleged violation). Another example would be where a
municipality changes its policies or makes an exception to its general policies in deliberately
choosing not to train an individual. See Brown, 219 F.3d at 458 & n.10 (finding that while no
“formal policy of training” existed, the record indicated that the municipality hired only
“trained and experienced” employees, though it had not adhered to that policy for the employee
at issue). In the normal case, however, if there is a failure to train, we should expect to see a
pattern of constitutional violations. If no such pattern can be shown, and no reason is given
why a pattern does not exist, then the need for training simply is not “so obvious” as to amount
to “deliberate indifference.”
31
City of Canton, 489 U.S. at 391–92.
32
Bryan County, 520 U.S. at 404.
16
No. 07-30443
mere ‘but for’ coupling between cause and effect.”33 The deficiency in training
must be the actual cause of the constitutional violation.
Accordingly, the District Attorney must not be held liable simply because
the culpable assistant district attorneys worked for him. “[A] municipality can
be found liable under § 1983 only where the municipality itself causes the
constitutional violation at issue.”34
That this heightened standard is vital to maintaining Monell’s prohibition
against vicarious liability in § 1983 cases has been clearly expressed by the
Supreme Court:
Where a court fails to adhere to rigorous requirements
of . . . causation, municipal liability collapses into
respondeat superior liability. As we recognized in
Monell and have repeatedly reaffirmed, Congress did
not intend municipalities to be held liable unless
deliberate action attributable to the municipality
directly caused a deprivation of federal rights. A failure
to apply stringent culpability and causation
requirements raises serious federalism concerns . . . .35
And in City of Canton, the Supreme Court further said:
To adopt lesser standards of fault and causation would
open municipalities to unprecedented liability under
§ 1983. In virtually every instance where a person has
had his or her constitutional rights violated by a city
employee, a § 1983 plaintiff will be able to point to
something the city “could have done” to prevent the
unfortunate incident. See Oklahoma City v. Tuttle, 471
U.S., at 823 (opinion of Rehnquist, J.). Thus,
permitting cases against cities for their “failure to
train” employees to go forward under § 1983 on a lesser
standard of fault would result in de facto respondeat
33
Fraire v. City of Arlington, 957 F.2d 1268, 1281 (5th Cir. 1992).
34
City of Canton, 489 U.S. at 385.
35
Bryan County, 520 U.S. at 415 (emphases added).
17
No. 07-30443
superior liability on municipalities—a result we
rejected in Monell, 436 U.S., at 693–694. It would also
engage the federal courts in an endless exercise of
second-guessing municipal employee-training
programs. This is an exercise we believe the federal
courts are ill suited to undertake, as well as one that
would implicate serious questions of federalism.36
This exacting standard is even more crucial in cases where the municipal policy
is said to be a failure to train. Because such a “policy,” standing alone,
implicates no constitutional violation, the “causal connection between municipal
policy and the deprivation of constitutional rights becomes more difficult to
discern.”37 Thus, a careful analysis is required to determine “whether a jury
reasonably could conclude that the [municipality’s] conduct was the moving force
in bringing about the constitutional violation.”38
To summarize, the requirements for imposing liability upon a municipality
for the individual acts of its employees are demanding. Relaxing these
heightened requirements would cause significant harm to the interests
underlying this demanding evidentiary principle: “adopt[ing] lesser standards
of fault and causation would open municipalities to unprecedented liability,”
“would result in de facto respondeat superior liability,” and would “engage the
federal courts in an endless exercise of second-guessing municipal employee-
training programs.”39 Therefore, we can hold a municipality liable only where
the evidence demonstrates “unmistakable culpability and clearly connected
causation” for the unconstitutional conduct of an individual employee.40
36
City of Canton, 489 U.S. at 391-92.
37
City of Springfield v. Kibbe, 480 U.S. 257, 267 (1987) (O’Connor, J., dissenting).
38
Id. at 268 (emphasis added).
39
City of Canton, 489 U.S. at 391–92.
40
Brown, 219 F.3d at 461.
18
No. 07-30443
DISCUSSION
A. Sufficiency of the Evidence—Culpability
The Brady violation here was a failure of one or more of the four assistant
district attorneys involved with Thompson’s armed robbery prosecution to turn
over the crime lab report to Thompson’s counsel.41 It is undisputed that the
District Attorney’s Office did not provide formal in-house training regarding
Brady.42 It is also undisputed that the assistant district attorneys were familiar
with the general rule of Brady that evidence favorable to the accused must be
disclosed to the defense.43 Thompson’s burden was to prove that Connick, the
41
Before the § 1983 trial, the parties stipulated that the failure to produce the lab
report constituted a Brady violation. As a stipulated fact, whether this actually constituted
a Brady violation is not now before the court.
After an extensive review of the record, this is the only Brady violation that has been
proven to have occurred during Thompson’s two trials. Thompson’s counsel had access to the
evidence lockup, where the physical blood evidence was clearly recorded on the evidence card
kept with the evidence. And Thompson’s claims of other Brady violations during his murder
trial lack merit. Some of these claimed violations were examined during Thompson’s initial
appeals and this court determined that they did not constitute Brady violations. See
Thompson v. Cain, 161 F.3d 802, 805-08 (5th Cir. 1998). Other claimed disclosure violations
have never been adjudicated to violate Brady.
Other than making conclusory assertions, Judge Prado fails to address any of these
other alleged violations in his opinion, or to indicate how non-disclosure of unrelated non-
Brady material is at all relevant to the need for, or Connick’s indifference to, Brady training.
