Cannon et al v. BP PRODUCTS NORTH AMERICA, INC.
Filing
97
MEMORANDUM AND ORDER denying 26 Opposed MOTION to Certify Class, granting 36 MOTION to Exclude Proposed Expert Testimony of Dr. Robert A. Simons, denying as moot 35 MOTION to Exclude Proposed Expert Testimony of Dr. Paul Rosenfeld (Signed by Judge Gregg Costa) Parties notified.(arrivera, 3)
UNITED STATES DISTRICT COURT
SOUTHERN DISTRICT OF TEXAS
GALVESTON DIVISION
KYLE CANNON, et al,
Plaintiffs,
VS.
BP PRODUCTS NORTH AMERICA,
INC.,
Defendant.
§
§
§
§
§
§
§
§
§
CIVIL ACTION NO. 3:10-CV-00622
MEMORANDUM AND ORDER
This proposed class action arises from allegations that numerous chemical
releases and emissions events that occurred at Defendant BP Products North
America’s Texas City Refinery1 after December 22, 2008, caused thousands of
surrounding residential properties to decrease in value.
Plaintiffs—five
homeowners in the Texas City area—bring common law claims of negligence,
trespass to property, and nuisance. They now seek certification of a class of:
All persons who own or have owned any piece of real property
classified as residential property, in the area (“Class Area”) identified
as affected by the air pollution plume of impact (“Plume”) modeled
by Dr. Paul Rosenfeld in his report of Jan. 9, 2012, and shown on
Figure ES.1 to Dr. Rosenfeld’s report (and attached as Exhibit B),
since December 22, 2008.
1
BP completed a sale of the Texas City Refinery to Marathon Petroleum Corporation on
February 1, 2013. The Court nonetheless will refer to the refinery as BP’s given its ownership
during the applicable period.
1 / 42
Docket Entry No. 26 at 20. The proposed class area includes roughly 14,300
residential parcels in Texas City and La Marque.
Plaintiffs’ motion requires the Court to determine whether the proposed
class meets the requirements of Federal Rule of Civil Procedure 23. But Plaintiffs’
motion also requires the Court to evaluate the reliability and sufficiency of their
two experts, Dr. Paul Rosenfeld and Dr. Robert Simons, upon whom they rely to
create their class model. Dr. Rosenfeld, an environmental chemist, conducted a
preliminary evaluation of air pollution emissions from the Refinery for 2009 and
2010. He modeled BP’s sulfur dioxide (SO2) emissions during that timeframe and
generated a plume of impact where the SO2 emissions reached a certain threshold.
The plume defines Plaintiffs’ proposed class area. Dr. Simons’s opinions form the
basis of Plaintiffs’ causation and damages theories. Dr. Simons, a real estate
economist, conducted a hedonic regression analysis, a real estate trends analysis, a
contingent valuation analysis, and property owner surveys, and concluded that
BP’s airborne chemical releases resulted in permanent economic losses to all
residential class properties, ranging between 5% and 20% of the property value.
As explained in more detail below, the Court finds that Dr. Simons’s
opinions are unreliable, and, accordingly GRANTS BP’s motion to exclude his
testimony. Left without a formulaic causation and damages model, Plaintiffs are
unable to show that questions of law or fact common to the class predominate over
2 / 42
individual ones, as is required by Rule 23(b)(3) under the present circumstances.
Accordingly, the Court DENIES Plaintiffs’ motion to certify.
I.
BACKGROUND
The Texas City Refinery, with more than twenty processing units and a
refining capacity of more than 460,000 barrels per day, is the third largest refinery
in the country. Docket Entry Nos. 34 at 5; 49 at 2. The Refinery has been in
operation since 1934. BP acquired it in 1998 as part of its merger with Amoco.
Though BP owned and operated the Refinery when Plaintiffs filed their complaint,
BP subsequently sold the Refinery to Marathon Petroleum Corporation on
February 1, 2013. See Press Release, BP, BP Completes Sale of Texas City
Refinery and Related Assets to Marathon Petroleum (Feb. 1, 2013), available at
http://www.bp.com/en/global/corporate/press/press-releases/bp-completes-sale-oftexas-city-refinery-and-related-assets-to-marathon-petroleum.html
(last
visited
Aug. 29, 2013).
Plaintiffs paint a picture of a plant that, at least in 2009 and 2010, was poorly
run and polluting at dangerous levels.
According to Plaintiffs, the Refinery
reported more total toxic air emissions in those years than any other refinery in the
United States. Docket Entry No. 49 at 3. The Galveston County Health District
purportedly received numerous odor, air quality, and property impact complaints
3 / 42
from residents who noted that fumes were burning the eyes, making them
nauseated, and depositing white, oily substances on their vehicles. Id.
While Plaintiffs’ claims cover all chemical releases from the Refinery,
including normal ones, Plaintiffs focus much of their complaint on the existence of
“emissions events,” which the Texas Administrative Code defines as “[a]ny upset
event or unscheduled maintenance, startup, or shutdown activity, from a common
cause that results in unauthorized emissions of air contaminants from one or more
emissions points at a regulated entity.”
30 Tex. Admin. Code § 101.1(28).
Plaintiffs allege that data from the Texas Commission on Environmental Quality
shows that in 2009–10, the Refinery had over 70 “reportable emissions events,”
i.e., emissions events that in any 24-hour period result in emissions exceeding
thresholds defined by the statute. Id. § 101.1(87) (defining “reportable emissions
event”); see also id. § 101.1(88) (defining thresholds for reporting). During those
emissions events, the Refinery purportedly released approximately 1,204,000
pounds of pollution over roughly 2800 hours. Docket Entry No. 49 at 4.
As BP points out, the reported emissions events “varied widely as to the
source within the facility, the type and quantity of substance emitted, the duration
of the emission, the cause, and the weather and wind conditions at the time of the
emission.” Docket Entry No. 34 at 18 (citing emission event reporting database).
But Plaintiffs highlight one event that it refers to as the “most notorious release,”
4 / 42
which lasted between April and May 2010 and started when a fire broke out on the
100-J compressor at the Ultracracker—a unit used for production of light fuels
such as gasoline. Docket Entry No. 49 at 4. According to Plaintiffs, BP shut down
the Ultracracker when it discovered the fire, but decided to restart the unit before
the compressor could resume operation. Id. Thus, with compressor down, the
Refinery could not safely process noxious chemicals and had to send them to a
flare, which Plaintiffs contend was technologically antiquated and vastly
inefficient. Consequently, the Refinery released 514,000 pounds of pollutants into
the air during the 40-day event. Id. Plaintiffs allege that BP was not only at fault
for restarting the unit prematurely, but also for failing to follow a number of
industry best practices that would have prevented the event.
In August 2010, before filing this suit, Plaintiffs’ counsel and other attorneys
filed a series of individual lawsuits in state court that were later consolidated as a
multidistrict litigation in Galveston County State District Court, In re MDL
Litigation Regarding Texas City Refinery Ultracracker Emission Event Litigation,
No. 10-uc-0001. As the case name suggests, the claims in that case cover BP’s
acts and omissions relating to the April/May 2010 emissions event. The state
petitions assert not only claims for property damage, but also for personal injury
based on benzene exposure.
Roughly 50,000 plaintiffs, including the named
Plaintiffs here, have joined that action, though the named Plaintiffs nonsuited their
5 / 42
property claims in the MDL prior to the class certification hearing before this
Court.2
On December 22, 2010, Plaintiffs filed this suit, alleging that chemical
releases and emissions events occurring after December 22, 2008 caused a
diminution in value for residential properties surrounding the Refinery. The suit
does not include personal injury claims and is not restricted to the April/May 2010
event. Plaintiffs assert three common law causes of action: negligence, trespass,
and nuisance. Plaintiffs amended their complaint on October 23, 2012.
