Opinion
Case No. 16-CV-12541
11-25-2020
JASON COUNTS, et al., Plaintiffs, v. GENERAL MOTORS LLC and ROBERT BOSCH LLC, Defendants.
Magistrate Judge Patricia T. Morris
OPINION AND ORDER GRANTING DEFENDANTS' MOTION TO ENFORCE SCHEDULING ORDER
On July 7, 2016, Plaintiffs filed a complaint against Defendant General Motors ("GM") alleging violations of the RICO statute and multiple state fraud statutes. ECF No. 1. Plaintiffs allege that
[i]n the wake of the major scandal involving Volkswagen and Audi diesel vehicles evading emissions standards with the help of certain software that turns off emissions controls when the vehicles are not being tested, reports and vehicle testing now indicate that General Motor's (GM) so called 'Clean Diesel' vehicle, the Chevrolet Cruze (Cruze), emits far more pollution on the road than in lab tests and that these vehicles exceed federal and state emission standards. Real world testing has recently revealed that these vehicles emit dangerous oxides of nitrogen (NOx) at levels many times higher than (i) their gasoline counterparts, (ii) what a reasonable customer would expect from a 'Clean Diesel,' and (iii) United States Environmental Protection Agency maximum emissions standards.Id. at PageID.12-13. Plaintiffs alleged, "GM has programmed its Cruze to turn off or otherwise limit the effectiveness of the emission reduction systems during driving conditions below 50°F and above 85°F, and emissions exceed U.S. limits by 1.8 to 13.8 times in other real-world driving conditions." Id. at PageID.14.
Plaintiffs further allege that "GM manufactures, designs, markets, sells, and leases the Cruze 'Clean Diesel' vehicle as if it were a 'reduced emissions' car that complies with all applicable regulatory standards, when in fact, this GM vehicle is not 'clean diesel' and emits more pollutants than allowed by federal and state laws—and far more than their gasoline fueled counterparts and far more than what a reasonable consumer would expect from a 'Clean Diesel.'" Id. at PageID.16. In the complaint, Plaintiffs indicate they used a portable emissions measurement system ("PEMS") to test the Chevy Cruze and found the emissions were 1.8 to 8 times higher in the Chevy Cruze than U.S. emissions standards. Id. at PageID.55-56. Specifically, Plaintiffs allege that they and other putative class members overpaid for their vehicles due to the promised emissions standards. See e.g., id. at PageID.64-65, 68, 74-75.
To furnish some historical context, on January 11, 2017, a third superseding information was filed against Volkswagen AG alleging Volkswagen committed one count of conspiracy from approximately May 2006 to approximately November 2014 to defraud the US and violate the Clean Air Act, one count of obstruction of justice, and one count of entry of goods by false statement. See Case No. 16-20394 at ECF No. 32. In a plea agreement dated March 10, 2017, Volkswagen admitted that they "(a) knew that the Subject Vehicles and the Porsche Vehicles did not meet U.S. emissions standards; (b) knew that VW was using software to cheat the U.S. testing process by making it appear as if the Subject Vehicles and the Porsche Vehicles met U.S. emissions standards when, in fact, they did not; and (c) attempted to and did conceal these facts from U.S. regulators and U.S. customers." ECF No. 68 at PageID.1441. Specifically,
VW AG employees . . . designed the VW defeat device to recognize whether the vehicle was undergoing standard U.S. emissions testing on a dynamometer (or 'dyno') or whether the vehicle was being driven on the road under normal driving conditions. . . . If the vehicle's software detected that it was being tested, the vehicle performed in one mode, which satisfied U.S. NOx emissions standards. If the defeat device detected that the vehicle was not being tested, it operated in a different mode, in which the effectiveness of the vehicle's emissions control systems was reduced substantially, causing the vehicle to emit substantially higher NOx, sometimes 35 times higher than U.S. standards.
On October 3, 2016, GM filed a motion to dismiss. ECF No. 12. Defendant argued Plaintiffs lack standing "because they do not allege a concrete and particularized injury." Id. at PageID.484-85. Second, Defendant argued Plaintiffs claims are preempted by the Clean Air Act ("CAA"), and so the claims should be stayed and referred to the EPA and California Air Resource Board. Id. Finally, GM argued Plaintiffs claims "do not meet basic pleading requirements." Id. Specifically, GM contended that "Plaintiffs' breach of contract claims do not allege any contract with GM or any breach by GM of a specific contract provision," the "fraud-based consumer protection and fraudulent concealment claims do not meet the requirements of particularity in Rule 9(b)" and that "[t]here are no factual allegations that any plaintiff actually saw, or relied on to his or her detriment, any specific representation that GM made about the diesel Cruze's emissions, much less the 'who, what, when, where, and how' required by Rule 9(b)." Id. at PageID.485.
On February 14, 2017, GM's motion to dismiss was granted in part. ECF No. 21. First, Plaintiffs were found to have standing based upon their "allegations that they overpaid for the vehicle based on GM's representations." Id. at PageID.764. Second, it was determined that Plaintiffs claims "are not preempted by the [C]AA" because "Plaintiffs' claims are not, as GM contends, contingent on proving that GM is in noncompliance with EPA emissions regulations." Id. at PageID.780. Third, "[g]iven the limited relevance of an EPA decision on the Cruze's regulatory compliance to Plaintiffs' claims and the significant delay a stay would produce, invocation of the primary jurisdiction doctrine would be inappropriate." Id. at PageID.781-82. Fourth, the parties agreed that Plaintiffs' breach of contract claims should be dismissed without prejudice, and this Court concluded there were sufficient facts alleged for the fraudulent concealment and consumer protection claims. Id. at PageID.782.
It was also decided that "Plaintiffs have standing to bring claims on their own behalf [and t]he question of whether they may bring claims on behalf of the unnamed class members is an issue that is properly addressed via a motion for class certification." ECF No. 21 at PageID.772-73.
On April 21, 2017, the first case management order was issued. ECF No. 28. On September 26, 2017, the Court granted the parties' request for an extension of the scheduling order and extended discovery until March 30, 2018. ECF No. 38. On May 9, 2018, a stipulation was entered extending the case management order a second time. ECF No. 92. On June 11, 2018, Plaintiffs filed an amended complaint adding Bosch LLC and Bosch GmbH as Defendants. ECF No. 93, 94.
