REPORT TO BOARD OF DIRECTORS OF GENERAL MOTORS COMPANY REGARDING IGNITION SWITCH RECALLS May 29, 2014 Anton R. Valukas JENNER&BLOCK Privil eged and Confide ntial: Protected by Anorney-Client Privilege and as Attorney Work Product Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product TABLE OF CONTENTS I. INTRODUCTION ............................................................................................................. I II. SUMMARY OF F ACTS ................................................................................................... 5 III. JENNER & BLOCK'S INVESTIGATIVE PROCESS .............................................. 12 IV. A. Mandate from GM 's CEO and Its Board ........................................................... 12 B. Document Collection and Review ....................................................... . c. Witness Interviews ............................................................................... ..... .. .... ... 14 D. King & Spalding's Role ....................................................................... ........... ... 15 ...12 BACKGROUND ............................................................................................................. IS A. General Motors ... .... ............................................................................ . .. ...... 15 B. GM 's Small Car Product Line ............................................................. . .. ........ 17 I. The " Delta Platform" of Vehicles ........................................... . .......... .. .. . 17 2. GM 's Marketing of the Cobalt ................................................. ............... 19 3. Cobalt and Delta Platform Sales and Profitability ................... ............... 21 c. GM in the 2000s .................................................................................................. 22 D. Preliminary Information Notices and Technical Service Bulletins E. Recalls ................................................................................................................. 24 F. The Ignition Switch and Airbag Tec hnology in the Vehicles at Issue . .......... .. .. . 24 ............... 24 I. The Ignition Switch ................................................................. . ...... 25 2. The Sensing Diagnostic Module ("SDM") and Airbag System ... 26 G. The Igniti on Switch and Its Impact on Airbag Deployment .............................. 29 H. Safety Regulations that Governed GM 's Vehicles .............................. . ...30 Privileged and Confidenfia/ Protec1ed by AUorney~ Client Privilege and as Allorney Work Product V. 1. Federal Motor Vehicle Safety Standards ................................ . ... 30 2. NHTSA 's Reporting Requirements ....................................................... 3 1 1997 - 2006: DEVELOPMENT, PRODUCTION, AND RE-D ESIGN OF THE IGNITION SWITCH ............................................................................................ 32 A. B. The Development of the Ignition Switch . ........................................... . ........ 34 1. 1997-2001 : The Specification for the Ignition Switch ........... . .......... .. .. . 35 2. 2001-2002: Earl y Changes to the Prototype Ignition Switch 3. 2001: Early Complaints During 4. September 2001: Component-Level Testing of the Ignition Switch.. 5. February 2002: As the Launch of the Ion Approaches, Ignition Switch Torque Tests Fail to Meet Specification ..................................... 45 6. May 2002: Despite Failing Test Results, The Ignition Switch Goes Through Production Part Approval Process ........................... .. .......... .. .. . 50 Vehicle~Level Testing ... .. . 41 ............... 42 44 The Ignition Switch In Production: Early Issues With the Ion and Cobalt ......... 53 I. Fall 2002: Ion Production Starts and Ignition Switch Problems Begin .. .. ................................................................................... . .............. . 54 2. 2003-2004: The Volume of No CrankINo Start Complaints Obscures Stalling Issue .......................................................................... 55 3. August 2004: The Cobalt Goes Lnto Production Using the Same Ignition Switch ....................................................................... .. ............... 57 4. Spring 2004: Issues With Moving Stalls in the Captured Test Fleet ...... 58 5. SummerlFall 2004: Early Reports of Cobalt Moving Stalls .... ............... 59 6. November 2004: Cobalt Moving Stall Is Classified As A NonSafety Iss ue . ............................................................................ . ............... 63 7. March 2005: November 2004 PRTS Closed With No Solution ............. 65 8. 2004-2005: GM Discussions With NHTSA Regarding Engine ~~ ........................................................................................................ n " Privileged and Confidenfia/ Protec1ed by AUorney~ Client Privilege and as Allorney Work Product c. VI. 9. Spring 2005: GM 's Brand Quality Group Learns of Moving Stall Issues ............................... .... ........................................ _ ......... . 75 10. The Cobalt VAPIR Committee Considers S hort~Tenn and Long~ Term Solution s to the May 17, 2005 PRTS ....... .. .. .. .. .. ... .. .. .. .. .. .. .. .. ......... 78 a. June 7, 2005 VAPfR Meeting and DeGiorgio's Ignition Switch Work ................................................................ . .............. . 79 b. June 14, 2005 VAP IR Meeting ................................................... 80 c. June 17, 2005 Ignition Switch Experiment at Milford Proving Grounds ......................................................... _ ............... 8 1 I \. June 2005: GM 's Product In vestigations Group Investigates Cobalt Moving Stalls in Response to Negative Media Coverage ............ 84 12. September 2005: Engineering Committees Reject Proposa ls to Change the Switch ................................................................... _ ............... 90 13. December 2005: GM Issues a Technical Service Bulletin Regarding «Information on Inadverte nt Turning of Key Cylinder, Loss of Electrica l Syste m and No DTCs." .............................. _.. .. .... .. .. 91 14. After Numerous Reports of Moving Stall s, Engineering Work to Resolve the Problem, Mu ltiple Committee Meetings, and a Brief Investigation by GM 's Safety Investigators, No Meaningful Action Is Taken ................................................................................... _ ............... 94 The Ignition Switch Part Change ........................................................ _ ............... 95 1. Late 2005 to Earl y 2006: Discussions About Changing the Spring and Plunger ............................................................................. _ .............. . 96 2. April 2006: DeGiorgio's Approval of the Ignition Switch Design Change .................................................................................... .. ............... 98 3. May 2006: DeGiorgio's Approval of the Design Change Without a Change in Part Number ........................................................ _ ............. 100 2006--2014: THE LONG INVESTlGA TlON INTO ROOT CAUSE ...................... 102 A. 2006: Litigation Into Fatalities Begins ............................................... _ .. .. .. .. .. .. . 103 I. Background on the GM Legal Department ............................. _ ............. 104 111 Privileged and Confidenfia/ Protected by A{{orney~Clielll Privilege alld as Allorney Work Product B. 2. January 2006: _ Roundtable ................................. . 3. September 2006: Evaluation ....................... ..... .. .. .. .. III 4. September 2006: Evaluation .................. ............. 11 2 5. October 2006: Crash ............................. . 2007: Outsiders Find the Solution But GM Remains Unaware I. 110 . .. 113 ............. 11 5 Trooper Young 's Reconstruction Report of Crash ..................................................................... ............. 11 6 C. O. 2. March 2007: N HTSA Expressed Interest in Cobalt NonDeployments ............................................................................ . ...... .. .. .. . 11 8 3. March-April 2007: The TSB Is Edited but the Edi ts Are Not Released ................................................................................................ 120 4. April 2007: The Indiana Universily Repon on _ Acc ident ................................................................................... ............. 121 5. May 2007: Case Evaluation ................................................................................. ............. 124 6. Apri l-September 2007: 7. July 2007: Continental SDM Ana lysis From _ 8. August 2007: Delphi Warranty Settlement Agreement ............. 127 2008: More Fatalities Occur and More Litigation Commences 128 Evaluation and Roundtable .... 125 Crash ............. 126 1. January 2008: Second Evaluation of_ and _ Matters .. .. .... ............................................................................. ..... .. ...... 129 2. July 2008: _ Roundtab le ........................................................ 130 2009: The Investi gation Continues and GM Goes Into Bankruptcy ............. 13 1 1. Late 2008/Early 2009: GM Learn s 2. February 2009: A New PRTS Report Is Opened and the Coba lt Key is Changed . ....................................................................... . ............. 132 IV Crash .. ............. 132 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product 3. E. May 2009: Meeting with Continental Regarding _ Crash SDM... .......................................................................................... 134 2010: A Different Recall and Outs ide Counsel Warns GM of Possible Punitive Damages .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. ............................ ... ... .. .. .. .. .. .. .. _.. .. .. .. .. .. .. 139 1. March 20 10: Cobalt Power-Steering Recall ......................................... 139 2. October 20 I 0: Warning of PUliiiJ.ve Damag.:s::::: Evaluation - The First ............. 140 143 F. 2006 to 20 I 0: Fai lure to Take Basic Steps ......................................... _ G. 2011: Another Engineering Investigation Begins ......... ... ... .. .. .. .. .. .. .. _.. .. .. .. .. .. .. 144 H. 1. January 2011: Pa lmer Presents t h e _ Case and NowakVanderhoef Requests a Meeting That Does Not Occur .......... _ ............. 145 2. July 2011: About PunililVe Damages 3. July 2011: Meeting to Kick Off the FPE Process ................................. 149 4. August 20 11: Stouffer and Sprague Meet to Transition Fi les _ ............. 154 5. Information Sharing Between FPA Engineers and the PI Group ........... 157 6. Summer to Fall 20 II : Stouffer Analyzes TREAD Data Eva luation and Another Warning ............. 148 ............. 158 20 12: The Search for the "Root Cause" ............................................. _ ............. 161 1. February 20 12: 2. March 2012: The Visit to the Da vison Junkyard and Re~ Di scovery of the TSBs ............................................................................................ 165 3. Apri l 2012: About Puniltive Evaluation .............. _ ............. 162 Evaluation and Another Warning oariia&es .. 167 4. May2012: Mechanical ve rsus Electrical RootCause ............ _ ............. 170 5. May 2012: The Second Trip to the Davison Junkyard ........................ 176 6. June 2012: Kemp Persuades Federico to Help ....................... _ ............. 178 v Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product I. 7. June 2012: The Stouffer lnvestigation Continues and He is Told There Are No Relevant Part Changes ..................................... . ........ 179 8. June 2012: The Shipp Expert Report Provides the Correct Answer ..... 180 9. Summer 2012: Stouffer's Investigation Considers Potential Root Causes .................................................................................................... 184 10. September 2012: The Red X Investigation .......................................... 186 11. October 2012: Stouffer Asks DeGiorgio about a New Switch .............. 189 12. October 2012: _ Eva luation, Providing the Correct Answer ... ... .. .. .. .. .. .. .. .. .. .. .. .. .. .. ............................ .. .. .. .. .. .. .. .. .. .. .. .. ... .. .. .. .. . 190 13. November 2012: The Team Exp lores a Mechanical Solution .............. 19 1 20 13: An Outsider Conclusively Demonstrates the Root Cause, But Still No Recall .................................................................................................. 194 1. January 20 13: DFSS Meeting ................................................ . 194 2. February 20 13: The FPE In vest igation is "Stuck" ................. . 196 3. Apri l 2013: GM Lawyers Review the _ Matter at the Roundtable ............................................................................... ............. 199 4. April 2013: Depos itions in t h e _ Case Begin and There is a " Bombshell" at DeGiorgio's Deposition ............................ . .. 199 5. May 2013: Kent ' s "Kick Off' Meeting ............................................... 200 6. May 2013: GM Lawyers and Outside Counsel Review The _ Matter Again .......................................................................................... 203 7. July 2013: Mall adi Presents Hi s Findings Orally ................................ 205 8. August 2013: _ 9. October 2013: Mall adi's Written Presentation ..................................... 207 10. November 2013: Stouffer Arrives at and Presents Hi s Conclusions, Which Are the Same as Trooper Young ' s, the IU Investigators' , and Erin Shipp ' s ............................................................ 209 I I. Conclusions: 20 II to 2013 ................................................................... 211 SRC Meeting ................................... . VI ..206 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product J. VII. The 2013- 2014 Recall Process ........................................................... _ .. __ .. __ .. __ .214 "M ini "~FPERC I. December 2013 Meeting ............................................ 215 2. December 17, 20 13 EFA DC Meeting _.. __ .. __ ............................ _ ............. 2 17 3. Deve lopments Following December 17,2013 EFADC Meeting .......... 220 4. January 3 1, 20 14 EFADC Meeting _.. __ .. __ .. __ ............................ _ ............. 221 5. Deve lopments Following January 31 , 2014 EFADC MeetinglFebruary 7, 2014 Recall Letter .. __ .. _.......................... _ ............. 224 6. February 24, 2014 EFADC Meeting _____ .___ .. __ .. __ .. __ .. __ .. __ .. __ .. __ .. _ 7. March 20 14 Service Parts Recall ....................................... . 225 . .. 226 GM'S LEADERSHIP, THE TONE AT THE TOP, AND GM CULTURE ............ 226 A. Current Senior Leadership .................................................................................. 227 B. The Board I. Board Oversight Process ......................................................... . ............. 233 a_ b. 2. 3. c. .__ .. __ .. __ .. __ .. __ .. __ .. __ .. __ .. __ .. __ .. __ .. __ .. __ .. __ ............................ _ ............. 2D The Full Board .__ .. __ .. __ .______________________ .__ .. __ .. __ .. __ .. __ .. __ .. __ .__ .. __ .. __ .. 233 I. Period ic Reports ............................................................ 233 II. Specific Reports ............................................................ 240 Board Committees ....................................................... . ............. 241 I. Public Poli cy Committee ................................ .. .241 II. Audit Committee .......................................................... 242 The Cobalt 19nition Switch .................................................... ............. . 244 a_ The Board's Lack of Knowledge ................................ . b. The March 5, 2009 BodylExterior SMT VPR Update Slide Deck .. __ .. __ .. __ .. __ .. __ .. __ .. __ .. ___ .__ .__________ .__ .. __ .. __ .. __ .. __ .. __ .. __ .. _ .. __ .. __ .. __ . 245 ..... __ .244 Other Board Activities ............................................................ . .......... .. . 247 Tone at the Top on Safety ................................................................... . VII .__ .248 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product D. VIII. Culture . ...... .. .. .. ................................................................................... . .......... .. . 252 1. Resistance to Raising Issues ..................................................... ..... .. .. .. . 252 2. The "GM Salute" and the "GM Nod" .................................................. 255 3. Failure to Share or Gather KnowledgelInformation Silos ... 4. " Root Ca use" ......................................................................................... 257 ..256 RECOMMENDA T10NS .............................................................................................. 258 A. Organizational Structure ...................................................................... ............. 259 B. Cultural Emphasis on Safety ............................................................................. 260 c. Indi vidual Accountability .................................................................... . .......... .. . 26 1 D. Communications Between and Within Groups ................................................ 262 E. Communications with NHTSA ........................................................................ 263 F. Role of Lawyers ... ............................................................................... . .......... . . 264 G. Interactions with Suppliers ................................................................................ 266 H. Data Storage, Retrieval and Analys is ............................................................... 266 I. Engineering Processes and Databases ................................................. . J. Product Investigation Process ........................................................................... 269 K. FPE Process ......................................................................................... . ............. 270 L. Policies and Training ............................................................................ ............. 272 M. Compliance, Auditing, and Oversight ............................................................... 274 N. Recordkeeping .................................................................................................. 274 o. Conclusion ......................................................................................................... 275 ... . 268 APPENDIX A: NHTSA REPORTING REQUIREMENTS ................................................ 277 APPENDIX B: GM'S VEHICLE INVESTIGATIONS PROCESSES ............................... 282 V111 Privileged and Confidenfia/ Protec1ed by AUorney~ Client Privilege and as Allorney Work Product APPENDIX C: GM'S APPROACH TO POTENTIAL SAFETY PROBLEMS THROUGH VEHICLE DEVELOPMENT PROCESS ............................................ 292 APPENDIX D: PRODUCTION PART APPROVAL PROCESS ("PPAP") ..................... 302 APPENDIX E: GM'S TREAD DATABASE ......................................................................... 306 APPENDIX F: CAPTURED TEST FLEET REPORTS ...................................................... 310 APPENDIX G: GLOSSARY .................................................................................................. 314 IX Privileged and Confidenfia/ Protec1ed by AUorn ey~ Client Privilege and as Allorney Work Product I. INTRODUCTION In the fall of 2002, General Motors ("GM") 1 personnel made a decision that would lead to catastrophic results - a GM engineer chose to use an ignition swi tch in certain cars that was so far below GM 's own specifications that it failed to keep the car powered on in circumstances that dri vers could encounter, resulting in moving sta lls on the hi ghway as well as loss of power on rough terrain a dri ver might confro nt moments before a crash. Problems with the switch's ability to keep the car powered on were known within GM's engineering ranks at the earliest stages of its production, although the circumstances in which the problems would occur were perceived to be rare. From the switch's incepti on to approx imately 2006, vari ous engineering groups and committees considered ways to resolve the problem. However, those indi viduals tasked with fi xing the problem - sophisticated engineers with responsibi lity to provide consumers with safe and re liable automobiles - did not understand one of the most fundamental consequences of the sw itch fa iling and the car stalling: the airbags wou ld not depl oy. The failure of the switch meant that dri vers were without airbag protection at the time they needed it most. This failure, combined with others doc umented below, led to devastating consequences: GM has identifi ed at least 54 fronta l-impac t crashes, involving the deaths of more than a dozen individuals, in which the airbags did not deploy as a poss ible res ult of the faulty ignition sw itch. The b e l ow~ s pec ifi ca ti on switch approved in 2002 made its way into a variety of vehicl es, including the Chevrolet Cobalt. Yet GM did not issue a recall fo r the Cobalt and oth er cars unti l 1 Except as otherwise specifically noted, the lenn "GM" is used Ihroughoulthis Report as follows. When reference is made 10 GM prior 10 July 10, 2009, the reference is to General MOlars Corporalion. On June 1, 2009, General Motors Corporation filed a Chapter II bankmplcy pet ilion in the United States Bankmptcy Court for the Southern District of New York. O n July 5, 2009, that court approved the sale of substantially all of the assets ofGencra l Motors Corporation to an entity that became known as Genera l Motors Company. That sale closed on July 10, 2009. When reference is made to GM from and aft er July 10, 2009, the reference is to General Motors Company. the purchaser of the assets. Privileged and Confidenfia/ Protected by AUorney~Client Privilege and as Allorney Work Product 20 14, and even then the initial recall was incomplete. GM personnel's inability to address the ignition switch problem for over 11 years is a hi story of failures. While GM heard over and over from various quarters - including customers, dealers, the press, and their own employees - that the car' s ignition switch led to moving stall s, group after group and committee after committee within GM that reviewed the issue fa il ed to take action or acted too slowly. Although everyone had responsibility to fix the problem, nobody took responsibility. It was an example of what one top executive described as the "GM nod," when everyone nods in agreement to a pmposed plan of action, but then leaves the room and does nothing. A critical factor in GM personnel ' s initial delay in fi xing the sw itch was their failure to understand, quite simply, how the car was built. GM had specifica lly designed the airbag system nol to deploy, in most circumstances, in the event that the ignition switch was turned to Off or Accessory, a deliberate and sensibl e decision made to prevent passengers from being injured by ai rbags in parked cars. In 2004, however, GM engi nee rs, faced with a multitude of reports of moving stall s caused by the igniti on switch, concluded that moving stalls were not safety issues because drivers could still maneuver the cars; they completely fai led to understand that the movement of the sw itch out of the Run position meant the dri ver and passengers would no longer have the protection of the airbags. As a result, GM personnel viewed the switch problem as a "customer convenience" issue - something annoying but not particularly problematic - as opposed to the safety defect it was. Once so defined , the sw itch problem received less attention, and efforts to fix it were impacted by cost considerations that would have been inunaterial had the problem been properly categorized in the fi rst instance. Indeed, in this same decade, GM 2 Privileged and Confidenfia/ Protec1ed by AUorn ey~ Client Privilege and as Allorney Work Product issued hundreds of recall s at great expense (including at times when its financial conditi on was prec ari o u ~) h eca u~e in th e great majority of in~tan ce~ , it correctl y determined or agreed that the issues that came to its attention impli cated safety and demanded prompt action . But in the case of the Cobalt, it did not do so. From 2004 to 2006, not one of the committees consi dering a fix for the switch - filled with engineers and business peopl e whose job was to understand how GM's cars were built and how d ifferent systems of the car interact - ever recl assified the problem from one of customer convenience to one of safety or demonstrated any sense of urgency in th eir efforts to fi x the sw itch. GM' s Product Investigations group, charged with identifyi ng and remedying safety issues, made the same mi stake; it opened and closed an in vesti gation in 2005 in the span of a month, findin g no safety issue to be remedi ed. As th e earl y committees failed to fi x the problem, acc idents and fataliti es in which airbags did not deploy began coming to GM personnel's attenti on, includ ing GM ' s in~h o use counsel and th e engi neers who worked with them. Those outside GM, including, in 2007, a trooper fro m the Wisconsin Safety Patrol and a research team from Indi ana Univers ity, figured out the connection between the switch and the airbag n on~depl oy m e nt . Yet, GM personnel did not. From 2007 on, as years passed and fatalities mounted, engineers investigating the n o n~ deployments and attempting to understand their cause were neither diligent nor incisive. The in vestigators fail ed to search for or obtain criti cal documents within G M' s own files , or publi cly available docum ents that helped n on~ GM personnel make the connection between the switch and ai rbag no n~d ep l oy m ent. Investi gators, convinced that the cause of the airbag 3 n o n~d e pl oym e nts Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product was a com plicated mystery still to be unraveled, failed to consider fully the si mple, and ultimately correct, cause: the switch that ca used cars to stall was turning th e power off and disabling the airbags just as cars were about to crash. Along the way, the investigators were misled by the GM engineer who approved the below-specification sw itch in the fi rst place; he had actually changed the ignition switch to solve the problem in later model years of the Cobalt, but failed to document it, told no one, and cl aimed to remember nothing about the change. While stumped by the inabili ty to determine why different model year Cobalts performed differently, the investigating engin eers nonetheless failed to take certain basic in vestigative steps, such as taking apart both poorly and properly functioning switches to compare the two. In 20 13, an outside expert working for a plaintiffs anorney took apart two switches and quickly found the cause it took GM years to determine. Throughout the entire II~ year odyssey, there was no demonstrated sense of urgency, right to the very end. The officials overseeing the potential fixes and investigations did not set timetables , and did not demand action. When, in December 20 13, GM ' s recall committee initially met to discuss whether to recall Cobalts and other vehicles with the swi tch, members of the committee deferred the decision for another six weeks to gather yet more infonnati on, in part because the presentation provided to them failed to alert them to fatalities. While the issue of the ignition switch passed through numerous hands at GM , from engineers to investigators to lawyers, nobody rai sed the problem to the highest levels of the company. As a result, those in the best position to demand quick answers did not know questions needed to be asked. 