The amended complaint differs from the original one in two central ways.
First, while the original complaint focuses on the Refinery’s problems with
benzene releases and resultant benzene exposure to the population, the amended
complaint barely mentions benzene. Compare Docket Entry No. 1 (mentioning
benzene 27 times), with Docket Entry No. 49 (mentioning benzene twice).
Second, the amended complaint proposes a new class definition: whereas the
original complaint proposed a class of individuals who owned real property in the
77590, 77591, or 77568 zip codes, the amended complaint proposes a class based
on exposure to sulfur dioxide emissions. Compare Docket Entry No. 1 at 14, with
Docket Entry No. 49 at 10. Specifically, the new class includes all individuals who
2
BP argues that the pendency of these tens of thousands of individual cases seeking both
personal injury and property damage relating to a BP emission event demonstrate that a class
action is not a “superior” method for adjudicating the claims in this case. The Court need not
decide that issue given its ruling on other grounds.
6 / 42
own or have owned residential property in the area “identified as affected by the
air pollution plume of impact [] modeled by Dr. Paul Rosenfeld . . . , since
December 22, 2008.” Docket Entry No. 49 at 10.
An understanding of Plaintiffs’ proposed class requires a familiarity with Dr.
Rosenfeld’s expert report.
Rosenfeld used AERMOD dispersion modeling
software to model the effects of the Refinery’s sulfur dioxide emissions on the
residents of the three zip codes listed in the original complaint. Docket Entry No.
26-1. The model was constructed using BP’s reported normal operating emissions
of SO2, which do not include emissions from emissions events.
Id. at 2.
Rosenfeld’s model generated an air pollution plume of impact showing where
residents were subjected to at least five incidents between 2009–10 during which
BP’s emissions caused a one hour 50 µg/m3 increase in ambient SO2
concentrations—the level at which two epidemiological studies relied on by Dr.
Rosenfeld identified a statistically significant risk of exposed populations
exhibiting asthmatic complications. Id. at 2–3. The report also noted BP’s history
of pollution and the existence of other air pollutants released from the Refinery,
including volatile organic compounds, nitrogen oxides, and particulate matter in
the plume area, but did not rely on those emissions or other alleged bad acts in
creating the plume.
7 / 42
Though Plaintiffs’ class boundary is based on exposure to SO2 emissions,
their causation and damages theory is not. See Plaintiffs’ Post-Hearing Brief,
Docket Entry No. 86 at 6 (“BP wants this case to be about health effects of a single
type of emission of one chemical. This case is about much more than that.”
(emphasis in original)). Plaintiffs’ amended complaint broadly identifies “BP’s
contamination” as the cause of property value diminution without explaining why
or how. They rely on their economic expert, Dr. Simons, for that explanation. See
Docket Entry No. 52 at 11 (noting that, at trial, Simons would “testify about
causation and damages” and “quantify[] the damages caused by BP’s pollution”).
Simons more specifically attributes the decline in property value to “general public
knowledge of BP’s extraordinary emissions.” Docket Entry No. 50-1 at 9. Simons
performed the following three analyses to estimate property value diminution:
a real estate trends analysis in which he compared the change in
median sales price per square foot between 2008 and 2011 for the
class area with the corresponding change for a control area
comprised of portions of Pasadena, Deer Park, and Baytown,
which Simons determined had similar houses with similar
proximity to refineries and industrial developments;
a hedonic regression analysis in which he compared real estate
sales in the class area after January 1, 2009 with sales prices in the
class area before 2009 and sales prices in the control area3 before
and after 2009. The model attempts to isolate the effects of a
particular disamenity, in this case hypothesized to be BP’s
3
The control area for the regression analysis was slightly larger than the one for the real estate
trends analysis, and included portions of Texas City and La Marque outside the plume area, as
well as the Harborwalk, Tiki Island, and Bayou Vista waterfront developments in Galveston
County. Docket Entry No. 52 at 6.
8 / 42
contamination, by holding all other factors—such as, lot size, year
built, bedrooms, bathrooms, swimming pool, foreclosure status,
and neighborhood characteristics—constant; and
a contingent valuation analysis and property owner survey in
which he surveyed non-class and class members, respectively,
about how the existence of emissions similar to the Refinery’s
would affect the likelihood of bidding and price of bids on
property.
See generally Docket Entry No. 50-1. The goal of Simons’ first two analyses was
to compare “housing price effects of living within an influence zone of reasonably
well-managed and appropriately maintained cluster of petroleum industry (the
control areas along the Houston Ship Channel) versus living near a poorlymaintained refiner (BP),” id. at 6, while the third analysis can more simply be
referred to as a survey. Based on the various analyses, Simons concluded that
BP’s emissions drove down property values in the class area by an average of 5–20
percent.
The Court must now decide whether class certification is appropriate under
Rule 23 and, to the extent it informs that decision, whether Plaintiffs’ expert
testimony should be excluded.
As explained both in the parties’ voluminous
briefing and at the two-day class certification hearing held on April 4–5, 2013,
Plaintiffs rely heavily on their experts in arguing for class certification. Without
Rosenfeld’s plume model, the class boundary would not exist.
And, without
Simons’s analyses, Plaintiffs would have to show causation and damages on an
9 / 42
individual property-by-property basis. For the reasons discussed below, the Court
excludes Simons’s testimony and, accordingly, denies class certification.
II.
STANDARD FOR CLASS CERTIFICATION
“The class action is ‘an exception to the usual rule that litigation is
conducted by and on behalf of the individual named parties only.’” Comcast Corp.
v. Behrend, 133 S. Ct. 1426, 1432 (2013) (quoting Califano v. Yamasaki, 442 U.S.
682, 700–01 (1979)). “[T]he party seeking certification [] bears the burden of
establishing that the requirements of Rule 23 have been met.” Bell Atl. Corp. v.
AT&T Corp., 339 F.3d 294, 301 (5th Cir. 2003) (citing O’Sullivan v. Countrywide
Home Loans, Inc., 319 F.3d 732, 737–38 (5th Cir. 2003)). Rule 23(a) imposes four
prerequisites to certify a class action: (1) a class “so numerous that joinder of all
members is impracticable”; (2) “questions of law or fact common to the class”; (3)
“claims or defenses of the representative parties [that] are typical . . . of the class”;
and (4) representatives that “will fairly and adequately protect the interests of the
class.” Fed. R. Civ. P. 23(a). These prerequisites are known as numerosity,
commonality, typicality, and adequacy. See Amchem Prods. v. Windsor, 521 U.S.
591, 613 (1997).
In addition to meeting all four prerequisites of Rule 23(a), a party seeking
class certification must also demonstrate at least one of the three conditions of Rule
23(b):
10 / 42
(1) litigating separate actions would create the risk of (a) inconsistent
rulings toward individual class members that would create
incompatible standards of conduct for the defendant or (b) rulings
with respect to individual class members that would impair the
ability of other individuals to protect their interests;
(2) the defendant’s conduct applies generally to the class such that
final injunctive or declaratory relief is appropriate as to the class
as a whole; or
(3) common questions of law or fact predominate over individual
questions and a class action is superior to other available methods
for fairly and efficiently adjudicating the matter.
Fed. R. Civ. P. 23(b).