42
The District Attorney’s Office stipulated to the fact that no prosecutor remembered
any formalized training on their Brady responsibilities.
43
The rule of Brady was directly written into the Louisiana Code of Professional
Responsibility:
With respect to evidence and witnesses, the prosecutor has responsibilities
different from those of a lawyer in private practice: the prosecutor should make
timely disclosure to the defense of available evidence, known to him, that tends
to negate the guilt of the accused, mitigate the degree of the offense, or reduce
the punishment. Further, a prosecutor should not intentionally avoid pursuit
of evidence merely because he believes it will damage the prosecution’s case or
aid the accused.
Articles of Incorporation of the La. State Bar Assoc., art. 16, EC 7-13 (1971); see also 21A LA.
REV. STAT. 213 (1974). And the American Bar Association included a rule on Brady when the
Model Rules of Professional Conduct were published in 1983. See MRPC R. 3.8(d) (1983) (“The
prosecutor in a criminal case shall . . . make timely disclosure to the defense of all evidence or
information known to the prosecutor that tends to negate the guilt of the accused or mitigates
19
No. 07-30443
policymaker for the Orleans Parish District Attorney’s office, was deliberately
indifferent to the need to train prosecutors in their Brady disclosure obligations.
As this circuit has recently recognized, “it is not enough for [Thompson] to
show that the [District Attorney’s Office’s] training program is, in a general
sense, wanting.”44 Instead, “the identified deficiency in a city’s training program
must be closely related to the ultimate injury.”45 The plaintiff must prove an
affirmative to the question: “Would the injury have been avoided had the
employee been trained under a program that was not deficient in the identified
respect?”46 Every prosecutor understood his general duty under Brady, so the
identified deficiency was not a failure to train on this general duty, but was
instead a failure to train on how to handle specific types of evidence such as the
crime lab report at issue.
There was evidence that some prosecutors doubted whether Brady itself
obligated the production of evidence that was not necessarily exculpatory.47 This
confusion seems to have arisen because the report itself had the potential to be
either exculpatory or inculpatory—depending on whether it matched Thompson’s
blood type.48 Accepting that there was no training on the Brady obligations
the offense . . . .”). So, in addition to being common knowledge to prosecutors, Brady was
written into the ethical rules as a duty incumbent upon every prosecutor.
44
Hinojosa v. Butler, 547 F.3d 285, 297 (5th Cir. 2008).
45
City of Canton, 489 U.S. at 391.
46
Id. (emphasis added).
47
This doubt is understandable considering that no court prior to 1985 appears to have
addressed the application of Brady under these circumstances. See infra note 52. The
testimony, however, at trial indicated that the District Attorney’s own policy was to turn over
such scientific reports, irrespective of Brady. Consequently, every prosecutor who testified
stated that he would have disclosed this crime lab evidence to the defense.
48
There is no indication that the prosecutors knew Thompson’s blood type at the time
of his attempted armed robbery trial.
20
No. 07-30443
pertaining to potentially exculpatory crime lab reports, we must determine
whether the need for that training was “so obvious” that a reasonable jury could
find that Connick was “deliberately indifferent” to that need.
Thompson did not show any pattern of similar Brady violations, and
instead relies exclusively on this single incident where prosecutors failed to
disclose his crime lab report. In another case before this court, we sustained the
district court’s conclusion that twenty-five years of records involving this District
Attorney’s Office (covering the time period of Thompson’s trial) revealed no
pattern of Brady violations.49 Connick testified that the District Attorney’s
Office handled tens of thousands of cases annually around this time, and that
in the ten years prior to Thompson’s case, only four convictions were overturned
based on Brady violations, none of which was similar to the violation at issue.50
So in only a minute number of cases were convictions overturned based on Brady
violations, and there was not a single instance involving the failure to disclose
a crime lab report or other scientific evidence. Connick was not alerted to a need
for further Brady training—especially not this specific type of
Brady training—by previous violations at the District Attorney’s Office.51
Nor has Thompson been able to refer us to a single reported opinion,
issued before this 1985 prosecution, from the Supreme Court, any of the circuit
49
See Cousin v. Small, 325 F.3d 627, 637 (5th Cir. 2003) (“[A] small number of cases
[with violations], out of thousands handled over twenty-five years, does not create a triable
issue of fact with respect to Connick’s deliberate indifference to violations of Brady rights.”).
50
According to Connick, none of these cases involved the intentional suppression of
evidence. In each case, the defense was aware of the existence of the evidence and requested
production. The trial court then, in each case, ruled in favor of the State. These rulings,
however, were all subsequently overruled by the Louisiana Supreme Court.
51
Judge Prado would refute this evidence with testimony from Connick that it was
possible that Brady violations may have occurred in other cases. Yet Thompson bore the
burden of proving Connick’s deliberate indifference; showing the mere possibility that other
violations may have occurred is not enough to show that Connick was put on notice that
additional training was required, let alone prove that the training was obviously necessary.
21
No. 07-30443
or district courts, or any state court that involved a similar Brady violation and
thus might have alerted Connick to the need for Brady training in this area.52
Thompson instead points to the following as substantial evidence that the
need for training in this area was “so obvious” that a failure to train constituted
deliberate indifference. First, Thompson argues that Connick testified that he
knew his prosecutors would frequently come into contact with Brady evidence.
Second, many prosecutors testified that the law regarding Brady contains “gray
areas.” Third, Thompson noted that several of the assistant district attorneys
were only a few years out of law school. Thompson also points to intra-office
discussions and opinions of various assistant district attorneys from 1999 and
later about whether the lab report was evidence covered by Brady.