4 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product As a whole, from beginning to end, the story of the Cobalt is one of numerous failures leading to tragic results for many_ As discussed below, many indi vidual s have substa ntial responsibility for th e delay in recalling the Cobalt. The se indi vidual s, as well as th e GM committees and groups that had responsibility for the Cobalt, fai led to demand action in the face of mounting injuries and fata li ties, to make themse lves or others accoun table, and to marshal the infonnation and expertise at their disposal to so lve a problem that brought hann to GM's customers. This report traces the hi story of the igniti on switch, from GM 's design and production of the ignition switch to its belated reca ll in 20 14, ultimate ly proposing recommendations to help avoid such a tragedy from ever occurring again. II . SUMMARY OF FACTS Jenner & Block LLP ("Jenner") was directed on March 10, 20 14, to detennine how and why it took so long for GM to reca ll the Cobalt. 2 Jenner was given unfettered access to witnesses and documents, and Jenner was asked for an unvarn ished account. Jenner was asked to compl ete this task on a very expedited timetable, recognizi ng, in thi s instance, it was critical to fi nd the truth quickly rather than explore every possib le avenue. 3 The evidence showed: From the outset, the Cobalt ignition switch had significant problems that were known to GM personnel. Designed to be a new generati on ign iti on sw itch first introduced in th e Saturn Ion, the switch was so plagued with prob lems that the engineer who designed it labeled it then "the switch from hell." The prototype switch performed so poorly that its enti re electrica l In Ihis repon, we sometimes usc "Cobalt" to refer to notjustthm model, but also other Delta and Kappa plmfonn vehicles thm shared the same ignition swi tch and that arc covered by Ihe ignition switch recall. These other models include Ihe Chevrolet HHR, the Pontiac G5 and Solstice, and the Smum Ion and Sky. 3 As part or this investigation, review of documents (discussed ill greater delail below) continues on a daily basis. To the extent that review yields additional relevant documents Ihat materially change the contents of this report, such infonnation will be provided to the Board. 2 5 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product concept needed to be redesigned. Upon completion, the redesigned switch fai led to meet the mechanical specifications for torque - that is, it required less force to turn the key than the designers had mandated. Erroneously believing that thi s low torque would not affect the car' s performance, an engineer - Ray DeGiorgio - approved go ing into production with a deviation from GM specificati ons. He permitted the switch to be produced with the low torque that would ultimately lead to accidents in which the airbags did not deploy. As it turned o ut, when finall y put into production, the redesigned switch initi all y failed for reasons entirely unrelated to the low torque. The igniti on switch often wou ld not start the car in cold weather, ca using a huge volume of customer complaints and a focus on fi xin g the switch so that cars would start. But starting the car was not the only issue. The low torque requi red to turn tile ignition switch led to its own set of problems. The Cobalt (and the Saturn Ion before it) was sta lling at highway speeds. From the outset, GM employees, customers, and members of the automotive press fo und repeatedly that they would hit the key fob or keychain with thei r knee, and the car would turn off'. GM received some of these reports before the Cobalt ' s launch , and others afterwards. There were scathing reviews in the press. There were customer complaints describing the moving sta lls and customers' sa fety concerns. The complaints were documented in GM electronic forms as a part ofGM's standard vehicle improvement process, and were duly consi dered by multiple committees of engineers and business people, including those with oversight over enti re vehicl e lines. These processes failed to reso lve the issue. The teams of experienced GM engineers on the committees that considered the Cobalt ignition switch did not recogni ze that th eir coll eagues 6 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product had designed the car so that turning the key to Accessory or Off would disable the airbags. 4 As a result and because numerous engineers believed that consumers could still sa fely maneuver cars that had stalled, the engineers categorized the problem as an issue of customer satisfaction, not safety. The decision not to categorize the probl em as a sa fety issue directly impacted the level of urgency with which the problem was addressed and the effort to resolve it. The problem of moving stalls and the ignition switch continued throughout 2005, and was described both within GM and in the med ia. In May and June 2005, reviewers from two newspapers, incl uding the New York Times , wrote about how they or a fam il y member had inadvertently turned a Cobalt off with their knees. One ofGM 's main safety lawyers e-mail ed a colleague trying to marshal evidence for the press that the risk was "remote" and "i nconsequential." He wrote that he did not want to be criticized for failin g to "defend a brand new launch." GM employees were upfront about the company's view about the complaints they had heard about moving stall s; they informed the New York Times of their view that moving stalls in the Cobalt did not pose a safety problem. At the same time, a different group of engineers within GM , the senior engineers charged with dealing with safety issues, attempted to detennine whether the problem could be easily replicated. The Di rector of Product Investigations, GM 's group of safety investigators, among others, succeeded in turning off the car with her knee. A more junior GM safety investigation engineer was tasked with investigating the issue, whi ch included compiling a fil e of customer compl aints, of which she fOllnd a number. The safety engineer presented the issue to a committee charged with reviewing sllch matters, and the committee closed the safety ·1 A single engineer involved in the review process told Jenncr interviewers that he raised this issue at the time, but neither contemporaneous documents nOT interviews of hi s collcagues corroborated his account. 7 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product investigation with no furt her action, also ignorant of the fact that turni ng the igniti on switch to Accessory disabled the airbags with potentiall y catastroph ic results_ In 2005, vari ous committees within GM conside red proposed fi xes, but those were rejected as too costly - a consideration that would not have mattered had the problem been consi dered a safety concern . Instead of fi xing the ignition switch, in December 2005, GM sent a notice to dea lers (a "Technical Service Bulletin," " bulleti n," or "TSB") warning that customers might complain of ignition c u t~offs caused by low igni tion torque, counsel ing that customers remove heavy items fro m their key rings, and offe ring an insert to the key that wou ld reduce the likelihood tha t the ignition switch would rotate inadverte ntly. That bulletin did not refer to the prob lem as "stalling," however, precisely because GM be li eved customers might associate stall ing with a safety problem. Onl y a customer who had already experienced a stall and who came into a dealer to complain wo uld get infonnati on about the proposed solutions. Other customers would remain unaware of the probl em, as well as GM's proposed solutions. In 2007, two years after GM closed its ini tial safety investigation, entities outside GM began payi ng attention as Cobalts crashed, ai rbags did not dep loy, and fata li ties occurred. In April 2007, a trooper of the Wisconsin Safety Patrol (Wisconsin's state police force) publi shed an accident reconstruction report on a Cobalt fatali ty in whi ch the airbags did not deploy. The trooper pointed out (a) that the igni tion switch appeared to be in Accessory; (b) that GM had issued a bulletin to dealers warning that low torque could lead to inadvertent turning of the ignition switch (he fo un d the bulletin on the NHTSA website); and (c) that airbags woul d not deploy when th e igni tion switch was turned to the Accessory position. Although an electronic 8 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product copy of thi s report resided in GM's fil es as of2007, no GM engineer investigating the igniti on sw itch reported seeing it until 2014 Over the years , Indiana Univers ity ' s Transportation Research Center and two plaintiffs' experts would also locate GM ' s Technical Service Bulletin and reach the same, correct conclusion regarding the connection between the ign iti on sw itch and ai rbag non~d e pl oyment. Within GM , however, the investigation langui shed. In 2007, a Field Performance Assessment ("FPA") engineer, John Sprague, whose j ob was to support GM ' s products liab ili ty defense team, was asked to track incidents of Coba lt airbag non-deployments in a spreadsheet. Sprague was given directions neither about a deliverabl e nor a time frame, hi ghlighting several themes that pe nneated GM personnel 's failed efforts to understand or solve the problem: lack of urgency, lack of ownership of the iss ue, lack of oversight, and lack of understanding of the consequences of the problem. The spreadsheet simpl y grew. Yet Sprague was not aware of important GM records of pri or problems wi th the igniti on switch (incl uding the bulletin sent to dealers and the pri or work done to propose soluti ons). Over time, however, through hi s work on various litigation matters, Sprague and his coll eagues in FPA came to recognize a pattern of n on~deplo yme nt of airbags in the Cobalt and hypothesized about a cause related to the electri ca l connections in the ignition switch. One of th e observations Sprague made was that the non~deployments stopped in Model Year (" MY") 2008 Cobalts. Sprague wondered whether there had been some kind of part change at that time that would exp lain why the non-deployments had ended. He was correct. In 2006, Ray DeGiorgio, the engineer who had approved the 9 below~spec i fication igniti on switch Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product origi nally, had authorized a change in the ignition switch that increased the torque required to turn the key. But when asked in 2009 and in the years that followed whether the igniti on switch had changed, DeG iorgio said that it had not. To thi s day, in informal interviews and under oath, DeGiorgio claims not to remember authorizing the change to the ignition switch or his decision, at the same time, not to change the switch 's part number. DeGiorgio ' s deliberate decision not to change the part number prevented investigators for years from learning what had actuall y taken place. By 20 11 , outside counse l, privy to the FPA engi neers' data, had repeatedly warned GM in~hou se counsel that GM could be accused of egregious conduct due to its failure to address the probl em of airbag non-deployment in the Cobait, and that such conduct might subj ect GM to liability, including puniti ve damages. Seeing the pattern of accidents regarding nondeployments, in 20 II , GM in-house lawyers ca ll ed fo r a meeting with engi neers and asked for the investigation to be taken over by GM's Product Investigations unit, which was charged with investigating potential safety issues. It was assigned to an investigator named Brian Stouffer. Once again, the investigation moved with no se nse of urgency. When he began the investigation, Stouffer was handed a dossier on the matter that included infonnation about the prior investigations into moving stall s, and he soon obtained the Technica l Service Bulletin. In 20 12, Stouffer was also given the 2007 Indiana Un iversity report and a report from a pl aintiff's expert, both laying out the correct answer. Stouffer, howeve r, discounted those reports, believing they did not answer the question before him , and, in doing so, ultimately prolonged the resoluti on of the problem. Among other things , he could not exp lain why MY 2008 Cobalts did not have a 10 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product problem with airbag n on~d ep l oyme nt , having also been misled by DeG iorgio ' s statement that the same ignition switch was in earl y and later mode l yea r ve hicles In the ensuin g two years, Stouffer ' s investigation proceeded slowly and without significant progress, ultimately becoming "stuck" by early 2013. It was only when a plaintitrs expert in a products liability case compared switches from pre~ and post~MY 2008 Cobalts by x- raying them - something GM investigators had never done - that GM engineers came to understand that th e early model year Cobalts had a different ignition switch than the later model year vehicles, and that the switch used in the earl y mode l years suffered from torque well below GM's specifications. Despite learning about what GM's outside counsel called a " bombshell" in Apri l 2013 , it was not until February 2014 that GM iss ued the first recall. In the interim, various additional investigative steps were taken that eventually confirmed what the plaintitrs expert had already made clear. Eventuall y, in December 20 13, the proposed recall made its way to the Executive Field Action Decision Committee (the " EF ADC"), the GM committee that considers recalls and whose members include three GM vice presidents, including its chief engineer. Still, no action was taken, because the chi ef engineer questioned the data demonstrating the problem and EFADe members were not presented with fatali ty information and therefore felt no sense of urgency to make a decision. It was not until approximately six weeks later that GM instituted its first recall of the ignition switch, and even that recall was incomplete based on the failure ofGM personnel to obtain and present all relevant information to the EF ADC II decisiol1~makers. Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product III. J ENNE R & BLOCK' S INVESTIGATIVE PROCESS This investigation covered a time period of more than fifteen yea rs, involved hundreds of witness interviews, and the review of millions of documents. Throughout the entire investigation, GM provided unfettered access and cooperation . A. Mandate from GM's CEO and Its Board GM's CEO, Mary Barra, and its Board directed Anton R. Valukas and Jenner to investlgate the circumstances that led up to the recall of the Cobalt and other cars due to the flawed ignition switch. We were directed to do as full and complete an investigation as was possibl e in th e short period allotted and to return to Ms. Barra and the Board with the unvarni shed truth about what happened, why it happened, and what GM should do to ensure that it never happens again. Jenner was al so asked to focus on the knowledge of specific seni or executives, as well as GM 's Board. Thi s report presents facts learned through the investigation requested by the CEO and the Board. While GM gave Jenner unfettered access and devoted enonnous resources to this effort, it is important to note what was not done. We have not sought to reconstruct acc idents or determine which injuries or fatalities were or were not caused by the safety defect in the Cobalt and other cars; that was not in our mandate, nor was it realistic in the time allotted.5 This report does not delve into any issues related to recall s other than those surrounding the ignition switch. B. Docll ment Collection and Review Jenner identified over three hundred document custodians for collection. Many of those custodians were identified earl y in the investigative process, and additional custodians were S Tbroughout this repon, whcn wc usc the tenn "defcct," "sa fety defect," " Ignition Switch defect" and thc li ke, wc are referring to the defect idcntificd by GM in its 573 lencr to NHT5A. GM Recall 573 Lellcr (Fcb. 7, 201 4) [DOC ID 000002796457; GMNHTSA000439255j. 12 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product added as new information and in vesti gative leads were obtai ned. Multiple sources of custodial and non-custodial data were collected and reviewed, including hut not limited to: ( I) forensically imaged hard dri ves; (2) server-based e-mail s and electronic share dri ves; (3) legacy electronic data coll ections; (4) hard copy documents; and (5) database co llections from various GM electronic databases, such as such as Global Document Management ("GDM") and Problem Resolution and Tracking System Plus ("PRTS+"). Search terms designed to identi fy the most relevant subset of information were developed and applied. Jenner also requested and reviewed hundreds of boxes of hard-copy documents from GM storage facilities. Those boxes were identified follo win g a search of indices for file names that were responsive to search terms. Potentially responsive documents were converted to electronic files. On April 10, 2014, Jenner requested that the ignition switc h manufacturer, Delphi Mechatronics (" Delphi"), provide documents rel evant to the approval and testing of th e igniti on switch, as well as all correspondence between GM and Delphi on the subject of th e switch. Delphi agreed to provide certain documents as we ll as the requested correspondence, under the condition that GM provide reciprocal documents and correspondence, to which GM agreed. As a result, GM and Delphi exchanged certain materials. As of May 28,2014, counsel for Delphi infonned Jenner that Delphi remained in the earl y stages of document review, still had not reviewed numerous documents, and had other rel evant materials it was not providing to Jenner. The volume of e-mail, other electronicall y stored documents, and converted hard copy documents is estimated at 23 terabytes of data. All data was maintained on four databases , housed and maintained on servers hosted by one of four professional forens ic firms: KPMG , 13 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product Ernst & Young, Iris Data Services, and Document Tech no logies, Inc. In total , Jenner coll ected in excess o f 4 I million documents. Documents were reviewed through two diffe rent processes: a document production process and an investigati ve process. In the document production process, after others conducted "fi rst level" review to identi fy documents potenti all y responsive to government requests, Jenner lawyers conducted a "second leve l" review and responsive documents were (a) prepared for production to vari ous governmenta l entiti es and (b) analyzed for investigative use. In parallel to the document production process, Jenner investigative teams also conducted targeted searches for doc uments. Any documents identified as relevant by Jenner investigative teams were directed to the document production process. C. \Vitness Interviews Jenner was provided with unlimited access to interv iew any GM employee. Every request for an interview ofa GM e mpl oyee was granted. A number of former GM employees and third parties were also interviewed as part of the in vestigation. Because former GM General Counsel Robert Osborne is emp loyed in an " of counse l" capacity at Jenner, an outside law firm unaffiliated with Jenner was retained to conduct, and did conduct, the interview ofMr. Osborne. The same law fi rm conducted the interview of form er GM General Counsel Thomas Gottschalk. In all , over 230 witnesses were interviewed over a period of approx imately 70 days. Because certain witnesses were interviewed more than once, the total number of interviews conducted was over 350. All interviews we re conducted by at least two Jenner attorneys, with the exception of the interviews of Mr. Osborne and Mr. Gottschalk, w hich were conducted by an outside fi nn in coordinati on with Jenner. 14 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product As part of the investigation, Jenner sent letters requesti ng the opportunity to interview plaintiff's counsel in a number of cases described herein. The letters were sent via electronic mail and via UPS overnight deli very on April 10, 20 14. Although UPS has confinned that the letters were deli vered, and the electronic mail was not returned as undeli verable, to date , none of the counsel contacted has responded to our request. D. King & Spalding's Role GM retained King & Spalding (" K&S") to assist in responding to requests from governmental agencies. Its lawyers assisted in searches for documents and conducted first level document review. K&S lawyers participated in a limited number of Our witness interviews, largely during the initial phase ofthe in vestigation. IV. BACKGROUND A. General Motors GM is one of the largest engineering and manu facturi ng enterprises in the world.6 For 77 years until 2007, and aga in in 20 11 , GM so ld more ca rs worldwide than any other company.' In 2013 alone, GM produced more than 9.7 milli on vehicles worldwide,s generating revenues of $ 155.4 billion and profits of$3.8 billion.9 The 2013 IndllSllJlWeek 1000, IndumyWcck, :l1 1~2, availahle al ww\v.industrywcck.comlrcsourccsliwlO00J2013 (last accessed May 27, 201 4) LDOC ID 000240753003 j. 7 Greg Gardner. GM Is World 's Largest Automoker Ollce More, Dctroit Free Press (Jan. 20. 2012), at I LDOC ID 0002407530061; Christine Tierney and Roben Snell, Toyola Overlakes GM in Global Sales , Detroit News (Jan. 22, 2008), at I rDOC ID 0002407860021. ~ Press Release, GM Delivered 9.7 Mi llion Vehicles Globally in 2013 , General Motors Company (Jan. 14,20 14), available al http://media.gm.com/ media/us/en/gm/ news. dctai 1htmllcontcntlPageslncwsluslenl20 141Jan! 01 14 201 3gm-global -sales hlml (last accessed May 29, 2014) I DOC ID 0002407530021. 9 Ge[~eral Motors Company, 2013 Annual Report, at 14 LDOC ID 000240786006]. 6 15 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product GM employs more than 210,000 employees in 396 facilities located in 30 countries. to As a result, the most senior GM executives have literall y tens of thousands of employees working under them. GM also has 21 ,000 dealers woridwide. 11 Along with its strategic partners, GM produces cars and trucks through II brands worldwide, four in North America. 12 (Prior to its bankruptcy, GM had several other brands in North America. l3 ) In 2013 alone, GM's brands sold over 9 million vehicles worldwide, in more than 120 countri es. 14 Of course, older models also remain on the road, meaning that the total number of vehicles currently in production vastly understates the fu ll number and variety of GM vehicles on the road at any given time. Each vehicle is itself a composite of numerous subsystems and component parts. A car includes a chassis with suspension, wheels, powertrain (composed of the drivetrain, engine and transmission), a coo ling system, an electrical system, a fuel system, an HV AC and climate control system, steering and braking systems, and an instrument panel and control s. Vehicl es have become increasingly compli cated over time, and now also routinely include power systems that operate locks, windows, doors, trunks, and headlights, as well as sophi sticated computerized systems controlling advanced safety features such as airbags, anti~lock brakes, electronic stability control , lane departure warnings, and parking ass ist, not to mention additional features lO Our Company, General Motors Company, at I, amilable al www.gm.com/ company/abolllGM/ouT_company.html (lasl accessed May 27, 20 14) rOOC lD 0002407530071Il Our Company, General MOlors Compa ny, at I, ami/able a/ www.gm.com/company/abollIGM/ollf_company.hunl (last accessed May 27, 2014) LOOC ID 000240753007J. 12 Our Company, General Motors Company, at 2, ami/able al www.gm.com/eompany/aboutGM/our_company.hlml (lasl accessed May 27, 2014) rOOC lD 0002407530071 ; General MOlors Company, 2013 Annual Repon, at 20 [DOC 10000240786006]. lJ See General Motors Corp., 2007 Annual Rcpon, al 63 rOOC lD 000236388003114 General MOlors Company, 2013 Annual Report, al 12 rOOC 10 0002407860061; Our Company, Genera l MOlors Company, amilable al www.gm.com/company/abouIGM/our_company html (lasl accessed May 27, 2014) I DOC ID 000240753007j. 16 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product like cmise control and web-based navigation and infotainment system s. Counting at the small est possibl e units, a single car can have literally tens of thousands of parts - all of whi ch must be manufactured or sourced by GM to make each of the more than 100 vehicles in its product line. l5 To fashion th ese parts into a car and se ll it in the market, GM 's employees perform a variety of functions, starting with research and devel opment and proceeding progressively to design , manufacturing, purchasing (component parts), marketing, sales, and customer service. B. GM's Small Car Product Line In the earl y 2000s, GM looked to develop a successful small car line to supplement sagging sales of its trucks and SUVs. As late as 2004, truck and SUV sales represented 60 percent of the vehicl es so ld by GM in the United States and represented a substantially hi gher profit ma rgin per vehicl e than smaller cars.l6 As gaso line prices escalated, however, sales of trucks and SUVs suffered in the United States. l7 This decrea sed demand for large vehicles made the success of smaller, more effi cient cars important to the Company's fi nancia l performance. l8 Sales of smaller, fuel-efficient cars were also important to GM 's ability to meet fede ral em issions regu lations. l9 I. The " Delta Platform" of Vehicles In developing vehicl es, including its new small car product line in the 2000s, GM grouped vehicles into " platforms." Vehicles in the same platfonn shared certa in base IS See John Stark, Product Ufecycle Managemelll: 21st Cenlllry Paradigm /or Prodllct Realisation (Spri nger 20(6), at 23 (,'There arc over 20,000 parts in a car") 16 General Motors Corp., 2004 Annual Report, at47 [DOC ID 000236389003 J; General Motors Corp., 2005 Atmual Report, at 5, 45, 48 LDOC ID 000236388002j. 17 Claudia Hirsch, Reality Check: US Auto Brands Fizzle in Jllne, Dealers Say , Market News Intemational (June 30, 2006) al 1,2 [DOC ID 000236387022]. J~ See General Motors Corp., 2005 Annual Report, at 52 fDOC ID 0002363880021 (expressing efTorts to "reallocatfe1 capital and engineering to support morc fuel-effi cient vehicles"). 19 See General Motors Corp. , 2005 Annual Report, at 30 [DOC ID 000236388002] ; Gap Assessment to Toyota Presentation lDraftJ (2005), at 19 LDOC 10 000240785002]. 