A district court must conduct a rigorous analysis of the Rule 23 requirements
before certifying a class. Gen. Tel. Co. v. Falcon, 457 U.S. 147, 161 (1982); see
also Amchem, 521 U.S. at 615 (requiring district courts to take a “close look at the
case” in making a Rule 23(b)(3) determination).
The “class determination
generally involves considerations that are enmeshed in the factual and legal issues
comprising the plaintiff’s cause of action.” Id. at 160. Thus, “[a]lthough class
certification hearings ‘should not be mini-trials on the merits of the class of
individual claims . . . going beyond the pleadings is necessary, as a court must
understand the claims, defenses, relevant facts, and applicable substantive law in
order to make a meaningful determination of the certification issues.’” Madison v.
Chalmette Refinings, L.L.C., 637 F.3d 551, 555 (5th Cir. 2011) (alteration in
original) (quoting Unger v. Amedisys Inc., 401 F.3d 316, 321 (5th Cir. 2005)); see
11 / 42
also In re Rail Freight Fuel Antitrust Litig.–MDL, --- F.3d ----, 2013 WL 4038561,
at *6 (D.C. Cir. Aug. 9, 2013) (“It is now indisputably the role of the district court
to scrutinize the evidence before granting certification, even when doing so
‘requires inquiry into the merits of the claim.’” (quoting Comcast, 133 S. Ct. at
1433)).
In that same vein, a district court’s “rigorous analysis” may necessitate the
evaluation of expert testimony. “Although courts are not to insist upon a battle of
the experts at the certification stage . . . , [i]n many cases, it makes sense to
consider the admissibility of the testimony of an expert proferred to establish one
of the Rule 23 elements in the context of a motion to strike prior to considering
class certification.” Unger, 401 F.3d at 323 n.6 (citations and internal quotation
marks omitted); see also Am. Honda Motor Co. v. Allen, 600 F.3d 813, 815–16
(7th Cir. 2010) (“The [district] court must also resolve any challenge to the
reliability of information provided by an expert if that information is relevant to
establishing any of the Rule 23 requirements for class certification.”); Sher v.
Raytheon Co., 419 F. App’x 887, 891 (11th Cir. 2011) (holding that “the district
court erred as a matter of law by not sufficiently evaluating and weighing
conflicting expert testimony on class certification” regarding the use of regression
modeling to determine property value diminution). For instance, in a securities
fraud action in which a showing of market efficiency was necessary to establish a
12 / 42
classwide theory of causation, the Fifth Circuit affirmed denial of class
certification in part because the expert testimony on market efficiency was
unreliable. Bell v. Ascendant Solutions, Inc., 422 F.3d 307, 314 n.13 (5th Cir.
2005).
And the Supreme Court recently found class certification to be
inappropriate when plaintiffs’ expert calculated damages based off of four theories
of liability of which three had already been dismissed by the trial court. Comcast,
133 S. Ct. at 1434–35. The Court rejected the view of the Court of Appeals that an
“attac[k] on the merits of the methodology [had] no place in the class certification
inquiry,” as well as the appellate court’s ruling that plaintiffs’ “assurances” that
they could fix the model at the merits stage would be sufficient for certification.
Id. at 1431, 1434 (alterations in original) (quoting Behrend v. Comcast Corp., 655
F.3d 182, 207 (3d Cir. 2011)); see also In re Rail Freight, 2013 WL 4038561, at *8
(“It is now clear [after Comcast] that Rule 23 not only authorizes a hard look at the
soundness of statistical models that purport to show predominance—the rule
commands it .”). And in one sense scrutiny of expert testimony being used to
show that a case is susceptible to class treatment seems less controversial than the
normal application of Daubert, because it does not intrude on the jury’s role given
that class certification is an issue for the court.
The Court thus turns to evaluating the expert testimony of Dr. Simons.
13 / 42
III.
ADMISSIBILITY OF DR. SIMONS’S TESTIMONY
A. Legal Standard
When considering expert opinions at the class certification stage, “court[s]
should rely on the admissibility standards for expert evidence as construed by the
Supreme Court in Daubert v. Merrell Dow Pharmaceutical, Inc., 509 U.S. 579
(1993) and Kumho Tire Co. v. Carmichael, 529 U.S. 137 (1999).” Hon. David
Hittner et al., Practice Guide: Federal Civil Procedure Before Trial, 5th Circuit
Edition ¶ 10:577.1 (2011); see also Am. Honda Motor Co., 600 F.3d at 816 (“[T]he
district court must perform a full Daubert analysis before certifying the class if the
situation warrants.”). Plaintiffs have the burden of establishing admissibility of
their experts by a preponderance of the evidence. Daubert, 509 U.S. at 592 n.10
(citations omitted).
Daubert identifies a nonexhaustive list of factors a district court should
consult in assessing the reliability of expert testimony: (1) whether the theory can
or has been tested; (2) whether the theory has been subjected to peer review and
publication; (3) the known or potential rate of error; (4) the existence and
maintenance of standards and controls; and (5) whether the theory has been
generally accepted in the relevant scientific, technical, or professional community.
Daubert, 509 U.S. at 593–94. Other guideposts have been articulated subsequent
to Daubert. Relevant to the Court’s analysis here, the Supreme Court in General
14 / 42
Electric Co. v. Joiner established the test of “fit” between the methodology and the
conclusions drawn, stating that a “court may conclude that there is simply too great
an analytical gap between the data and the opinion proffered.” Gen. Elec. Co. v.
Joiner, 522 U.S. 136, 146 (1997) (citation omitted); see also Fed. R. Evid. 702
Advisory Committee Notes (2000 Amends.) (listing factors relevant to the Daubert
inquiry, including “[w]hether the expert has adequately accounted for obvious
alternative explanations” and “[w]hether the expert has unjustifiably extrapolated
from an accepted premise to an unfounded conclusion”).
B. Analysis
Dr. Simons’s opinions fail to meet the standards set forth in Daubert and its
progeny.
As explained below, not only are specific aspects of Simons’s
methodologies flawed, but his overarching theory of damages is disconnected from
Plaintiffs’ causes of action of negligence, trespass, and nuisance which are limited
to a particular time period beginning in late 2008.
1. Emissions Levels in the Control Area
Simons’s real estate trends analysis and hedonic regression analysis are both
premised on a comparison between the class area and the control area.4 The
control area is comprised primarily of portions of Pasadena, Deer Park, and
4
As Simons noted at the class certification hearing, the regression analysis is essentially a “more
refined” version of the real estate trends analysis. Docket Entry No. 75 at 20. Along those lines,
he acknowledges that the regression analysis should “carry more weight” than the other
methodologies he used. Id. at 25, 37.
15 / 42
Baytown—areas that, like the class area, are in the greater Houston metropolitan
area, are along the Houston ship channel, and contain industrial facilities. Docket
Entry No. 75 at 38–39.
As Simons acknowledged at the class certification hearing, “the purpose of a
control area in a regression model is to give you some basis to isolate and value the
characteristic that you’re trying to value.” Id. at 39. Simons admits that “the
characteristic that distinguishes the class area from other properties outside the
boundary is that Dr. Rosenfeld has modeled a certain level of sulfur dioxide
emissions within the class area.” Id. But he qualifies his response by noting that
his regression model did not isolate the effects of sulfur dioxide emissions, but
more generally isolated the effects of “the activity of releases from BP, including
exceedances and other things that Dr. Rosenfeld will discuss,” because “sulfur
dioxide is just a proxy for all the air pollutants from the plant.” Id. at 39–40; see
also Simons Rebuttal Declaration, Docket Entry No. 50-1 at 6 (“While residents or
potential buyers of property would not necessarily be expected to have knowledge
of specific sulfur dioxide concentrations, they would likely have knowledge of
these general factors.”).