On June 11, 2018, a request for summons was filed for Bosch GmbH and Bosch LLC. ECF No. 98. On August 13, 2018, Bosch LLC filed a motion to dismiss Plaintiffs' amended complaint, which GM joined. ECF Nos. 108, 109. The motion to dismiss was denied. ECF No. 122. On November 30, 2018, a stipulation was granted extending the case management order a third time. ECF No. 125. On April 30, 2019, a fourth stipulation to extend the scheduling order was entered. ECF No. 166. Throughout discovery, a plethora of discovery motions were referred to Magistrate Judge Morris. See e.g., ECF Nos. 259, 291. Fact discovery concluded on July 31, 2019. ECF No. 166 at PageID.12838.
On October 8, 2019, a stipulation was entered extending the case management order for a fifth time. ECF No. 284. On December 23, 2019, a stipulation was entered amending the scheduling order a sixth time. ECF No. 309. A show cause order was entered on February 14, 2020. ECF No. 316. The order stated that summons for Bosch LLC and Bosch GmbH were issued on June 11, 2018 and Bosch LLC waived service, but no proof of service was entered for Bosch GmbH. Also, at that point in the case, fact discovery had been closed for more than six months. Plaintiffs were directed to show cause why Bosch GmbH should not be dismissed. In response, Plaintiffs filed a notice of voluntary dismissal. ECF Nos. 317, 318, 319.
On February 27, 2020, a stipulation was granted modifying the case management and scheduling order a seventh time. ECF No. 323. On April 6, 2020, the parties submitted a stipulation to suspend deadlines in the case for at least six weeks due to the COVID-19 pandemic. ECF No. 329. This Court recognized the unprecedented challenges brought on by the pandemic but rejected the stipulation because neither party explained that they attempted to use all available means to continue discovery electronically. Id. Subsequently GM filed a motion to stay all deadlines, including discovery, for six weeks due to the difficulty of conducting electronic depositions and counsel's restricted access to documents and witnesses. ECF No. 330. On April 13, 2020 the requested stay was denied, but a six-week extension of deadlines was granted. ECF No. 332.
On June 29, 2020, this Court rejected the parties' further stipulation to extend dates to allow for rebuttal expert testimony. ECF No. 333. The stipulation explained that total expert reports were 500 to 650 pages long and that Plaintiffs sought time for rebuttal expert testimony on the topics of engineering, damages, and marketing. Id. at PageID.17867-71. Plaintiffs identified nine subtopics under engineering, three under damages, and four under marketing. Id. The stipulation was rejected because discovery had been ongoing for three years and "[t]he parties have not furnished a compelling justification for an additional third round of expert reports and additional testimony." ECF No. 333 at PageID.17866. After the stipulation was rejected, neither party filed a motion for rebuttal experts, nor provided an additional explanation for their desire for rebuttal experts. On August 28 and 29, 2020, Plaintiffs and Defendants filed Daubert motions and motions for summary judgment. ECF Nos. 337-49.
On September 28, 2020, Defendants filed the instant motion to enforce the Court's Scheduling Order and for expedited briefing, which is the primary subject of this Opinion and Order. ECF Nos. 351, 352. Expedited briefing was directed. ECF No. 353. On October 9, 2020, Plaintiffs filed a motion for leave to file a sur-reply and attached their proposed sur-reply. ECF No. 363. On October 13, 2020, this Court suspended the Daubert and summary judgment motion response deadlines pending this Court's determination on the motion to enforce the scheduling order. ECF No. 367.
Throughout this case, as earlier noted, multiple motions to compel, motions to strike, and motions to seal were filed and adjudicated by the undersigned and Judge Morris. See e.g., ECF Nos. 65, 67, 74, 84, 141, 164, 198, 199, 294.
I.
Defendants contend that on September 4 and 8, 2020—a week after summary judgment motions and Daubert motions were filed—Plaintiffs disclosed supplemental testing by their PEMS testing expert, Mr. Juston Smithers. ECF No. 351 at PageID.25138. Defendants argue that the additional testing disclosure is outside the bounds of FRCP 26(e) and should be excluded under FRCP 37(c)(1). Plaintiffs believe the supplemental disclosures are appropriate. Specifically, Plaintiffs argue Defendants' motion is premature, supplemental testing is proper under Federal Rule of Evidence 702, the supplemental testing is necessary to rebut Defendants' experts, and it is properly within the bounds of FRCP 26(e). ECF No. 360.
A.
i.
On August 29, 2020, GM filed an omnibus motion to exclude the expert opinions of Juston Smithers, Kirill Levchenko, Venkatesh Shankar, and Edward Stockton, ECF No. 344. Mr. Smithers is Plaintiffs' PEMS emissions expert and Mr. Levchenko is Plaintiffs' software expert. Defendant offered the testimony of Ryan Harrington to criticize Plaintiffs' experts. Defendant argues in its motion that
Smithers' opinions are based on testing a single used Cruze vehicle that: (i) failed the EPA emissions test that its alleged 'defeat device' was designed to pass, (ii) had numerous mechanical problems that interfered with it emissions controls, (iii) was subject to an open recall related to emissions control technologies at issue in this case, and (iv) may have a host of other issues about which Defendants and this Court will never know because Smithers' team deleted (spoliated) critical vehicle code information just before GM was set to inspect the vehicle. His attempt to 'extrapolate[] from this minuscule sample an inference that one or more defeat devices exist in all of the Subject Vehicles' is inherently unscientific, unreliable, and requires exclusion at the onset. These reliability missteps are compounded by Smithers' (or Plaintiffs' attorneys') inexplicable decision not to test another Cruze—not even any of the named Plaintiffs' own vehicles.ECF No. 339 at PageID.19317 (footnote omitted) (emphasis in original). Defendant further argues that Mr. Levchenko
never examined nor tested any named Plaintiff's vehicle. Instead, he examined selected portions of the Cruze's software functions and calibration values, claiming to 'find' two features that purportedly could result in higher emissions under certain driving conditions. But . . . GM disclosed these long ago to the EPA and CARB . . . when the Cruze was certified. Levchenko is a computer scientist who specializes in cybersecurity and has no education or training in automobile mechanics, diesel or gasoline engines, emissions controls, or emissions regulations. He therefore cannot evaluate the engineering explanations that GM disclosed to EPA and CARB in the certification process and opine that these disclosed strategies are 'cycle beaters.'Id. at PageID.19318.
ii.