17 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product architectural design e lements. 2o The " De lta Pl atform" of vehi cles included the Saturn Ion, Chevrolet Cobalt, Chevrolet HH R, and Pontiac G5. 2l Certai n model years of each of th ese vehi cles contained the igni tion swi tch that led to GM's reca ll s in 2014 (hereinafter " Ignition Sw itch,,):22 I. Saturn /on: The Saturn Ion was the fi rst car in the Delta Platform to launch. It was a small sedan, first offered in MY 2003. 23 GM subseq uently aimed to improve the Ion 24 by integrating design elements from the Coba lt. Production ended with MY 2007. 25 2. Chevrolel Coball: The Chevrolet Cobalt lau nched in MY 2005. 26 It replaced the 27 Cava lier, but was designed to have a higher price poi nt than its predecessor. GM ended production of the Coba lt in June 20 10. 28 3. Chevrolet HHR: The C hevrolet HHR was a station wagon, launched in MY 2006, 29 and offered through MY 201 1. 4. Pontiac G5: The Ponti ac G5 was fi rst introduced in Canada for MY 2005, and was released in the U.S. in MY 2007. The 20 10 model was the last sold to th e pUblic.30 The Ignition Sw itch used in the Delta Platform was also used in certai n model years of Kappa Platfonn vehicles - th e Saturn Sky and Pontiac So lstice: 3l J&B Interview of James Zito, May 14, 2014. GM Program Architecture Brand (Mt,lotmal'let of this swIkh I'I()r PPAP status. II. I O'V .. rtq\lU ~.~ Nola: I ) S....tlet1 desiQn wi require detent mod,l\eations lor !lie GMX 001 Program. (along WIth any OII'ref rOior J PCB deslgn moc:ts ntw:e$ury to meellhe CTS lequlreme~ .) Z) StanclardiZition 01 GMX 3201357 switch designs nor: pl'obable unleu above requirements lie mel 3) Ulimat~ry . if 11M! TALC Groupoon noI buy . ot! on the feel of the swi!cflllum chal1g8S 10 the design will be required. Ray (tirM of ltrelWllen from 11&1) DeGiOfgiG 170 E-mail from Raymond DcGiorgio, GM, to Erik R. Malison, Delphi, el al. (Feb. 19, 2002) LDOe ID 000077721033 ; GMNHTSA0003371071 17 r E-mail from Raymond DcGiorgio, GM, to Erik R. Mattson, Del phi, el al. (Feb. 19, 2002) [DOC TO 000077721033; GMNHTSA000337107l; J&B Interview of Raymond DcGiorgio, May 7-8, 2014. 172 E-mail from Raymond DcGiorgio, GM, to Erik R. Malison, Del phi, el al. (Feb. 19, 2002) [DOC ID 000077721033 ; GMNHTSA000337107] ; e·mail from Erik R. Mattson, Delphi , to Raymond DcGiorgio, GM, el al. (Feb. 19, 2002) LDOe ID 14042800393158; GMNHTSAOO044148 1j. 48 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product Mattson asked for clarification regarding the Spec ification requirements: " Where in the CTS [Component Technical Specification] is there an exact requirement of these forces? A ll I see is a genera l curve that is marked ' This is a TARG ET curve only, actual curve to be furnished by supplier after GM Engineering approval. ",173 As noted above, however, by 2002 , the "TARGET curve" language had been removed from the Specification and replaced w ith a specific requirement of20 N~cm w ith a +/~ 5 N-cm to lerance. Mattson further wrote that he was "not saying it is impossible to change the detent forces , but it does have an impact on timing and our suppliers w ill not do it for free .,,174 Mattson a lso rep li ed that the «Talc samples were 9.6 N~ cm. The new producti on intent vers ion of the switch has 9.5 N-cm. We fee l thi s is a match of the TALC switch. ,,175 Mattson concluded: " We can revise thi s again but we all need to be aware of the impacts in timing, cost, and possible other issues that mi ght be created when we are thi s close to PPAP. ,,176 We have not located a response fro m DeGiorgio. When interviewed, DeGiorgio acknowledged that Delphi had not achieved the required torque for the Ignitio n Switch. 177 Given the switch 's history of electrical failures, however, he was hesitant to make any changes that might j eopardize the functionality of the switch 's electrical architecture. 178 Because he beli eved the Ignition Switch had performed properly and without incident duri ng the numerous vehicle-leve l tests conducted on the prototype lon,179 E-mail from Erik R. Mallson, Delphi , to Raymond DeGiorgio, GM, et al. (Feb. 19, 2002) IDOC 14042800393158 ; GMNHTSAOOO4414811. 174 E-mail from Erik R. Mallson, Delphi , lo Raymond DcGiorgio, GM , el al. (Feh 19,2(02) [DOC 14042800393158 ; GMNHTSAOOO44148I j. 175 E-mail from Erik R. Mattson, Delphi , to Raymond DeGiorgio,GM, et al. (Feb. 19, 2(02) LDOC 14042800393158; GMNHTSAOOO4414811. 176 E-mail from Erik R. Manson, Delphi, to Raymond DcGiorgio, GM , el al. (Fcb. 19, 2002) rOOC 14042800393158; GMNHTSAOOO4414811. 177 J&B Intcrview of Raymond DcGiorgio, May 7-8, 2014. ]78 J&B Interview of Raymond DeGiorgio, May 7-8, 2014. ]79 Issues noted during Captured Test Fleet ("C IT') testing or the Lon are set Ollt in Appendix F. 173 49 ID 10 lD ID Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product OeGiorgio approved production of the switch - even though the switch 's torque was below the Specification _ISo OeGiorgio explained that he signed hi s e~mail message " tired of the switch from hell " because he was frustrated by the numerous electrical issues exh ibited by the Ignition Switch and the amount of time and energy he had spent reso lving them. lsi Following these discussions with Mattson, subsequent torque tests reported in 2002 showed even worse results; torque values to rotate from Run to Accessory ranged from as low as 4 N-cm or 5 N-cm up to 11 N_cm. 182 Like the January report, the April report states "Not OK" next to eac h of the res ults. 18J GM does not have records of receivi ng these test results around the time they were conducted. However, as with the January 2002 AOVP&R report, under standard procedures, OeGiorgio would have received and reviewed them as the engineer responsible for the Ignition Switch. 184 After interviewing hundred s of witnesses, we have not identified any GM personnel , other than OeGiorgio, who received or reviewed these test results, or knew (prior to 2013) that the Ignition Switch fai led to meet the Specification when it was approved for production in 2002. 6. May 2002: Despite Failing Test Results, The Ignition Switch Goes Through Production Part Approval Process. On May 3, 2002 , GM personne l approved shipment of the Ignition Switch through its Production Part Approval Process, known as the " PPAP.,,185 PPAP is the process through which J&B Interview of Raymond OeGiorgio , May 7-8, 2014. J&B Interview of Raymond OeGiorgio , May 7-8, 2014 182 GM AnalysislOevelopmenllValidation Plan & Report (May 2 [, 2002), at 3, 9, 12-13,21 , 27, 39, 50, 61 LOOe ID 120198-00028IJ. The full results for the torque required to tum from Run to Accessory were 5 N-cm, 5 N-cm, 4 Nem, 7 N-em, 10 N-em, atld II N-em. IS3 GM Analysis/Devc\opmcnl/Validation Plan & Repoll (May 21, 2002), at 3, 9, 12-13, 21 , 27, 39, 50, 61 rooe TO 120 198-000281]. 184 J&B Interview of Raymond OeG iorgio, May 7-8, 201 4. 185 PPAP Repoll (May 3, 2002) (Delphi version) [DOC ID 120198-000677]; PPAP Repoll (May 3, 2002) (GM version) LOOe lD 000005779025; GMNHTSAOOO607040j. ISO l SI 50 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product parts are tested, validated, and ultimately released for production. The PPAP package should contain, among oth er things, two key authori zations: one from the supplier and one from GM. The Commodity Validation Sign Off, GM Form 3660 (" Fonn 3660") signifies G M's engineering approval for a part to ship,186 and the Part Submi ssion Wa rrant ("PSW") represents the supplier's confirmation that the parts being shipped comply with GM's requirements. l S7 No complete PPAP package for the Ignition Swi tch approva l in 2002 has been located in either GM ' s or Delphi ' s fil es. According to GM 's G loba l Supplier Quali ty Manual and the standardized PPAP guidelines adhered to by GM , the onus is on the suppli er - Delphi - to compi le and maintain the complete PPAP "package.,,188 Delphi has not provided, and has been unable to locate, a compl ete PPAP package for the Ignition Switc h. 11l9 The fina l engineering drawing would have been a part of the PPAP package. 190 According to DeGiorgio, in 2002, he had the authority to approve the Igniti on Switch for production - and did in fact approve it - even though it did not comply wi th all Specification requi rements, without needing anyone else to sign off on hi s approval dec ision. 191 There are inconsistent accoun ts o f whether practices at the time required someone else al so to approve the J&B Interview orRalfNickel, March 26, 2014. AIAG PPAP Manual (3d ed. Aug. 2000), al 90 100C 10 0001232526801. 188 GM Global Supplicr Quality Manual (Sept. 2008), at 22 ~23 rOOC ID 0000497400031; AIAG PPA P Manual (3d ed. Aug. 2000), at 2 ~3 [DOC ID 000123252680] 189 Typically, GM would store the Form 3660 and Part Submission Warrant in the Global Quality Tracking System ("GQTS"). GM Global Supplier Quality Manual (Sept. 2008), at 22 LOOC 10 000049740003 J. GM has searched its database and other document sourccs extcnsively and has not located this documentation; because the requisite database that stores Fonn 3660s automatica lly deletes infonnation from the system one year after a part ceases to be in production, it is not li kely that the original Fonn 3660 from 2002 for the Ignition Switch still exists in GM ' s fil es. 1&8 Illlerview of Brenda Van Eman, April 15, 201 4. 190 AIAG PPAP Manual (3d ed. Aug. 2000), at 3 [DOC ID 000 123252680]. 191 J&B Interview of Raymond OeGiorgio, May 7~8 , 2014 . 186 187 51 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product sw itch and whether GM req ui red doc umentation of a deviation from Specifica ti ons. l 92 However, no evidence has been uncovered that suggests that any GM employee, other th an DeG iorgio, knew, prior to production, that the Ignition Switch approved fo r the Ion and Cobalt was below . 193 spect"filcatton. Ulti mately, with the knowledge and approval of DeGiorgio, the Ignition Swi tch went into production despite having a torque value well bel ow the requirements set fort h in the Specification. 194 Thi s explains how the switch came to be in a GM ca.r in the first place. The fact, however, that a below-specificat ion ignition switch had been authorized was not known to the engineers who were investigating the Cobalt 's Ignit ion Switch over the next decade; indeed, it is not until 201 3 that GM engineers, lawyers, and investi gators came to understand that the )92 Two GM employees who were validation engineers for switches (including ignit]on switches) during the time periods of 2004 and 2007, stated that - prior to 2004 - there were not validation engineers for ignition switches, and the design release engineer for the switch had authority to approve a switch for production. J&6 Interview of Michael Berden, May 27, 2014; J&B Interview of Eugene Carnago, May 15,2014. Other employees, who did not have validation responsibi lity for swi tches, thought olher approvals may have been required in the early 2000s for component\cvel sign-off into production. J&B Interview of Emie McCutchen, May 16, 2014; J&B Interview of John Hendler, April 25, 2014. )93 An October 7, 2001 '2 67 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product "sure" solution was not seriously considered because it did not address the torque issues in the Ignition Switch and wou ld have been an ex pensive fi X. 264 Initiall y, both the CPIT and the VAPIR committees di scussed the key head design change as what they called a "containment" solution. 26~ However, on March 9, 2005, a few days after a second CPIT meeting, the November 19, 2004 PRTS was closed "with no action. " The report reflects that it was closed pursuant to the "directive" of the Cobalt Program Engineering Manager ("PEM ,,).266 Altman , who was the PEM at the time, sa id that he did not have the authority to close the PRTS without action, and suggeste d se nior Coba lt program team members Doug Parks (Chief Engineer), Lori Queen (Vehicl e Line Executive), or Queen 's direct report, 267 Sarah DeVries (Vehicle Line Director) made the decision . Queen denied that she made the deci sion to close thi s report, and suggested Altman did have authority because the issue was a 269 severity level 3. 268 Ne ither Parks nor DeVrie s recalled closing the PRTS. Under the GM policy in effect at the time, Altman did not have authority to close the PRTS himself; the CPIT had to approve the c losure of a non-safety related PRTS when no action was taken. 270 The interviews here showed a troubling di savowal of responsibi lity made possible by a proliferation of committees. It is an example of what witnesses called the "GM salute," a crossing of the arms and pointing outward towards others, indicating that the responsibility J&B Interview of David Tm~h , March 17, 201 4; J&B Interview of David Tm ~h , April 30, 2014 265 PRTS NI72404 (Nov. 19, 2004), at 10 LDOC ID 000001771143; GMNHTSAOOOI42658J . 266 PRTS NI72404 (Nov. 19, 2004), at 10 LDOC ID 000001771143; GMNHTSAOOOI42658J . 267 J&B Interview of Gary Altman, March 14, 2014. Deb Lund and Sarah DeVries confinncd that the PEM could not close the PRTS by itselr. J&B Interview of Deb Lund, May 14 & 16, 2014; J&B Interview of Sarah DeVries, May 13 & 16, 2014. 268 J&B Interview of Lori Queen, May 12,2014. 269 J&B Interview or Doug Parks, May 1·2, 201 4. J&B IllIcrview of Sarah DeVries, May 13 & 16, 2014. 270 "Closed Without Action" Approval Guidel ines By Issue Type & Closure Code LDOC ID S006897 _000000629J. 264 68 Privileged and Confidenfia/ Protec1ed by AUorney~ Client Privilege and as Allorney Work Product belongs to someone else. Here, because a committee was "responsible ," no single person bore responsihility or was indi vidually accountahle. ~ - . .... . . .. . . . CMXOOI ffiM's directive we an: dosin& this PRTS ~th no aclton.. The main reason .. areas rollowia&: I.AII pouible .oIutions were presented in CPIT and VAPIR:. a. The lead-tnnt tOr all the .oludons illOO long. b. 1be tooli ng cost and P)e(:e price.re too high. e. None oflhe soIutlolU secms to tully countcrmeasUR: the possibility of the key beiDa turned (igr\illioa nnn oft) d"';Dg drivil'lg. ThLlS none orlhe ~Iutiord represents an acceptable business cue. 03l09flOO5 • Blendi Sullaj As shown above, the November 19, 2004 PRTS stated that "none of the soluti ons represents an acceptable business case" - a standard phrase used by G M personnel for closing a PRTS without action.271 David Trush, the lead eng ineer on the ignition cylinder, explained that to present an "acceptable business case," a soluti on should so lve the issue, be cost effecti ve, and have an acceptable lead time to implement the change. 272 In this case, according to Trush, the proposed changes were not implemented because none of them was guaranteed to resolve the problem completely.273 Lori Queen confimled that a discretionary, or non-safety related, part change had to clear a seri es of financial hurdles before it would be approved. 274 Engineers who proposed a discretionary change had to make out an "acceptable business case" before the GMNA PRTS+ C losure Codes (Close wlout Action) (Effective Oec. 2007) [OOC ID GMCB-000000977300l m J&B Interview ofOavid Trush, March 17,2014; J&B Interview ofOavid Tmsh, April 30, 2014. m J&B Interview ofOav id Trush, March 17, 2014; J&B Interview ofOavid Tmsh, April 30, 2014. 274 J&B Interview of Lori Queen, May 12, 201 4. 271 69 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product financing for the change would be approved. The speci fie financial hurdles included cost, lead time, in vestment, and whether the proposed change would fix a problem?75 Here, the decision to close the PRTS without action was a direct consequence of the decision to classify the problem as one of "convenience" rather than "safety." As a direct consequence of their failure to understand the safety defect in the Coba lt, the engineers put the issue into a category of problems for whi ch cost was a relevant consideration. Without exception, the engineers involved with the Coba lt said that, if they had understood that rotating the key to Accessory would cause the airbags not to deploy and had considered the issue one of safety, the PRTS would not have been closed without action. 276 Witnesses sa id the presence of any sa fety concerns would have changed the ana lys is entirely, mooting the cos t~b e n e fit analysis of the "business case.,,277 Because the issue was not designated as a safety concern - severity level 1 - at the outset, the consequence was no action by GM. And the initia l classification of the issue as one of "convenience," once made, was not likel y to be di sturbed. As Lori Queen stated, if an issue was not identified as a safety issue at the time it was presented to the CP IT, which had responsibility for reviewing and approving severity designations, there was unlikely to be any check or review of that initial determinati on. 278 Queen stated: " If I was an engi neer, a nd defin ed something as a customer satisfaction issue, I don't think anyone would have questioned that, at the program- m J&B Interview of Lori Queen, May 12, 2014. J&B Interview of Gary Altman, March 14, 2014; J&8 Interview of John Hendler, April 25 , 2014; J&6 Interview of Joseph Manson, April 22, 2014; J&8 Interview of Lori Queen, May 12, 20 14; J&6 Interview of David Tmsh, March 17, 2014. m J&B Interview of John Hendler, Apri l 25, 201 4; J&8 Interview of Joseph Manson, April 22, 2014 ; J&6 Interview of Lori Queen, May 12, 2014; J&6 Interview of David Tmsh, March 17, 2014. 278 J& B Interview of Lori Queen, May 12, 201 4. 276 70 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product team level, or above, unless it was truly obvious that somethi ng more serious was going on. ,,279 In practice then, the initial input of an engineer on an issue, whi ch may not be full y understood because it had not yet been investigated, effecti vely beca me the default for determi ning the level of severity of a prob lem. The VA PIR was, however, a committee that incl uded a cross~section ofVehjcle System Engineers who had expertise in the various systems within the vehicle (including electrical, powertrain, chass is, and others), as well as a safety engi neer, rather than engineers focused si mpl y on one part of the car - as engineers such as DeGiorgio and Trush did . The purpose of having a committee w ith engineering expertise is to identi fy whether a n issue in one part or system may impact other fun ctions in the vehicle. Nonetheless, this committee simpl y deferred to the origi nal designation that thi s was not a safety issue. There is no evidence that anyone on the CPIT or the VAPIR either understood or at least considered that turning the key to Accessory would disable th e airbags. The failure of the committees to fulfill their role of understanding the interaction of the vari ous systems in GM cars and to review and consider the severity designation resulted in th e Igniti on Switch being treated as a convenie nce issue, with its solution dependent on cost. When a cost~e ffecti ve sol uti on was not found, the safety defect was not addressed. Here, as elsewhere in the story of the Cobalt, the structure within GM was one in whi ch no one was held responsible and no one took responsibility. Wh ile people were responsible for being on the committee, they were insulated from indi vidual responsibil ity fo r action. 279 J&B Interview or Lori Queen, May 12, 2014. 71 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product 8. 2004-2005: GM Discussions With NHTSA Regarding Engine Stalls As noted above, the engineers working on the Cobalt did not view a moving stall as a per se safety problem. That view was shared by others in the automotive industry in thi s time peri od,280 and also was the subj ect of discussions between NHTSA and GM around the same time that the Coba lt team was working on issues rel ated to the Ignition Switch (2004~05).28 1 According to Gay Kent, the then~Director of Product Investigations, NHTSA was concerned with engine stalls across the automobile industry and several safety recalls were issued for engine stall issues in that time period. 282 Kent sa id NHTSA' s new focus was to establi sh a fram ework with which to evaluate stalling conditions. 283 In December 2004, GM noted internally that it was "[c]urrentl y debating three engine sta ll conditions at this time.,,284 None o f these conditions related to the Cobalt. 285 Kent, Bill Kemp, Keith Schultz, and others engaged with NHTSA in the late spring of 2004 regarding engine stalling more broad ly. Kemp ' s notes fro m these conversations suggest that NHTSA agreed with GM ' s view that engine stalling was not a per se safety issue. 286 Kemp's contemporaneous notes also indi cate that GM represented that, with respect to engine 280 See, e.g. , SalsifZ v. Nasser, 208 F.R.D. 589, 593 (E.D. Mieh. 2002). In this maller, Ford Motor Co. had concluded tilat moving stalls did not pose an unreasonable tilrcatto motor vehicle safety whcn it dccided Ilotto report multiple stalls created by its mounting a "thick Iilm ignition switch" directly onto the engine motor block. Aftcr repeated analysis and discussion, Ford' s Policy and Stratcgy Committec detcnnined that thc stalling issue did not risc to the level ofa reponable safety problem. ld. 281 E-mail from Gay Kent, GM , 10 GM emp loycc~ and NHTSA repre~en1alive~ (Jan. 2, 2(05) [DOC lD 000007486261 ; GMNHTSA000218516J; J&B Interview of Gay Kent, March 18, 20 14. 282 J& B Interview of Gay Kent, March 18, 2014. 283 J&B Interview of Gay Kcnt, May 16, 2014. 284 Bob Lange & Douglas Wachtel, GM, Field Pcrfonnance Evaluation (FPE) Sensilivity and Recall Lessons Learned - 2004 Presenta tion (Dec. 4, 2004), at 66 rOOC ID 000000626923; GMNHTSA000218550l 285 J&B Interview of Gay Kcnt, May 16, 2014. 286 William Kemp. Draft Talking Points for Queen-Medford Conversation on Mond:'IY. May 24, 2004 [DOC ID 000045764417J. 72 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product stall s, GM focused on the frequen cy with which sta lls occurred; GM personnel also emphasized that, in the spec ific c ar under discussion , there had been "[n]o crashes, no injuries. ,,287 On May 17, 2004, during a NHTSA vis it to the GM Milford Proving Grounds, GM gave a presentation titled "Engine Stall & Loss of Assist Demonstration. ,,288 The purpose of the demonstration was "[t]o establish through demonstration and data that an engine stall is not, per se, a safety defect. ,,289 The dri ving demonstration permitted NHTSA officials to experience whether a vehicle could be controlled after a veh icle stall on a variety of different road courses. ,,290 Airbag deployment was not addressed in the slide dec k. At a June 3, 2004 meeting w ith NHTSA, GM presented its perspecti ve on engine stalls specifically, that those occurring on acceleration require more rigorou s review. OM also represented to NHTSA that in assessing a given stall, it considered severity, incident rate, and warning to the driver. 291 Kemp 's notes from thi s meeting indi cate NHTSA told OM that, in a case w here the number of fa il ures was " inord inately high," the factors should be considered but did not necessarily " immunize" a manufacturer from conducting a sa fety recall .292 On June 11 , 2004, Kent followed up with a letter to NHTSA, which included responses to questions NHTSA asked GM at the June 3, 2004 presentation. One such question was the 287 William Kemp, Notes fro m Discussion wi th ROil Medford (May 19, 2004), at 2 [DOC ID 00004577568 1]; William Kemp, Draft Points for Settlement of Saab 9-3 Stalling Condition [DOC ID 000045764403 J. 288 Engine Stall & Loss of Assist Demonstration (May 17, 2004) [DOC ID 000130987046-5 11289 Engine Stall & Loss of Assist Demonstration (May 17, 2004) rDOC ID 0001309870481 . 290 Engine Stall & Loss of Assist Demonstrat ion (May 17, 2004) [DOC ID 0001309870491; J&B Interview of Gay Kellt, March 18, 2014. 291 Gay Kent, Engine Stall Presenlalion (June 3, 2004), al 9 [DOC ID 000087173734]. 292 William Kemp, Notes (June 3, 2004), al 2 [DOC ID 000045775064]. 73 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product number o f safety recalls (3) and other field actions (17) that GM had conducted in the past 20 . sta 11 mg. · 293 years ,lor engme As time progressed, GM continued to deve lop it s framework for assessing en gine stall ing issues. In March 2005 , GM Product Investigations, the group of engineers with responsibility for safety issues, drafted a stall problem. 294 multi~factor framework for assessing the safety impact of the engine Factors relevant to the framework included whether a vehicle could be restarted after a stall , whether the stall occuned when the vehicle was moving or parked , whether the dri ver retained control over power steering and brakes, and whether the driver received any warning signs before the stall occurred ?95 As pan of the framework, GM established a guideline that 20-30 incidents per thousand vehicles ("JPTV") over a three-year period was a reasonable rate for an engine stall problem. This guideline does 110t appear in the written presentation given to NHTSA at the time. Wachtel reported that Kent orall y infoffiled NHTSA of thi s guideline. 298 297 296 Doug Kent recalls discuss ing the 20-30 IPTV with NHTSA as "guidance," but did not reco ll ect presenting it as a " GM guideline.,,299 We have uncovered no documentary evidence as to whether NHTSA specifically accepted GM 's approach to stall s in 2004 or 2005 , though NHTSA's own approach can be seen 293 Lctter from Gay Kent, GM, to Kathleen C. DeMeter, ODI-NHTSA (Junc 11 , 2004), at 1 rOoe ID 000087 173763] 294 W. Kauffman, GM, Applying Stalling Assessment Framework (March 2005), at 1 lODC ID 000005586008 ; GMNHTSA0003290 IOJ; J&B Interview of Gay Kent, March 18, 20 14. 295 J&B Interview of Gay Kcnt, March 18, 2014; W. Kauffman, Applying Sta ll ing Assessment Framework (March 2005) ]DOC ID 000005586008; GMNHTSAoo0329010l 2% J&B Interview of Gay Kcnt, March 18, 2014. 297 Engine Stall Prcsentation, GM and NHTSA QTRLY Rcview (March 8, 2005) rDDC ID 0000090885781298 J&B Interview of Gay Kent, March 18, 2014. 299 J&B Interview of Gay Kent, May 16, 2014. 74 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product by its actions. 3OO There is evidence that NHTSA, in fact, may not have accepted GM ' s supposed "guideline" In th e s ummer of2013, NHTSA questioned GM 's use of the rate of20 IPTV over three years. 30l Additionally, there is no documentary evidence that anyone at GM was keeping track of IPTV for moving stalls for the Cobalt, much less reporting that infonnation separately to NHTSA. There was no formal policy in place at GM regarding recording IPTV figures or monitoring cars over time.302 Finally, there is no evidence that anyone at GM or anyone at NHTSA was considering the circumstance presented by the Cobalt - that the moving stall was caused by an inadvertent shut-off of the ignition , which resulted in airbag non-deplo yment. 9. Spring 2005: GM's Brand Quality Group Learns of Moving Stall Issues After the Cobalt program team closed the November 19, 2004 PRTS with no action taken, additional complaints of Cobalt stall s and inadvertent Ignition Switch shut-offs continued to come into GM's Brand Quality Group. The engineers in that group similarly failed to connect the problems with the Ignition Switch to loss of airbags. and therefore failed to classify the problems as a safety concern. 