This type of evidence amounts to no more than general observations that
would apply to any area of law and any number of district attorney’s offices
throughout the country. All district attorneys know that Brady issues—along
with many other areas of constitutional law—are routine matters that their
assistants handle every day. Every district attorney knows that nearly all issues
he deals with are shaded with “gray areas,” whether they concern Brady, search
and seizure, Miranda, evidence of a defendant’s other crimes, expert witnesses,
sentencing, or many more.53 Incorrect prosecutorial decisions in any of these
areas may lead to later reversal of convictions. Nearly all district attorney’s
offices employ prosecutors only a few years out of law school. That there were
different opinions about Brady evidence, or any other issue that may be raised
52
The only case pointed to as similar is a Tenth Circuit case decided in 1995, a decade
after Thompson’s trial, which held that physical evidence that may have had exonerating blood
on it should have been disclosed. See Smith v. Sec’y of N.M. Dep’t of Corr., 50 F.3d 801 (10th
Cir. 1995). This is the only case even arguably on point decided since Brady was issued in
1963.
53
While we might often wish for simple-to-apply, bright-line rules, they are the
exception rather than the norm, especially within our constitutional jurisprudence.
22
No. 07-30443
among lawyers, should surprise no one. All of this evidence involves generic
generalizations—not the type of exacting evidence required to show that Connick
and the District Attorney’s Office were deliberately indifferent to an obvious
need to further train its professional prosecutors. To the extent that this
evidence could be used to show that the municipality’s training was, in a general
sense, wanting, similar evidence could also support a deliberate indifference
finding against any prosecutor’s office for nearly any error that leads to a
reversal of a conviction.
We cannot accept the argument that generalized failure to train evidence
sustains a finding of official deliberate indifference. In Pineda v. City of
Houston,54 Judge Higginbotham squarely rejected a plaintiff’s argument that
“‘because the City has admitted that specialized training is required for officers
in such situations [specialized narcotics investigations], there is sufficient
evidence that the training was inadequate.’ . . . No butterfly will emerge from
this hollow chrysalis of an argument.”55
Pineda relies on City of Canton, which displayed utmost caution toward
generalized failure to train evidence:
Neither will it suffice to prove that an injury or accident could have
been avoided if an officer had had better or more training, sufficient
to equip him to avoid the particular injury-causing conduct. Such
a claim could be made about almost any encounter resulting in
injury, yet not condemn the adequacy of the program to enable
officers to respond properly to the usual and recurring situations
with which they must deal. And plainly, adequately trained officers
occasionally make mistakes; the fact that they do says little about the
training program or the legal basis for holding the city liable.56
54
291 F.3d 325 (5th Cir. 2002).
55
Id. at 334 n.35.
56
City of Canton, 489 U.S. at 391 (emphasis added).
23
No. 07-30443
Because this case concerns the actions of licensed attorneys who have
independent professional obligations to know and uphold the law, there is even
more reason than in City of Canton or Pineda not to rely on generalized
statements about lack of training. Training is what differentiates attorneys
from average public employees. A public employer is entitled to assume that
attorneys will abide by the standards of the profession, which include both
ethical and practical requirements. Thus, prosecutors are personally responsible
as professionals to know what Brady entails and when to perform legal research
to understand the “gray areas.”57 To hold a public employer liable for failing to
train professionals in their profession is an awkward theory. By analogy, it is
highly unlikely that a municipality could be held liable for failing to train a
doctor it employed in diagnostic nuances. Mere nostrums about training in
Brady, a basic due process principle of criminal procedure, will not suffice.
While Thompson has failed to produce any specific evidence relating to the
obvious need for training on the identified deficiency, there is sufficient evidence
to show that the need for training in this area was in fact not “so obvious.” This
situation—with scientific evidence providing a blood type or other indicator of
the perpetrator’s identity—occurs every time a crime lab report is prepared and
blood or other scientific evidence has not been taken from the defendant. And
yet Thompson failed to show any similar failures to disclose crime lab reports
from this District Attorney’s Office either before his trial or since. The alleged
failure to train extended over a long period of time, during which hundreds or
thousands of crime lab reports were prepared.
As a matter of probability, if violations were the “highly predictable
consequence” of a failure to train, then we would expect to see more than just
one violation in hundreds or thousands of cases. Thompson has, as a legal
57
See Cousin, 325 F.3d at 638 (“[P]rosecutors exercise independent judgment in trying
a case, and they have the legal and ethical obligation to comply with Brady.”).
24
No. 07-30443
matter, failed to prove that his violation was the “highly predictable
consequence” of failing to train prosecutors. This means that the need for
training was not “so obvious,” and thus that Connick was not “deliberately
indifferent” to Thompson’s constitutional rights.58
Indeed, three witnesses testified that the District Attorney’s Office had a
policy of disclosing all crime lab reports. Dubelier testified that turning over
crime lab reports was “standard operating procedure in the office.”59 Williams
testified that “[y]ou have to turn over any scientific evidence that you have.”
And Assistant District Attorney Val Solino testified that turning over crime lab
reports was “the policy in the office.”60 While there was some confusion as to
Brady obligations, every single witness who was asked stated that they would
have disclosed the crime lab report had they known about it. This testimony was
uncontradicted and unimpeached.61 Because the District Attorney’s Office had
58
As our court held in a similar case: “Because Burge failed to establish the existence
of a single prior Brady violation, let alone demonstrate a pattern of violations sufficient to
demonstrate deliberate indifference on the part of the sheriff, we find that no reasonable jury
could have concluded that Sheriff Strain in his official capacity was deliberately indifferent to
Burge’s right to a fair trial.” Burge, 336 F.3d at 373.