In actuality, Simons does not, and cannot, know exactly what characteristic
he isolated with his regression model—it could have been sulfur dioxide
emissions, exceedances, events, bad press about the Refinery, or any other
16 / 42
difference between the class area and control area that was not accounted for in his
model, including non-BP related variables like neighborhood crime rates or the
effects of Hurricane Ike. See infra Part III(B)(3).
But even assuming that Simons was able to isolate BP’s conduct from all
other relevant variables, his model is still flawed because Plaintiffs fail to show
that SO2 emissions have been worse in the class area than the control area since
December 2008. Simons admits that he did not assess SO2 levels in the control
areas and that he “hop[ed] they’re less, but [had] no knowledge of what they are.”
Docket Entry Nos. 38-6 at 26–27; 75-1 at 8.5 After a critique from one of BP’s
experts, Dr. Rosenfeld analyzed SO2 emissions in the control area in his rebuttal
report and concluded that “concentrations of SO2 in Texas City/La Marque are
elevated above levels typically measured in the Pasadena region with statistical
significance.” Docket Entry No. 51-1 at 5. But Rosenfeld’s control area analysis
is flawed and unreliable because he measured emissions levels for the control area
using air monitors ranging from 5.7 to 20.7 miles away from the center of the
control areas, while he measured emissions levels for the class area using air
monitors not only within the class boundary, but within the boundary of the
Refinery itself. Docket Entry No. 59-1 at 6, 11–12. As BP’s toxicologist expert,
Dr. Phillip Goad, points out, “chemicals disperse (that is, the concentration
5
Simons testified that he “certainly [did] not know” what the level of PAHs in the control area
were. Docket Entry No. 38-6 at 27.
17 / 42
decreases) with distance when released into the air and receptors or monitors
located at difference distances can have vastly different results, reflecting unique
emissions sources within the local area.” Id. at 6. Goad conducted a more applesto-apples analysis by calculating SO2 emissions from the top SO2 emissions
sources within a certain radius6 of the class and control areas. Goad’s study
revealed that the SO2 emissions from within these control area boundaries were
roughly six times higher in 2009 and more than eight times higher in 2010 than the
SO2 emissions from within the respective class area boundaries. Id. at 7, 16. This
is not surprising given that one of the control areas, Baytown, is home to the
“largest petroleum & petrochemical complex in the United States.” ExxonMobil,
Baytown
Area,
About
Us,
http://www.exxonmobil.com/NA-
English/about_where_ref_bt_aboutus.aspx (last visited Sept. 12, 2013); see also
Docket Entry 26-1 at 6 (showing that the Baytown Refinery had 2,344,831 pounds
of Toxics Release Inventory air emissions for 2009–10 versus 2,515,337 pounds
for the Texas City Refinery).
The Court is persuaded by Goad’s analysis and finds that Plaintiffs have not
shown that SO2 emissions levels were worse in the class area than in Simons’s
control area. Despite Plaintiffs’ admonition that this case is not about the release
of one type of chemical, a distinction in SO2 emissions between the class area and
6
Goad used a radius of 7.4 miles—the distance from the Park Place monitor to the center point
of the closest control area. Docket Entry No. 59-1 at 7.
18 / 42
control area is key to Simons’s regression and real estate trends analyses. See
Docket Entry No. 51-1 at 5 (statement from Rosenfeld in his rebuttal report that his
analysis of SO2 levels in the control areas “serves to support the opinions of Dr.
Roby Simons”). The Court agrees with BP’s expert Dr. Thomas Jackson: “If
[Simons] found a delta, or price difference, you have to be able to say it’s due to
this characteristic that differs between the two areas.” Docket Entry No. 76-3 at
14.
As the proposed class definition demonstrates, Plaintiffs use SO2 emission
levels to define the boundaries of BP’s liability and show the reaches of BP’s bad
acts.
Though Plaintiffs may not be arguing that SO2 emissions caused the
diminution in property value, they are still using SO2 emissions as a proxy for
other pollutants and to show the reach of wrongful conduct. Simons testified that
he looks at SO2 “as a proxy for the whole soup of chemicals that were released
from the plants, and [] a good way to measure a certain area,” Docket Entry No. 75
at 14, and Rosenfeld testified that “the SO2 plume [] serves as a proxy for all the
other contaminates released by BP,” Docket Entry No. 75-4 at 188. Assuming that
to be true, if the class area and control area have similar levels of SO2 emissions,
or if the control area has higher levels of such emissions, then the difference
between property-value change in the two areas is very likely the result of
19 / 42
something other than BP’s wrongful conduct.7
Accordingly, with respect to
Simons’s regression and real estate trends analyses, the Court concludes that “there
is simply too great an analytical gap between the data and the opinion proffered.”
Joiner, 522 U.S. at 146 (citation omitted).
2. Flawed Comparison in Regression Analysis
In addition to its unsupported assumption regarding SO2 emissions in the
control area, Simons’s regression model is also structurally flawed by failing to
compare the change in class area property values that occurred after December 22,
2008 with the change in control area property values after that date. As Dr.
Jackson explains, regression models are designed to compare a subject group
before and after a relevant date with a control group before and after the same date
in order to measure the effect of some event or circumstance in the subject group.
See Docket Entry No. 38-4 at 30–31 (citing W. Rogers, Errors in Hedonic
Modeling Regressions: Compound Indicator Variables and Omitted Variables, The
Appraisal Journal 208–13 (April 2000)). Rather than perform such a comparison,
7
At the class certification hearing, Plaintiffs backtracked on the use of SO2 emissions as a
proxy, calling it “underinclusive” of other releases. Docket Entry No. 76-2 at 33. But if that
were the case, Plaintiffs have not provided an alternative comparison (such as odors, bad press,
or other pollutants) between the class area and control area. Thus, the same causation hole
would remain: in order to draw any conclusions from his model, Simons would have to know
that his control area did not share the same characteristic as his class area. See Docket Entry No.
76-3 at 14 (testimony from Dr. Jackson at the class certification hearing agreeing that “if
[Simons] were to pick another characteristic of the class area [besides SO2] he would have to
know, in order to draw any conclusions from a regression model, that his control area didn’t
share that other different characteristic”).
20 / 42
Simons compared the class area in the after period with the subject area in the
before period and the control area in the before and after periods.
Figure 1: Proper comparison for a regression model.
Figure 2: Comparison used by Simons.
While the scientific literature critiques such an approach as “biased due to
omitted variables,” see id., common sense explains why it is unreliable. Under
Plaintiffs’ theory of the case, BP’s conduct became worse after December 2008,
causing property values to decline.8 Their theory therefore requires some change
in BP’s conduct starting in December 2008, because the ordinary operation of the
refinery—which has been in operation for more than 75 years—would have
already been factored into Plaintiffs’ initial purchase price of their homes. For
8
The two-year statute of limitations explains Plaintiffs’ theory; Plaintiffs filed their suit on
December 22, 2010.
21 / 42
instance, Plaintiff Genaro Ramirez bought his home in Texas City in 2006. See
Docket Entry No. 26 at 35. If the stigma associated with the Refinery already
existed in 2006, he would have purchased his house at a discount then and would
not have been injured by the discount in his property value that still existed during
the class period. Cf. LaBauve v. Olin Corp., 231 F.R.D. 632, 677 n. 96 (S.D. Ala.