300 In 2007, NHTSA indicated its approach when it denied a vehicle defect petition eonceming stalls in Ford Taunts and Mercury Sable vehicles, finding that only 3 out of 228,000 vehicles cxperienced stalling and thmthose 3 stalls occurred at low speeds and did not result in any crashes, losses of control , or high fisk events. In reaching this conclusion, NHTSA looked at such fac tors as incidents per thousand vehicles; speeds at which stalling occurs; type of opemtion during which stalling occurs; whether the vehicle can quickly be restarted after stalli ng; whether the stalling affects steering functions ; whether the stalling affects braking functions ; and any crashes or other unsafe events lhal may have res-uhed from the sla1ling. See NHTSA Denia l ofMo[or Vehicle Defect PeL, 72 Fed_ Reg_ 73973-01 (Dec. 28, 2007) LDOC LD 000240799004j. 301 In his interview, Doug Wachtel stated that Cannen Benavides mentioned the 20-IPTV-over-threc-years rate to NHTSA in summer of2013 and they " beat [herl up" about i\. According to Wachtel, NHTSA said that they do not endorse a 20-IPTV-over-thrce-year rate and asked Benavides how GM came up with [hat number. J& B Interview of Douglas Wachtel, May 14, 20 14; e-mail from Benavides, GM , to Douglas Wachtel and othcrs, GM (Aug. 20, 2013) (Benavides states that whcn she "visited NHTSA a few weeks back they wcre clear Ihat [sicllO point oulto me Ihey do not prescribe to a 20 IPTV at 3 years") [GMCB-OOOO00836377; GMNHTSA 100 105461 ]. 302 J&B Interview of Gay Kent, May \6, 2014. 75 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product For exampl e. in March 2005, Jack Weber, a GM engineer, reported that during " heel-toe down shifting" in a Cobalt SS with a manual tran smission (a hi gh~performance Cobalt model), his knee contacted the key fob and key ring, which caused "pulling on the key to move it to the 'Off' position."JOJ Weber also noted that the key fob "levered around the steeri ng column cover and turned the ignition off. ,,304 The problem did not occur when the fob was removed from the key. Steven Oakley. a Brand Quality Manager, opened a Field Performance Report to address this issue, and the report assigned the issue the lowest severity level: four - "annoyance or continuous improvement. ,,30s Oakley stated that severity level 4 was the default setting and he did not change it. 306 Oakley initially thought an inadvertent sh ut-off was a sa fety issue, but Gary Altman, the PEM for the Cobalt program team, and other engineers told him it was not, and he deferred to them . 307 In May 2005 , a customer demand that GM repurchase hi s Cobalt was forwarded to Oakley. The complaint was that the Ignition Switch shut off during normal dri ving conditions with no apparent contact between the driver's knee and the key chain or fob. 308 Oakley forwarded this information internally at GM , stating that the Ignition Switch " goes to the off position too easily sh utting the car Off.,,309 This e-mai l chain was forwarded to DeGiorgio with a E-mail from Jonathan L. Webcr, GM , toRajivMcllla, GM, elal. (March 9, 2005), at 22 (attached to FPR0793/200SIUS) [DOC ID GMHECOOOOI9677[. 304 E-mail from Jonathan L. Wcbcr, GM, to Rajiv Mchta, GM, elal. (March 9, 2005), at 22 (attachcd to FPR079312005IlJS) [DOC ID GMHECOOOOI9677] 305 FPR 0793 /2005/lJ S (March 9, 2005) LDOC ID GMHECOOOOI9677J ; GM CPIP and PRTS+ Process Overview, at 14 [DOC lD S006967 000206103 J. 306 J&B Intcrview ofSlcvcn Oaklcy, April 10, 2014. 307 J&B Intcrview ofStcvcn Oaklcy, April 10, 2014. j lJtl E-mail from Stcvcn Oakley, GM, to Amalld Dcssiricix, GM (May 2, 2005) rOoe ID 000077753011; GMNHTSA0003374831. 309 E-mail from Steven Oakley, GM, to Amalld Dessirieix, GM (May 2, 2005) [DOC ID 000077753011 ; GMNHTSA000337483 J. 303 76 Privileged and Confidenfia/ Protec1ed by AUorney~ Client Privilege and as Allorney Work Product request for an Igniti on Switch "at the high end of the tolerance spec.,,310 However, the field representati ve to whom th e customer complained expressed skepticism that a switch with hi gher torque could be located. Oakley explained: The fi eld rep will swap the parts if we wa nt them to. He is concerned that this will not correct the conditi on, as he feel s several stock cars at the dealership have about th e same level of effort for the switch. They would like to have a column sent to them that we have some kind of confidence is better than what they are taking out. Again, if you just want a swap out we can do this, but without the ability to measure the effort, I have a hard time persuading them this will actually fi x the car.311 10 ~.-...v$'GMIGMCOG M ~. &ob,tct: ..., At.: IJI'VItt AoIq_ TIM fitl:l rtpwil ..... p!he ptlt$ if .... '/1111"11 ,"m 10', Htb""","," twol 1tn'fril1lC:ll ~"" .... , OnQCiOn, n htlttl$ _rtl MlXII WI1.I .... 2 J&B Interview of Douglas Wachtel, March 18, 201 4. 363 J&B Interview of Gay Kent, May 16, 201 4; J&B Interview of Maureen Foley-Gardner, March 26. 2014. 364 J&B Interview of Maureen Foley-Gardner, March 26, 201 4. 365 J&B Interview of Douglas Wachtel, March 18, 201 4. 359 360 86 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product (U ntil April 20 13, the Product Investigations group had 8 in vestigators; now it has 40.)366 Wachtel's team looked at earl y data from the field and fo und 14 incidents that might have been related to the Ignition Switch. 367 The PI group also tried to recreate the problem th emselves. 368 Wachtel and Gay Kent obtained a Cobalt and drove around GM 's property in Warren. Kent had a long and heavy key chain, and was able to knock the ignition from Run to Accessory by moving her leg so that her jeans caused fri ction against the fob. 369 Wachtel could reproduce the phenomenon even more easily, but still only by co ntacting the key cha in rather than hitting . heroa. d TIO bumpsmt Notwithstanding the multiple media reports, the customer complaints, and being able to rep licate the issue in the fi eld, the PI team concluded that the Ignition Switch probl em was not of such a magnitude as to require a recall , either for safety or customer service. On June 28, 2005, the same day that the VAPIR was fin ishing its consideration of engi neering fixes, the PI team held an Investigation Status Review (" ISR") meeting at whi ch Elizabeth Kiihr presented on the Ignition Switch. 371 Following that meeting, Kiihr e-ma il ed Manzor stating: "We are currently not planning on continuing this issue in the FPE process.,,372 According to Kiihr, at the time of the in vestigation, she did not think the problem was occurring frequently; she noted that drivers still J&B Intcrvicw of Mark Johmon, May 7, 201 4; J&B Intcrvicw ofCanncll Bcnavidcs, April 22, 2014. Interview of Elizabeth Kiihr, May 15, 2014; TREAD Search Results (June 28, 2005) LDOC ID 000005586004; DOC ID 000005586005; DOC ID 000005586006]. 368 J&B Interview of Douglas Wachtel, March 18, 2014. 369 J&B Interview of Douglas Wachtel, March 18, 2014. n o J&B Interview of Douglas Wachtel, March 18, 201 4. 371 J&B Interview ofElizabcth Kiihr, May 15, 2014. 372 E-mail from Alberto Manzor, GM, to Joseph Manson, GM, er a/. (June 28, 2005) [DOC ID S007 11 6_00000826IJ (quoting e-mail from Elizabeth Ki ihr). 366 367 J&B 87 Privileged and Confidenfia/ Protec1ed by AUorney~ Client Privilege and as Allorney Work Product had manual steering and brakes availab le after a stal1. 373 She and her P I coll eagues, GM ' s lead safety in vestigators, fai led to understand that h1ming off the ignition also di sabled th e airbags They did not consider the circumstance that would be repeated in years to come: customers in off~roa d sihlations losing power due to the low torque of the Ignition Switch just before a crash - when ai rbags are needed the most. Instead, the PI team supported issuing a Technical Service Bulletin - a notice to dealers , but not consumers - about the issue, which GM did in December 2005. 374 Product Investigations also supported changing the key design from a slot to a hol e sty le key.37s The evidence demonstrates that DeGiorgio was alerted to the press problems as we ll. Delphi internal communications from June 2005 indi cate that " Ray" asked for "force di splacement curves" on " ignition switches (main ly the Delta)." The Delphi engineer stated, "Cobalt is blowing up in their fac e in regards to turning the car off with the dri ver's knee.,,376 From: Coniff, John B sent: TlJI'Sday, June 14, 2005 2:28 PM To: Svoboda, Thomas E; Lin, George J Subject: Force displacement curves Tom, who has the capabitity 10 run fOrce displacement curves on all the Ignition swi:ches (mainly the Della) is it Condura or D.G.? Ra1 is requesting this inlOrmalion. Cobalt is blowing up in their face with the driver's knee. Please Ie I me know. r. regards to turning \tie car off Thanks m J&B Interview of Elizabeth Kiihr, May 15,2014. m J&B Interview of Douglas Wachtel, March 18, 2014; Service Bullet in 05~02~ 35007 (Dec. 2005) rDOC ID 1402 130 1709348; GMHEC000329773l m J&B Interview of Douglas Wachtel, March 18,2014. 376 E-mail from John B. ConifT, Delphi, to Thomas E. Svoboda, Delphi, ef 01. (June 14, 2005) [DOC 10 00005 1786007j. 88 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product DeG iorgio clai ms not to recall hearing about any of the negative press, nor whether he had made such a comment to Delphi _J77 We have fo und no documents within GM refl ecting such a communication. Days after th e ISR, a GM c ustomer fi led a comp laint about a 2005 Coba lt prone to moving stall s. J78 The customer co mplaint states: "2005 Chevrolet Cobalt experienced problems with total loss of the electrical system and the vehicl e sta lli ng .. The consumer stated the ignition sw itch was poorly install ed. Even with the sli ghtest touch the vehicle will shut off in motion." Attached to the comp laint is a letter fro m the customer to GM Customer Service that states: This is a sa fety/recall iss ue if ever there was one. Forget the bu lletin. I have found the cause of the problem. Not suggested causes as li sted in bu lletin. The problem is the ignition turn switch is poorly installed. Even w ith the slightest touch, the car will shut off whil e in motion. I don ' t have to li st to you the safety problems that may happen, besides an accident or death , a car turning off whi le doing a high speed must cause engine and other problems in the long haul. I am forwarding this letter to [N HTSA] as I firml y believe that this ignition switch needs to be recall ed, reexamined and corrected.379 J&B Interview or Raymond DeGiorgio , May 7~8 , 2014. Customer complaint (June 29, 2005) rooe ID 000014669078; GMNHTSAOO0540683l We arc also aware or one Beller Business BUTeau arbitrator dec ision mandating that GM repurchase a Cobalt rrom a customer who complained orintennittent stal ling, because " unexplai ned sta lling ora vehicle in trafTic cenainly constitutes a serious sarety ha7.-ard." BBB arbitration decision or Nonna Lovaco (Feb. 27, 2006) [DOC ID 377 J78 EX PORTJM000042544]. 379 Customer complaint (June 29, 2005) LDOe ID 000014669078; GMN HTSAOOO540683 j. 89 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product J""A'Z9,lCOS Cb_~_ ~Mo!on Corp_ ,4 ro IkrI JJI7G I' DmoIt,Ml4Wl-! I7G Rlt, 2OII5CHn'YCOBALT YIN, J(; IAJ.5l.J I>- CIlltDmM'Service: .... In)' ,"Y\oQ:COIIr~ ~ I cq:rIIIIad.a_r/;'J tlfIarloa bid CImIrftd (tbrDa tbIl:1o) wl'lhm)' _lOO3 CotIIJI- lI.Inia oS'wNJe 1b 1lIinI.:1111 ..... ~ oIote ID aft KcIoWI61\ I drlYIU_ nil b.~recall iuucitewrtberewu 0lIl:, FOIJCI tile lNDetiA. I M\'c lbuDd tho~ of 1M problem... Not PIQH'kd ClUICI u lilted La bullctia. Theprobllll11t tb8 ipitioe. mm rwitchia poorly 1GC:aJI.e4. sv- ~ 1bo I1I&hat IDaeh, 1ba QIl wW Ihat off wt.Ue ia. moCkc. J doG, .,,, to Jill to }'OIl tha tIft(y prebl. . . . mq ...." blew.. loll .cew.t 01' daJf\. • CG' Nmi.D& ottwtlfle dotq • tUp IIpttd mlIIt eqp IlHI. oIbIr probIcmI in the kq haul. 1lI_ I n ~ .... ktIDr to Ihe NdonaJ H1atrwa:r TnfIIc s.ky Mmmi....mn _ I t1rmJy believe that dllIlphioGttartl'lW'!teh Mt.dI to be tet.al1od., ~ nlVileil ed oom:otod. '- DrMf. n. ... MOT'*..~... Despite this complaint, the Ign iti on Switch problem conti nued to be categorized as a customer convenience issue rather than a safety issue. On July 12, 2005 , GM updated its prior Preliminary Informa.tion to dealers (not customers) , adv isi ng that " Engineeri ng has come up with an insert for the key ring, so that it goes from a ' slot' design to a hole. As a result, the key ring can not move up and down the slot any longer, it can only rotate on the hole. In additi on, the previous key ring has been replaced with a small er, 13mm design. ,,38o 12. September 2005: Engineering Committees Reject Proposals to Change the Switch In September 2005, engineering teams conside red whether to replace the problemati c Coba lt Ignition Switch altogether with the GMT 191 switch that DeGiorgio had been advocating. The motive to make the change, according to Lori Queen (the Vehicle Line Engineer for Small Cars), was that the Cobalt team identifi ed a customer sati sfact ion issue with the Ignition Switch 31!O Preliminary Informat ion PIC3421A (July 12. 2005) [DOC ID 000124908092 ; GMNH TSA000458336j. 90 Privileged and Confidenfia/ Protec1ed by AUorn ey~ Client Privilege and as Allorney Work Product and had fi gured out how to fi x it. 38 1 Queen did not reme mber why the proposed use of the GMT 19 1 switch was rejected; her review of contemporaneous documents led her to conclude that it was rejected as being too expensive and would not result in offsetting changes in savings on warranty costs:182 Once again, GM personnel 's mi sclassifi cation of the issue as one of conveni ence, rather than safety, resulted in a solution being derail ed by cost. In rej ecting the GMT t 91 switch proposal , the C ross ~ VAPIR committee (see Appendix B) concluded th at th e proposed key head design change would be an adequate fi x - a fix that ended up being cancelled in the coming months. 383 13. December 2005: GM Issues a Technical Service Bulletin Regarding " Information on Inadvertent Turning of Key Cylinder, Loss of Electrical System and No DTCs" Having failed to identify the airbag non-deployment connection and having decl ined to treat mov ing stalls as a safety issue, GM personnel proceeded with parti al field solutions. One of those was the dec ision to send a Technical Service Bulletin in 2005 to dealers. A Technical Service Bulletin ("T SB") is also a communi cation to dealers (not consumers), and generally provides infonn ation both about a probl em and a potential so lution. TS Bs are provided to NHTSA and available on the NHTSA website. In December 2005, GM issued a TSB entitled: " Informati on on Inadvertent Turning of Key Cylinder, Loss of Electrical System and No DTCs.,,384 Like the prior Preliminary Information, the TSB explained to deal ers that GM now had an insert to change the key from a slot to a hole and a new smaller 13 mm key ring. The TSB applied to fi ve models: 3~1 J& B Intcrview orLori Quccn, May 12,20 14. J& B Intcrview orLori Quccn, May 12,20 14. 383 J& B Interview or David Tmsh, March 17, 2014; EWO 521 309 (June 9, 2005) [DOC 10 $007116 000008319]. 384 Service Bulletin 05-02-35007 (Dec. 2005) LOOe ID 14021301709348; GMH EC000329773j. 382 91 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product • 2005-2006 Chevrolet Cobalts; • 2006 Chevrolet HHRs; • 2005-06 Pontiac Pursuits; • 2006 Pontiac Solstices; and • 2003-2006 Saturn lons. 385 The main text of the TSB was virtuall y identical to the prior Preliminary Informations from earl ier in 2005. in describing a "potential for the driver to inadvertently turn off the ignition due to low ignition key cylinder torque/effort.,,386 Like the prior Preliminary lnformation, the TSB also warned that drivers should "remo v[e] unessential items from their key cha in," and explained that engineering had developed an insert for the key head, designed to change the slot to a hole, and noted that GM would be providing smaller key rings. 387 Unlike the Preliminary Information that preceded this Service Bulletin388 or an earlier Service Bulletin request drafted by Steve Oakley, the Technkal Service Bulletin did not describe the problem as involving a "stall.,,389 According to Oakley, the term "stall" is a " hot" word that GM generally does not use in bulletins because it may raise a concern about vehicle safety. which suggests GM should recall the vehicle. not issue a bulletin. 39o Service Bulletin 05-02-35007 (Dcc. 2005) [DOC ID 14021301709348; GMHEC000329773 j. Preliminary inronnation, Engine Stalls, Loss Electrical Systems, and No OTCs (Feb. 28, 2005) [DOC ID 000000000067; GMNHTSAOO0272962]. 381 Servicc Bulletin 05-02-35007 (Dec. 2005) [DOC ID 14021301709348; GMHEC000329773J. 388 Preliminary infonnation, Engine Sialls, Loss or Electrical Systcms, and No OTCs (Feb. 28, 2005) [DOC 10000000000067; GMNHTSAOO0272962j. 389 Servicc Bulletin 05-02-35007 (Dec. 2005) rOOC ID 14021301709348; GMHECOOO329773l; Preliminary infonnmion, Engine Stalls, Loss of Elcctrical Systems, and No OTCs (Feb. 28, 2005) rOOC ID 000000000067; GMNHTSA000272962l; Preliminary Inronnalion (July 12, 2005) rOOC ID 14032000467431 ; GMNHTSA000002397l; Preliminary Inronnalion (Sept. 15, 2005) rOOC ID 14032000467433; GMNHTSA000002399]. 390 J&B Interview of Steven Oakley, April 23, 2014. 385 386 or 92 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product Others agreed that GM is sensitive to using the word "stall " in a service bulletins and clo~el y ~crutini ze~ any bulletin that does include "~tall~» a~ a ~ymptom 391 Other~ at GM confirmed that there was concern about the use of "stall " in a TSB because such language mi ght draw the attention of NHTSA. 392 Oakley said that at times, he included " hot" words to draw attenti on to an issue fro m Product Investigation s personnel who review Service Bulletins before release. 393 Oakley also noted, however, that he was reluctant to push hard on safety issues because of his perception that his predecessor had been pushed out of the job for doing just that. 394 As discussed above, discussions with Altman and other engineers all eviated Oakley' s initial concern that the Ignition Switch presented a safety issue. 395 Oakley's draft with the " hot" language went to Product Investigations for review. Again, the engineers in Product Investi gations failed to make the link between the Ignition Switch position and airbag n on~dep l oyment and did not el evate the Cobalt's Ignition Switch as a safety issue into the process to consider recall s.396 In stead, it exc ised the " hot" word "stall " from the bulletin, and considered the matter c1osed. 397 As a result, the Techni cal Service Bulletin did not solve the problem. In order for a GM vehi cle owner to learn of the suggested service fix (i.e. , removal of excess weight from the key chai n) and to obtain a key insert plug and smaller key ring, (I) the customer had to experience a J&B lntcrvicw of Douglas Wachtcl, March 18, 201 4; J&B Intcrview of Keith Mi.kkclson, April 23, 2014. J&B Interview ofCannen Benavides, May 19.2014: J&B Lntcrview of Gay Kent. May 16, 2014. 393 J&B Intcrview ofStcvcn Oaklcy, April 23, 2014. 394 J&B Intcrview ofStcvcn Oaklcy, May 7, 2014. 395 J&B Interview ofStcven Oaklcy, April 10, 201 4. 396 J&B Interview of Douglas Wachtel, March 18, 201 4. 397 J&B lntcrview ofStcvcn Oaklcy, April 23, 2014; Preliminary Infonnation, Enginc Stalls, Loss of Electrical Systcms, and No DTCs (Fcb. 28, 2005) rOOC ID 000000000067; GMNHTSA000272962l; Scrviee Bullctin 05~02~ 35007 (D". 2005) [DOC ID 14021301709348; GMHEC000329773]. 391 392 93 Privileged and Confidenfia/ Protec1ed by AUorn ey~ Client Privilege and as Allorney Work Product movi ng stall ; (2) the customer had to go to the dealership to complain of th e issue; (3) the techni c ian at the dealership had to diagnose the issue properl y; and (4) the techni cian had to search GM's bulletin database to identify the appli cable bull etin without using the te rm "stall" since the word would not appear in the TSB as issued. The odds were not with the consumer. Moreover, when GM issued the TS8 , the prior Preliminary Informations (which had accurately used the word "stall ") were removed from the dealer database as obso lete. Although GM created over 10,000 key plug inserts, onl y approximately 430 were ever requested by customers. Moreover, as di scussed above, although GM made the key insert availabl e to consumers of previously purchased vehicl es, it did not, at the same time, change the key for cars that were rolling off the assembl y line and those yet to be produced. That fi x for future production vehi cles, approved by the VA PIR in June 2005, was not impl emented in 2005 . GM was hav ing a pri ce d ispute and quality probl ems with its suppli er, Ortech, whi ch was goi ng out of business. Those disputes were delaying other changes to the igniti on cylinder that GM persomlel beli eved were more important ; in September 2006, GM personne l abandoned the change to the key head, citing those delays. Thus, even the " band~ a id" that GM engineers thought they were applying was not implemented fo r new cars . Thi s ended GM's efforts to impl ement the key design and key ring change in the 2005 and 2006 time period. 14. After Numerous Reports of Moving Stalls, E ngineering \-Vork to Resolve the Problem, Multiple Committee Meetings, and a Brief Investigation by GM's Safety Investigators, No Meaningful Action Is Taken The fact that keys were being inadvertently turned to Accessory was certainly no secret. Consumers, the medi a, and GM 's own empl oyees reported the problem on numerous occasions, and how it led to moving stalls. GM engineers repli cated the stall s multipl e times in the field, 94 Privileged and Confidenfia/ Protec1ed by AUorn ey~ Client Privilege and as Allorney Work Product and kn ew full well - from the outset - that the igniti on turned to Accessory because of low torque, a below-Spec ifi cation torque approved by DeGiorgio_ Whil e e ngineers deve loped shortterm and long-term solutions for the problem, none of the engineers on the multiple, crossdi sc iplina ry committees that looked at the stalling probl em made th e s impl e connection that GM had designed the car so that turning the key to Accessory disabled the airbags. The engineers made a basic mistake. They did 110t know how their own vehicle had been designed. And GM did not have a process in place to make sure someone looking at the issue had a c omplete understanding of what the failure of the Ignition Switc h meant for a customer. As a result, those looking a t the stall s caused by the inadvertent turning of th e Ignition Sw itch rated the probl em a " convenience" issue rather than one of "safety" - an error that led to enormous consequences for many. GM personnel furth e r compounde d this error by prov iding infonnation to its dealers that obscured the problem - removi ng the word "stall" fro m its TS B prec isely because that word mi ght trigger customer's concern s about sa fety. As a result of these multipl e failures, and the failure of DeGiorgio to prov ide critical infonnation to those reviewing the issue, Cobalts stayed on the road with a safety defect that could have been addressed many years earli er. C. The Ignition Switch Part Change Unbeknownst to the GM engineers di scussing potential so lutions to the problem of low torque in the Ignition Switch, DeGiorgio was communicating with De lphi about how to increase torque in the Ignition Switch for future model years. These discussions led to a soluti on to the probl em of low torque in certain MY 2007 Cobalts and all Cobalts from MY 2008-2010. 95 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product The fact that GM personnel in vestigatin g the Ign ition Switch in future years did not know about the fix, and that DeG iorgio himse lf says he does not remember it, had a significant effe ct on GM's response in ensuing years. In the end, as described below, the hidden implementation of a solution to the problem addressed the issue for drivers of MY 2008~20 I 0 Cobalts (and some MY 2007 Cobalts), but made it more difficult for GM engineers to determine what was causing problems for the drivers of earlier model s. As a result, drivers of earlier model year Cobalts and other vehicles continued to drive cars with safety defects. I. Late 2005 to Early 2006: Discussions About Changing the Spring and Plunger Documents reflect that, in late 2005 (months after the bad press) and early 2006, DeGiorgio discussed with Delphi putting a stronger spring and plunger into the Ignition Switch. 398 In January 2006, a Delphi engineer sent DeGiorgio an e~mail informing him that, " per our conversation in December, I sent you the 24 Delta Ignition Switch samples for vehic le testing. These switches contain the new PCB design and also the stronger Catera detent spri ng ~plunger. ,,399 As noted in the January 2006 e~mail, one of the solutions di scussed was using a spring from the new Cadillac Catera SRX (GMT 265) sw itch in the Cohalt. Throughout 2005, DeGiorgio had heen working on improving the electrical architecture of the ignition switch for 398 E-mail from John B.ConifT, Delphi, to Raymond DcGiorgio, GM (Sept. 22 , 2005) IDOe ID 000012168115 ; GMNHTSA0002639 I 81; c-mail from Raymond DcGiorgio, GM \0 John B. eonifT, Dclphi, el al. (Sepl. 22, 2005) [DOC ID 000012168115; GMNHTSA0002639 I 8]; c-m:l il from John B. COlliff, Delphi, [0 R:lymond DcGiorgio, GM (Sept. 22, 2005) [DOC ID 000012168115; GMNHTSAOO0263918J ; e-mail from George J. Lin, Delphi, \0 Raymond DeGiorgio, GM, John B. Coniff, Delphi , ef al. (Sepl. 28, 2005) lDOe ID 000012168174; GMNHTSA000406699l; e-mai l from GeorgeJ. Lin, Delphi, 1O Raymond DeGiorgio, GM , John B. ConifT, Delphi, el al. (OCI. 4, 2005) rDOe ID 000012168224; GMNHTSAOOO263926l; e-mail from George J. Lin, Delphi , 10 Raymond DcGiorgio, GM , John B. Coni fT, Delphi, el 01. (Oct. 20, 2005) rDOC ID 000012168344; GMNHTSA00040670 11. 399 E-mail from Arturo Alcala, Delphi, to Raymond DeGiorgio, GM, John B. Coniff, Delphi, el al. (Jan. 6, 2006) lDOe ID 000051786002; GMNHTSAOO0257777J, 96 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product the Catera SRX ; in fact, DeGiorgio had worked on an earlier version of the switch in 1999 for the Cadillac Catera car (the GMX 320). In early 2006, DeGiorgio and Delphi ex pressly discussed using the spring and detent plunger from the Catera in the Cobalt Ignition Switch. 4OO The evidence does not show, however, whether importing the Cate ra spring into the Ignition Switch in the Cobalt (and other Delta and Kappa platfonn vehicles) was Delphi 's idea or DeGiorgio's. Regardless, the plan took root. Following discussions with DeG iorgio, Delphi produced a new version of the Cobalt Ignition Switch with two changes, one electrical and one mechanical. The electrical one was a "new PCB [Printed Circuit Board] design," intended to address electrical problems with the Ignition Switch that continued to cause the vehicle not to start (the no crank/no start problem). The mechanical change was a "stronger Catera detent spring~plunger, " a change intended to increase the torque required to tum the Ignition Switch. 