Mr. Smithers received a Bachelor of Science in Chemical Engineering from the University of California, Berkley. ECF No. 337-10 at PageID.18526. He is currently the Chief Technology Officer of 44 Energy Technologies Incorporated. Id. For this case, he tested a 2015 Cruze diesel and a 2015 Cruze gasoline engine. Id. at PageID.18529. The Cruze diesel vehicle was Certified Pre-Owned, accident-free, and had 16,290 miles at the time of purchase. Id. at PageID.18544. Mr. Smithers tested the diesel Cruze vehicle with a chassis dynamometer and a portable emissions measurement system (PEMS). Id. at PageID.18546. The PEMS testing occurred in temperatures between -2 and 103° F and on flat and graded roads in the city and highway. Id. at PageID.18546-47. He concluded that, according to his testing, "the emissions quantified on the test cycles during the certification process are not representative of emissions in normal operation and use" and "emissions are often as high as 36 times the relevant standard and in excess of emissions from the comparable gas vehicles" due to software manipulation. Id. at PageID.18531.
Dynamometer testing is essentially placing the vehicle on a treadmill in a regulated, systematic, and therefore repeatable set of conditions. The EPA has multiple tests that are conducted using the dynamometer, including the FTP-75, HWFET, US06, SC03, and Cold CO. ECF No. 337-10 at PageID.18539-43. PEMS tests emissions while the vehicle is being driven with the testing apparatus on board the vehicle. Dynamometer testing simulates vehicle exhaust while PEMS testing tests the actual exhaust emissions. ECF No. 337-10 at PageID.18548. Mr. Smithers provides that "the analyzers used in chassis dynamometer testing and PEMS testing have virtually identical levels of accuracy and are subject to the same strict requirements for calibration and drift." Id. at PageID.18548.
During the PEMS testing, Mr. Smithers found that the diesel vehicle averaged 4.1 times the federal standard of NOx emissions in city driving conditions. Id. at PageID.18553. Mr. Smithers explained that he filtered out trips less than 3.6 miles and downhill and uphill grades from the data prior to his analysis to better match real world driving to the dynamometer testing. Id. at PageID.18558. On the highway, the NOx emissions were 2.4 times the federal standard on average. Id. at PageID.18555. Only data from flat roads and speeds below 65 mph were analyzed, and "cold starts, hot starts, and active regenerations" were removed from analysis. Id. at PageID.18561. The NOx emissions were significantly higher than federal standards at low temperatures (15° F and lower) and high temperatures (90° F and higher). Id. at PageID.18562.
"Oxides of nitrogen (nitric oxide and nitrogen dioxide, collectively called NOx) are of particular concern [in emissions] because NOx pollution contributes to nitrogen dioxide, particulate matter in the air, and reacts with sunlight in the atmosphere to form ozone." ECF No. 337-10 at PageID.18527.
For the chassis dynamometer, Mr. Smithers conducted multiple EPA tests, the FTP-75, HWFET, and Cold CO test cycles. Id. at PageID.18570. In his report, Mr. Smithers states that the results from the dynamometer testing "confirm that the vehicle is in proper working order and that the emission control systems are functioning as designed." Id.
Mr. Smithers believes the PEMS testing results demonstrate GM's use of "defeat devices"—that is, "an auxiliary emission control device (AECD) that reduces the effectiveness of the emission control system under conditions which may reasonably be expected to be encountered in normal vehicle operation and use" unless certain requirements are met. Id. at PageID.18571 (quoting 40 C.F.R. § 86.1803-01). He argues that there is an online dosing defeat device (which treats emissions after they exit the engine) and a high and low ambient temperature defeat device (which treats emissions prior to leaving the engine).
An AECD is defined as "any element of design which senses temperature, vehicle speed, engine RPM, transmission gear, manifold vacuum, or any other parameter for the purpose of activating, modulating, delaying, or deactivating the operation of any part of the emission control system." ECF No. 337-10 at PageID.18571.
a.
According to Mr. Smithers, there are two dosing strategies used by GM to regulate NOx emissions, model dosing and online dosing. Dosing strategies are "meant to achieve the proportional injection" of NOx and diesel exhaust fluid ("DEF") in the selective catalytic reduction ("SCR") aftertreatment system. The SCR system treats NOx and other emission pollutants after they leave the engine system but before they are released through the tailpipe. Id. at PageID.18575. Dosing is "designed to reduce engine-out emissions of NOx with a high degree of effectiveness." Id. at PageID.18572-73. Mr. Smithers believes model dosing is legally permissible, but that online dosing is a defeat device.
Model dosing "targets a constant concentration of ammonia (NH3) on the surface of the [SCR] catalyst for a given temperature" "to maintain maximum conversion efficiency." Id. In model dosing, a computer algorithm determines the amount of ammonia needed for the catalyst in an attempt to reach "exact stoichiometric proportion" of NH3 (ammonia) to NOx. Id. Generally, "[t]he target ammonia load decreases as SCR temperature (a reflection of exhaust gas temperature) increases." Id. A simplified image of model dosing from Mr. Smithers report (Id. at PageID.18574) is below.
Image materials not available for display.
Figure 10-2: Simple representation of "model dosing" SCR operation. Every molecule of ammonia used by the reaction to reduce NOx is replaced to maintain a constant concentration of ammonia on the catalyst. The ammonia to NOx ratio (alpha) is 1.0 in this example.
Online dosing, as explained by Mr. Smithers, occurs when the catalyst is overwhelmed with NOx particles, resulting in greater NOx emissions. Id. at PageID.18574-75. The "capacity of the catalyst has been exceeded" so the proportional introduction of DEF will not result in the ideal conversion as in model dosing, but rather unreacted ammonia will "slip" through the catalyst and result in "ammonia slip" being emitted. Id. When online dosing is used, DEF injection rates are reduced, thereby reducing DEF consumption. Id. at PageID.18577. Mr. Smithers argues that "GM and Bosch use online dosing in many circumstances where it is not technically justified [and i]nstead they use it as a means of reducing consumption of DEF and extending the DEF tank refill interval" which results in higher NOx emissions. Id. at PageID.18573. In the Chevy Cruze, online dosing is turned on "if the 70 second filtered average NOx rate from the engine is 22 mg/s or greater." Id. at PageID.18588. However, it "remains active until engine out NOx drops below 9 mg/s." Id. The following image from Mr. Smithers' report demonstrates a simplified perspective of online dosing. Id. at PageID.18575.