2005) (“[I]f plaintiffs were proceeding on a ‘stigma’ theory, surely any stigma and
associated property devaluation would have attached no later than 1983, when the
Olin plant was declared (amidst much fanfare and publicity) to be a Superfund
site.”).
Comparing the class’s “before and after” property values with the control
group’s “before and after” property values is thus key to determining the effect of
BP’s purported change in conduct after December 2008. Simons’s regression
model is unable to properly account for that change because it fails to reflect price
differences between the class and control areas during the before period. Under
Simons’s model, which indicates a 4.7% diminution in value, property values
could have been increasing in the class area in 2009–10 relative to the control
group and the 4.7% diminution could be attributed to higher “before” property
values in the control group. A tweak of the model to account for the proper
comparison indicates this is exactly what happened. BP’s expert respecified the
model to include all four categories (subject and control, before and after) without
22 / 42
making any other changes. Docket Entry Nos. 38-4 at 31; 76-3 at 36. He found
that, in contrast with Simons’s conclusion of a 4.7% loss, the tweaked model
“indicate[d] that sales prices for single-family residential properties in Simons’
subject area were 4.02% less than his control areas in the period prior to January 1,
2009 and they were .34% less in the period after January 1, 2009 . . . for a net gain
or improvement of 3.68%.” Docket Entry No. 38-4 at 31.
Accordingly, the Court finds that Simons’s regression model was improperly
constructed and that his conclusion regarding a 4.7% diminution in property value
is unreliable.
3. Inability to Isolate BP’s Wrongful Conduct
Simons’s regression model fails in one other way even if it is able to show a
4.7% diminution in property value during the relevant time period. Though the
model holds a number of variables constant between the class and control areas,9
the Court finds that he still fails to isolate BP’s wrongful conduct as the cause of
the identified diminution.
First, related to the Daubert-related factor concerning “[w]hether the expert
has adequately accounted for obvious alternative explanations,” Fed. R. Evid. 702
Advisory Committee Notes (2000 Amends.), the Court notes that seemingly
9
The full list of variables is included in Exhibits 6-1–6-4 of his expert report and includes lot
size, square footage, bedrooms, bathrooms, pool, garage, cooling system, heating system,
foreclosure status, SAT scores, and distance to airports, highways, and railroads.
23 / 42
important variables are missing from Simons’s analysis.
This is problematic
because “failure to include a major explanatory variable that is correlated with the
variable of interest in a regression model may cause an included variable to be
credited with an effect that actually is caused by the excluded variable.” Federal
Judicial Center, Reference Manual on Scientific Evidence 314 (3d ed. 2011). Such
missing variables here include neighborhood crime rates and location in or outside
of the floodplain, but the Court will only focus on one by way of example—the
effects of Hurricane Ike. Hurricane Ike hit the Texas Gulf Coast in September
2008, just a few months before the relevant, and “will likely go down in history as
the most costly and destructive storm ever to hit Texas.” Jack Colley, Foreword to
the Federal Emergency Management Agency’s Hurricane Ike Impact Report, at ii
(December 2008). The effects of the storm lasted well into Plaintiffs’ class period,
with an estimated $3.4 billion in damage to housing. Id. at 17.
But Simons did not include Hurricane Ike impact as a variable in his
regression model; thus, the 4.7% unexplained loss that his model generated could
have easily been generated from variances in Hurricane Ike effects, such as
increases in insurance rates for hurricane–vulnerable properties, as from proximity
to a poorly-managed refinery.
And such variances in Hurricane Ike effects
between the control areas and class area are not just hypothetical. Whereas the
bulk of Pasadena is located inland, Texas City and La Marque are much closer to
24 / 42
the Gulf and Galveston Bay, though parts are protected by a levee. Id. at 36–37. It
logically follows that property values would be affected in different ways in the
different areas. Moreover, the likelihood that events other than BP’s emissions
would have caused any property value diminution is significant given that, even as
Dr. Rosenfeld admits, the air quality in Texas City has been improving over the
last ten years including during the class period. Docket Entry No. 75-4 at 24–25.
The failure to account for such factors deems Simons’s methodology unreliable.
See Bazemore v. Friday, 487 U.S. 385, 400 & n.10 (1986) (Brennan, J., joined by
all other Members of the Court, concurring in part) (“Normally, failure to include
variables will affect the analysis’ probativeness, not its admissibility. . . . There
may, of course, be some regressions so incomplete as to be inadmissible as
irrelevant . . . .”); see also Phillips v. Am. Honda Motor Co., 238 F. App’x 537,
542 (11th Cir. 2007) (affirming exclusion of expert testimony due to model’s
failure to control for significant alternative sources of temperature variation).10
Second, even if Simons’s model could accurately attribute the alleged
property value diminution to the Refinery or to BP or to extraordinary emissions, it
could not specifically attribute the diminution to the conduct for which BP would
10
While the regression analysis is missing some key controls, it has far more controls than the
real estate trends analysis. As Simons testified at his deposition, “there’s a lot of factors in real
estate trends that are not controlled for, explicitly, say, numbers of bedrooms and all the other
things that go with it that regression does control for.” Docket Entry No. 38-6 at 26; see also
Docket Entry No. 75-2 at 31 (noting at class certification hearing that trends analysis is not as
good as regression at holding things constant).
25 / 42
potentially be liable. For instance, Simons ties the diminution to general public
knowledge of BP’s extraordinary emissions, where knowledge is derived from
press reports, direct observations by people in the class area, and word of mouth.”
Docket Entry No. 50-1 at 9. But the regression model has no way of separating the
stigma from the nuisance, trespass, and negligence that Plaintiffs claim in this case
from any other BP-related stigma, especially considering that during the time
period in question BP was responsible for the largest accidental oil spill in
history.11 See Ponca Tribe of Indians of Okla. v. Cont’l Carbon Co., 2009 WL
5842042, at *8–9 (W.D. Okla. Jan. 23, 2009) (finding diminution calculations from
hedonic regression analysis to be “unhelpful and not relevant to [] nuisance
claims” in part because they were “not directly tied to the emission events that
support Plaintiffs’ claims in [the] lawsuit”); cf. Robert Simons, When Bad Things
Happen to Good Property, Environmental Law Institute (May 2006) (“Many
courts have denied recovery for stigma damages finding that there is no
interference with plaintiffs’ rights when there is no causal connection between the
injury of plaintiffs and the unreasonable conduct of the defendant.” (citations
omitted)). This fact was highlighted during the class certification hearing when
one of the named Plaintiffs was asked which particular things relating to the
refinery caused people to want to move out of his neighborhood and he responded:
11
An intentional oil spill caused by Iraqi forces in Kuwait during the Persian Gulf War is the
only larger spill.
26 / 42
“[T]he oil spill, which is different from this, but a lot of things – a lot of people,
whether it’s based on fact or not, assume that – a lot of negativity towards the
plants still from BP.” Docket Entry No. 75-5 at 29.
Similarly, even if there is a property value decline directly correlated with
extraordinary emissions, as opposed to other BP stigma, there are too many
sources of adverse environmental impacts within the class area to isolate and
evaluate property value impacts from any one source. As BP’s expert explained in
his report, there are several other refineries and heavy industrial facilities near the
Refinery, including Marathon Petroleum Company, Valero, and Sterling
Chemicals, and 43 companies within the proposed class area that emit EPA criteria
pollutants and are listed in EPA’s AIRS database. Docket Entry No. 38-4 at 14.