4 01 An internal Delphi document indi cates that this switch design - with a longer detent s pring~plunger - was the same as the longer detent spring-plunger design origina ll y drafted by Delphi in 2001 .402 In other words, thi s option had been avai lable when the Ignition Switch had first been des igned. E~lll a il frOIll Eduardo P. Rodriguez, Delphi , to Raymond DeGiorgio, GM , ef al. (March 24, 20(6) LDOC ID DLPH_DOJ_0000913 J; spreadsheet referring to Catera spring and plunger in Cobalt ignition switch LDOC ID DLPH_DOJ_00009241 , attached to e-mail from Eduardo P. Rodriguez, Delphi , to Raymond DeGiorgio, GM , ef al. (March 31 , 20(6) [DOC TO DLPH_DOJ_OOO0923l; e-mail from Eduardo P. Rodriguez, Delphi , to Raymond DeGiorgio, GM, ef al. (March 31 , 2006) [DOC JD DLPH_DOJ_0000927]; e-mail fmmRaymondDcGiorgio. GM. to Eduardo P. Rodriguez, Delphi, ef al. (April 5, 2006) LDOC ID DLPH_DOl_000927J. 401 E-mail from Arluro A lcala. DelphitoRaymondDeGiorgio. GM. John B. Coniff, Delphi, et al. (Jan. 6, 2006) [DOC ID 000051786002; GMNHTSAOO0257777l ·102 Drawing 741-76307-T rDOC ID GMHECOOOO03206l ; 200 I Long Detem Spring Drawing, Drawing 741-79378 (200 1) [Ex. A.3.a(2) 2001 Long Detent Spring Drawingl; 2001 Short Detent Spring Drawing, Drawing 741-75259 (200 1) [Ex. A.3.a (I) 200 1 Shon Detent Spring Drawing]; e-mai l from Amero Cuervo, Delphi, to Ly le Miller, Delphi (Oct. 29, 2013) [DOC ID 000004253527; GMNHTSAOO0223906]. There is no indication that GM received these documents in 200 J . 400 97 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product In earl y 2006, Delphi employees continued to correspond with DeG iorgio regarding the changes to the Ignition Switch_ The main focu s of Delphi 's correspondence was on the electrical changes, but the e-mails also discllssed the modification in the switch to the stronger "Catera Spring/Plunger. ,,40] An internal De lphi document from January 2006 specificall y described the change to the spring as a request from GM "to be in specification according [to] the GM spec for the torque forces," an acknowledgment that the current production Ignition Switch was out of specification and that DeG iorgio was now focu sing on meeting the Specification. 404 2. April 2006: DeGiorgio' s Approval of the Ignition Switch Design Change On Apri l 26, 2006, DeG iorgio approved the redesigned Ignition Switch by signing what is ca lled a Form 3660, giving Delphi permiss ion to begin manufacturillg the redesigned sw itch. 405 The Form 3660 stated, «[nJew detent plunge r (Catera spring/pIUJlger) was implemented to increase torque force in switch.',4Q6 Each Form 3660 has to link bac k to a master work order, and thi s one did as we ll. But the work order to which it was linked was only for the electrical improvements to the Ignition Switch; the work order did not include the change to the spring and plunger. 40 7 As the work order stated, and DeGiorgio confirmed, there was no cost 408 associated with the change to the Catera spring. The new Ignition Switch required greater torque to tum the key than the original switch and, as became clear in future years, did not lead E~m.ail from Anuro Alca la, Delphi, 10 Raymond DcGiorgio, GM, John B. Coni iT, Dclphi, el al. (Jan. 6, 20(6) LDOe ID 000051786002; GMNHTSA000257777J; e-mail from Raymond DeGiorgio, GM, to Arturo Alcala, Delphi (Jan. 9, 20(6) rDOC ID 000051786002; GMNHTSAOO0257777j. ·II).! Delphi Change Request Fonn (Jan. 15 , 2006) rooe ID [20198-0003531·105 Fonn 3660 (Apri l 26, 20(6), at 3 rOOC ID 000004253529; GMNHTSA000223924]. 406 Fonn 3660 (April 26, 20(6), at 3 rDOC ID 000004253529; GMNHTSA0002239241407 EWO 302726 (Feb. 19,2004) rOOC ID 000000000080; GMNHTSAOO02206671 ·108 J&B Interview of Raymond OeGiorgio, May 7-8, 2014; EWO 302726 (Feb. 19, 2004) [DOC ID 000000000080; GMNHTSA000220667 J. ·103 98 Privileged and Confidenfia/ Protec1ed by AUorney~ Client Privilege and as Allorney Work Product to the non-deployment of air bags in crashes. Thus, the change that actually succeeded in resol vlng the Igniti on Switch problem by increas ing the torque (a) was made at no cost, (h) llsed a part that had been available in 200 1, and (c) had no corresponding work order, resulting in no one findin g it until shortly before the reca ll. I::> coriad 51& _ _ GM DECI~ON : - Ed:IW:I~ PhOI\t No." o..d:' Ro~t-~lfbtl9oo~ D ' ,"*,111 PnxlIlc:I Eng f,f"''''''8a'I:t G,""N'.l1'(pteNI~b. GM 1IIr.harIzGd Ti..• ...a..o _ _ New Pi::80C1N<11!11TWig1~ 10" " - tnj .... iiiiii"lIlIiOIIiiaI. ~ t_ _ iii '*ow) ..., ..~ . . eo:-..nC 0 ~~SCI:l7 ."... SISlr>-()ft OOmpiel. [K] -"' ~ Delphi 's Part Submiss ion Warrant was signed a month later.409 Delphi documents suggest that the new Ignition Switch went into production sometime after June 26, 2006. 410 DeGiorgio states that he does not remembe r any of these events, including th e di sc ussions with Delphi about increasing torque, the use of a Catera spring he had previously worked on, or the authorization of the change. ·109 4 10 Delphi Part Submission Warrant (May 26, 2006) [DOC ID 120198-000363]. PRTS 199612006/US (May 22, 2006) lODe 10 1402130173 1670j. 99 Privileged and Confidenfia/ Protec1ed by AUorney~ Client Privilege and as Allorney Work Product 3. May 2006: DeGiorgio's Approval of the Design Change Without a Change in Part Number Although the c:::> e::: L po ....... 1 design of the Igniti on Sw itch changed, the part OElPttl MECHATRQNIC SYSTEMS Engineering Change Suppot1 lnformallon Cover Page $l l ~.doe I number remained th e ,...., Iolll00lioMiOl . same. An internal . bout T.etwM~ e......-on· Tl'ootlnfon"l\lllon MIPPO" !No io ...... . ,1M1on (1ft NoewPCS 12861211 R... S..-w::I "'- N_ o.t.nt Pk.onget 7. 1.79378 , \IIU P«kwl'l GM ~in le T.It .COO6- l l1. OU'ItIiIty 3 bn'Ies life _ Delphi document dated diI " TCIIq.IiI.nd Af9tt tett wn petfoIm in OG w1Ih WOfIl. Request OS00571 -. May 27, 2006, expressly ~ o.GIcxglo I GM RtiE) ~"IO 10 1• • 6" ehIrQt~ ~GM pA Ht~idntn 8PPfOYII WCh GM3960 Ie ., .... II boIh stated that " Ray [sic] DeGiorgio ... agree[ d] '* 'PC" p!IoIo Isn:rnsrcg JII1(I bW9 in Mg:lt*ioos VMhv..y.~IOg«. betI« pta • • oct dOc ... .... ..... ~ POT mMling ..... ~ nee ....... 10 to implement change gel . , . CO"i-.cI pice tIeCaIM Ihe ()oQ f.1kuI • tel in the ... dosed 10 30% without changing GM pin. He provides hi s approval. ,.4 11 As Mary Barra has explained, it was "Engineerin g 101 " that, given the signi ficance of th e change to the ignition switch , the part number should have been changed. 4 12 But DeG iorgio did not change it. Part numbers are the means by whi ch GM and its suppl iers keep track of the components in GM vehicles. GM had a policy that required an engineer to change the part number for any change that impacts a part 's " fit, form, or function .'.413 Here, the change to the spring in the Ignition Switch changed the part's fun ction, but DeGiorgio chose not to change the part number. Fonn 3660 Cover Memorandum (May 26, 2006), at 1 (enclosing a copy of the Fann 3660 for approval) rDOC ID 000004 253529; GMNHTSA0002239241. 412 J&B lntervicw of Mary Barra, April 14, 2014. oH3 Sect ion No. G3.!.! ofGM ' s Global Product Description System - " Vehicle Pari Number Assignment" (Feb. 8, 2003), at 2 [DOC ID 000039358889; GMNHTSAOO0553412j. ·HI 100 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product GM 's policy further provided that an engineer could seek an exception to the part number requirement (i) if certain criteria are met, or (ii) if the criteria are not met, with approval of the Change Approval Board. 414 The facts here did not meet the criteria for an exception, and DeGiorgio did not seek authorization of the Change Approval Board to proceed with the same part number. Design Release Engineers, like DeGiorgio, are responsible for detennining whether a part change requires a new part number. 415 There are exceptions to the part number change rule, none of which DeGiorgio availed himself. We found no proposal not to change the part number, no use ofa temporary work order (under which he could change a part while operating temporarily under the old pan number) , and no other scrutiny of any kind. For his pan, DeGiorgio remembers none of this and offers no explanation for why he did not change the part number. Whatever DeG iorgio 's reasons for not changing the part number, the failure to make the change had serious consequences. A !though the change effectively cured the problem of low rotational torque in the Ignition Switch and addressed the safety problem in future cars, it did nothing for cars built prior to the break~point.41 6 Worse, as crashes continued to occur and GM in vestigators tried to find the cause, the failure to change the part number led the investigators in future years to a wrong conclusion. When they incorrectly concluded there had not been a change in the Ignition Switch between MY 2008 model s and earlier, they incorrectly concluded Section No. G3 . I.l ofGM's Global Product Description System - " Vehicle Pan Number Assignment" (Feb. 8, 2003), at 3~4 rOOC TO 000039358889; GMNHTSAOO05534l2]. 41) J&B Interview of Vickie Teetsel, April 9, 2014. 41 6 In September 2006, Raymond DeGiorgio purchased an MY 2007 Cobalt for his son. It is not clear, based on the timing orthe purchase, w hether the Cobalt had the old version orthc Ignition Switch or the improved version orthe Ignition Switch with the new spring authorized by DeGiorgio. ,114 101 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product that the Ignition Switch was not to blame. Had others at GM known that the Ignition Switch had been c hanged during MY 2007 ,4 17 it is highly likely they would ha ve concl uded th eir in vesti gation much more swiftly and reca lled the Cobalt and other cars earlier. VI. 2006 - 2014: THE LONG INVESTIGATION INTO ROOT CAUSE As described above, by June 2006, engineers at GM had been addressing the issue of moving stall s in the Cobalt for almost two years. Still , they did not view the problem as a safety issue a nd did not understand the relation ship between airbag non-deployment and th e ease with which the Igniti on Switch could be rotated to Accessory. And , notwithstanding a decision to change the key head, no solution had, in fact, been impl emented. (Although DeGiorgio had approved a redesign of the switch in 2006, it did not address the MY 2005, MY 2006, and certain MY 2007 vehi cles.) By this time, GM personnel had publicl y announced that th ere was no sa fety issue with Cobalt moving stalls, and GM 's group of sa fety engineers, the Product Investigations (" PI") group, had detennined no further investigation was warranted.418 From 2006 to 2011 , GM's PI group did not focus on issues related to th e Cobalt Ignition Sw itch. Beginning i.n 2006, however, GM attorneys and a different group of engineers working m Thcrc is no cvidcncc that DcGiorgio to ld others at GM , including cngineers on the Cobalt programtcam, about thc spring changc to thc Ignition Switch that he authorized in April 2006. Howcvcr, documcnts subsequcnt to the changc indicatc that a handful of GM engi neers in othcr dcpanmell\s received infonn ation describing the change. For cxample, a June 30, 2006 e-ma il from Delphi to DcGiorgio described the changes to thc switch, including that the "detent plunger is implemented to increase torque forces to be within specification," and included four other GM engineers on the e-mai\. E-ma il from Eduardo P. Rodriguez, Delphi, to Raymond DeGiorgio, GM, el at. (June 2, 2006) [DOC ID DLPH_DOJ_00009701. Si milarly, an unrelatcd PRTS rcpon was opcned in Junc 2006 by a Saturn supply quality cnginccr that referred back 10 DcGiorgio's changc to Ihc switch, and specifically nOled: "The delelll plunger torque force was increascd. This changc was trcated as a black box tier 3 design change and was approved via a 3660 document SigllCd offby thc GM ORE." PRTS N202832 (Junc 13,2006), at 10 LDOC ID 14030700024180; GMNHTSA000020327; GMHECOOOO20630l Thcsc individuals werc not involved in the invcstigalions that ensucd in the coming years, nor did they hold a position, like DcGiorgio's, with responsibi lity for the Ignition Switch. 418 J&B Intcrview of Gary Altman, March 14,2014; J&B Intervicw of Doug Parks, March 18, 2014; PRTS N 172404 (Nov. 19, 2(04) [DOC ID 0000017711 43; GMNHTSAOOOI42658 1]; Alan Adler, GM Statement on Chevrolet Cobalt Inadvenent Shutoffs (Junc 13, 2005) [DOC ID S006878_000015062; GMNHTSA300005313 J. 102 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product in collaboration with them began to see accidents and claims in Ions and Cobalts that involved non~deplo ym ent of air bags and that, they wou ld ultimately conclude, related to th e Ignition Switch. As discussed below, the ongoing work of Field Performance Assessment CFPA") engineers made slow progress, in no small part because they failed to search for or find infonnation relevant to the problem of airbag non-deployment that was either public or actually in GM 's own files. The FPA engineers, accustomed to focusing on individual claims or lawsuits and explanations of individual crashes, did not even consider searchin g for this information, which might have materiall y advanced their investi gation. The failure to locate and use thi s information, the failure to elevate issues in a timely manner, and the narrowness with which each group - lawyers and engineers alike - approached their rol es, are important reasons why the investigation proceeded slowly. A. 2006: Litigation Into Fatalities Begins For the most part, the first Cobalt and Ion airbag GM 's Legal Staff in late 2005 and 2006.41 9 The first _ n on~dep l oyment cases bega n reaching death and . head injuries, involved accidents in which GM would not have expected the airbags to deploy. But in the third, the death engineers believed that the airbags should have deployed, and despite "extensive ana lysis," the GM lawyer working with the GM ' s Lcgal DcpanmclI( receivcd not icc ofthc fi rst Ion non-dcploymcnt claim in January 2004 [or- - ., who was in a 2004 SatUnl Ion. Lclter from Donglas Brown, GM, ]o Yverre Young, ESIS (Jan. 28, 2004)""[D0C'i'D 00000 17 14004; GMNHTSA000200613 J; J&B Interview of Douglas Brown, March 19, 20 14. While GM ' s claims administrator ESIS handled this claim, Doug Brown and Jaclyn Palmer were consulted, and the claim was ultimately dcnicd. Mcmo from Yvct~S I S, to Jaclyn Palmcr, GM (Jan. 13, 2005) rDOC ID 000001714013 ; GMNHTSA000200638l _ _ is thc fi rst Cobalt claim that was brought to GM Lcgal's allcntion. GM opened the matter in September 2005. Leller from Kristy Gibb, GM , 10 Douglas Brown, GM (Sept. 9, 2005) LDOC ID 000001660004; GMNHTSA000200670J. GM Legal had the engineering analysis completed by January 2006, in timc for a January 31 , 2006 SClllcmCI1\ roundtablc mccting. GM , Activity Notcs fo nn, Filc No. 50 1661 (Jan. 3 1, -H9 2006) IDOC ID 000001660023 ; GMNHTSAOO0200717j. 103 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product assistance of outside counsel experienced in airbag liti gation reported to colleagues that " the . have no so I·d · I exp Ianat10n . .,,420 engmeers I tee hmea I. Background on the GM Legal Department Before addre ssing these specific accidents, we provide some background on how the GM Legal Department (al so referred to as " GM Legal Staff' or " GM Legal") dealt with product liability cases, particularly when they intersected with safety issues. Thomas Gottschalk was General Coun sel ("GC" ) of GM from 1994 until August 2006. He was succeeded by Robert Osborne. 421 In July 2009, Michael Millikin succeeded Osborne as Gc. 42 2 During the relevant period, the General Counsel of GM North America reported to the GC. That post was held by Chris Johnson until October 2008, by Michael Robinson from October 2008 to September 2009, by Fred Fromm from 2009 to 2011 , and by Lucy Clark Dougherty from March 20 II to the present. Both the lawyers in charge of safety issues and the lawyers in charge of product litigation reporte d to the General Counsel ofGM North America. Two of Clark Dougherty 's direct reports are Bill Kemp and Larry Buonomo. 423 Kemp, Counsel for Global Engineering Orga nization, is widely regarded as GM ' s most knowledgeable, experienced, and trusted safety lawyer. 424 He has been at GM for decades. Buonomo's title is Practice Area Manager - Global Process & GM. Senlcmell\ Roundtable/Case Summary, D'/ale oj v. GM - PL: 511176 (Oct. 3, 2006) rOOC ID 000001714429; GMNHTSAOO02974361. 4·~1 Cotsiri los, Tighe & Streiekcr, Poulos & Campbell, Ltd . Interview of Roben OsbonlC, May 12, 2014. m J&B Interview of Michael Millikin, April 4, 201 4. m J&B Interview of Lucy Clark Dougherty, April 16, 2014. 424 J&B Interview of Lawrence Buonomo, April 16, 20 14. <120 104 Privileged and Confidenfia/ Protec1ed by AUorn ey~ Client Privilege and as Allorney Work Product Liti gat ion , where he oversees, among other things, product litigation cases. Before Buonomo, a GM lawyer named M ichael Gruskin oversaw product litigation cases.425 In the area of safety, the two function s can intersect. Indeed, Kemp and hi s colleague Deb N owak~ Vanderhoef sit on one of the two committe es that detennines whether and at what 426 price to settle product li ability lawsuits. A reason for that assignme nt is to ensure that infonnation from lawsuits finds its way into GM 's sa fety fun ction, that is, to the eng ineers who make safety dec isions. 42 7 Their presence is a practi ce spec ifically designed to defeat information silos , a nd to enSure important infonnati on makes its way to eng ineers. 428 Product litigati on staffattomeys have frontlin e responsibility for managing laws uits and claims that are not in litigation, called not~in~ s uit-matt e r s , or NISMs. When necessary to defend lawsuits or NISMs, the product litigation staff attorneys retain outside co unsel. Outside counsel is requi red to provide a written evaluation of all lawsuits and NISMs to GM soon after receiving a matter. The purpose of these "early" evaluations is to determine whether a case is a trial cand idate or a case th at should be settled. Outside counse l is expected to prov ide updated case evaluations at key poi nts througho ut the litigation - including after the case is assigned a tri al date . Two groups of non-l awyers dedi cated to assisting GM product li tigators and o utside counsel in defending product claims are Field Performance Assessment ("FPA") eng ineers and ES IS cla ims admini strators. FPA engineers are ass igned to gather information and assess 4·~) J& B J& B m J& B 428 J& B 4 26 Intcrview Intcrview Interview Interview ofLawrcnce Buonomo, April 16, 201 4. of William Kcmp, April 17, 2014; J&B Intervicw of Dcborah Nowak-Vandcrhocf, May 1, 2014. of ROllald Porter, April 30, 201 4; J&B Interview of JaclYIl Palmer, April 24, 2014. of Michael Gmskin, May 12, 2014. 105 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product technical issues in lawsuits and NISMs. 429 The FPA group is distinct from Product Investigations and the FPE process. 430 FPA engineers conduct their own techni cal assessments, which might include reviewing police reports and medical records, interviewing witnesses, inspecting vehicles, and anal yzing SDM data. 43l FPA engineers share thei r technical assessments with product litigation staff attorneys and outsi de counsel, assist in responding to plaintiffs' di scovery requests, and may testify as experts or 30(b)(6) w itnesses. FPA engineers' technical assessments are the lawyers' primary source of technical information for the earl y case evaluations, and are a critical factor in the eva luation of settlement dec isions. 432 GM ' s claims administrators are employees of a vendor called ESIS , work solel y on GM claims, and have offices at GM. 4B . ESIS employees are responsible for conducting field in vestigations and processing NISM claims aga in st GM .434 ESIS field investigators photograph the vehicle, conduct interviews, and, with penniss ion, download SDM and other diagnostic data out of a vehicle 's computer. 435 ESIS forwards thi s information to GM lega l staff. ES IS also negotiates settlements and presents claims at meetings for settlement evaluation. 436 ESIS acts at the direction of the GM product liti gation staff attorneys, and the claims administrators meet with the GM attorne ys on a weekly basis. GM has a structured settlement process by which GM decides whether and for how much a case should be settled. At the lowest monetary level, product litigation staff attorneys were J&B Inlerview of Jennifer Sevigny, Ma rch 27, 2014; J&B Imervicw of Douglas Brown, March 19,20 14; J&B Intervicw of Keith Schultz, March \8, 201 4; J&B Interview of Katby Anderson, March 19-20,2014. 430 J&B Interview of Keith Schultz, March \8, 201 4. 431 J&B Interview of Douglas Brown, March 19, 2014. m J&B Interview of Kathy Anderson, March 19-20, 2014; J&B Intcrview of Douglas Brown, March 19, 2014. 4JJ J&B Interview of Kathy Anderson, March 19-20,2014. 434 J&B Interview of Dan Derriek , March 31 , 2014; J&B Interview of Annclte Rigdon, May 14, 2014. 435 J&B Interview of Ryan Jahr, April 1, 201 4; J&B Intervi ew of Jaclyn Palmer, May 8, 2014. 436 J&B Interview of Annette Rigdon, May 14, 2014; J&B Interview of J,uneia Pricc, May 12, 2014. 429 106 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product vested with $ 100,000 in settlement authority.437 Settlements of between $ 100,000 and $ 1.5 million (a limit whi ch was eventually increased to $2 milli on) required approval at a committee known as the " Roundtable." The Roundtable Committee met weekly, was led by the Liti gation Practice Area Manager, and all product litigation staff attorneys were invited to attend. Settlement offers between $2 and $5 million required approva l ofa group ca ll ed the Settlement Review Committee ("SRC"), wh ich met monthl y, and was chai red by the head of global litigation. Members of the SRC included both th e GC ofGM North America and Kemp .438 Cases over $5 million req uire approva l by the General Co unsel. When a case was before the Roundtable or the SRC, the responsible product liti gation 4V staff attorney would present hislhe r case. ) At the end ofa presentation to either group, the Roundtable and SRC committees voted. 44o The chair was the ultimate decision maker. 441 Mi chael Gruskin chaired both the Roundtable and SRC committees fro m August 2007 to March 2012. Larry Buonomo chaired both committees beginning on March 8, 2012. In terms of case volume, the SRC considered an average of 1.4 cases per meeting in 2012 and 1.3 per meeting in 20 13. The Roundtables considered an average of 3.4 cases per meeting in 2012 and 3.76 per meeting in 20 13. Michael Gmskin, GM, Legal Staff Settlcment Review Proccsses, March 23, 2009 LDOC lD 0 _00016584 J. The current range for the SRC is $2 million to $5 million; earlier it was $ 1.5 million 10 $5 million. If there were no matters to consider, the meetings were cancellcd. Michael Gmskin, GM, Legal StaffSeulement Review Processes, March 23, 2009 rDOC ID 0 00016584]. 439 J&B Interview of Lawrence Buon o l~o, April 16, 201 4. ·1.10 J&B Interview of Lawrence Buonomo, April 16, 20 14. 4~1 J&B Interview of Lawrence Buonomo, April 16, 20 14. 437 438 107 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product The main functi on of the Roundtable was to generate better claim eva luations and settl ement forecasts through a " peer review" stmcture_ 442 But a number ofGM lawyers reported that it had a second function as well: to spot trends indicating safety issues. Product liti gation attorney Ron Porter stated that it was well recognized that a goal of the Roundtable and SRC process was to identify potential safety issues and refer them to engineers.44 3 Hi s product litigation colleague Jacl yn Palmer sa id that attorneys discussed potential product safety or accident trends at Roundtables on occasion, and the Roundtable Com mittee referred issues to GM engineers for a follow~up investigation. 444 Not all GM lawyers, however, agreed with this view. Buonomo, for example, said that it was not the Roundtable 's function to spot trends and that if a lawyer had to flag a trend, then the system had already fai led.4 4.'i To put thi s issue in a specific context, at a Roundtable di scuss ion on the Cobalt airbag non~deplo yment issue in 2012, a junior lawyer reca ll ed asking whether there should be a recall .446 He was told that the issue had already been rai sed with engineering, that th e engineers were working on it, and that they had not come up with a so lution. This lawyer got the " vibe" that the lawyers had "done everything we can do. " We have discovered no forma l written policies governing how the settlement committees should handle safety issues. Multiple witnesses, however, described how things worked in practice. In practi ce, the mechanism for getting safety issues from GM Legal to GM Engineering was through Bill Kemp. For example, fonner GC Thomas Gottschalk described Michael Daar, GM , el a/. , Outside Counsel Guidelines Sub Group - Early Case Evaluation and Peer Review Process, Feb. 27, 2009 [DOC ID 0000057 87998]. 4H J&B Interview or Ronald Poner, Apri l 30, 201 4. 4H J&B Interview or Jaclyn Palmer, April 24, 201 4. ·1~5 J&B Interview or Lawrence Buonomo, April 16, 20 14; J&B Interview or Lawrence Buonomo, M.1Y 13, 2014. 4~6 J&B Interview or Nabeel Peracha, May 21 , 201 4. ·142 108 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product how the primary responsibi lity for spotting trends belonged to engineering, which had access to a host of different inputs ranging from warranty claims to customer complaint"_ But, said Gottsc halk, when lawyers did become aware of a safety issue, they were to raise the issue, and a common way of doing that was to go to Kemp or hi s colleague Deb Nowak~Vanderhoef - who . ... h . 447 wou Id , m turn, raise It wit engmeers. GM 's expectations for its lawyers - including the expectation that lawyers elevate certain issues to superiors - were reflected in a 2003 memorandum from then~GC Gottschalk to "All Attorneys.'.448 Gottschalk reminded the lawyers who worked for him , " If you as an attorney are aware of any threatened, on-going, or past vio lation of a federal, state or local law or regulation ... it is your responsibility to respond appropriately .,.449 Gottsc halk discussed factors in determining how to respond, including the seriousness of the matter, the existence of threat of harm to others, the lawyer's degree of knowledge of the situation, and the lawyer's level of experience and position within GM. Gottschalk wrote that while some matters could be addressed directly by the attorney, in other cases "it may be more appropriate to report the situation to your supervisor .. _so that it can be handled at a higher level." Gottschalk also di scussed what to do if one's superiors had concluded that appropriate action had been taken in response to a perceived problem, but the more junior lawyer disagreed: Tfyou believe ... that the conclusion is wrong, you should continue to seek an appropriate resolution. It is your duty to bring the situation to the attention of 44 7 Cotsiriios, Tighe & Streicker, Ltd. Interview of Thomas GOllschaIk, May 12, 2014. 