59
Dubelier testified that: “If this report was not turned over, I didn’t see it.” He stated
that: “I prosecuted thousands of cases when I was a DA and turned over thousands of these
types of reports. If I had the report, I would have turned it over.” Dubelier continued by
stating that it was standard procedure to disclose crime lab reports: “[W]e were obligated to
turn over a crime lab report. That’s the way it was. That was standard operating procedure
in the office. There are, again, records, I am sure, in the office of hundreds, if maybe not more
cases where I turned over these type of reports. So my practice would have been to turn over
the report.”
60
Solino was the Federal Rule of Civil Procedure 30(b)(6) representative in this case for
the District Attorney’s Office. He was not employed in Connick’s office at the time of the
Thompson attempted armed robbery trial. Solino testified that “I would have expected a
prosecutor to turn that lab report over, period. That’s what I would have done and that’s what
I would [have] expected them to have done under the policy in the office as I understood it at
the time.”
61
See Reeves, 530 U.S. at 150–51. Judge Prado states that this evidence of a policy of
disclosing scientific reports is contradicted by uncertainty about whether the report at issue
was Brady material. But evidence of an open-file policy would not be contradicted by
25
No. 07-30443
a policy of disclosing all crime lab reports, there was no need to train on the
specifics of which reports would or would not be covered by Brady. Just as a
municipal policymaker could not be found deliberately indifferent to citizens’
Brady rights if the policymaker established clear policies—such as an open-file
policy—to protect those rights, Connick cannot be considered “deliberately
indifferent” to a Brady violation based on a failure to disclose a crime lab report
when his employees generally understood that they had to disclose exactly those
types of reports.
For these reasons, under Monell, City of Canton, and Bryan County, the
evidence in this record does not support the conclusion that Connick was
deliberately indifferent to an obvious need for training. Consequently, the
District Attorney cannot be held liable for the failure by his employees to
disclose this crime lab report.
B. Sufficiency of the Evidence—Causation
Nor does the diffuse evidence of Brady misunderstanding among several
assistant district attorneys satisfy the causation requirements relating to the
violation at issue here. The specific question we ask is whether Connick’s failure
to provide in-house training was the moving force behind the failure to turn over
the lab report. Stated differently, the question is whether, under the teachings
of City of Canton and Bryan County, unfamiliarity with Brady obligations with
respect to this lab report was the actual cause—the moving force—of this
constitutional violation. This standard of causation must be established by
uncertainty over whether any particular piece of the file was Brady material. In fact, many
District Attorney’s Offices have blanket disclosure policies specifically so that prosecutors need
not make difficult Brady decisions. Judge Prado then goes on to state that this evidence is
contradicted by evidence of a policy not to disclose certain police reports and witness
statements. But there is nothing contradictory in having a policy to disclose one type of report
alongside a policy not to disclose an entirely different type of report. Judge Prado is
contradicting apples with oranges; there is simply no evidence in the record which refutes the
testimony of these witnesses that there was a policy of disclosing lab reports and
scientific evidence.
26
No. 07-30443
substantial evidence. If Thompson did not submit substantial evidence that the
failure to produce the lab report was caused by confusion over Brady, the jury
could not have reasonably concluded that the lack of training was the direct
cause of Thompson’s injury, and judgment as a matter of law is required.
The record leaves many unanswered questions about the actual cause of
this constitutional violation. It is, however, clear that four assistant district
attorneys were involved in the failure to produce the lab report. They were
Bruce Whittaker, James Williams, Gerry Deegan, and Eric Dubelier. The failure
to produce the lab report lies with one or more of these assistant district
attorneys. We turn now to examine what the record reflects as to these
assistants.
Whittaker, as the armed robbery “screener,” was responsible for initially
reviewing the police file on Thompson, deciding whether to prosecute, and
assigning the case to the correct division. He “screened” Thompson’s file on
February 25, 1985—approximately five weeks after Thompson had been
arrested. The police report indicated that evidence had been collected from the
crime scene that possibly contained the perpetrator’s blood. Accordingly,
Whittaker wrote “May wish to do blood tests,” on the screening action form.
After Whittaker screened the case, Dubelier, an experienced assistant
district attorney, was assigned as lead prosecutor, and Deegan was assigned as
the junior assistant on the case. Dubelier had no independent recollection of the
armed robbery prosecution that had occurred twenty years earlier, but the
record indicates that he and Deegan handled most of the pre-trial work.
Williams did not become involved with the case until March 11 when Dubelier
asked him to handle a pre-trial evidentiary hearing. During this hearing,
Williams, based on his review of the screening action form, announced to the
court and Thompson’s counsel that the prosecution intended to test Thompson’s
blood. However, no such test was ever ordered by the prosecution. Later,
27
No. 07-30443
sometime before the April 11 trial,62 Dubelier asked Williams to take over trial
responsibility for the case.
There is no testimony in the record from Deegan, who died several years
before Thompson instituted this action. The record does contain, however, an
affidavit and testimony from his colleague and close friend Michael Riehlmann.
According to Riehlmann, shortly after Deegan was diagnosed with terminal
cancer, Deegan confessed “that he had intentionally suppressed blood evidence
in the armed robbery trial of John Thompson that in some way exculpated the
defendant.”63 Deegan’s confession is, in part, supported by the evidence card, a
card on file that identified the physical evidence in the case. The card indicates
that the morning the trial was set to begin, Deegan checked out the blood
evidence and never returned it.
As for the lab report itself, the record is not clear who ordered the testing.