Consequently, “[i]t would be virtually impossible to isolate and analyze the source
of any such emission, let alone to segregate the property value impacts, if any,
from a single facility or emission event.”
Id. at 14–15.
Simons even
acknowledged at his deposition that “there could be, and probably is, another plant
in this but it seems logical there’s some other sources of pollution and they should
possibly have some, also, responsibility.” Docket Entry No. 38-6 at 26.
Based on all of these problems, Simons is unable to reliably or formulaically
calculate Plaintiffs’ damages in this case.
27 / 42
4. Real Estate Trends Analysis Not Tied to Class Area
As discussed above, Simons’s real estate trends analysis is intended to
compare real estate sales patterns in the class area and the control areas from a
period prior to the events investigated to a period when knowledge of the event
became widespread in the community. Specifically, Simons attempted to compare
the median sales price per square foot for transactions in the control area and the
class area for 2008 and 2011 and concluded that property values decreased at a
higher percentage in the class area than in the control area.
But Simons analyzes the wrong class area. The purported class area used in
his analysis is the old, zip-code based area proposed in Plaintiffs’ original
complaint, rather than the plume-based area proposed in the current complaint. As
Simons admits, the three zip codes contain properties not included in the plumebased boundary. Docket Entry No. 38-6 at 11. BP’s expert Jackson calculates,
and Plaintiffs do not dispute, that 24.15% of the parcels in the original class area
are not included in the current one. Docket Entry No. 38-4 at 28 n. 61.
As such, Simons’s real estate trends analysis is not a reliable indicator of
property value diminution in the current class area. Cf. Comcast, 133 S. Ct. at
1433 (“[A] model purporting to serve as evidence of damages in this class action
must measure only those damages attributable to that theory.”). Plaintiffs assure
the Court that they can fix the model at the merits stage if needed. “But such
28 / 42
assurance is not provided by a methodology that identifies damages that are not the
result of the wrong.” Id. at 1434; see also In re Rail Freight, 2013 WL 4038561, at
*8 (rejecting plaintiffs’ attempts at saving their damages model when plaintiffs
failed to rerun the model). Given that Plaintiffs’ bear the burden of establishing
the reliability of their expert and that a properly-run model could change Simons’s
results (as happened with changing the temporal comparison in the regression
model), the Court determines that the real estate trends analysis fails as a
mechanism for showing causation or damages on a classwide basis.
5. Contingent Valuation Analysis
As Plaintiffs explain, a “contingent valuation analysis is a survey-technique
that attempts to value things that do not typically have a market price, such as the
presence or absence of environmental contamination.” Docket Entry No. 50 at 7.
For Simons’s contingent valuation analysis, he retained a professional survey firm
to call 400 homeowners in the zip codes surrounding Baytown, Deer Park, and
Pasadena, which Simons describe as demographically similar to the class area.
The interviewer presented scenarios of a house “very similar” to their current one
but located in four different areas—one by a business park, one by a closed gas
stations that had leaking underground storage tanks, one by a landfill or hydraulic
fracturing site, and one by a petroleum refinery intended to mirror BP’s Refinery—
and asked the homeowners the likelihood that they would make an offer on the
29 / 42
homes on a scale of –3 to +3 and the highest bid they would make on the home.
Docket Entry No. 38-5 at 88–105. Simons’s results showed a 64% drop in demand
and a 20% average discount for the top half of potential bidders.12 From those
results, Simons predicted that the refinery caused property values to decline in the
rough range of 5% to 20%. Id. at 33.
As an initial matter, the Court notes its uncertainty that even Plaintiffs would
argue that contingent valuation on its own—without the reinforcement of a
regression or real estate trends analysis—would serve as a reliable calculation of
damages. Simons admits that it does not carry as much weight as his regression
analysis, Docket Entry No. 75 at 25, and Plaintiffs contend that none of the
analyses should be viewed in a vacuum, Docket Entry No. 50 at 2. See also
Docket Entry No. 50-1 at 12 (noting that “researchers have advocated the use of
buyer surveys in conjunction with analysis of actual sales”).
A debate exists in the scientific community about the validity of contingent
valuation as a methodology for assessing market discounts associated with real
estate disamenities. While BP’s expert Dr. Jackson states that it is “not a generally
accepted valuation method within the appraisal field,” Simons notes that at least
nineteen authors have published contingent valuation studies in the peer-reviewed
12
Simons used two slightly different versions of the survey, which described the refinery
scenario slightly differently. In the first there were detectable levels of PAHs in the air
conditioning filter system; in the second, there was a nearby public school that had air quality in
the worst 5% of the U.S. For simplicity, only the first set of results are shown above.
30 / 42
real estate literature. Compare Docket Entry No. 38-4 at 22, with Docket Entry
No. 50-1 at 11–12.
Jackson presents a number of problems with contingent
valuation: it is not as reliable as the existing transactional data; hypothetical bias
may exist because the respondents do not have to bear the consequences of their
decisions; it does not incorporate many factors that go into a home purchase; and
respondents may be biased or not understand the scenarios. Docket Entry No. 38-4
at 23–25.
But regardless whether contingent valuation is a reliable methodology in
general, Simons’s contingent valuation analysis, standing alone, is unable to serve
as a reliable or formulaic causation and damages model in this case. First, the
analysis suffers from many of the same control problems described above. For
instance, in the survey’s refinery scenario, a public school within five blocks of the
house was ranked as being in the worst five percent for air quality in the U.S.
Docket Entry No. 38-5 at 31–32. But the survey did not inform the respondents in
Deer Park and Pasadena that they already lived near schools ranking in the worst
one percent for air quality in the nation. See Docket Entry No. 38-4 at 21–22;
USA Today, The Smokestack Effect: Toxic Air and America’s Schools,
http://content.usatoday.com/news/nation/environment/smokestack/index
(last
visited Sept. 12, 2013) (listing nine schools in Deer Park and Pasadena being in the
top percentile of pollution, including a school ranked sixth worst in the nation).
31 / 42
Likewise, the refinery in the survey is described as having had a malfunction that
led to the release of over 500,000 pounds of pollutants, but the survey does not
inform its Baytown respondents that they live near a plant that emitted over 1.25
million pounds of pollutants in 2009. See Docket Entry Nos. 38-4 at 21–22; 26-1
at 6.13
Second, and relatedly, a problem exists with respect to the survey’s refinery
scenario in which “PAHs were found in the air conditioning filter system.” Docket
Entry No. 38-5 at 31. Neither Simons nor Rosenfeld performed any testing of air
conditioning systems in the control area to search for PAHs, thereby calling into
question the study’s results. Additionally, as BP’s expert Dr. Goad explains,
“PAHs are ubiquitous compounds that would be expected in ‘detectable’ levels in
the community soils and in living space surface dust of homes,” Docket Entry No.
39-1 at 22 (citing academic literature), meaning both that the respondents likely
had PAHs in their air conditioners and that the existence of PAHs is not the
disamenity that the survey makes it out to be.14 Finally, the significance of the
13
The survey also suffers from the opposite problem. After Simons began his survey, he added
three new Houston zip codes within the Sam Houston Tollway/Belt 8 Loop, which do not share
the same demographics or proximity to a refinery as the class area. See Docket Entry No. 75-1
at 5 (testimony from Simons that “it’s important for the control group to have the attitudes of the
people that live there that are accepting the petroleum”). Unlike the residents of Texas City, the
residents of those zip codes did not already have proximity to a refinery factored into the
purchase price of their home.