4~~ Thomas A. GOllschalk, Memorandum re Standards of Professiona l Conduct of Auomeys (Aug. 4, 2003), at 1 [DOC 1000136806990]. Thomas A. GOllschalk, Memorandum re Standards of Professional Conduct of Attomeys (Aug. 4, 2003), at 1 [DOC 10 00 136806990j. ·1.19 109 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product yo ur supervisors or their supervi sors, as necessary. If yo u believe that they have not addressed it appropriately or if you feel that bringing it to their attention would be futile, you should pursue it higher in the organization - if necessary, to me as General Counsel. Michael Gruskin , a long-time GM attorney who headed the product litigation team during part of the relevant period, understood Gottschalk's direction to apply to safety issues. 45o In interviews, Osborne and Mi llikin, like Gottschalk, said that they expected their lawyers to elevate to them significant safety issues.45 \ 2. The . 2006.4 52 January 2006: Roundtable matter was reviewed by GM lawyers at a Roundtable meeting on January 3 1, was not wearing a seatbelt when she lost _ control of her 2005 Chevy Cobalt on a residential street on July 29,200 5. Her car struck several small trees and a larger tree ; paramedics extricated her from the vehicle, but she di ed en route to the ho spital.45 3 The SDM recorded the vehicle 's power mode status as "Accessory" at the time of the crash. 454 FPA engineers Manuel Peace and Kathy Anderson and Doug Brown of the GM Legal Staff were ass igned to the matter. 455 Records show that the allegation was that the airbag did not deploy when it should have. Anderson, however, opined that the airbag was not expected to deploy because of the nature of the crash. First, it was th e car's front right corner that made contact with the tree, rather than the J&B Imcrvicw of Michael Gm ~ kin , May 28, 2014 J&B Interview of Michael Millikin, May 29, 201 4; Interview of Robert Osbome. May 7. 2014. 4 52 GM. Activity Notes fom1, File No. 50 1661 , Jan. 3 1, 2006 LDOC lD 00000 1660023 ; GMNHTSA000200717J. m Cal span Corporation, Calspan On-Site Air Bag Non-Dcp loymcnt Invcstigation, Case No. CA05-049, Dec. 12, 2006 rDOC ID GMCB-000000073786; GMHEC10OO26303]; GM, Activity Notes fonn, File No. 501661 , Jan. 31, 2006 LDOC ID 000001GG0023 ; GMNHTSAOO02007l7J. 4 54 Crash Data Retrieval Systcm, . SDM Data, Sept. 14, 2005 rDOC ID 00000 16600 II ; 450 451 GMNHTSA0002006881. 4 55 GM. Activity Notes fom1, File No. 50 1661 , Jan. 3 1, 2006 LDOC lD 00000 1660023; GMNHTSA0002007 I 7]. 110 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product car colliding head on. 456 Second, the car had slowed gradually because it hit several small trees before com ing to a stop, and one would not expect a frontal airbag deployment in those circumstances. 457 The matter was presented to the GM Roundtable.4~8 Anderson stated because the airbag was not expected to deploy, the fa ct that the vehicl e's 4s9 power mode was in Accessory was not a focu s of her in vestigation. The case was settled. 3. September 2006: Six months after th e . Evaluation Roundtab le, GM Lega l Staff recei ved a written bri efin g about anothe r non-deployment case on September 7, 2006.460 On February 26, 2006, was injured when a drunk dri ver ran a red light and hi t _ 2004 Saturn Ion on the 461 Because this was a side impact, the GM engineers concl uded that the dri ver's airbag side. should not have dep loyed. 462 Unlike the Cobalt, the SDM for the Ion was not designed to and did not record ignitio n status at the time of a crash. Thi s case was stayed at the time of bankruptcy in 2009 and did not settle. J&B Interview of Kathy Anderson, Mareh 19-20, 201 4. J&B Illl erview of Kat hy An der~on , March 19-20, 2014 4 $8 J& B Interview of Ka thy Anderson, March 20 14. 4 59 J& B Interview of Ka thy Anderson, March ·160 Bowman and Brooke LLP, Case tVlII"",,,on: 2006) [DOC ID 00020668611 0]. 4~1 Bowman and Brooke LLP, Case Evaluation: 2006) IDOe ID 00020668611 OJ. ·162 Bowman and Brooke LLP, Case Evalu ation: 2006) IDOe ID 00020668611 OJ. 456 4S7 III v. General MOlars and (Sept. 7, v. General M Olars and (Sept. 7, v. General Molars alld (Sept. 7, Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product 4. September 2006: Evaluation died after her 2004 Saturn Ion left the road at high speed on July 4, 463 2004, went over a low curb, braked, and then struck a large utility pole head on. did not deploy, and although she had been wearing her seatbelt, _ The airbag was found dead at the 464 scene. On September 25, 2006, the Dykema law firm submitted a case evaluation to GM for the _ case.46S The assigned FPAs were the same as in the _ fata lity, Manuel Peace and Kathy Anderson. They identifi ed thi s crash as one in which there should have been an airbag deployment, and that the dep loyment likely wou ld have saved _ life. The Ion SDM was not designed to capture the ignition switch position at the time of the crash (as the Cobalt's SDM did). It was, however, expected to record the most recent " near deploy event.'.466 But it failed to make such a recording, and the FPA engineers were unable to develop an explanation for this failure. 467 Anderson considered the hypothesis that the vehic le lost power early in the crash event, but that idea was di scounted be cause the SDM should have had power for 150 ms after the vehicl e lost power and thus shou ld have recorded the evel1t. 468 There is no evidence that Peace, Anderson , in ~ hou se counsel, or outside counse l were aware of the prior PRTS reports concerning the Ign ition Switch that the Cobalt and Ion shared, Dykema Gossett LLP, Fi le No. 51176 (Sept. 25, 464 Dykema Gossett LLP, File No. 51 176 (Sepl. 25, 465 Dykema Gossett LLP, File No. 51176 (Sept. 25, ·166 Dykema Gossett LLP, File No. 51176 (Sept. 25, 41>1 Dykema Gossett LLP, File No. 51176 (Sept. 25, ·168 Dykema Gossett LLP, File No. 51176 (Sept. 25, 463 Early Evaluatioll Report: tstate oj v. GM and Satllrn 2006), at I IDOC 10000001714421 . Early Evaluation Report : Estale oj 1'. GM and SOIlIm 2006), m 1 [DOC 10 00000171 442- . Early Evaluation Report : Esta/e oj II. GM and Satllrn 2006), at 1 [DOC 10 000001714421; MNHT AOO02007 17J. Early Evaluation Report: Es/Ole oj _ _ 1'. GM alld Sa/11m 2006), at 1 rDOC ID OOOOOI 71442~2007 1 71 Early Evaluation Report: Estale oj _ _ 1'. GM alld Sa/11m 2006), at 6 [DOC lD OOOOOI71442~2007 1 7J. Early Evaluation Report : Esta/e oj _ _ II. GM and Satllrn 2006), at 6 [DOC lD OOOOOI71442~2007 1 7J. 112 Corporation, GM Corporatioll, GM Corpora/ion, GM Corporalioll, GM Corpora/ioll, GM Corpora/ion, GM Privileged and Confidenfia/ Protec1ed by AUorn ey~ Client Privilege and as Allorney Work Product or the 2005 TSB discussing how a dri ver could inadvertently turn off the Ignition Switch with a knee. Peace, one of the assigned FPA engineers, had no reco llecti on of the _ case when interviewed.469 He did not recall ever seeing the 2005 TSB, and could not opine on whether it would have made a difference to his evaluation of the case.470 GM sta ff attomey Doug Brown presented the _ case to a Roundtable meeting about a week afte r receiving Dykema 's evaluati on, on October 3, 2006.471 He reported that: Despite extensive analysis, the engineers have no solid technical explanati on. The engineers agree that I) the airbags ... should have deployed; 2) the SDM did not record th e crash event, for unknown reasons; ... and 4~ so n ab l y likely that deployment of the driver airbag would have prevented _ death in thi s 472 accident. Brown also pointed out how this n on~depl oyment would be "problematic" gi ven a standard GM test: " [T]he vehicl e damage is remarkably simi lar to the test vehi cle damage in a GM pole test where the bags deployed. Discovery of thi s test will also be probl ematic for GM's defense.,.473 The Roundtable granted settlement authority, and GM settled the case.474 5. October 2006: C rash On October 24, 2006, a crash occ urred in which a 2005 Cobalt left the road and struck a telephone box and two trees.475 were killed; the driver, J& B Interview of Manuei Peace, May 8, 2014. J& B Interview of Manuel Peace, May 8, 201 4. 4 71 Douglas Brown, GM, Settlement Roundtable/Case Summary, Estate o~ II. GM - PL: 511176 (OCI. 3, 2006) rOOC ID 00000171 4429; GMNHTSAOOO297436l 472 Douglas Brown, GM. SClllcmcnI RoundlablcJCase Summary, Estate oj II. GM - PL: 511176. (Oct. 3, 2006), at 2lDOC ID 00000171 4429; GMNHTSAOO0297436j. 473 Douglas Brown, GM, Settlement Roundtable/Case Summary, Estate oj II. GM - PL: 51 1176, (OCI. 3, 2006), at 2 rODe ID 0000017144 29; GMNHTSAOO0297436l <174 Douglas Brown, GM, SClIlcmcnt Roundtablc!Case Summary, Esfate oJ~ II. GM - PL: 511176, (Oct. 3, 2006), at 2lDOC ID 00000171 4429; GMNHTSAOO0297436J; e-I~Brown to Tim G. Jaeger (Nov. 21 , 2006) LDOC ID 000001714470; GMNHTSA000297475j. m Tom Ellingsworth, GM, Claim form: File # 624620, _ et 01. [sic] (Feb. 28, 2007) [DOC ID 000003680368; GMNHTSA000284395J. 469 4 70 113 Privileged and Confidenfia/ Protec1ed by AUorn ey~ Client Privilege and as Allorney Work Product was severely injured. _ was sitting in the front passenger seat. She was not wearing her seatbelt, and the airbag did not deploy. Thi s crash first cam e to GM's attenti on on November 15,2006, through a TV reporter 's inqui ry. Alan Adler e-mai led Dwayne Davidson, Senior Manager for TREAD Reporting at GM , and others, copying Gay Kent, Jaclyn Palme r, Brian Everest, and Doug Wachtel, with the subject line "2005 Cobalt Air Bags - Fatal Crash; All eged Non-Deployment," asking whether anyone knew about the acc ident and other airbag no n~d e ploy ment incidents invo lving the Cobalt. Adler noted: «Reporter asking for response from GM by the end of day Wednesday on what we know about air bag iss ues in ' 05 Cobalt.'>476 Several of the recipients responded to the e-mai l and prov ided avai lable data on Cobalt fronta l airbag c1 aims.471 It is unclear how GM responded to the reporter, but the story ultimately ran, and Adler did not think additional communi cati on with the station was necessary .478 Adler does not remember any response, and none has been located in a search of his fil es. The day after the _ crash - the timing was coincidental - GM issued an updated Technical Service Bulletin on " Inadvertent Turn ing of Key Cy linder, Loss of Electrical System and No DTCs" in the Cobalt and other vehicl es. The updated Bulletin extended the December E~ mail from Alan Ad ler, GM , to Keith Schultz, GM, el al. (Nov. 15, 2006) [DOC ID 000049639345; GMNHTSA000285 148 J. 477 Specifically, Dwayne Davidson conducted a TREA D database search Ihal yielded over 700 records of field repons and complai nts, which he offered to summarize. E-ma il from Dwayne Davidson, GM , to Ala n Adler, GM, el al. (Nov. 15.2006) [DOC ID 000049759326; GMNHTSA000276789]. DOllg Wachtel reviewed exi~t ing fie ld actions involving the Cobalt and recommended that GM acquire the EDR data. E-mail from Douglas Wachtel, GM , to Alan Adler, GM, el af. (Nov. 15, 2006) [DOC ID 000049759327; GMN HTSAOO0276790J. Jaelyn Pa lmer, a GM airbag lawyer, passed Adler's meS5<1ge along to Doug Brown, another GM airbag lawyer, so thm he eOlild be prepared for any potential claims related to the crash. E-mai l from Jaclyn C. Palmer, GM, to Alan Adler, GM, el al. (Nov. 15, 2006) [DOC ID 000049759330 ; GMN HTSA000276791]. Chris Janik provided a summary of the two Cobalt frontal airbag elaims in the NHTSA database. E-mail from Christopher Janik, GM, to Douglas Wachtel, GM , el af. (Nov. 15, 2006) I DOC ID 000049759328; GMNHTSA000285149-50j. 4 78 E-mail fromAlan Ad ler. GM . to Douglas Brown. GM. el af. (Nov. 20, 2006) [DOC ID 000009\09885]. 476 114 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product 2005 Bulletin to include Model Year 2007 for all the models previous ly included, and al so added the 2007 Saturn Sky 479 Otherwise, the October 25, 2006 Bulletin was the same as the one issued in December 2005.480 Like the 2005 Bulletin, thi s version omitted the word "stall." B. 2007: Outsiders Find the Solution But GM Remains Unaware In 2007, two analyses of the _ crash, one by Wi scons in State Trooper Keith Young and another by Indi ana University researchers, clearly stated that the movement of the ignition switch from Run into Accessory may have caused the airbag non~deplo ym ent - the correct answer that GM engineers did not accept until January 20 14. Although Trooper Young's report was in GM's legal department fil es as of February 2007, GM lawyers and engineers working on airbag n o n~dep l oy m ent cases did not learn of its existence until a few months ago. They did not learn about the Indiana University study until 20 12 when a plaintiff's ex pert brought it to th eir attention. They were also unaware ofGM's 2005 and 2006 TSBs on the ignition sw itch issue - which, aside from being in GM ' s own files, were publicl y ava ilable on the NHTSA website, where Trooper Young found them - and the PRTSs. GM, Technical Scrvice Bulletin, "Infonnat ion on Inadvenent Turning of Key Cylinder, Loss of Electrical System and No DTCs" (Oct. 25, 2006) [DOC ID 000001771028; GMHEC000138614l ·180 GM, Service Bulletin " Infonnation on Inadvenent Turning of Key Cylinder, Loss of Electrical System and No DTCs" (Dec. 2005) [DOC ID 14021301709348; GMHEC000329773] ; GM, Technical Service Bulletin "Infonn ation on Inadvertent Turning of Key Cylinder, Loss of Eleclrical System and No DTCs" (Oct. 25, 2006) I DOC I D 000001771028; GMHECOOOI38614 1. Some copies of the October 25, 2006 Teclmical Service Bulletin (05-02- 35 -007A) were printed wi th a July 1, 2011 modification date. The Technical Service Bulletin, however, was not actually updated and/or published in July 2011 (or anytime after October 25, 2006). The July 1, 20 11 modification date was the result ofGM ' s June 2011 migration of data in the Electronic Service Infonnation database (which maintained the Technical Service Bl111 el in ~) from anlngre~ data ba~e 10 an Oracle database. As pari oflhe migration, GM transferred data regarding the "owner" of Technical Service Bulletins. Because some "owners" of the Technical Service Bu lletins had le n the company, including Ihe owner of the October 25 , 2006 Technical Service Bulletin, GM assigned a new employee 10 serve as the "owner" of the Technical Serve Bulletins as part of the migration. This change automat ically resulted in a ncw modification date of July 1, 2011 for Tec hnical Service Bulletins with new owners. Technical Service Bulletins with a new owner printed after July 2011 therefore renected a modification date of July 1, 20 II. GM fixed this problem in March 20 14 when it migrate£! data to the GM Data Center in Warren, Michigan. Accordingly, Technical Service Bulletins printed after March 2014 should renect the correct modification date(s). -179 11 5 Privileged and Confidenfia/ Protec1ed by A Uorn ey~ Client Privilege and as Allorney Work Product I. ~0 07: Trooper Young's Reco nstru ction Report of Cras h _ The Wisconsin State Patrol (Wi sconsin's state police force) issued a "Collision Analysis & Reconstructi on Report" ("the Wisconsin State Patrol report") on February 14, 2007, about the _ crash. INVEST IGATIVE SUMMARY The following :5 lJ1tt~~ ;11111: opinkms and /nt-renee:! of this IllUthor. Tht yare ba.sed upon the Inform.tion revltwed to date. The.s • .stetemen~ are ""untte to" reMon.bl. degree of a eientifle et iUlnty: • None ofttie cx:ct4)8Ilts were wealfng their saraty bens at the time orthe crash. The two lTonl seat allbllgs did nol deploy. II appears lhat the Ignition S'MIch had somehow been turned from the IU"I pos/IIon 10 accessory prior to the .colisbn with the ~ ~t!ili~ "" . IIII!~IIII.I!II!~ drivelS' ifceme • stabJs was a va rld inslrucOOn perm[t _ was violating the restrictions on her inslrudion permit by operallngo a vehicle v.fthout a quarried licensed drt/er In 1M front seal The front tires cl the vehicle did not have the requt ed mln1rrum legal lire tread depth of 2132 Inch in two or more places on each front tire. The speed calculated fmm the vault at the driveway of 48 mph ard Ihe crosh speed at the trees of 38 mph do not malch !he iriOlmatioo reeoroed in Ihe five n eonds of pre-<:rash data on the SDM; thrs may be due to power loss. 1M Authored by Trooper Keith Young of the Techn ical Reconstru ction Unit, the report set out why Trooper Yo ung believed the airbags did not deploy. He wrote: • "The ignition switch on the . . . vehi cl e appears to have been in the accesso~ pos iti on when it impacted the trees preventing the airbags from deploying.' 81 He noted th at a search of the NHTSA website revealed fi ve complaints o f 2005 Cobalts turning off while being dri ven, three of which «talk about th e knee or leg touching the igniti on or key chain causin g the engine to turn Off. ,,482 • Trooper Young then cited the October 2006 technical service bulletin whi ch "discusses the potential for the driver to inadvertently turn off the ignition due to Keirh A. Young, Technical Reconstruction Unit, Wisconsin State Patrol Acadct~ Analysis & Reconstruction Report, Feb. 14, 2007, attached to GM Claim fonn (File # 624620, _ _ et al [sic]; Feb. 28, 2007), at 91 DOC ID 000003680368; GMN HTSAOO0284395J. 482 Keir h A. Young, Technica l Reconstruction Unit, Wisconsi n State Parrol Acadet~ Analysis & Reconstruction Report, Feb. 14, 2007, attached to GM Claim fonn (File # 624620, _ _ et at [sic l; Feb. 28, 2007), at 91 DOC ID 000003680368; GMN HTSAOO0284395j. ·181 11 6 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product low key cylinder torque/effort.,ri83 Therefore, he wrote: "It appears li kely that the vehicles' [sic] key turned to Accessory as a result of the low key cylinder torque/effort." • In hi s investigative summary, whi ch he sa id was "accurate to a reasonable degree of sc ientifi c certainty," he wrote: "The two front seat a irbags did not deploy. It appears that the ignition switch had somehow been turned from the run position to accessory prior to the colli sion with the trees." On November 15, 2006, the ES IS opened a "rumor fi le" on the _ c rash - a fi le type opened for the purpose of tracking incidents that are not in active litigation and for whi ch no claim against GM has been made. 4 84 In February 2007, ES IS Claims Administrator Kri sty Gibb received a copy of the report. 48S Forensic evidence shows that Gibb saved the report in GM Legal's electronic fi les on March 2, 2007. 486 The forensic evidence furt her shows that prior to 20 14, the electronic copy of th e report was not accessed by anyone in GM Legal, with the exception ofa GM legal assistant whose responsibiliti es included locating claim documents fo r prod uction to N HTSA in co nnection with GM ' s TREAD Act reporting duties and who accessed the document in May 2007. 487 Dwayne Davidson, Senior Manager for TREAD reporting at GM , stated that he obtained a copy of Trooper Young's report contemporaneously in the 2007 timeframe, fro m someone a.t GM Legal and then provided it to NHTSA in connection with GM ' s quarterly death and injury report in Keith A. Young, Technical Reconstruction Unit, Wisconsin Slate Patrol Acadet~ Analysis .& att,tched 10 GM Clai m fonll (File # 624620, _ _ et at lsiel; Feb. 28, Reconstruction Report, Feb. 14, 483 2007), at 9 rODe ID 000003680368; GMNHTS!\OOIJ28'!39:;]. GM, Data Sheet rc _ _ Crash ID 000001714569]. Tom Ellingsworth,~ form: Lsic] (Feb. 28, 2007) LDOC ID 000003680368 ; GMN HTSA000284395 J; GM, Data Crash (saved March 2, 2007), at 17 LDOC ID 0000017 145691·186 J&B Intervicw ofMing-Liang Chi, May 14, 2014. 4~1 Keith A. Young, Technica l Reconstruction Unit, Wisconsi n State Patrol Acadet~ Analysis & Reconstruction Report, Feb. 14, 2007, attached to GM Claim fonn (File # 624620, _ _ et at Lsie]; Feb. 28, 2007), at 9 1DOC ID 000003680368; GMNHTSAOOO284395 1; Kimberly M. Nothnagel, Resume (saved Sept. 4, 484 485 2009) [DOC ID 000078367937]. 117 Privileged and Confidenfia/ Protec1ed by AUorn ey~ Client Privilege and as Allorney Work Product 2007.488 However, none of the GM lawyers and engineers working on Cobalt matters recall being aware of thi s report until 2014 2. March 2007: NHTSA Expressed Interest in Cobalt Non-Deployments NHTS A directly expressed an interest in Cobalt airbag n o n~d epl oy m en ts to G M on March 29, 2007, when a group of G M engineers, including Gay Kent and Brian Eve rest, attended a Quarterl y Review meeting at NHTSA' s headquarters. 489 During the course of that meeti ng, or during a break, NHTSA offi cials told the GM representatives th at they had observed a number of air bag non~deployments in Cobalt and Ion ve hicles. 49o According to GM witnesses, NHTSA made no formal request and did not ask GM to report bac k to it about the nondepl oyment issue ,491 There is al so no evidence that NHTSA indicated that it viewed the airbag n o n~de plo yme nt as being caused by anything rel ated to the Igniti on Sw itch. We have discovered no document refl ecting a furth er request or foll ow ~up from NHTSA, a nd witnesses could recall no suc h follo w ~up from NHTSA. According to Everest, when the GM personnel returned to Detroit, Kei th Schultz, then Manager of Internal Investi gations in Product ln vesti gations, directed that Everest and John Sprague, an FPA airbag engineer, to compil e infonnation on Cobalt and Ion N1SMs and lawsuits, and asked Dwayne Dav idson to pull the T READ data for similar instances. 492 As a result of thi s instruction, Sprague began compiling an Excel spreadsheet listing the vari ous J& B Intcrview ofDwaync Davidson, May 6, 201 4. E-mail from Doll gl a~Wac h tel . GM . t oC hri ~to ph erJ an ik . GM . el al. (March 27, 2007) [DOC ID 000001597661 ; GMNHTSA000002849J; J&B Interview o f Brian Everest, March 27, 2014. At this point in time, GM held quarterly meetings with NHTSA w here they would have roundtable discussions in which they worked through trends that NHTSA was seeing in G M vehicles and reviewed opcn investigations. 1&B Intcrview of Kcith Schultz, Apri l 24, 20 14; J& B Intcrview of Gary Dowd, May 8, 20 14. 490 J& B Intcrview of Brian Evcrest, March 27, 201 4. 491 J& B Intcrview of Gay Kent, March 18, 2014; J&B Interview orMan krinsky, April 16,2014. 492 J& B Interview of Brian Everest, March 27, 201 4. For a description of the TREAD database, see ;lIfra Appendix E. ·188 ·189 11 8 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product Cobalt accidents and non_deployments. 493 The purpose of Sprague's tracking spreadsheet was to look for trends 4 94 He does not reca ll being given a deadline or timetabl e, or a specifi c deli verable.495 No one specifically asked him to track Cobalt non-deployments.4 % He generall y remembers sharing hi s spreadsheet with Everest, likely upon Everest's request. 497 He does not remember sharing the spreadsheet at any fomlal meeting.498 Witnesses have inconsistent recoll ections as to whether the Product Investigations group became invo lved in the Coba lt airbag non-dep loyment issues at thi s stage. Everest reports that, in Apri l 2007, the FPA group transitioned the Cobalt airbag matter to the PI group, where it was taken on by an engineer named Eri c Buddrius.499 Docume nts in Buddrius's files indicate he was working on the issue, and a May 4, 2007 Investigation Status Review (" ISR") Presentation Planning Worksheet states that Buddri us was scheduled to present on an issue described as "Cobalt/Ion Airbag (NHTSA discussion item).,,500 Buddrius had no recollection of J&B Interview of Brian Everest, Marc h 27, 201 4; J&8 Interview of John Sprague, March 26, 2014; J&6 Interview of John Sprague, May 9, 2014; J&B Interview of John Sprague, May 15, 2014; J&B Interview of John Sprague, May 23, 2014. m J&8 Interview of John Sprague, March 26, 201 4; J&B Interview of John Sprague, May 9, 20 14; 1&B Interview of John Sprague, May 15, 20 146; J&B Interview of John Sprague, May 23, 2014. 495 J&8 Interview of John Sprague, May 23, 2014. 496 J&8 Interview of John Sprague, March 26, 201 4; J&B Interview of John Sprague, May 15, 2014. 497 J&B Interview of John Sprague, May 15, 201 4. 498 J&B Interview of John Sprague, May 15, 2014. Sprague SC I up a system whereby Annette Rigdon (an ES IS claims administrator) would fo rward him - and not ot her FPA engineers - any claims related to non-deployments, and [hose claims would fonnally be assigned [0 Sprague 499 J&B Interview of Brian Everest, March 27, 201 4. During a late r interview, Everest said Buddrius may not have been fonnally assigned to investigate the issue. J&B Interview of Brian Everest, April 29, 2014. 500 E-mai l from Douglas Wachtel, GM, to Elizabeth A. Bardowcl l, GM (May 1, 2007) mOC ID 000001596837; GMNHTSA2584851. The e-mai l and auached agenda state that Buddrius was scheduled to present "Cobalt/lon Airbag (NHTSA discussion itcm)"); e-mail from Keith Schultz, GM , to Brian Everest and John Spraguc, GM (May 3,2007) rDOC ID 000049652057; GMNHTSAOO02730851 ("We arc planning to have a brief discussion on the Cobalt/Ion Air Bag non-deployment issue tomorrow as part of our bi-weekly Investigation Status Review. You are both weleome to join us for this discussion ... in fac t it may be helpful if at least aile of you can .... "). 493 11 9 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product in volvement. SOI 3. In Ma r ch-Ap ril 2007: T he TSB Is Edited h ut the Edits A re Not Released March~April 2007 , GM 's technical bu lletin group proposed publishing a revised version of the TSS that would change the subject line to include the word "stalls" - the very term deleted from the December 2005 TSs. '02 The title proposed was: " Information on lnadvertent Turning of Key Cylinder, Loss of Electrical System, Hesitation, Stalls and No DTCs Set."S03 Produc t Investigations rejected the proposed revision on April 10, 2007.~04 However, on April 24,2007 , Doug Wachtel (Senior Manager - Intemal lnvestigation , Product Investi gations) provided hi s approval to "go ahead and add the word 'stall ' to the symptoms section o f the bulletin."sos One of Wachtel's direct reports indicated to Wachtel and others in an e~mai l that the change to the bulletin was proceeding, but that he was awaiting warranty data and infonnation on buybacks. so6 Wachtel later forwarded thi s e~mai l chain to Gay Kent. so7 Despite Wachtel's approval, GM has no record of pub li cation of the 2007 TS8 . No witness has been able to explain why the revised TSS was not publi shed, and it appears that GM personnel mi ssed an opportunity to improve the information it provided to dealers. 50 1 J&B Interview of Eric Buddrius, April 29, 2014. Stouffer told Jenner & Block that he went to Buddrius in 2011 for infonnation on the investigation and Buddrius had no file and no recollection of any infonnation from it. J&B Interview of Brian $toulTer, May 9, 201 4. 502 GM. Bulletin Tracking System - Corporate Review, for Bulletin Number 05-02-35-007 (April 10, 2(07) [DOC ID 000126448004; GMNHTSA0003162931. 503 GM. Bulletin Tracking System - Corporate Review, for Bulletin Number 05-02-35-007 (April 10, 2(07) [DOC ID 000 126448004; GMN HTSA0003 I 6293] S04 GM. Bulletin Tracking System - Corporate Review, for Bulletin Number 05-02-35-007 (April 10, 2(07) [DOC ID 000126448004; GMN HTSA000316293 J. 505 E-mai l from Douglas Wachtcl, GM, to Dan Fernandez, GM , el al. (April 24, 2007) [DOC ID 000001600410; GMNHTSA0002190061. :>U6 E-mai l from Mickcy Sabol, GM ,to Dan Fernandez, GM , el 01. (May 15, 2007) rDOe ID 000001600410; GMNHTSA0002190061. 