The report was dated April 9, just two days before trial, and addressed to
Whittaker. Whittaker recalled seeing the report and placing it on Williams’s
desk. Because he was only the screening prosecutor, Whittaker was not
responsible for turning it over to the defense. Williams and Dubelier both
testified that they never saw the report. Riehlmann’s account of Deegan’s
confession is not clear whether Deegan was referring to the lab report, to the
actual blood evidence, or to both.
The statements of these four assistant district attorneys—the only
prosecutors who had any involvement in the armed robbery case—provide very
little information regarding the lab report. What is clear from the record is this:
first, Whittaker received the report a few days before trial; second, no prosecutor
62
Williams testified that Dubelier asked him to take over the case just a few days before
trial. Dubelier, however, could not recall when he asked Williams to take over, and he testified
that it would have been very odd to make such a change only a few days before trial.
63
At trial, Riehlmann’s testimony was more equivocal: he stated that Deegan told him
“that he had failed to turn over stuff that might have been exculpatory.”
28
No. 07-30443
ever turned it over to Thompson’s counsel; and third, the report did not appear
again until it was discovered fourteen years later.
Thompson based his case upon a single causation theory: that one or more
of the assistant district attorneys involved in Thompson’s prosecution decided
not to turn over the report because they did not know that they were legally
obligated to produce it and that training sessions on Brady would have avoided
this incident.64 To prove his theory, Thompson must present substantial
evidence from which a jury reasonably could conclude that the failure of Connick
to provide training sessions on Brady was the actual cause of and the moving
force behind the failure to produce the report. The precedents require
substantial evidence of direct causation. This standard demands more than
evidence of confusion over Brady’s “gray areas” in the District Attorney’s Office.
Finally, Thompson must establish that this lack of understanding would have
been remedied by an in-house training program.
Thompson’s brief fails to point out any such evidence to sustain municipal
liability. As best we understand his brief, the only arguments he makes
regarding causation are these: (1) the record supports the conclusion that these
four prosecutors knew about the blood evidence and yet failed to disclose it;65 and
(2) the jury was free to reject Connick’s theory of a single rogue prosecutor. Even
if we accept both of these assertions as correct, they still fail completely to
establish that the Brady violation was caused by unfamiliarity with Brady. And
64
Yet, we must observe that the record does not reflect whether a decision to produce,
or not produce, the report was ever made, or whether the report was misplaced or overlooked,
or whether one or more of the four prosecutors assumed it would be handled by someone else.
65
For example, Thompson’s brief states, “It is also not surprising that at least four
prosecutors—Dubelier, Williams, Whittaker, and Deegan—knew about the blood evidence yet
each failed to disclose it.” And, “Dubelier, the special prosecutor in charge of both cases, knew
but chose not to reveal there was blood evidence.” And, “Based upon that evidence, the jury
was free to infer that both Whittaker and Williams received the crime lab report, but did not
produce it.”
29
No. 07-30443
because Thompson bore the burden of proof, he had to do more than simply
assert that the jury was free to reject Connick’s explanations for the violation.
Thompson had to put forth substantial evidence supporting his own theory of
causation: that the assistant district attorney (or attorneys) responsible for the
constitutional violation did not understand Brady, that this lack of
understanding caused the failure to produce the report, and that Brady training
could have resolved this lack of understanding.
We have reviewed the record for any such evidence. First, it contains
evidence that Williams, when asked if Brady material includes documents that
could be used to impeach a government witness, incorrectly replied “No.”66
Second, the 1987 policy manual from the District Attorney’s Office could be read
to imply that Brady evidence need only be produced when the defense requests
it and it fails to note that impeachment evidence is also included under Brady.
Third, Solino and Connick, after the report was discovered, contended that the
lab report was not subject to Brady as such because Thompson’s blood type was
unknown and the report thus had no exonerative effect on Thompson’s guilt.
Fourth, although Williams stated unequivocally that all technical or scientific
reports, like the lab report, were required to be turned over to a defendant, he
also testified that this obligation did not necessarily arise from Brady because
the report was not necessarily exculpatory.67
66
See Giglio v. United States, 405 U.S. 150 (1972) (holding that Brady requires the
production of evidence that could be used to impeach a government witness).
67
Thompson seeks to bolster the evidence of a link between the lack of formal
Brady training and non-disclosure of the lab report with evidence of other allegedly illegal
failures by the prosecutors to turn over evidence to his counsel in the murder case. This
attempt fails for several reason. First, as has been noted, supra n.41, none of the other non-
disclosures has been held to violate Brady—or any other legal rule. It is a non-sequitur to say
that failure to train on Brady had anything to do with failure to disclose non-Brady evidence.
Second, without proof of the illegality of the nondisclosures, Thompson cannot rely on them to
prove systemic lack of training about Brady obligations. Third, the more general notion of a
“culture of indifference” toward a district attorney’s disclosure obligations is a description
30
No. 07-30443
The record fails to establish, by substantial evidence, that the actual cause
and moving force behind the constitutional violation of not producing the lab
report was the failure of the District Attorney to have in-house training sessions
on Brady. For example, an assistant district attorney’s confusion regarding
whether Brady applied to impeachment evidence may show a need for
enlightening this assistant but is irrelevant here because the lab report clearly
was not impeachment evidence and would not have been turned over on that
basis. The policy manual, although incomplete in its instructions on Brady
evidence and post-dating Thompson’s trial by several years, does little to
establish the necessary direct causal link, and the jury concluded in its verdict
that the violation was not due to an established municipal policy.