Image materials not available for display.
Figure 10-3: Simple representation of "online dosing" SCR operation. There is no stored ammonia on the catalyst to react with NOx. Only the ammonia injected can react in real time. The ammonia to NOx ratio (alpha) is 0.5 in this example.
On highway driving, Mr. Smithers found that 15.3% of miles traveled had over 22 mg/s of engine out NOx emissions and 63% had more than 9 mg/s emissions. For city driving, 4.6% of miles had more than 22 mg/s and 26% were over 9 mg/s NOx emissions. Id. The report highlights multiple tests when online dosing was activated when engine out NOx emissions ("emissions emitted from the engine") were greater than 22 mg/s, but because the disable mechanism requires 9 mg/s, online dosing continued longer than necessary, even after engine out NOx emissions were below 22 mg/s. Id. at PageID.18532, 18588-98.
Mr. Smithers' report explains that manufacturers are required to disclose the presence of any AECD to regulators. Emission Increasing AECDs (EI-AECDs), AECDs approved by the regulatory body, "reduce[] the effectiveness of the emission control system under conditions which may reasonably be expected to be encountered in normal vehicle operation and use; and the need for the AECD is justified in terms of protecting the vehicle against damage or accident." Id. at PageID.18598-99. Mr. Smithers contends that the GM online dosing should have been submitted as an EI-AECD by GM, but even then, the strategy is not justified. Id. He continues,
The strategy [of GM's EI-AECD] is designed not by the limits of the chemistry and physics of the system but rather by the particular timing and features unique to the US06 [test] in order to avoid a significant impact on NOx emissions, and the strategy is tailored in such a way that online dosing duration is far longer in real world driving, with resulting excessive emissions.Id. He further argues "the disablement threshold is not dictated by chemistry and physics, but rather by the goal of producing a low tailpipe NOx number on the US06." Id. at PageID.18603.
b.
Second, Mr. Smithers argues that the high NOx emissions at high ambient temperatures
are the result of the use of a technically unjustified adjustment to the EGR function [exhaust gas recirculation which reduces NOx emissions from leaving the engine] at high ambient temperature; an undisclosed cycle beating strategy for the SC03 test presented as representing the high ambient temperature adjustments to EGR; and an undisclosed inaccurate vehicle model for ambient temperature that results in the systematic over-estimation of ambient temperature and the activation of EGR reductions when actual ambient temperatures are lower than those modeled by the vehicle.Id. at PageID.18607. He explains that "[a]lthough tailpipe NOx emissions are always the composite result of all the emission control devices (injection timing, EGR, and SCR), the high emissions at ambient temperatures above 86°F are driven primarily by adjustments to the EGR rate which, in turn, trigger online dosing." Id. at PageID.18607. The Cruze does not calculate ambient air temperature directly, but "rather [it is] calculated based on the intake air temperature and [a] computer model that accounts for other parameters." Id. at PageID.18613. In his PEMS testing, Mr. Smithers attached a temperature probe to the top of the vehicle to directly measure ambient temperature. Id. He then compared the calculated ambient air temperature from the Cruze with the temperature probe. Mr. Smithers also reported higher NOx emissions from lower ambient temperatures based on the EGR system. Id. at PageID.18630-40.
iii.
Plaintiffs' software expert is Mr. Kirill Levchenko, ECF No. 339-2. He is an Associate Professor at the University of Illinois at Urbana-Champaign and works for the Department of Electrical and Computer Engineering. Id. at PageID.19577. In 2008 he earned his Ph.D. in Computer Science and Engineering from the University of California, San Diego, focusing on "networking, computer security and related technologies." Id. He reviewed two "potential cycle-beating strategies implemented in software,"
(1) manipulating the vehicles' ambient air temperature model resulting in low emissions on an SC03 dynamometer test but significantly higher emissions during real-world driving and (2) an alternate Diesel Exhaust Fluid (DEF) dosing mode in the Selective Catalytic Reduction (SCR) system that decreases the amount of DEF used at the expense of increased emissions.ECF No. 339-2 at PageID.19574. Mr. Levchenko tested an EDC17 Engine Control Unit (ECU) from a 2014 diesel Chevy Cruze. Id. at PageID.19574. He explains that the ECU "is a computer that controls nearly all aspects of a vehicle's engine and exhaust treatment, including mechanisms designed to reduce harmful tailpipe emissions," including the position of the Exhaust Gas Recirculation valve and the SCR system. Id. He states that "[c]omputerized engine control gives carmakers precise moment-to-moment control over how a vehicle behaves under different conditions," such as "vary[ing] the amount of exhaust gas recirculated based on outside air temperature or change how much DEF is injected into the exhaust based on the temperatures of the exhaust gas." Id. at PageID.19574-75. He formed two opinions based upon review of software documentation from GM and calibration data from Bosch.
First,
[T]he vehicles do not measure ambient air temperature directly. Instead, they estimate the outside air temperature using several other sensors, including temperature sensors in other parts of the engine. The ECU software in the 2014 Chevy Cruze diesel cars is biased to produce lower temperature estimates during a high-temperature SC03 test cycle when that cycle is carried out within 5 hours of being operated at normal ambient temperatures. Because exhaust gas recirculation is reduced at temperatures above 30 °C, and turned off at temperatures above 42 °C, keeping the outside air temperature estimate artificially low during the SC03 test causes the 2014 Chevy Cruze diesel vehicles to produce lower NOx emissions during the test than under real-world driving conditions that are otherwise equivalent to the test.Id. at PageID.19576. In support of this opinion, Mr. Levchenko argues "[m]odeled ambient temperature is lower during the SC03 test because of a combination of three mechanisms that appear to have been intentionally abused by [a GM or Bosch employee] in response to concerns over increased NOx emissions during SC03 testing." Id. at PageID.19586-87. First, the "modeled outside air temperature is smoothed using a low-pass filter" where "the time constant for a decreasing temperature is 3 seconds, while the time constant for increasing temperature is over 1073.7 seconds" allowing the model to slowly adjust to a higher temperature during the SC03 testing. Id. Second, the "ECU is configured to store the current modeled temperature in non-volatile memory" so that "it initializes its modeled temperature estimate to the stored value from the previous operating cycle if the previous cycle was fewer than 5 hours ago." Id. at PageID.19587-88. This is relevant because "the SC03 test takes place within 2 hours of FTP/HWFET/US06 testing" which are conducted at a lower ambient temperature than the SC03 test. Id. Third, the low-pass filter stops when the vehicle speed is less than 34.2 mph. Id. "Because 79% of the SC03 test is at speeds below this threshold, the model only updates for 126 seconds during the 600-second test [which c]ombined with the time constant and memory mechanisms . . . this keeps the model reporting substantially lower ambient temperatures." Id.