507 E-mail from Douglas Wachtel, GM, toGay Kent, GM (May 15 , 2007) [DOC ID 000001600410; GMNHTSA0002 I 9006 J. 120 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product 4. April 2007: The Indiana University Report on _ Accident Also in 2007, the Indiana Univers ity (" IV") Tran sportation Research Center issued a report ("the Indiana University study") commissioned by NHTSA on April 25 titled "On-Site Air Bag Non~ Deployment Investi gation" for t h e _ crash, the same fata l accident that was the subject of the report by the Wi sconsin State Patrol..508 Although at least one person was able to find the IV study on NHTSA's website,509 GM personnel did not. Moreover, GM personnel made no efforts to systematicall y monitor NH TSA ' s website for such information and therefore did not obtain it until years later. More broadl y, we have seen no evidence that anyone from GM asked for or sought out publicl y available informati on that would have been rele vant to the investigation of the Cobalt.510 Indeed, many GM employees involved with the Cobalt investigation became aware of these public ly available reports only recentl y.511 We are unaware of any formal system at GM for seeking out or tracki ng publicl y avai labl e information rel ated to crashes or investi gations done by others of crashes involving GM vehicles (other than the work GM does to compl y with its TREAD obli gations). 512 GM did become awa re of some public cases through the informal Indiana Univcrsity Transportlltion Rcscarch Center, On-Site Air Bag Non-Deployment Invcstigation (April 25, 2007) [DOC ID 000001791070; GMNHTSAOOO223985 j; Keith A. Young, Technical Reconstruction Unit, Wisconsin State Patrol Aca~o n Analysis & Reconstruction Report, Feb. 14, 2007, attached to GM Claim fo nn (File # 624620, _ _ et al [sicj; Feb. 28, 2007) [DOC ID 000003680368 ; 50S GMNHTSA0002843951. As describcd infra, in Junc 2012, thc plaintirr s cxpcn in an airbag non-deployment case indieated that she found thc Indiana Univcrsity study on the NHTSA websiTc. We have not hecn able to con finn if the repon was accessible on NHTSA 's website in 2007. SIO J&B Interview of CarnIen Benavides, May 19, 201 4; J&B Interview of Lawrence Buonomo, May 13, 201 4; J&B Interview of Elizabcth Zatina, May 14, 2014. SII J&B Intcrview ofDwaync Davidson, May 6, 2014; J&B Intervicw of Eric Buddrius, April 29, 20 14; J&B Interview of Deborah Nowak-Vanderhoef, May 1, 2014; J&B Interview of Dale Furney, May 6, 2014; J&B Interview of Mark Johnson, May 7, 2014. SI2 J&B Interview of CarnIen Benavides, May 19, 201 4; J&B Interview of Lawrence Buonomo, May 13 , 201 4; J&B Interview of Keith Schultz, March 18, 201 4 ; J&B Interview o f Mickey Sabol, April 24, 2014. 5(}9 12 1 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product "rumor" tracking system by which ES IS in vestigators or other GM Legal Staff wou ld start fi les on cases that were not form all y involved in li tigation but potentially could lead to li tigation_51 3 But rumor fi les were noted by som e as being hard to track, difficult to access, and not easil y searchable. 5 14 The IU researchers noted that in the accident they were examining, the power mode was recorded as Accessory, and, indeed, that the deputy sheriff who responded had fo und the ignition sw itch "jammed" in the Accessory position.5 15 The investigators wrote: It is poss ible the ignition switch could have been knocked to the "Accessory" position by the dri ver' s leg or knee at the time of the vault . Th is investi gatio n revea led that inadvertent contact with the ignition switch or a key chai n in the 2005 Chevro let Cobalt can in fact result in engi ne shut~down and loss of power. 5 16 The IU researchers then descri bed how they had examined GM's October 2006 TSB and reported: "The bulletin indicates that there is a potentia l for the driver to inadvertently tum off the ignition due to low ignition key cy linder torque/effort.,,51 7 Researchers then reported th at they had identified " at least" six compla ints on the NHTSA website 51 8 relating to the engine shutting off and loss of power when the Ignition Switch or key chain was contacted by the driver. m J&8 Interview of Ryan Jahr, Apri l 1, 2014; J&B Interview of Cannen Benavides, May 19, 2014; J&B Interview of Dwayne Davidson, May 6, 2014; J&B Interview of Lisa Staccy, May 8, 20 14; J&B Interview of Ronald Poner, May 13, 2014; J&8 Interview of Jamcia Price, May 12,2014; J&8 Interview of Annette Rigdon, May 14, 2014; J&B Illlerview of Gay Kent, May 16, 2014. SI4 J&8 Interview of Jaclyn Palmer, May 8, 201 4; J&8 Interview of Ronald Poner, May 13, 2014; J&B Interview of Cannen Benavides, May 19, 2014. SIS The physical position of the ignition switch would later become critically important. For years, GM investigators were reviewing data thm rencctcd the ignition switch position, but because they lacked data regarding the actual key position, they were skeptical about whether the daTa were accuraTe. In panicular, they staTe thaT data showi ng the ignition switch in "run" was a barrier for years to their understanding that this was a mechanical issue rather than an electrical issue. J&8 Interview of Brian Stouffer, May 9, 2014. SI6 Indiana University Transportation Research Cellter, all-Site Air Bag Non-Deployment Investigation (April 25, 2007), at 7 rDOC ID 000001791070; GMN HTSAOO02239851. ~11 Indiana University Transportation Research Cellter, all-Site Air Bag Non-Deployment Investigation (April 25, 2007), at 7 rDOC ID 000001791070; GMN HTSAOOO223985l SI8 Indiana University Transportation Research Center, On-Site Air Bag Non-Deployment Investigation (April 25, 2007), at 7 [DOC ID 000001791070; GMNHTSAOO0223985J. 122 Privileged and Confidenfia/ Protec1ed by AUorney~ Client Privilege and as Allorney Work Product Some of the complaints reported a simpl e "brushing" of the key chain or touching of the ignition sw itch was all that was required for the engine to shut ofr SI9 The researchers concluded' It is not known what role, ifany, thi s may have played in the non~deplo yment of the air bags. Such a determination would most likely require an analysis ofthe air bag system and ignition wiring schematic in order to determine if in fact the air bag is capable of deploying when the ignition is switched from the ' on ' position to the ' accessory' position. 52o INDIANA UNIVERSITY TRANSPORTATION RESEARCH CENTER School of Public and Envirottnl:!Ql Affairs 122 W.,. Second Sc=I ~.1ndia'Ia ~ 74(1). 1 ~1 ("l'W·)IIOt ..... ("lIm·JIIl ON·SITE AIR BAG NON-DEPLOYMENT INVIlSTIGA TION CASE NU M8ER - IN-06-OJ} lOCATION · WISCONSIN VEm aE • 2005 OIEVROLET COBALT CRAS!! DATE · o.dIor lOO6 down. A sean:h of the NIITSA. Orl'ke of Defccts Investiption (001) ..."d) sile. ~lainl tab. A:\'caled al leas! sb romplainl$ (ODI id<:nli(lC~ion 11.1IIl1)c11: 10144299. 1014$9$9. 10129121 . 101323)S. 10IS I~. :m;I 10197022) re-latina 10 !he tn&ine 5Irunirc 0IT:m;I1oss of po"'l'r in Ow::vrokl Cobahs ...·hen the ignibon swil(:h Or ko:y dwn waJ conlltted by the dri ver. Somt oftb:~Wrl$ o:ponc:d a si~k: .brushin&- oflb: key dwn or IWchin& Qflb: ienition 5WilCh ....::as all thu"'"2$ reII J&B Interview of John Sprague, March 26, 201 4; J&8 Interview of John Sprague, May 27, 2014. The signed 3660 was stored in GM 's Global Quality Tracking System ("GQTS"). 612 J&8 Interview of John Sprague, March 26, 201 4. 613 J&8 Interview of John Sprague, March 26, 201 4. 606 137 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product OeG iorgio's office in the Vehicle Eng ineering Center in Wa rren, Michigan.614 Sprague remembers tellin g DeG iorgio why he wanted to speak w ith him, incl uding the fac t that the vehi cles appeared to be losing power while driving.61~ Sprague said he brought a spreadsheet showing work orders related to the Ignition Switch to th is meeting with DeGiorgio and the two of them looked at it together. 616 Sprague asked about work orders from the relevant time period to make sure he understood whether any changes to the switch had occ urred. 617 OeG iorgio told Sprague that the MY 2008 change to the Coba lt Ignition Switch was a change only to the antitheft system. 618 DeG iorgio said there had been no change to the switch that would have affected the power mode shutting Off. 619 OeG iorgio d id not te ll Sprague that there were any new Ignition Switch changes unaccounted for on his spreadsheet and did not reference anything about changes to the detent plunger that would have affected the torque required to turn the key.62o DeGiorgio's statements diverted . . 6'1 investigators. - J&B Interview of John Sprague, March 26, 201 4. J&B Interview of John Sprague, March 26, 2014. 616 J&B Interview of John Sprague, March 26, 201 4. 617 J&B Interview of John Sprague, March 26, 201 4. 618 J&B Interview of John Sprague, March 26, 201 4. 619 J&B Interview of John Spraglle, March 26, 201 4 620 J& B Interview of John Sprague, March 26, 2014. Sprague says he spoke with DeGiorgio again a half dozen times at most. J&B Interview of Jolm Sp rague, March 26, 2014. 621 Also in 2009, Tom Mercer, a senior engineer on ai rbag systcms, approached DeGiorgio after he had lcamed about the problems ofnon~deployment of airbags in the Cobalt frOIll Lisa Stacey of the FPA group. He told DeGiorgio about two crashes and asked whether there was something different about the Cobalt igni[ion switch that would result in some type of electrica l fa ilure. DeGiorgio did not mention anything about the low torque in the switch or the prior instances of stalls caused by inadvertent tuming of the ignition from Run to Accessory. J&B Interview of TIlOmas Mercer, May 2, 2014. 614 615 138 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product E. 2010: A Different Recall and Outside Counsel Warns GM of Possible Punitive Damages In 20 10, GM issued a power steering recall for MY 2005 ~20 I 0 Cobalts, even though GM had stated publicl y in 2005 that it did not view a moving stall at highway speeds to be a safety issue. GM al so rece ived its first warning from outside counsel that the company was at risk of being subj ect to punitive damages due to the unre solved investigation into repeated instances of Cobalt airbag non~deploy ment s . 1. March 2010: Cobalt Power-Steering Recall On February 19, 2010, NHTSA sent a letter to Gay Kent informing her that NHTSA was going to "investigate all egations of electric power steering (EPS) system failure in .. . 2005 through 2009 Chevrolet Cobalt vehicl es. , ,622 In March, GM issued a safety recall of certain 2005~2010 Cobalt models for a power-steering defect, unconnected to the problem of low torque in the Ignition Switch. 623 In notable contrast to GM 's deci sion in 2005 that moving stall s, which involved a loss of acceleration plus a loss ofpower~steering, plus a loss of power breaks (as well as loss of airbags, which the engineers failed to understand), were not a safety issue and did not compel a recall , in 20lO, GM issued a recall for loss ofpower~steering alone. The 2010 safety recall did not cause anyone to reconsider the prior categorization of moving stalls as n on~ safety issues, or produce a sense of urgency in addressing the Ignition Switch problems; moreover, when Cobalt dri vers brought their cars into the dealer in 20 10 to have the power~ s teerin g fix ed, the problem with th e Ignition Switch was not remedi ed. Letter from Jeffrey L. Quandt, NHTSA , to Gay Kent, GM (Feb. 19, 20 10) fDoc ID 14030700008565; GMHECOOOOOSO IS]. 623 Letter from Gay Kent, GM, to Daniel C. Smith, NHTSA (March 1, 20 10) [Doc ID 000006051812 ; GMNHTSA000265363 J. (>22 139 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product Desp ite the recall , at least some at GM still did not consider the loss of power-steering a safety probl em_ GM internally noted that power-steeri ng was normally a customer sati sfacti on issue, but the "Cobalt was handled in a di ffe rent manner based on GM's desire to obtain qui ck resolution and closure of the government investigation.,,624 Alan Adler, GM's manager for safety comm uni cations, remembered that GM had in itially been planning to categorize the electric power~steeri n g issue as a c ustomer satisfaction issue, but as a result of the congressiona l scruti ny of Toyota and the unintended acceleration issues, it was agreed that GM should iss ue a safety recall before the hearings so that "we would not get mentioned or dragged in to the Senate. ,,625 2. October 2010: Evaluation - The First 'Yarnin g of Punitive Damages In 2010, GM received its first warning fro m outside counsel that it was at risk of being subjected to punitive damages based on the non-dep loyme nt of airbags in the Coba lt. In October, King & Spalding (" K&S") prepared its first case evaluation for the _ matter. 626 died of blunt brain trauma after her 2006 Cobalt sideswiped a Volkswagen and then hi t a tree 2009. 627 _ head~on at 2:29 am on December 3 1, was wearing her seat be lt, and her Cobalt's airbag did not deploy.628 July 19, 2010 PowerPoint presentation titled "GM/NHTSA Mecting Technieal Briefing" at 13 [DOC 10 00000 1584943], auached 10 e~mail from Slcphcn G_Gehring, GM, 10 DOllglas Wachlel, GM (JlIly 19, 2010) [DOC ID 000001584940J. 625 J&B Interview of Alan Adler, May 2, 2014. 626 Lel\cr from Harold E. Franklin, Jr., Ki ng & Spa lding, 10 Jaclyn C. Palmer, GM , and Annel\e Rigdon, ES IS (OCI. 7, 2010) rDOC ID 000001662352 ; GMNHTSAOO02111491 "21 Lel\cr from Harold E. Franklin, Jr., Ki ng & Spa lding, 10 Jaclyn C. Palmer, GM , and Annel\e Rigdon, ESIS (OCI. 7, 2010), at 2 LDOe lD 000001662352 ; GMNHTSAOOO211149J. 628 Letter from Harold E. Franklin, Jr., King & Spalding, to Jaclyn C. Palmer, GM, and Annette Rigdon, ES IS (OCI. 7, 2010), at 2 LDOe lD 000001662352 ; GMNHTSAOOO211149J. 624 140 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product According to the SDM data, the vehicle 's power mode status was in the Off position at the time of the accident 629 The _ matter came into GM as a NISM assigned to Jaclyn PalmeL Kathy Anderson and John Sprague were the FPA engineers. The evaluation in 20 lOis the first case eva luation to note what K&S lawyers memo ri alized as a "sensing anomaly" in Cobalt vehicles.63o According to the case evaluation, Anderson identified this crash as one in which there should have been a deployment and that a "sensing anomaly" may have prevented deployment. 631 According to Anderson, who never recei ved the case eva luati on, she had not identified a "sensing anomaly" as the potential cause of the non~deplo y ment. but rather the "anoma ly" was the Ignition Switch "contact bounce" theory that Sprague and Churchwell had been testing, but were unable to prove. 6 _12 K&S wrote that the failure of the airbags to deploy because of a "sensing anomaly" made thi s case difficult to defend: The severe frontal impact and damage profile will be fonnidab le obstacles to any technica l justification for the nondeployment. The prel iminary determination by our airbag expert that a sensing "anomaly" prevented deployment is clearly the most chall enging aspect of this case. 633 Data Report File Infonnation (March 10, 2010) LDOC ID 00000 1662309; GMNHTSA0002 I09371Case Evaluation (OCI. 7, 2010), aI 8[ DOC ID 000001662352; GMNHTSA000211149]. Case Evaluation (Oct. 7, 201 0), al 8 rDOC ID 000001662352; GMNHTSA00021 1149l; J&B Anderson, May 27, 20 14. J&B Interview of Kathy Anderson, May 27, 201 4. The case evaluation further stated that "[tJhe SDM calibration was changed for the 2008 model year and she lAndersonJ has not seen this anomaly in the 2008 model year or subsequent Cobalts." In fact , Anderson had located a sensing repon lhat indicated were had been a c hange from an electronic front sensor system to a raw data scnsor system in MY 2008 Cobalts. According to Andcrson, however, that change had nothing to do with the ignition swi tch "contact bounce" theory or the airbag non-deployment and she believes that K&S eonflated the two concepts in writing the case evaluation. J&B Interview of Kathy Anderson, M~ 633 _ _ Case Evaluation (Oct. 7, 2010), at 12lDOC ID 000001662352; GMNHTSA00021 I 149J. 141 Privileged and Confidenfia/ Protec1ed by AUorney~ Client Privilege and as Allorney Work Product A significant plaintiffs verdi ct is li kely under these unusual circumstances in which an apparent malfunction (sensing system a nomaly) prevented a irbag . t h·IS severe fimnta I Impact. . 634 depIoyment 10 Indeed, coun sel warned that the " facts and circumstances surrounding the investigation" into the "anomaly" - that is that there were prior and sti II unreso lved instances - could even lead to the award of punitive damages: .. . th e facts and circumstances surrounding the investigation into the sensing system " anomaly" that may be present in some Cobalts could provi de ferti le ground for laying the foundation for an award for puniti ve damages, resulting in a significantl y larger verdict. 635 We recommeod that General Moton LLC attempt to J'C$Olve this case: at this time. In Ugbt oftbc foregoing tcdmical usc mc:utI by our ~ our UDdcntandin& ofthc factual background. and input from Tent counsel (M»o reports that Nashville an::ajurisdictions are not "judicio! hcllholeo"). a .......figun: vetdict remains probable. Furthenn."., Ibe """ 0Ild circumstances sumrundinS the iDvestiptioo into the aeasina system "momaly" that may be ~ in some Cobelts could provida fertile arouod for laying the fOlJDdstion for III aMId for puniijve ........ !016 J&B Intcrview orLawrence Buonomo, April 16,2014. 647 J&B Intcrview or Michael Gruskin, April 17, 201 4. 648 Jadyn Palmer, GM, Settlemenl Review CommittecJCase Summary (Jan. 11 ,20 11), at I [DOC lD 000006234239J. 645 146 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product cases fro m this week 's SRC, I know several of us would like to see some additional background about an ' anomal y' . " 649 According to both Kemp and Nowak-Vanderhoef, at some time in 20 11 Lucy Clark Dougherty attended a meeting at which airbag non-deployme nt was di scussed and gave Kemp a direction as to how to proceed. Clark Dougherty was new to the company; she joined the SRC and moved into her role as Kemp and Nowak-Vanderhoef s boss in March 20 11. 650 Although Kemp could not remember at what meeting thi s occurred, he described how Clark Dougherty directed him to have engineering look into the ign itionlai rbag deployment issue.65 1 Kemp recalled this as having occurred at the end of an SRC meeting; he said it was " uncommon" for a technical issue to be directed to him at an SRC meeting, but that it did happen from time to time.652 Kemp 's recoll ection was corroborated by Nowak~Vanderhoef, who similarl y remembered Clark Dougherty directing Kemp to have e ngi neers investigate thi s issue. 653 For her part, Clark Dougherty stated that she did not remember such a conversation wi th Kemp, but she did not rul e it out. She sa id it was possible that she had the conversation with Kemp hut did not then understand the gravity of the siruation or the full facts. 654 Regardless of when or whether such a direction occ urred, no meeting occurred in January 2011 pursuant to Nowak~Vande rhoef s e~ mail or a direction from Clark Dougherty. Instead, the E-mail from JacylnPalmer. GM.to DebNowak-Vanderhoef. GM (Jan. 14, 201 1) IDOC ID 000001745032 GMNHTSA000220386l; J&B Interview of Jaclyn Palmer, April 24 , 20 14. 6S0 Beforejoining GM in 2010, Clark DougheTly had been in government service and before thm private practice 651 J& B Interview of William Kemp, April 17. 201 4. Kemp described the meeting as an SRC meeting. The only SRC mceti! l J in! 2011iin which we have verified that Cobalt airbag non-deployment was discussed. however. concerned Clark Dougherty had 110t been assigned to Detroit by then. 6S2 J&B Interview 0 William Kemp, April 17, 2014. 0J Nowak-Vanderhoef reported that Clark Dougherty gave Kemp lhis instmction at an SRC meeting in 20 11 at which Jaclyn Palmer presented, but then said she was uncertain and confused about the date. J&B Interview of Deborah Nowak-Vanderhoef, May 1, 2014. 654 J& B Interview of Lucy Clark Dougherty. April 16, 2014. 649 147 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product meeting occurred in July 2011. Witnesses could not explain why six months passed before the meeting took place, but th e delay again highlights the lack of urgency in addressing the issue .65s Palmer bel ieves that in the interim she talked to Sprague and Everest about obtaining Cobaltrelated data and sharing the data with Kemp, Nowak~ Vanderhoef and Carmen Benavides, Director of Product Investigations, Safety Regulations, Field Performance Assessment, and TREAD. 2. J uly20 1l : About rurllll1,e D'a mlag'" Eva lu ation a nd Another War ning On July 26, 20 11, a little over six months after the _ submitted a case evaluation for the _ SRC meeting , K&S matter. 656 As in _ Palmer was the GM lawyer and Anderson and Sprague we re the FPA engineers on the _ 13,201 1 matter. On February was driving her mother 's 2007 Chevro let Cobalt when she lost control of the vehicle, traveled off the roadway and ran road. 657 The airbag did not deploy.658 _ head~on into a tree on the side of the suffered significant injuries and was hospitalized for nine days .659 The SDM data showed that at the time of the accident the vehicle's power ' A ccessory. 660 mo de status was III Interview of William Kemp, May 9, 2014. Kemp did not recall the dates of either the SRC or the July meeting. 656 Letter from Harold Franklin, King & Spa lding, to Jaclyn Palmcr, GM , and Anncttc Rigdon, ESIS (July 26, 2011) [DOC ID 000005185394]_ 657 Letter from Harold Franklin, King & Spa lding, to Jaclyn Palmer, GM , and Annette Rigdon, ESI$ (July 26, 201 I) [DOC ID 0000051853941. 658 Letter from Harold Franklin, King & Spa lding, to Jaclyn Palmcr, GM , and Annellc Rigdon, ESI$ (July 26, 2011) (DOC ID 0000051853941. (>:>9 Letter from Harold Franklin, King & Spa lding, to Jaclyn Palmcr , GM , and AnnCIlC Rigdon, ESI$ (July 26, 20 II) [DOC ID 0000051853941. 660 Letter from Harold Franklin, King & Spa lding, to Jaclyn Palmer, GM, and Annette Rigdon, ESIS (July 26, 20 11), at 8 l DOC ID 000005 185394J. 6551&8 148 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product K&S ' s July 26, 2011 report again highlighted the "sensing anomaly" about which it had twice before written in the _ evaluations , noting that jurors wou ld expect the airbags to deploy in this case and even GM ' s own expert could not rule out the sensing "anomal y" as the 66 J K&S wrote about how Sprague had seen thi s problem before in "some Cobalt cause. vehicles" and further noted that " the fact that the SDM data indicated that the car was in accessory mode at the time of the accident is clearly the most challenging aspect of this case.',(j62 K&S warned of the poss ibility of puniti ve damages beca use of "the facts and circumstances" surro unding the Cobalt investi gation: In addition, the facts and circumstances surrounding the investigation into the sensing system "anomaly" that may be present in some Cobalts could provide fertile ground for laying the foundation for an award ofEun iti ve damages, resulting in a signifi cantl y larger verdict. 63 At the time, K&S , in~hou se counsel, and the FPA engineers were not aware of the 2005 TSB , the PRTS reports that led to the TSBs, the Wisconsin State Patrol report, or the Indiana University study. The _ matter was reviewed at an August 3, 201 1 Roundtable, attended by Michael Gruskin and Jaclyn Palmer (who presented).664 The Roundtable summa ry stated: The vehicle power mode status was recorded as Accessory which indicates the sensi ng algorithm could have been di sabled from deploying the ai rbags. This would have to be continned by sending the SDM to the supplier for further interrogation . John Sprague indicates it is also possible the airbags did not 661 Letter from Haro[d Franklin, King & Spa [ding, to Jaclyn Palmer, GM , and Annettc Rigdon, ESIS (July 26, 201 I), at 2 l DOC ID 000005 185394J. Lctter from Harold Franklin, King & Spa lding, to Jaclyn Palmer, GM , and Anllclle Rigdon, ESIS (July 26, 2011), al 14 [DOC ID 0000051853941t.M Letter from Harold Franklin, King & Spa lding, 10 Jaclyn Palmer, GM , and Annelle Rigdon, ESIS (July 26, 2011), at 14 LDOC ID 000005 I 85394J. 664 General MOIOrs, Settlement Roundtab le Case Summary (Aug. 3, 20[ 1), at [-2 [DOC ID 000001662599 GMHNTSA0002 12972 J. 662 149 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product deploy. because of the extended crash pulse. The subject vehic le was s u~ect to the Cobalt power steering recall, which was perfonned prior to the crash. 65 3. July 2011: Meeting to Kick Offthe FPE Process In July 2011 , GM lawyers finall y met with the Product Investigation team concerning the ongoing Cobalt airbag non~deployment matter. 666 According to participants, this was the meeting that Nowak-Vanderhoefhad first requested in January after t h e _ SRC. 667 According to Bill Kemp, thi s was an "unusual " meeting. It was not, he said, in the normal courSe to pull that many people together to initiate the FPE process, and, indeed, he had never before seen a meeting like it. 668 For thi s issue, though , Kemp said he "wanted to make sure senior management had eyeballs on thi s and not let it flow through the normal process.,,669 FPA supervisor Everest similarly reported that he understood the purpose of the July 27 , 2011 , meeting was for GM Legal to express to PI that certa in Cobalt non~depl oy me nt cases represented a sa fety concern and that PI urgently needed to detennine the root cause of the problem. 67o Palmer too expressed fru strati on that the process did not seem to be moving forward, and played an important role in putting together the prese ntation. 67 t 665General Motors, Settlement Roundtable Case Summary (Aug. 3, 2011), at 1-2 rDOC ID 00000 1662599 GMHNTSA0002129721. 666 Invitation from Jadyn Palmer, GM, to Canncn Bcnavides, GM , ef al. (Junc, 27, 20 11 ) rDOC ID 0000017450201. Invitees included GM al1omey~ Deborah Nowak-Vanderhocfand Bill Kemp. Product Investigations per~onncl Cannen Benavides and Doug Wachtel, and engineers Matthew Jerinsky, James Churchwell, Brian Everest, and John S~rague. According to Palmer' s invitation, "FPA will be the primary presenter during the meeting." 66 J&8 Interview of Jaclyn Pa lmcr, April 24, 2014; J&B Intcrview of William Kcmp, May 9, 20 14; J&8 interview of Dcborah Nowak-Vanderhoef, May 1, 2014. QtXI J&8 Interview of William Kemp, May 9, 2014. 669 J&8 Interview of William Kemp, May 9, 2014. 670 J&8 Interview of Brian Everest, March 27, 2014. 671 J&B Interview of Brian Everest, March 27, 2014; J&B Interview of Jaclyn Palmer, April 24, 20 14. 150 Privileged and Confidenfia/ Protec1ed by AUorn ey~ Client Privilege and as Allorney Work Product In advance of the meeting, Pa lmer met with Sprague to review the non-deployment cases and in cidents and to di sc uss what would be presented to PL672 Palmer and Spra gue settled on presenting three Cobalt non-deployment incidents: (I ) the 2005. _ crash; and (3) the 2009 _ crash.673 • crash; (2) the 2008 was selected because it was the incident that NHTSA had brought to GM's attention in March 2007, and because data downloaded from the . expected. 