Thus, even assuming that Connick was deliberately indifferent to a need
for training, Thompson failed as a matter of law to show that the lack of training
was the actual cause of the constitutional violation.68 Therefore the judgment
should be reversed and rendered for the defendant.
C. Jury Instructions
The jury was probably misled in its decision by the district court’s plainly
erroneous jury instructions. After several hours of deliberations, the jury sent
out a single question:
rather than a legal rule. It has no support in caselaw and is inconsistent with the lack of
pattern evidence in this case. Finally, if there was or could be municipal culpability founded
on a “culture of indifference” to Brady violations, it speaks to the District Attorney’s
policy—but the jury rejected Thompson’s unconstitutional policy theory!
68
Judge Prado states that we have ignored evidence about causation. But a careful
review of the panel opinion reveals that while Judge Prado discusses extensively how the jury
was free to reject Connick’s alternative arguments, he fails to point to substantial evidence
from which the jury could conclude that a failure to train was the moving force behind this
violation. Considering the record as a whole, as we must, and keeping in mind that Thompson
bore the burden of showing such substantial evidence, we conclude that Thompson has failed
to present sufficient evidence for any reasonable jury to conclude that Connick’s deliberate
indifference was the moving force behind this Brady violation.
31
No. 07-30443
What does “Deliberate” Indifference mean? Does it mean intentional
or would “Failure to monitor” be considered Deliberate?
The district court responded that:
“Deliberate Indifference” does not necessarily mean intentional, but
does require more than mere negilgence [sic] or even gross
negligence.69
The district court’s answer lacked “concrete accuracy,”70 and instead defined
deliberate indifference as something less than intent but more than
negligence—a nebulous answer that failed to sufficiently inform the jury of the
controlling law.71 Nor was there any clarification in the original jury
instructions, which failed to state that municipal liability requires a “conscious”
or “deliberate” choice on the part of the policymaker.
By failing to clearly define “deliberate indifference,” the instructions were
confusing and misleading. The instructions erroneously omitted the clear and
obvious requirement laid down by the Supreme Court in City of Canton that
“[o]nly where a failure to train reflects a ‘deliberate’ or ‘conscious’ choice by a
municipality—a ‘policy’ as defined by our prior cases—can a city be liable for
such a failure under § 1983.”72 This en banc court has previously held that
69
Nowhere did the district court give the legal definitions for “negligence” or “gross
negligence.” So while this instruction provides as least some minor guidance to the trained
legal mind, it provides extraordinarily little clarity to the average juror.
70
United States v. Stevens, 38 F.3d 167, 169–70 (5th Cir. 1994) (“When a deliberating
jury expresses confusion and difficulty over an issue submitted to it, the trial court’s task is to
clear that confusion away with concrete accuracy.” (quotation omitted)).
71
Even Judge Prado’s arguments and selective quotations, while not fully reflecting the
case law of either the Supreme Court or this circuit, would have provided greater guidance to
this uncertain jury than the answer given by the district court.
72
489 U.S. at 389 (internal quotation omitted). See also City of Oklahoma City v.
Tuttle, 471 U.S. 808, 823 (1985) (plurality) (“[T]he word ‘policy’ generally implies a course of
action consciously chosen from among various alternatives; . . . evidence [must] be adduced
which proves that the inadequacies resulted from conscious choice—that is, proof that the
policymakers deliberately chose a training program which would prove inadequate.” (emphasis
added)); Estate of Davis ex rel. McCully v. City of North Richland Hills, 406 F.3d 375, 383 (5th
32
No. 07-30443
deliberate indifference is a “lesser form of intent.”73 The district court failed in
its duty to incorporate these governing rulings when it issued jury instructions.
Although the plain error standard limits appellate review here because no
proper objection was made to the instructions, we believe that standard was
met.74 The district court’s failure to correctly instruct the jury was clearly and
obviously inconsistent with the law, and this error affected the Defendant’s
substantial rights by allowing liability without any actual municipal
“culpability.”
Finally, plain error is to be corrected if the “error seriously affects the
fairness, integrity or public reputation of judicial proceedings.”75 “The jury
system is premised on the idea that rationality and careful regard for the court’s
instructions will confine and exclude jurors’ raw emotions.”76 As noted above,
the Supreme Court has specifically warned that reducing the standards of fault
in municipal liability cases would “impose de facto respondeat superior liability”
and raise “serious questions of federalism.”77 Correctly instructing the jury on
the applicable standard of fault is particularly important in municipal liability
cases which involve the public purses of our cities and local governments. In this
case, nebulous jury instructions authorized a verdict manifestly unfair to these
Defendants and plainly inconsistent with the Supreme Court’s and our relevant
Cir. 2005) (“[A] showing of deliberate indifference requires that the Plaintiffs show that the
failure to train reflects a ‘deliberate’ or ‘conscious’ choice to endanger constitutional rights.”
(quotation omitted)).
73
Doe v. Taylor Indep. Sch. Dist., 15 F.3d 443, 453 n.7 (5th Cir. 1994) (en banc).
74
See Puckett v. United States, 129 S. Ct. 1423, 1429 (2009); see also Wright v. Ford
Motor Co., 508 F.3d 263, 272 (5th Cir. 2007).
75
Puckett, 129 S. Ct. at 1429 (quotation and alteration omitted).
76
CSX Transp., Inc. v. Hensley, 129 S. Ct. 2139, 2141 (2009).
77
See City of Canton, 489 U.S. at 392; see supra notes 35–36 and accompanying text.
33
No. 07-30443
precedent. The verdict also undermines the will of Congress expressed in §
1983.78
We urge this point in further explanation of this unjustifiable verdict and
to discourage other district courts from making similar mistakes.