Second, Mr. Levchenko opined,
At any given time, the vehicle calculates the amount of DEF it needs to use using either the normal dosing or the online dosing formula. However, the online dosing formula limits the NOx conversion efficiency to at most 65%, leading to increased NOx emissions. Moreover, 2014 Chevy Cruze diesel vehicles switch to online dosing when the amount of NOx entering the SCR system increases above 22 mg/s and does not leave online dosing mode until the amount of NOx entering the SCR system falls below 7.5 mg/s.Id. at PageID.19576-77.
iv.
Defendant's expert, Mr. Harrington, is a Principal at Exponent, "an engineering and scientific consulting firm," and has "more than 20 years of experience working in the automotive industry and the federal government." ECF No. 339-7 at PageID.19871. He earned his Bachelor of Science in Mechanical Engineering from the University of Nebraska and a Master of Science in Automotive Engineering from the University of Michigan, Ann Arbor. Id. He reviewed Mr. Smithers report and underlying testing data and Mr. Levchenko's report. Mr. Harrington did not conduct his own testing on vehicles, but rather analyzed Mr. Smithers testing data. He criticized Mr. Smithers' testing methodology as well as his conclusions. In his report, Mr. Harrington opines, among other things,
3. Mr. Smithers' comparison of on-road PEMS measurements to federal certification standards is invalid because, among other things, on-road measurements obtained from PEMS are not directly comparable to highly regulated and prescriptive standardized dynamometer test measurements, Mr. Smithers oversampled extreme temperature conditions (low and high), and Mr. Smithers' data segmentation is flawed.
4. Extrapolating the results from PEMS testing conducted on a single vehicle to an entire population of vehicles is invalid and inconsistent with sound engineering judgment and practice. This is especially true where the one vehicle tested had a
number of problems, including: a poorly documented NOx sensor replacement, abnormally high NOx results when tested on standardized dynamometer test cycles, and uncertainty regarding the presence of DTC codes, the impact of a significant leak in the emissions after-treatment system, air conditioner use during testing, and vehicle maintenance.ECF No. 339-7 at PageID.19879.
5. Mr. Smithers' comparisons between the Diesel Test Vehicle and the Gasoline Test Vehicle is an apples-to-oranges comparison that renders any conclusions invalid.
6. Auxiliary Emissions Control Devices (AECDs) are a typical aspect of vehicle design that are used to modulate and control systems that impact vehicle emissions. GM clearly disclosed the AECDs to the EPA and CARB, in detail, as part of the Subject Vehicles' certification process, including the Online Dosing AECD and the EGR AECD.
7. Mr. Smithers misinterpreted normal variation in diesel vehicle emissions caused by extreme operating conditions such as high and low temperatures, increased engine load (for example, from road grade, additional vehicle payload, vehicle speed, electrical load, air conditioning use, wind speed and/or aerodynamic drag, etc.), and combinations of these conditions. His reliance on these variations to conclude there is a defeat device is methodologically flawed and unreliable.
8. I have not identified any evidence of a defeat device or cycle detection behavior in the operation of the Online Dosing AECD, the OAT model [the calculated ambient air temperature process], or increased engine loads. Mr. Smithers' opinions to the contrary are unsupported and unreliable because, among other things, they reflect the failure to consider valid engineering reasons for the AECDs. Further, 'fixing' the three 'problems' that Dr. Levchenko identified regarding the Subject Vehicles' OAT Model would actually make the model less representative of the true outside air temperature and, under certain conditions, may result in increased NOx emissions.
B.
Expert reports must be supplemented when required under Rule 26(e). Fed. R. Civ. P. 26(a)(2)(E). Under that rule, a report must be timely supplemented "if the party learns that in some material respect the disclosure or response is incomplete or incorrect, and if the additional or corrective information has not otherwise been made known to the other parties during the discovery process or in writing." Id. at 26(e)(1)(A). For expert witnesses in particular, "the party's duty to supplement extends both to information included in the report and to information given during the expert's deposition." Id. Supplemental disclosures under FRCP 26(e) "must be made at least 30 days before trial." Id. at 26(a)(3)(B). A party may file objections to the supplemental disclosures within 14 days. Id.
The "main purpose of Rule 26(a) is to ensure that parties can properly prepare for cross-examination and that 'deposition testimony can proceed with parties already armed with the expert's report, so as to be able to evaluate the opinions to be offered.'" Tanner v. Grand River Navigation Co., Inc., 2015 WL 8310291, at *3 (E.D. Mich. Dec. 9, 2015) (quoting Beller ex rel. Beller v. United States, 221 F.R.D. 696, 700 (D.N.M. 2003)).
"Rule 26 should not be read so narrowly to prevent an expert witness at trial [or at summary judgment] from (1) rebutting the analysis by another expert or (2) clarifying his or her opinion." Hochstein v. Microsoft Corp., 2006 WL 8066573 at *5 (E.D. Mich. 2006). Therefore, the court must "determine initially what, if any, of Plaintiff's expert's declarations are 'new.'" Id. Specifically, "the Court should determine initially whether the expert is attempting to rebut the analysis of another expert or to clarify his or her position, comporting with the 'general scheme of the report.'" Id. "A harmless violation is one that involves an honest mistake, combined with sufficient knowledge by the adversary." Tanner v. Grand River Navigation Co., Inc., 2015 WL 8310291, at *2 (E.D. Mich. Dec. 9, 2015) (citing Roberts ex rel. Johnson v. Galen of Va., Inc., 325 F.3d 776, 783 (6th Cir. 2003)).