14
Along those same lines, it is doubtful that the refinery problems discussed in the survey
scenario would have the same impact on actual property values as they do in the contingent
valuation analysis because property purchaser do not have perfect information. Simons cites
32 / 42
PAHs in air conditioning filters is questionable given that Dr. Rosenfeld did not
identify BP as their source; indeed, vehicle exhaust, solvents, cleaning products,
and cigarette smoke are common sources of PAHs. Id. at 21.
Third, it is unclear how the contingent valuation analysis, standing alone,
could calculate damages on a classwide basis.
While the survey contained
scenarios in which the refinery was half a mile from the hypothetical house and
1.75 miles away from the hypothetical house, the analysis does not explain how to
calculate damages for houses that are other distances away from the refinery,
including the houses of three of the five named Plaintiffs’ and much of the putative
class, which are located more than 1.75 miles away from the Refinery. Along
those same lines, the survey fails to account for the fact the certain houses are more
exposed to contamination based on their location in relation to the Refinery. Thus,
if wind patterns point primarily to the north/northwest, a house a mile northwest of
the plant may suffer different damages from a house a mile southeast of it. See
Docket Entry No. 39-1 at 11 (discussing effects of meteorological conditions).
Item 6 of the Texas Seller’s Disclosure of Property Condition, which requires a seller to disclose
any “condition on the property which materially affects the physical health or safety of an
individual.” Docket Entry No. 38-5 at 82. But a review conducted by BP’s expert Dr. Jackson
shows that of the 49 publicly-available seller’s disclosure forms for properties currently listed in
the class area, none identified any reportable condition on their disclosure. Docket Entry No. 384 at 34.
33 / 42
For these reasons, Simons’s contingent valuation analysis is not a reliable or
formulaic model to establish causation or calculate damages for the class.15
6. Conclusion
Based on the above, the Court concludes that Dr. Simons’s testimony and
reports fail to meet the standards propounded by Daubert and its progeny and are
therefore inadmissible.
IV.
EFFECTS OF SIMONS’S EXCLUSION ON CLASS CERTIFICATION
As discussed in Part II, supra, a plaintiff seeking to certify a class must
satisfy all four prerequisites of Rule 23(a)—numerosity, commonality, typicality,
and adequacy—as well as one of the three conditions set forth in Rule 23(b). The
Court need not address Plaintiffs’ ability to meet the prerequisites of Rule 23(a),
because, for the reasons set forth below, the Court finds that Plaintiffs are unable to
satisfy Rule 23(b) without the testimony of Dr. Simons.
A. Rule 23(b)(3): Predominance
A party seeking certification under Rule 23(b)(3) must show “both (1) that
questions common to the class members predominate over questions affecting only
15
Simons also conducted a separate survey of putative class members. Regardless of that
survey’s reliability, it provides no support to Plaintiffs in terms of class certification and raises
more questions than it answers. For instance, in response to the question, “Do you think
emissions/contamination from the BP Refinery has caused a change in your property values?”
only 68% answered “yes.” Docket Entry No. 38-5 at 34. And only 62% indicated that
contamination issues affected the use and enjoyment of their property. Id. at 34–35. Thus, the
survey not only fails to provide a formulaic way to determine causation or damages for the class,
but it also shows that such an endeavor may not be possible.
34 / 42
individual members, and (2) that class resolution is superior to alternative methods
for adjudication of the controversy.” Bell Atl. Corp. v. AT&T Corp., 339 F.3d 294,
301 (5th Cir. 2003). The Court here will only focus on the predominance inquiry,
which “tests whether proposed classes are sufficiently cohesive to warrant
adjudication by representation” and is “far more demanding” that Rule 23(a)’s
commonality requirement. Unger v. Amedisys Inc., 401 F.3d 316, 320 (citation
omitted).
The predominance hurdle “requires district courts to consider how a trial on
the merits would be conducted if a class were certified.” Madison v. Chalmette
Refining, L.L.C., 637 F.3d 551, 555 (5th Cir. 2011) (citations and internal
quotation marks omitted). “This, in turn, entails identifying the substantive issues
that will control the outcome, assessing which issues will predominate, and then
determining whether the issues are common to the class, a process that ultimately
prevents the class from degenerating into a series of individual trials.”
(citations and internal quotation marks omitted).
Id.
As a general rule, a “mass
accident” is “not appropriate for a class action because of the likelihood that
significant questions, not only of damages but of liability and defenses to liability,
would be present, affecting the individuals in different ways,” thus necessitating
multiple, separately-tried lawsuits. Fed. R. Civ. P. 23(b)(3) advisory committee’s
note; see also Steering Comm. v. Exxon Mobil Corp., 461 F.3d 598, 604 (5th Cir.
35 / 42
2006) (finding that case involving an oil leak and a three-day fire did not have “any
exceptional features that warrant departing from the general rule”).
Plaintiffs argue that they fall within the exception to the general rule because
numerous common issues exist including the standard of care, BP’s exercise of
care, the quantity of the Refinery’s emissions, the effects of BP’s emissions on air
quality in the class area, the effect of air pollution on property values, and whether
liability for exemplary damages. Docket Entry No. 26 at 42–44. They illustrate
their point by proposing a trial plan composed of the following four phases in
which the jury or Court would consider: (1) common issues as to liability and
individual issues, if any, as to causation and damages for the representative
plaintiffs; (2) exemplary damages; (3) causation and damages issues for the absent
class members, with the possibility of the Court creating subclasses; and (4) the
calculation of punitive damages for the class as a whole based on the ration
established in phase two. Id. at 27–29.
But Plaintiffs’ trial plan—and their class theory as a whole—are highly
dependent on Dr. Simons. They rely on Dr. Simons to (a) prove on a classwide
basis that BP’s wrongful conduct (through theories of negligence, nuisance, or
trespass) caused a diminution in property value; and (b) calculate damages
formulaically. See id. (proposing that at phase three plaintiffs would “present
damages on a classwide basis, following the methodology of Dr. Simons”); Docket
36 / 42
Entry No. 52 at 11, 13 (“The damages analysis will be done formulaically, by Dr.
Simons’s analysis.”); Docket Entry No. 86 at 12 (“Dr. Simons has demonstrated
that a common method of proof can be used to evaluate the diminution of property
value caused by the proximity of a refinery which is poorly run.” (emphasis
omitted)); Docket Entry No. 87 at 3 (arguing that a regression model can provide
evidence of causation).
Plaintiffs provide no alternatives to Simons’s methodologies to prove
causation or damages, and the Court cannot envision how a class action trial would
operate without his testimony. Presumably, each of the roughly 14,300 putative
Plaintiffs would have to prove damages by presenting appraisal figures before and
after December 22, 2008 and would have to prove causation by presenting
evidence the BP’s wrongful conduct, and not some other source, caused the
diminution in their property value. See Steering Comm., 461 F.3d at 602 (noting
that separate types of individualized proof would be necessary to prove property
damage and devaluation stemming from an oil leak). As explained in Steering
Committee, which involved only a single accident rather than years’ worth of
operations, chemical exposure is not straightforward or uniform. Id. at 603. If
Plaintiffs’ proved causation and damages for one plaintiff, they would still have to
make the same proof for all the others.