CONCLUSION
Judgment as a matter of law “is a method for protecting neutral principles
of law from powerful forces outside the scope of law—compassion and
prejudice.”79 We fully appreciate that Thompson has suffered a horrible wrong
inflicted by agents of the government and that in many cases the principal would
be responsible for the acts of these agents. But as Judge Wisdom counseled
when overturning a jury verdict, “[i]n reviewing [a] . . . case when the plaintiff
has been injured grievously, hard as our sympathies may pull us, our duty to
maintain the integrity of substantive law pulls harder.”80 The Supreme Court
has stated clearly and emphatically that the liability of municipalities is limited
to cases where a municipal action caused the constitutional violation. The
plaintiff must show the “requisite degree of culpability” on the part of the
municipality—deliberate indifference to an obvious need for training—and must
demonstrate a “direct causal link” between the failure to train and the
constitutional violation.81
78
See Bryan County, 520 U.S. at 400 (“[I]n enacting § 1983, Congress did not intend to
impose liability on a municipality unless deliberate action attributable to the municipality itself
is the ‘moving force’ behind the plaintiff’s deprivation of federal rights.”).
79
Rutherford v. Ill. Cent. R.R. Co., 278 F.2d 310, 312 (5th Cir. 1960).
80
Turner v. Atl. Coast Line R.R. Co., 292 F.2d 586, 589 (5th Cir. 1961). Judge Prado
rightly puts great faith in the Seventh Amendment right to a jury trial. But we also have a
duty to maintain the integrity of our substantive law. Regrettably, Judge Prado has chosen
not to engage with many of these arguments, or with the substantive law more generally,
despite the Supreme Court’s clear warnings that we must preserve the heightened evidentiary
standards of municipal liability. See supra notes 35–36 and accompanying text.
81
Bryan County, 520 U.S. at 404.
34
No. 07-30443
Thompson failed to produce substantial evidence to support his claim that
the District Attorney was deliberately indifferent to an obvious need for training
of his staff. And he failed to produce adequate evidence of causation to show
that the failure to train was the actual cause and moving force behind the failure
to produce the lab report. Thompson has, in short, failed to meet the heightened
standards for culpability and causation imposed by Monell, City of Canton, and
Bryan County, and we would therefore reverse the district court’s judgment.
35
No. 07-30443
PRADO, Circuit Judge, with whom KING, WIENER, STEWART, and ELROD,
Circuit Judges, join:
“The right of jury trial in civil cases at common law is a basic and
fundamental feature of our system of federal jurisprudence which is protected
by the Seventh Amendment. A right so fundamental and sacred to the citizen,
whether guaranteed by the Constitution or provided by statute, should be
jealously guarded by the courts.” Jacob v. City of New York, 315 U.S. 752, 752-
53 (1942).
The panel opinion thoroughly explains why the evidence the jury heard in
this case is sufficient to support its verdict. See Thompson v. Connick, 553 F.3d
836 (5th Cir. 2008). Judge Clement’s dissent (“the dissent”) to this court’s order
affirming based on a tie en banc vote, however, overlooks much of the evidence
the jury heard and ignores the standard of review that we apply to jury verdicts.
By reading the dissent, one would be hard pressed to even realize that a
jury rendered the verdict in this case. At the outset, the dissent attempts to
explain the standard of review but fails to acknowledge the deference we must
accord to a jury’s verdict. We have repeatedly admonished that
our standard of review with respect to a jury verdict is especially
deferential. As such, judgment as a matter of law should not be
granted unless the facts and inferences point so strongly and
overwhelmingly in the movant’s favor that reasonable jurors could
not reach a contrary conclusion.
Flowers v. S. Reg’l Physician Servs. Inc., 247 F.3d 229, 235 (5th Cir. 2001)
(internal quotation marks and citations omitted). “A jury verdict must be upheld
unless there is no legally sufficient evidentiary basis for a reasonable jury to find
as the jury did.” Int’l Ins. Co. v. RSR Corp., 426 F.3d 281, 296-97 (5th Cir. 2005)
(internal quotation marks omitted). We must view the evidence the jury heard
with this deferential standard in mind.
36
No. 07-30443
A review of the full record—as laid out in the panel opinion—reveals that
the dissent is merely quibbling with the jury’s factual findings. See Thompson,
553 F.3d at 843-46. This oversteps our bounds as an appellate court. The
dissent presents nothing more than a skewed version of the facts in favor of the
District Attorney’s Office. Its approach is directly contrary to the rule that we
must view all evidence and draw all reasonable inferences in favor of the jury’s
verdict. See United States v. Miles, 360 F.3d 472, 476-77 (5th Cir. 2004); Am.
Cas. Co. of Reading, Pa. v. Myrick, 304 F.2d 179, 182 (5th Cir. 1962) (“[I]n
reviewing a jury’s verdict a court may not substitute its judgment on the facts
for the jury’s determination simply because inconsistent and uncertain
inferences are equally supported by the proof.”). The dissent thus ignores the
maxim that the jury is allowed to accept or reject competing evidence. See
Reeves v. Sanderson Plumbing Prods., Inc., 530 U.S. 133, 150 (2000) (noting that
in reviewing a jury’s verdict an appellate court “may not make credibility
determinations or weigh the evidence”).