"District courts have broad discretion to exclude untimely disclosed expert-witness testimony." Pride v. Bic Corp., 218 F.3d 566, 578 (6th Cir. 2000). If a party does not comply with the requirements of Rule 26(a) or (e), then it is subject to the automatic and mandatory sanction of Rule 37(c)(1). See Dickenson v. Cardiac and Thoracic Surgery of Eastern Tenn., 388 F.3d 976, 983 (6th Cir. 2004).
Rule 37(c)(1) provides, "[i]f a party fails to provide information . . . as required by Rule 26(a) or (e), the party is not allowed to use that information or witness to supply evidence on a motion, at a hearing, or at a trial, unless the failure was substantially justified or is harmless." Id. The exclusion is "automatic and mandatory . . . unless non-disclosure was justified or harmless." Dickenson, 388 F.3d at 983.
The Sixth Circuit adopted a five-factor test to determine if an omitted or late disclosure is substantially justified or harmless.
(1) [T]he surprise to the party against whom the evidence would be offered; (2) the ability of that party to cure the surprise; (3) the extent to which allowing the evidence would disrupt the trial; (4) the importance of the evidence; and (5) the nondisclosing party's explanation for its failure to disclose the evidence.Howe v. City of Akron, 801 F.3d 718, 748 (6th Cir. 2015).
C.
Before addressing the parties' arguments, a timeline of the events referenced is instructive. On October 2019, Mr. Smithers (Plaintiffs' emissions expert) submitted his expert report. ECF No. 351 at PageID.25155. On November 2019, GM conducted in-use vehicle testing and in December 2019 GM shared the information with the EPA. ECF No. 360-8.
On May 20 2020, Defendants deposed Mr. Smithers. ECF No. 351 at PageID.25155. On June 5, 2020, Defendants shared their own expert report. ECF No. 351 at PageID.25155. Between June 10 and 16, 2020, the parties internally agreed to extend the scheduling order to allow time for rebuttal experts. ECF No. 351-3 at PageID.25183-87. The stipulation was submitted to the Court on June 23, 2020. ECF No. 351 at PageID.25156-57. On June 16, 2020, Plaintiffs purchased a second Chevrolet Cruze. ECF No. 351 at PageID.25152. On June 29, 2020, the parties' stipulation to allow rebuttal experts was denied. ECF No. 333.
On July 7, 2020, however, Plaintiffs purchased a third Chevrolet Cruze. ECF No. 351 at PageID.25152. Between July 15 and August 26, 2020, Plaintiffs' tested the second and third Cruze vehicles and conducted additional testing on a Cruze gasoline vehicle. ECF No. 360-2 at PageID.26787; ECF No. 360 at PageID.26760. On August 28 and 29 the Daubert and motions for summary judgment were filed. ECF Nos. 337-49.
On September 4 and 8, Plaintiffs disclosed the supplemental testing to Defendants for the first time. ECF No. 351 at PageID.25150. On September 9, Plaintiffs discovered GM's undisclosed EPA in-use vehicle testing while working on a different matter. ECF No. 363-1 at PageID.27543. On September 25, 2020, Plaintiffs notified GM about the in-use EPA data. ECF No. 360-8. On September 28, 2020, Defendants filed the motion to enforce the scheduling order and on October 2, 2020, GM provided the in-use data to Plaintiffs. ECF Nos. 351, 360-8.
D.
As an initial matter, Defendants explain that this Court retains jurisdiction over enforcement of its own orders. ECF No. 351 at PageID.25162. The Sixth Circuit has held that a "district court always has jurisdiction to enforce its own orders" and "manage its proceedings." McAlpin v. Lexington 76 Auto Truck Stop, Inc., 299 F.3d 491, 504 (6th Cir. 2000) (citing and quoting Kokkonen v. Guardian Life Insurance Co. of America, 511 U.S. 375, 379-80 (1994)). Additionally, "[i]t is the province of the Court to set deadlines in this case," including for expert disclosures. Tanner v. Grand River Navigation Co., Inc., 2015 WL 8310291, at *3 (E.D. Mich. Dec. 9, 2015). Plaintiffs do not dispute that this Court has jurisdiction to enforce its own scheduling orders.
E.
On October 9, 2020, Plaintiffs filed a motion for leave to file a sur-reply and included a copy of the proposed sur-reply. ECF No. 363. The motion was fully briefed by the end of the day. ECF Nos. 365, 366. Plaintiffs seek to file a sur-reply to "clarify the timeline concerning these materials [regarding the in-use data] and to show that their concealment was not harmless." ECF No. 363 at PageID.27528. Due to the importance of the question of whether Plaintiffs' recent expert discovery should be included, Plaintiffs will be granted leave to file a sur-reply.
F.
Defendants explain that the additional data "is completely new work, with new tests of new vehicles resulting in greater testing data than originally produced during the whole of discovery." ECF No. 351 at PageID.25167. Plaintiffs dissent and maintain the additional data fits within the confines of FRCP 26(e). Their response identifies four arguments. Each will be addressed in turn.
i.
Plaintiffs argue that this Court is unable to determine whether material is "new" "unless and until Plaintiffs offer these materials." ECF No. 360 at PageID.26762-63. Even though this Court has not reviewed the details of the additional data, the parties have provided sufficient information to make a determination regarding admissibility. Plaintiffs admit they purchased two new vehicles, conducted testing on them, and provided the information to Defendants after the motions for summary judgment were filed, without informing Defendants of the supplementary testing. Plaintiffs stand by their decision to test only one vehicle initially. ECF No. 360 at PageID.26772.
In support, Plaintiffs cite to a motion in limine order from West v. Wilke, 2019 WL 4741861 (E.D. Mich. 2019). West v. Wilke was an employment termination case and motions in limine were decided on the eve of trial. Plaintiffs try to analogize the two cases stating,
Modelling the Court's approach in West [waiting until the Court has additional context to make a determination] will allow Plaintiffs to respond to GM's allegations in the context in which Plaintiffs actually offer the supplemental material, rather than based on hypothetical reasons that—according to GM—support a blanket prohibition for any purpose in this litigation.ECF No. 360 at PageID.26762. However, the question in West v. Wilke regarded the credibility of a trial witness. In this case, Defendants have objected to the doubling of testing data shared more than four years into the case. There is no question of credibility to be decided here. Plaintiffs clearly outline why they created the additional data. The question of whether the data is supplementary under FRCP 26(e) may be decided in advance of the Daubert and summary judgment motions.
ii.