37 / 42
Though individualized damages issues will not always preclude class
certification, “where individual damages cannot be determined by reference to a
mathematical or formulaic calculation, the damages issue may predominate over
any common issues shared by the class.” Id. at 602 (citations omitted); see also
Bell, 339 F.3d at 306 (holding that “class certification is not appropriate” because
plaintiffs “failed to demonstrate that the calculation of individualized actual
economic damages, if any, suffered by the class members can be performed in
accordance with the predominance requirement”); O’Sullivan v. Countrywide
Home Loans, Inc., 319 F.3d 732, 745 (5th Cir. 2003) (“In light of the individual
calculation of damages that is required, the district court abused its discretion in
certifying [plaintiffs’] claims.”); Robertson v. Monsanto, 287 F. App’x 354 (5th
Cir. 2008) (holding that individualized issues of causation and damages precluded
certification and noting that “[a]lthough the alleged cause of the plaintiffs’ injuries
is a single incident . . . each plaintiff still must show that [defendant’s] negligence
in causing the gas leak was proximately connected to the specific injuries
complained of”).
Accordingly, without the aid of Simons’s testimony, Plaintiffs are unable to
show that “the questions of law or fact common to class members predominate
38 / 42
over any questions affecting only individual members.” Fed. R. Civ. P. 23(b)(3).
Thus, class certification under Rule 23(b)(3) is denied.16
B. Rule 23(b)(1) and (b)(2)
Plaintiffs also argue that they meet the conditions for certification articulated
in Rule 23(b)(1) and (b)(2). Indeed, a plaintiff need only satisfy one of the three
conditions of Rule 23(b) in order to obtain certification. Nonetheless, for the
reasons below, the Court does not find certification to be appropriate under (b)(1)
or (b)(2).
Rule 23(b)(1) applies when separate actions by individual class members
would create a risk of “establish[ing] incompatible standards of conduct for the
party opposing the class,” or where individual adjudication “as a practical matter,
would be dispositive of the interests of the other members not parties to the
individual adjudications or would substantially impair or impede their ability to
protect their interests.” Fed. R. Civ. P. 23(b)(1). Rule 23(b)(2) applies when “the
party opposing the class has acted or refused to act on grounds that apply generally
to the class, so that final injunctive relief or corresponding declaratory relief is
16
BP makes numerous arguments why Plaintiffs’ fail the predominance inquiry independent of
whether Simons’s testimony is excluded. For instance, they argue that their affirmative defenses
of preemption and limitations require individualized inquiries; that each of the 70+ emissions
events presents individual liability issues because each event affected each property differently;
that the nuisance claim requires individualized proof of substantial interference with the use of
property; and that a certified class would be overwhelmed by subgroups based on such things as
proximity to the plant and property type. Given the Court’s broader finding that Simons’s
analysis is unreliable, it need not decide the specific issues.
39 / 42
appropriate respecting the class as a whole.” Fed. R. Civ. P. 23(b)(2). “Classes
certified under (b)(1) and (b)(2) share the most traditional justifications for class
treatment—that individual adjudications would be impossible or unworkable as in
a (b)(1) class, or that relief sought must perforce affect the entire class at once, as
in a (b)(2) class.” Wal-Mart Stores, Inc. v. Dukes, 131 S. Ct. 2541, 2558 (2011).
Rule 23(b)(1) and (b)(2) are framed for situations in which class treatment is more
clearly called for than in (b)(3) situations, and, accordingly, do not provide the
same protections as a (b)(3), such as the opportunity to opt out or notice of the
action. Id. (citations omitted).
Plaintiffs’ justification for certification under (b)(1) and (b)(2) are both
based on the fact that they are seeking an injunction that would require BP to
change the operation of its Ultracracker Flare. See Docket Entry No. 26 at 37–
40.17
Plaintiffs’ arguments fail for two reasons. First, the injunctive relief upon
which Plaintiffs base their arguments no longer appears to be part of this case. BP
sold the Texas City Refinery to Marathon Petroleum Corporation on February 1,
17
Specifically, Plaintiffs’ amended complaint seeks an injunction: (1) requiring BP to calibrate
all flowmeters and to conduct a relative accuracy test audit on the gas chromometer associated
with the ULC Flare; (2) requiring BP to implement and enforce a policy on proper steam to gas
ratios for operation of the ULC Flare; (3) requiring BP to test the ULC Flare after those steps
have been completed to ensure that BP is actually operating the ULC Flare in such a manner as
to actually achieve 98% destruction efficiency; and (4) prohibiting BP from using the ULC Flare
unless and until the previous three steps have been completed, and thus that BP can demonstrate
that the ULC Flare can be operated at a 98% destruction efficiency. Docket Entry No. 49 at 16.
40 / 42
2013. Since BP no longer owns the Refinery, it clearly would not be able to
implement changes to the Ultracracker Flare there. Thus, as Plaintiffs’ counsel
recognized at the class certification hearing, “at least on the current state of the
pleadings, that moots [Plaintiffs’ request for] injunctive relief.” Docket Entry No.
75 at 7.
Second, even if the injunction were still part of the relief Plaintiffs seek,
certification would still be inappropriate under (b)(1) and (b)(2) because the
individualized monetary relief that Plaintiffs seek is not merely “incidental to the
injunctive or declaratory relief,” as required by the case law interpreting Rule
23(b). Wal-Mart, 131 S. Ct. at 2557 (“[A]t a minimum, claims for individualized
relief . . . do not satisfy [Rule 23(b)(2)].” (emphasis omitted)); see also Allison v.
Citgo Petroleum Corp., 151 F.3d 402, 415 (5th Cir. 1998) (“[M]onetary relief
predominates in (b)(2) class actions unless it is incidental to requested injunctive or
declaratory relief.” (citation omitted)); Altier v. Worley Catastrophe Response,
LLC, 2011 WL 3205229, at *14–15 (applying Wal-Mart and Allison in the context
of Rule 23(b)(1) and collecting cases).18
The Fifth Circuit has defined incidental damages as those “that flow directly
from liability to the class as a whole on the claims forming the basis of the
18
Though Wal-Mart and Allison are specific to Rule 23(b)(2), their reasoning applies with equal
force to Rule 23(b)(1) because both Rules contain the same due process concerns for a
mandatory class. See Wal-Mart, 131 S. Ct. at 2558–59 (explaining that “individualized
monetary claims belong in Rule 23(b)(3)” due to the procedural protections attending the (b)(3)
class including predominance, superiority, mandatory notice, and the right to opt out).
41 / 42
injunctive or declaratory relief.”
Allison, 151 F.3d at 415 (citing statutorily
mandated awards as an example of incidental damages). The damages should “not
[be] dependent in any significant way on the intangible, subjective differences of
each class member’s circumstances.” Id.
Because each of the putative Plaintiffs seeks individualized damages for
property value diminution—which vary significantly from plaintiff to plaintiff and
do not flow directly from their request for injunctive relief—the Court determines
that Plaintiffs fail to meet the requirements of certification under Rule 23(b)(1) and
(b)(2) as propounded in Wal-Mart and Allison.
V.
Conclusion
For the reasons discussed above, the Court rules that:
BP’s Motion to Exclude Proposed Expert Testimony of Dr. Robert
A. Simons (Docket Entry No. 36) is GRANTED;
Plaintiffs’ Motion for Class Certification (Docket Entry No. 26) is
DENIED; and
BP’s Motion to Exclude Proposed Expert Testimony of Dr. Paul
Rosenfeld (Docket Entry No. 35) is DENIED as moot.
The Court will set a conference to discuss how the case shall proceed.
IT IS SO ORDERED.
SIGNED this 30th day of September, 2013.
___________________________________
Gregg Costa
United States District Judge
42 / 42
Disclaimer: Justia Dockets & Filings provides public litigation records from the federal appellate and district courts. These filings and docket sheets should not be considered findings of fact or liability, nor do they necessarily reflect the view of Justia.
Why Is My Information Online?