For example, the dissent states that evidence regarding whether Connick
had a policy to disclose crime lab reports was “uncontradicted and
unimpeached.” See dissent at 25-26. However, ADA James Williams testified
that he did not have a Brady duty to disclose crime scene technician reports and
equivocated regarding the disclosure of blood reports if he did not know the
perpetrator’s blood type, which is the exact situation he faced when prosecuting
Thompson. R. at 2353-54 (suggesting that the crime lab report was not Brady
material “because I didn’t know what the blood type of Mr. Thompson was, and
I didn’t know what the blood type of Mr. LaGarde [the victim of the robbery]
was”). Similarly, Williams explained that Connick had a policy not to disclose
certain police reports and witness statements, contradicting his testimony that
he had to turn over all written reports generated in a case. Compare R. at 2027,
with R. at 2354. Val Solino, the DA’s Office’s official Rule 30(b)(6)
37
No. 07-30443
representative, stated that under Connick’s Brady policy, even as later
memorialized in a 1987 manual, an ADA in Connick’s office would not have had
to produce a crime lab report if he did not know the defendant’s blood type. R.
at 2874-75. As another example, the dissent minimizes the evidence the jury
heard regarding the District Attorney’s other Brady violations in this and other
cases. The dissent manages to highlight Connick’s testimony that the Louisiana
Supreme Court had overturned only four convictions based on a Brady violation
during his tenure as District Attorney, but fails to acknowledge Connick’s
concession that because most Brady violations typically do not lead to published
opinions, Brady violations had likely occurred in other cases as well. R. at 2823-
25. Similarly, by discounting the numerous other examples of nondisclosure
that occurred in this case (in addition to the nondisclosure of the crime lab
report)—examples that Thompson presented to the jury without objection—the
dissent ignores the fact that the jury could consider these other events1 as
further proof of the need for and Connick’s indifference to training on Brady.
In another attempt to overturn the jury’s verdict, the dissent simply
ignores evidence about causation. As discussed fully in the panel opinion, the
evidence permitted the jury to find that Connick’s deliberate indifference caused
the Brady violations in this case. See Thompson, 553 F.3d at 853-57. Thus, the
only way the dissent can reach its desired result is by departing from our
deferential standard, reading the record selectively to support its position, and
substituting its own judgment for that of the jury’s.
1
These other examples included, inter alia, the nondisclosure of eyewitness statements.
As this case evidences, many Brady violations are not uncovered until years after the event,
if they are ever uncovered. In Kyles v. Whitley, 514 U.S. 419 (1995), the Supreme Court
concluded that the nondisclosure of similar eyewitness statements by this very office within
a year prior to Thompson’s first murder trial violated Brady. These statements did not come
to light until five years after Kyles’s trial.
38
No. 07-30443
Finally, the dissent acknowledges that Connick did not preserve his
objection to the jury instructions but still attempts to support a conclusion that
the district court plainly erred in its instruction on deliberate indifference. As
the panel opinion explains, however, the district court’s instructions were legally
correct—and certainly not so “clearly” or “obviously” wrong as to constitute plain
error. See Thompson, 554 F.3d at 859-63.
Deliberate indifference and intent are not synonymous. They are instead
separate, albeit sometimes legally equivalent, concepts. See Bd. of the County
Comm’rs of Bryan County v. Brown, 520 U.S. 397, 419 (1997) (“Deliberate
indifference is thus treated, as it is elsewhere in the law, as tantamount to
intent, so that inaction by a policymaker deliberately indifferent to a substantial
risk of harm is equivalent to the intentional action that setting policy
presupposes.”). One need not show actual intent for a jury to find liability under
the deliberate indifference standard. In Hope v. Pelzer, 536 U.S. 730 (2002), the
Supreme Court clarified that the fact-finder can infer deliberate indifference
“from the fact that the risk of harm is obvious.” Id. at 738 (citing Farmer v.
Brennan, 511 U.S. 825, 842 (1994)); see also Farmer, 511 U.S. at 841 (noting that
the test from City of Canton v. Harris, 489 U.S. 378 (1989), for deliberate
indifference is an “objective standard”). This court also has stated that
deliberate indifference is “a lesser form of intent.” Doe v. Taylor Indep. Sch.
Dist., 15 F.3d 443, 453 n.7 (5th Cir. 1994) (internal quotation marks omitted)
(emphasis added). A finding of deliberate indifference is thus tantamount to a
finding of intent in the context of municipal liability, but a finding of deliberate
indifference does not require a finding of intent—which is exactly what the
district court instructed the jury in light of this precedent. See Thompson, 553
F.3d at 859-63. The dissent improperly adds this intent element, thereby
recasting the meaning of deliberate indifference so as to assign error to the
district court.
39
No. 07-30443
At bottom, the dissent seeks to retry this case through the appellate
process. This approach abdicates this court’s duty to uphold a jury’s verdict
unless the facts point so strongly in Connick’s favor that no reasonable jury
could rule to the contrary. See Flowers, 247 F.3d at 235. Indeed, the fact that
reasonable judges on this court view the evidence differently suggests that these
factual disputes were for the jury to resolve. As the extensive discussion in the
panel opinion demonstrates, there was ample evidence to allow the citizens of
this New Orleans jury to find for Thompson. Of course, this is an extraordinary
case with extraordinary facts. Allowing this judgment to stand will not subject
municipalities to widespread liability, as a holding that the need for training was
“so obvious” and the lack of training “so likely” to create a constitutional
violation will apply only in the rare instance. This is that rare case. The jury
heard substantial evidence that the District Attorney’s Office provided no Brady-
specific training, despite the known risk of the exact type of systemic
nondisclosure that the failure to train caused here. Acknowledging the proper
standard of review and viewing the jury’s verdict in the correct deferential light
compels us to uphold the jury’s decision.
40
41