In their second and third arguments, Plaintiffs argue the additional testing falls under FRCP 26(e) and is within the general scheme of the original data. Defendants counter by asserting that the testing is not supplementary, but new.
Second, they argue "supplemental disclosures are appropriately offered for Daubert/702 purposes in situations like this where reliability challenges cannot be known before Defendants' expert disclosures are served." ECF No. 360 at PageID.26749. Third, Plaintiffs advance that the additional data "responds to criticisms of Defendants' experts and clarifies an expert's opinion [and therefore] may be offered to help the Court assess the substance of the parties' arguments at the summary judgment stage." Id. --------
Plaintiffs explain that Mr. Smithers "could not have anticipated all of the specific criticisms, nor the unsupported and untested nature of those criticisms" of his initial report. ECF No. 360 at PageID.26752. Specifically, Plaintiffs argue that GM tested a single vehicle for EPA conformity certification, so they could not have expected GM would argue testing one vehicle would be insufficient for this litigation. Id. at PageID.26754-55. Second, Plaintiffs argue there is no evidence of spoliation or manipulated data and that "cross-correlating PEMS measurement systems with a dynamometer" is not industry standard practice as Mr. Harrington asserts and the only way to correct him was to run the tests "demanded by Mr. Harrington." Id. at PageID.26755-57. Third, Plaintiffs argue Mr. Smithers could not have anticipated that "dynamometer test results obtained during two in-use compliance programs" would be used by Mr. Harrington "to demonstrate that the subject test vehicle in Mr. Smithers' report has NOx emissions performance that is inconsistent with the in-use population." Id. at PageID.26756-57. Therefore, the supplemental testing was necessary "to rebut Mr. Harrington's mischaracterization of the data." Id. Fourth, Plaintiffs assert they "could not have anticipated that Defendants would demand that Mr. Smithers' test protocols adhere to regulatory dynamometer test conditions or European on-road testing requirements that have boundary conditions unrelated to U.S. diesel vehicle emissions during real-world driving." Id. at PageID.26758. Fifth, Plaintiffs explain they "could not have guessed that GM's expert would rely on" "PEMS testing conducted by GM in 2015 and 2016 and representations it made to CARB and EPA as a result of those tests." Id. at PageID.26759.
After laying the foundation of the criticisms leveled by Mr. Harrington, Plaintiffs assert "the only reliable way to assess their validity was to conduct additional PEMS tests." Id. at PageID.26760. They explain,
PEMS results on two additional Cruze Diesel vehicles showed substantially the same expected emissions behavior as the original tests. And dynamometer tests confirmed that the vehicles were not somehow anomalous. Finally, testing the PEMS machine against the dynamometer showed that the PEMS equipment was operating normally.Id. (internal citations omitted). Further, the results from the data "amount to a total of 15-20 spreadsheets of processed PEMS data for each vehicle and a Power Point summary of the results." Id. The remaining data "contains quality assurance results for the testing equipment, raw data that is duplicative of the processed data, vehicle maintenance histories, and vehicle photos." Id. at PageID.26761. Plaintiffs conclude that they "do not offer the supplemental data in support of a brand-new theory of liability or to demonstrate a new and improved testing methodology." Id. at PageID.26764. They explain that Mr. Smithers "employed the same methodology and ran the same tests on new vehicles in response to various criticisms by Defendants' experts that the original test vehicle was not representative of Cruze Diesel vehicles in the market." Id.
Defendants argue that Mr. Smithers "had several years to design a rigorous and defensible vehicle emissions test program before he prepared his original expert report." ECF No. 351 at PageID.25167. The additional data "is completely new work, with new tests of new vehicles resulting in greater testing data than originally produced during the whole of discovery." Id.
While Defendants analogize this case to Auto Club Group Insurance Co., 2017 WL 3263355 (E.D. Mich. 2017), Plaintiffs in this case offered a timely and complete expert report. The supplemental data was not offered to address a barebones initial report. That said, Plaintiffs made an intentional decision to conduct extensive expert testing on one Chevy Cruze vehicle. Throughout this case, Plaintiffs were aware it was a putative class action with potentially thousands of class members. ECF No. 94 at PageID.6310. Both parties have aggressively litigated the case over the past four years—from dispositive motions to discovery disputes. Fact discovery was open for three years. Plaintiffs knew the stakes and stand by their opinion that testing a single vehicle was sufficient to prove their case. Defendants' decision to use one vehicle to demonstrate conformity to EPA certification while arguing that Plaintiff's one vehicle is insufficient for Plaintiffs to defeat a motion for summary judgment may be inconsistent. However, Plaintiffs cannot use this fact to justify their belief that Mr. Harrington's argument about the use of a single vehicle being insufficient was unexpected.
Second, Plaintiffs have a right to challenge Defendants' experts. That said, any questions of spoliation or manipulation of data should be included and addressed by the Daubert motions—not introduced through belated rebuttal testing. The scheduling order did not provide for rebuttal experts, primarily because the parties did not initially seek it. Plaintiffs, obviously, now strongly believe that rebuttal expert witness testing was necessary. To address this concern, Plaintiffs could have informed opposing counsel of their intention to conduct additional testing and filed a motion to extend the scheduling order to allow time to conduct rebuttal testing and depositions prior to dispositive motions being filed. Neither occurred.
Plaintiffs argue the additional data simply responds to Defendants' expert and does not offer any new opinions. However, in the same breath, Plaintiffs also argue that they needed to conduct additional testing to confirm their methods and discredit Defendants' theories because they could not have foreseen Defendants' challenges. The additional data cannot be both a simple continuation of Plaintiffs' earlier arguments (a need they could have anticipated) and simultaneously respond to unexpected attacks by Defendants (a need they say could not have been anticipated).
Even though the additional data is in a similar format to the original data set, Plaintiffs chose to test one vehicle and now have decided four years into the case to purchase two new vehicles and conduct additional tests to rebut Defendants' expert. This data doubles the amount of emissions tests conducted and data collected in the case. It triples the number of diesel vehicles that have been tested. This is not a few extra tests Plaintiffs ran after Defendants' expert provided his report. Rather Plaintiffs concluded that they needed to completely re-run tests on new vehicles to ensure their original testing was correct. Plaintiffs cannot now expect that doubling the amount of data and tripling the number of diesel testing vehicles would be considered supplementary.
iii.
Plaintiffs also spend several paragraphs discussing recently discovered GM in-use testing data for a 2015 Chevy Cruze vehicle that was previously not disclosed. However, they do not explain how this non-disclosure is relevant to their argument that their additional data is supplementary. ECF No. 360 at PageID.26761-62, 26768-69. The discovery of the undisclosed evidence occurred in September 2020. The new cars were purchased in June and July 2020 and testing was conducted in July and August 2020. According to the timeline, the newly discovered in-use data could not have been a predicate for the new testing.
It is clear the new data is not a proper supplemental expert disclosure under FRCP 26(e). As such, an analysis of FRCP 37(c)(1) is necessary.
G.
When analyzing if an expert report should be excluded, the first factor to consider is whether the new evidence is a surprise to the opposing party. Plaintiffs purchased the two vehicles in June and July and conducted testing in July and August 2020. To their credit, the purchase and testing of the vehicles occurred soon after Defendants' expert report was provided in May 2020. However, Plaintiffs did not inform Defendants of their intention to purchase and test additional vehicles. Instead, they emailed Defendants seeking concurrence in a stipulation to allow for time for rebuttal experts. Then, after the stipulation was denied, Plaintiffs did not inform Defendants of their ongoing additional expert discovery. Plaintiffs' disclosure after Daubert and summary judgment motions had been filed was untimely by any measure.
The second factor is the ability to cure the surprise. Defendants could seek a short deposition of Mr. Smithers, and in fact, Plaintiffs have offered to make Mr. Smithers available for a rebuttal deposition. ECF No. 360 at PageID.26761. However, this case has been ongoing for more than four years and summary judgment and Daubert motions have already been filed. Any potential to cure the surprise to Defendants would require Mr. Smithers' rebuttal deposition, as well as at least one round of sur-reply briefing for the summary judgment and Daubert motions to allow the additional evidence to be fully briefed. Plaintiffs argue that the hearing date is scheduled for February 24, 2020 and that there is plenty of time to correct any surprise. ECF No. 360 at PageID.26770. However, the date was selected to provide sufficient time for the Court to manage its docket, as well as the parties' expectations on the timing of a decision, not to encourage parties to continue conducting discovery.
Third, Plaintiffs argue that a trial date has not yet been set and therefore allowing the evidence would not delay trial. However, the fact that a trial date has not been selected does not change the fact that this case has been ongoing for more than four years and that Plaintiffs conducted significant additional testing without informing opposing counsel or seeking leave of court. As mentioned in the second factor, inclusion of the new data would push back the time this Court needs to address the underlying motions, thereby delaying a future trial.
Fourth, Plaintiffs' key argument for the additional testing is that they need to rebut Mr. Harrington's baseless accusations and mischaracterizations of the original testing data. These are questions that can be resolved in addressing the Daubert motions. Plaintiffs had an opportunity to test multiple vehicles if they believed the number of vehicles would be an issue and can argue against Mr. Harrington's analysis of their testing methodology in their Daubert motion. The fact that Defendants have attacked Plaintiffs' decision to test one vehicle and have allegedly misconstrued the proper way to interpret and test certain emissions data is not a proper reason for Plaintiffs to double the number of diesel vehicles tested and conduct additional testing to reaffirm their original position.
Fifth, Plaintiffs only explanation for the new data is that "the material is directly responsive to criticisms that they could not anticipate until June 2020 and to in-use data that GM withheld from Plaintiffs and GM's own expert until October 2020." ECF No. 360 at PageID.26772. Again, Plaintiffs do not explain how the withheld in-use data furnishes a rationale for introducing the new data. The two vehicles were purchased in June and July and were tested in July and August 2020. Plaintiffs did not discover the missing in-use testing data until September 2020. Additionally, some of the criticisms of Mr. Smithers could have been foreseen, including attacks on the number of vehicles tested and GM's use of PEMS testing data from 2015 and 2016. Even if every allegation could not have been anticipated, Plaintiffs could have sought concurrence from Defendants or moved for an amended scheduling order during the summer. However, neither occurred. Plaintiffs waited until after two motions for summary judgment and two Daubert motions were filed before disclosing the new testing to Defendants.
As this Court explained in another case, "there is no reason to believe that [Defendants] should have expected [Plaintiffs] to disclose approximately eighty-five pages of new expert reports and opinions a month before the dispositive motion deadline." Auto Club Grp. Ins. Co. v. Omega Flex, Inc., 2017 WL 3263355, at *3 (E.D. Mich. Aug. 1, 2017). Here, Plaintiffs doubled the number of diesel vehicles being tested which resulted in more than 5.7 gigabytes of data and did not disclose the ongoing testing to the opposing party until after motions for summary judgment were filed. The extremely late disclosure, the limited ability to cure the surprise without further delay, and the necessity for additional briefing if the new testing were allowed requires prohibiting Plaintiffs from relying upon the new data in the responses and replies to the motions for summary judgment and Daubert motions.
II.
Accordingly, it is ORDERED that Defendants' Motion to Enforce Scheduling Order, ECF No. 351, is GRANTED. Plaintiffs' may not utilize the testing data collected from July 15 to August 26, 2020 in their response to Defendants' Motions for Summary Judgment and Daubert Motion, or their reply to Plaintiffs' Daubert Motion.
It is further ORDERED that Plaintiffs' Motion for Leave to File Surreply, ECF No. 363, is GRANTED.
It is further ORDERED that responses for Motions for Summary Judgment and Daubert Motions are due on or before December 22, 2020.
It is further ORDERED that reply briefs are due on or before January 14, 2021. Dated: November 25, 2020
s/Thomas L. Ludington
THOMAS L. LUDINGTON
United States District Judge
Id. at PageID.1442-43. One such method was a "steering wheel angle recognition" feature because the steering wheel was not turned during dynamometer testing. Id. at PageID.1449-50. Additionally, the vehicle was designed to minimize the urea and water solution injected into the exhaust gas system, which reduced the necessary tank size to the benefit of a larger trunk—and greater emissions. Id. at PageID.1445.