674 _ SDM indi cated that an airba g deployment would have been too, was selected because the data downloaded off ofth e SDM indicated an airbag deployment would have been expected. _ _ was also selected because the SDM had been analyzed by Continenta l in 2009 and fou nd to ha ve had a disabled crash sensing algorithm at the time of the crash - a fact consistent with FPA's working theory that the non-deployments were caused by a change in the power-mode message received by the SDM.675 Finally, _ was selected because it was the incident that prompted the meeting in the first pl ace and because pi ctures of the_ crash showed a particularly significant impact. 676 Witnesses describe attendees of the meeting as including, among others, Kemp, Palmer, Nowak-Vanderhoef, Sprague, Wachtel, Sevigny, and Carmen Benavides.677 Although numerous J& B Interview of John Sprague, May 15, 20 14. J& B Interview of John Sprague, May 15, 201 4. 674 J& B Interview of John Sprague, May 15, 201 4. 675 J& B Interview of John Sprague, May 15, 2014. 676 J&B Intervicw of John Spraguc, May 15 , 2014 ('77 J& B Interview of Jaclyn Palmer, April 24, 201 4; J&B Interview of John Sprague, March 26, 20 14 ; J&B Interview of Brian Everest, March 27, 20 14. Invitation from Jac1yn Palmer, GM, to Cannen Benavides, GM, el 01. (June, 27, 2011) [DOC ID 000001745020], Invitees included GM anomeys Deborah Nowak-Vanderhoef and Bi ll Kemp, Product investiga tions personnel Cannen Benavides and Doug Wachtel , and enginccrs Mal1hcw Jcrinsky, James Churchwell, Brian Everest, and John Sprague. Invitation from Jaelyn Palmer, GM, to Brian Everest, GM, el al. (June, 27, 201 1) LDOC ID 0000021 49006 GMNHTSA000220996J. Benavides stated that at the time, she was focused on why the airbag did not deploy - and therefore on potentia l problems with (he SDM - rathe r than on the ignition. She also said tbat she understood that the legal department was handling (he airbag non-deployment issue 672 673 15 1 Privileged and Confidenfia/ Protec1ed by AUorn ey~ Client Privilege and as Allorney Work Product witnesses described WatcheJ' s presence - and comments that he made - Wachtel states he did . 678 not attend t he meetmg. At the meeting, Palmer presented the fa cts of the _ _ and _ crashes. 679 Then Sprague presented fi les and pi ctures from all three incidents and rev iewed what FPA knew at that time, as well as the questions that remained unanswered.68o Although he does not specifically remember, Sprague believed that fa taliti es were di sc ussed; the deaths were apparent based on the fil es shared and pictures shown.68 1 The urgency o f the situati on should have been manifest. Sprague explained that there was a theory that the airbag n o n~de plo y m en lS were connected to the SDM receiving a message that the veh icle power mode was in Accessory or Off. 68 2 One witness recall ed that Sprague had thought these power~mode messages might be connected to the Igniti on Switch.683 Several witnesses reported that, at the meeting, Wac htel, the Senior Man ager of Product Investigations, expressed reluctance to take on the issue by opining th at the incident rate was not high.684 According to Palmer, Wachtel deni ed that PI had previously been alerted to the PI non~ and that she should not discuss the content of the meeting with others to ma intain anomey-client privilege. J&B Interview ofCannen Benavides, March 25, 2014. 678 J& B Interview of Douglas Wachtel, May 22, 2014. 679 J& 8 Interview ofCan nen Benavides, March 25, 20 14. 680 J& 8 Interview of Jenn ifer Sevigny, March 27, 201 4; J&8 Interview of Deborah Nowak-Vanderhoef, May I, 20 14; 1&8 Interview of John Sprague, May 15, 2014. 681 J&8 Intervicw of John Spraguc, May 15, 2014 682 J& 8 Interview of Jennifer Sevigny, March 27, 201 4. 683 J& 8 Interview of Jennifer Sevigny, March 27, 201 4. 684 J&8 Interview of Deborah Nowak-Va nderhoef, May 1, 2014; 1&8 Interview of Brian Everest, March 27, 2014. Wachtel docs not recall attending this meet ing and stated that, in any event, he would have placed lillie emphasis on the incident rate because of the seriousness of the crashes. J&8 Interview of Douglas Wachtel, May 22, 20 14. Kemp, too, remembered someone making an observation abom a low incident rate, but could not remember who. According to Kemp, he would have dismissed such a statement because he considered the potential severity ofthc accidents in addition to tllC incident ratc. In addition, said Kemp, the view was "inconsistent with my vicw atthc 152 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product dep loyment issue, which she believed was untrue. 68S Both Palmer and Everest reponed being frustrated by Wachtel and th e PI group's response.686 As noted above, Wachtel deni ed attending the meeting at all , saying that he was told after the meeting to assign a n investigator to the matter. 687 According to witnesses, Kemp instructed Wachte l to put the Cobalt matter through the investigation process. 688 Everest reported that Wachtel appeared to understand and accept Kemp's instructions, and agreed to ass ign an investigator. 689 Following the July 27 meeting, Wac hte l confe rred with Sprague. 69Q Wac htel met with Sprague, Everest, and Jim Churchwell to di scuss PI openi ng an investigation.691 He th en assigned PI investigator Brian Stouffer to investigate the matter and asked him to co nfer with Sprague. According to Everest, as a res ult of the meeting, the Cobalt investigation was "on [Wachtel's] list, but not top of hi s list.,,692 Accord ing to Stouffer, Wachtel never told him that th is was an urgent matter. Although Sprague continued to be involved in the matter, it was at th is moment that a fonna l PI investigation began. In sum, the h and~off to PI took place because GM lawyers were frustrated with the lack of progress being made by FPA, a group of engineers whose foc us was on individual products li abil ity cases, not safety investigations. By the account of each lawyer, they were seeing cases time pIllS the direction I had been given," a reference to his report to liS that Clark Dougherty had directed him to have PI investigate the issue. J&8 Interv iew of Wi lliam Kemp, May 9, 20 14. 685 J&B Interview of Jaelyn Palmer, April 24, 201 4; e~ll1ail from Jaclyn Palmer, GM , to Douglas Wachtel , GM (July 27, 2011), forwarding an e-mail fromErieBlI dd ri ll ~ . GM . lo J aclynPalmer. GM(Jan .11 . 2011) {DOC ID 00000 1745053 GMNN HTSA000220997]. 686 J& B Interview of Jaclyn Pa lmer, April 24, 201 4; J&8 Intcrview of John Sprague, May 15, 2014. 687 J&B Intcrview of Douglas Wachtel, May 14, 2014. 688 J&B Intcrview of Brian Evcrest, Marc h 27, 2014. ~9 J&B Intcrview of Brian Evcrest, March 27, 2014. 690 J&B Intcrview of Douglas Wachtel, March 18, 2014. 691 J&B Interview of Douglas Wachtel, May 14, 2014. 692 J& B Interview of8rian Everest, Marc h 27, 2014. 153 Privileged and Confidenfia/ Protec1ed by AUorn ey~ Client Privilege and as Allorney Work Product they did not understa nd, and wanted something more do ne . Indeed, the very fact of the "unusual " meeting to ki ck off an FPE process bespoke a sense of urgency. But the meeting took six months to schedule - an unexplained delay that meant that more accidents were occurring whil e the investi gation did not move forwa rd. For their part, the lawyers felt they had done their job by emphasizing the importance of the issue to th e engineers and moving the issue from the FPA enginee rs to the PI g roup . But, faced with a pattern of crashes that had resulted in fataliti es and an unexplained "anomaly" that affected the depl oyment of airbags, they did not at the same time elevate the issues to the General Co unsel and do not appear to have insisted on a quick and concrete timetable for the sa fety investigation . Moreover, any urgency wa s lost at the moment o f the hand-off to the PI group. Wachtel (who was one of the engineers able to recreate the Cobalt moving stall in 2005 before PI closed its in vestigation) demonstrated he was di sinclined to open an investi gation at all and then never imparted a sense of urgency to the investigator he assigned. Wachtel' s conduct is a demonstration of what Mary Barra descri bed as th e " GM nod" : When everyone nods in agreement to a proposed plan of action, but then leaves th e room and does littl e. 693 Witho ut a timetable or a supervisor demanding action, the ensuing PI investigati on proceeded at an inex.cusably slow pace. 4. August 201 J: Stouffer and Sprague Meet to Transition Files In the fall of 2011 , after Wachtel assigned Stouffer to the investi gation, Sprague met with Stouffer and described that he had been tracking Cobalt non-deployment crashes in w hi ch 693 J& B Interview orMary Barra, April 14, 201 4. 154 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product available data suggested an airbag should have deployed.694 Sprague explained a working th eory that th e non-deployments were caused by the S DM rece iving a power mode message of Off or 695 Sprague further noted that the nonAccessory irrunediately before or during the crash. deployment issue was not present in the MY 2008 and later CobaJts.696 Stouffer understood from thi s early meeti ng that Sprague had previously exami ned all changes related to the Igniti on Switch. 697 At or around the time of their initial meeting, Sprague provided Stou ffer with a flash dri ve o f relevant materi als. 698 Although Stouffer's and Sprague ' s respective memories differ on whether the fla sh drive included a copy of Sprague's then-current tracking spreadsheet,699 at a minimum the drive contained information pertaining to the non-dep loy ment events underl ying Sprague's tracking spreadsheet. related to the Ignition Switch. 7OO Sprague al so provided Stouffer with a li st o f work orders 701 69~ J&B Interview or Brian Stouffer, May 9, 201 4; J&B IllIerview o f Brian Stouffer, May 24, 20 14; J&B Interview of John Sprague, May 26, 20 14. J&B Interview or Brian Stouffe r, April 4, 201 4. 696 J&B Interview or Brian Stouffe r, April 4, 201 4. 697 J&B Interview or Brian StoufTer, May 9, 201 4. 698 J& B Interview of Bria n Stouffer, April 4, 201 4; J&B Interview of John Sprague, March 26, 20 14. 699 Compare J&B Interview or Brian StoufTer, April 4, 2014, wilh J&B Interview of John Sprague, March 26, 2014. Stouffer believes a version of Sprague's spreadsheet was included ollthc nash drive when it was initially provided to Stouffer in or around August 2011. A review of the nash drive's contents, however, reveals that the earliest tracking spreadsheet on the nash dri ve is dated to March 2012 - several months after Stouffer's initial meeti ng wi th S£rague. 7 Sprague has no speci fic reeolleelion of including Ihe spreadsheci on Ihe nash drive when he inilially provided il to Stouffer. He nonethe less remembers providing infonnation thai supported the contents of his tracking spreadsheet. J&B Interview of John Sprague, March 26, 201 4. 701 J&B Interview or Brian Stouffer, May 9, 201 4. Stouffer did nOl recall whether Sprague provided him with electronie versions orthe EWOs or whether Stouffer col lected copies of the EWOs after meeting with Sprague. While Stouffer recalls having copies of relevant PRTSs (including N 172404 and N I 82276) on the nash drive, it is unclear whether those PRTSs were provided by Sprague. J&B Interview or Brian Stouffer, May 9, 2014. Sprague lacks direct access to the PRTS database and has never seen PRTSs relevant to the ignition switch. J&B Interview of John Spmgue, May 15, 2014. 695 155 Privileged and Confidenfia/ Protec1ed by AUorn ey~ Client Privilege and as Allorney Work Product Desp ite the fail ed experiment at the Milford Proving Grounds, Sprague expla ined to Stouffer that he had a working theo!)' (initially deve loped in mid~ 2 00 9) that a "contact bounce" caused the Ignition Switch to "open up" and send an Accessory or Off signal to the S DM, 702 resulting in non~deployment of the airbags . Sprague described one fatality involving a girl who drove away from a party and crashed the ve hicl e she was drivi ng. 703 During the course of hi s investigation, Sto uffe r learned about one additio nal fatal ity, but it occurred in the 2005-2006 ti meframe, and Stouffer did not be lieve that fatalities were continuing to occur.704 Shortl y after hi s meeting w ith Sprague, Wachtel gave or referred Stouffe r to the fil e that had been created by FP E investi gating engineer Elizabeth Kiihr in 2005 . In obtaining that fil e, 705 Kiihr's six ~yearStouffer immediately gained access to infonnation that Sprague never had. old file contained the scathing press articl es at the Coba lt launch relating to moving stall s, customer complaints, and a copy of a Februa!)' 2005 Preliminary Infonnation on engine stall in the Cobalt. 706 According to Stouffer, he did not view the materi als he received as particu larly significa nt because they did not explain two phenomena menti oned by Sprague that confounded Stouffer throughout hi s in vesti gatio n: I. The non-deployment phenomenon was absen t in MY 2008 and later Cobalts and hi s be lief that no relevant part changes had occurred to expl ain thi s, and 2. In almost half of the non-deployment cases, the SDM recorded the Ignition Swi tch as being in " Run ." J& B J& B 7(14 J& B 705 J& B 706 J& B 11)2 7IJJ Interview Interview Interview Interview Interview of Brian StouITer, April 4, 201 4. of Brian StouITcr, April 4, 201 4. of Brian StoliITcr, May 9, 201 4. of John Sprague, May 23, 2014. of Brian StOll ITer, May 9, 201 4. 156 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product These two fa cts led Stouffer to beli eve that the low torque required to turn the Ignition Switch was not th e «root ca use" of the airhag non~deployment -'07 Thus, notwithstanding the conclusions drawn by the Wisconsin State Patrol report and the Indiana Uni versity study (the former of which Stouffer was never aware, and the latter of which he only became aware in July 20 12). Stouffer's focus on these two facts paralyzed hi s ability to fully analyze the problem or revisit hi s preliminary conclusions. 5. Information Sharing Between FPA Engineers and the PI Group Earl y in our investigation, we were told that Sto uffer was denied access to information by Sprague because it was privileged. thus hampering Stouffer's investigation of the issue . Wachtel identified lack of informati on and data sharing by the FPA group, of which Sprague was a member, as one of th e reasons the Cobalt airbag non-depl oyme nt investigation took so long. 708 Given the potentially serious implications of this stateme nt, we investigated th e issue exhaustively, including conducting multiple interviews of the two central figures , Stouffe r and Sprague. Stouffer said that he was never denied information he sought for privilege reasons. 709 Although he noted he mi ght not have access to documents,710 he does not reca ll ever being told by Sprague that he could not have access to certain infoffil ati on because it was protected by privilege. 7Jl Sprague has no specific recollection of rebuffi ng a Stouffe r request on privilege grounds, and explained that, if Stouffe r asked for information that was contained in a privileged J&S J&S 7()9 J&S 710 J&B 711 J&B 1(11 7tm Interview of Sri an Stouffer, May 9, 2014. Interview of Douglas Wachtel, March 18,2014. Interview of Sri an Stouffer, May 9, 2014. Interview of Brian Stouffer, Apri l 4, 2014. Interview of Brian Stouffer, May 9, 2014; J&B IllIcrview of Brian Stouffer, April 4, 2014. 157 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product document or source, Sprague would have provided the in formation to Stouffer, just not the . If_'" pn·vi·1eged docum ent itse Jennifer Sevigny, the head of FPA, described her group 's approach to pri vilege in a similar way: Because FPA works with GM attorneys on lawsuits and lega l claims, much of the infonnation and materi al the group has is pri vi leged. If other GM employees, such as PI investigators, make a request for infonnation fro m FPA for use in a PI investigation, FPA routes the request through the lega l department for approva l. Sevigny said that, when such requests for informat ion are made, the legal department has generall y been able to accommodate them, and FPA has in the past provided infonnation in response to such requests .713 In other words, if Stouffer requested information from Sp rague, Sprague would give it to him.71 4 6. Summer to Fall 2011: Stouffer Analyzes TREAD Data According to Stouffer, his first investigative step was to retrieve and anal yze the relevant TREAD data to identify any additional airbag non-dep loyme nt incidents that matched the pattern observed by FPA. 7 15 Stouffer looked broadly for reports of airbag nOl1~deployments with . • • InJunes. 7 16 J&B Interview of John Sprague, May 23, 2014. J&B 11l1erview of Jennifer Sevigny, Mll rch 27, 2014 Dne wilness reponed lhm FPA engineers arc discouraged from sharing infomlation about specific lawsu its, and another said that there is an "i mperfect transfer from FPA engineers to FPE engineers due to privi lege concerns." J&B Interview of Kathy Anderson, April 24, 20 14; J&B Intervicw of Ronald Poner, April 30, 2014. 714 SlOulTer did not ask for, nor was he given, materi als gencratcd by lawyers that wcrc prescnted to thc Roundtable or SRC. Those materials, howcver, relied on the technical analyscs ofSpraguc and others of the FPA group, and, as notcd above, Spraguc did provide techn ical infonnationto Stouffer. 115 J&B Interview of Brian Stouffer, Apri l 4, 201 4. 116 J& B Interview of Brian Stouffer, Apri l 4, 201 4. 112 113 158 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product From the TREA D database, Stouffer retrieved instances of airbag no n~deplo yment for the Cobalt from 2005 to 2007. 717 Stouffer, however, did not pull TREAD data on deaths and injuries, which is stored in a different part of the TREAD database. 71 8 Stouffer did not have access to the death and injury database, although he could have asked someone else to run those searches for him. 71 9 He did not know of the database's existence at the time. In January 2014, Davidson conducted full TREAD data searches and found additional fata lities and serious injuries connected to the Cobalt.no As many witnesses have noted, the TREAD database is extremely difficult to use and search, potentially resulting in valuable data being missed by those in vestigating potential safety or other problems. See Appendix E. GM has a dedicated team that performs such searches upon request. In this case, Stouffer's failure to fully search the TREAD data caused him to miss key pieces ofinfoffilation, serious injuries, and fata lities associated with the Cobalt. 721 Stouffer's 20 II search for TREAD data also included a search for airbag nondeploym ents in MY 2005-2007 Ion s. However, Stouffer mi stakenly left out MY 2003~2004 Ions, causi ng him to mi ss several accidents, including accidents in vol ving fataliti es. 722 After the J&B Interview of Sri a n Stouffer, April 4, 201 4. Stouffer did not pull TREAD data for the Ion unli12012. He did not recall whether he pullcd thc HHR data in 2011. In addition 10 the airbag bucket TREAD report, Stouffer also pullcd thc Coh.1il power Ira in buckel , ignilion bucket , and electrical bucket from TREAD Stouffer pulled thi!>; dma because he was aware of the 2005 engine stall complaints and PRTSs. 718 J&B Interview of Dwayne Davidson, A pril 22, 201 4. 719 J&B Intcrview of S ri an Stouffcr, May 9, 2014. 720 J&B Intcrview ofDwaync Davidson, April 22, 2014; J&S Intervicw ofCanncn Bcnavides, May 25 , 2014. 721 J&B Intcrview of S ri a n Stouffcr, April 4, 201 4; J&S Intervicw of Dwaync Davidson, April 22, 2014; J&B Interview ofCanncn Benavides, March 25, 201 4. m J&B Interview of Brian Stouffer, April 4, 2014. In 2012, Stouffer pulled engine stall data for MY 2003 -2007 Ions, but did not search fo r Ion airbag non-deployment data for these model years. 717 159 Privileged and Confidenfia/ Protec1ed by AUorn ey~ Client Privilege and as Allorney Work Product Cobalt recall was issued on February 7, 2014, Davidson pulled the full data for the Ion and di scovered Ion fataliti es that Stouffer had missed_723 According to Stouffer, he completed hi s first analysis of TREAD data on November 17, 2011. 724 He updated and summarized the TREAD data aga in on Marc h 12, 2012. 725 After he rev iewed the TREAD data, Stouffe r identifi ed what he recollects to be approximately 50 additi onal reports that he believed might fit the criteria of the crashes identified to date. 726 He shared this list of incidents with Sprague and asked Sprague to identi fy whether any on the li st were crashes where Sprague would have expected the ai rbags to deploy but th ey did not. 727 Sprague responded to Stouffer that, in hi s view, none of the crashes on the li st fi t that description. 72R According to Stouffer, from August 2011 to March 201 2, he di scussed th e Cobalt in vesti gation with Wachtel peri odicall y, and Wachtel never told him to prioriti ze the in vesti gation or proceed at a faster pace.729 At the time, Stouffer did not personall y feel that the investigation should be moving at a fas ter pace. 730 Stouffe r understood that he was investigating crashes that had occurred several years ago, and hi s understanding was that th e Coba lt n on~ deployments were not continuing to occur whil e he in vestigated thi s matter. 73 ! According to Kemp , he repeatedly spoke to Wachtel about the progress of the investi gation. Kemp said that J& B Interview ofCannen Benavides, March 25, 2014. J& B Interview of Brian Stouffer, April 4, 201 4. m J& B interview of Brian Stouffer, Apri l 4, 201 4. 726 J& B Interview of Bria n Stouffer, April 4, 201 4. Those criteria were, according to Stouffer: the vehicle had had an off-road incident prior to the crash, there had been a fro nt-end crash, and Sprague thought the airbags probably should have deployed based on Sprdgue' s review of the available data. 1&B Interview of Brian Stouffer, May 24, 20 14. 727 1&8 Interview of Brian Stouffer, May 24, 2014. m J& 8 Interview of Brian Stouffer, May 24, 2014; 1&B Interview of John Sprague, May 26, 2014. 729 J& 8 Interview of Brian Stouffe r, April 4, 201 4; J&B Interview of Brian Stouffer, May 9, 2014. 730 J& 8 Interview of Brian Stouffe r, April 4, 201 4. 731 J& B Interview of Brian Stouffe r, April 4, 201 4. 723 724 160 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product he had a "couple" of conversati ons in whi ch he told Wachtel that GM needed to "get on this. ,,732 Kemp sa id that Wac htel repeated ly told him variati ons of: "There's something goin g on here we don't know.',733 In sum, however, at the beginning of20 12, a full year after GM was warned of the possibility of punitive damages, and the GM legal team had requested a meeti ng to move an investigation into possible safety issues in the Cobalt fOfW'ard, the investigation had moved not at all. There were no timetables, no meetings scheduled, and no progress. A new investigator was looki ng at the same data as Sprague, and he also had some additional infonnation that had been in GM's fil es for years. But littl e e lse had changed, and GM was no closer to address ing the safety defect in the Cobalt H. 2012: The Search for the "Root Cause" In th e first part of 20 12, mechanical engineers and electrical engi neers debated how best to proceed. In the meantime, Cobalts with a safety defect stayed on the road, while GM engineers held more meetings. In June 20 12, the plainti ff's expert in an airbag f]on~deployment case located the Indiana University study and cited it in her report, bringing it to the attention of GM engi neers and lawyers working on the Ignition Switch issue for the first time. The expert, Eri n Shipp, concluded that the airbag n o n~d eployme nt s were being ca used by the low torque of the Igniti on Sw itch. Stouffe r discounted the Indiana Uni versity study and the Shipp report because they did not explain airbag n on~dep l oyments in cases in whi ch the Ignition Switch was in Run, and because they did not explain why the non~deployments J&B Interview or William Kemp, May 9, 2014. m J&B Interview or Wi ll iam Kemp, May 9, 2014. 732 16 1 were limi ted to pre~MY 2008 vehicles. Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product Having failed to make progress in hi s investigation, Stouffer enlisted the help of the Red X team , which likewise fail ed to make progress in the in vestigation_ I. February 2012: Evaluation On February 24, 20 12, K&S submitted to GM lawyer Ron Porter its first case evaluation for the . matter. 734 The . 10, 2010, in Georgia. case arose out of a two~car crash that occurred on March was driving her 2005 Cobalt through heavy rain when she encountered stand ing water in the roadway. _ who the accident report said was driving too fast for the conditions, lost control of her car, veered sideways into the southbound travel lane, and was struck on the passenger side by another vehicle. _ Coba lt then traveled rear first off the roadway and dropped approximately fifteen feet into a creek with rising water from the rain. According to the SDM data, at the time of the accident, the vehicle ' s power mode status was in Accessory. 735 The plaintiffs counsel never all eged at any time during the litigation that the front airbag should have deployed. On the contrary, plaintiffs expert agreed that it should not have. though the airbag was not at issue , the in iti al K&S evaluation - as in the _ _ 736 Even and cases - pointed out " the Cobalt likely lost power approximately three seconds before the The evaluation also reported extensive conversations with Sprague about the Ignition Switch being in the Accessory position, and why that might be: The SDM data also shows that the car was in "accessory" rather than the " run" mode immediately prior to the crash. Brian Everest and others at FPA have Case Evaluation (Feb. 24, 2012) rOOC ID 000003895797; GMNHTSA000252849j. SOM Crash Data (March 22, 2011), at 48 rOOC ID 0000028 186 12; GMNHTSA000224895]. of Kelly Kennett (May 14, 2013), at 20 [DOC ID 000001771 172 ; GMNHTSAOOOI 44290l Case Evaluation (Feb. 24, 2012), at 19 LoOe ID 000003895797; GMNHTSA000252849j. 162 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product looked at the data and concluded that the car could have transitioned from " run" mode to "accessory" mode in one of three ways: (I) the car's after~market security system could have interrupted power in the ignition switch so that the car transitioned to accessory mode; (2) there could have been a "bounce" in the ignition switch caused by external factors like standing water or some other internal "anomaly" in t~ion that caused the transition from the run to the accessory mode; or (3) _ could have intentionally or inadvertently turn ed the key from the run position to the accessory position as she was losing control of the car. Because the third option seems the least likely, our analysis at this point has focused on the after-market security system and a possible "bounce" anomaly in the Cobalt's ignition system .738 K&S observed that GM had seen this phenomenon before: "F PA engineers have seen instances in 2005~07 model year Cobalts where rough road conditions cause 'bounce' in the igniti on sw itch.,,739 FPA supervisor Everest (who worked on t h e _ technical evaluation) disputed that he or so meone on his team could have drawn any definitive causal connections at the time the _ evaluation was drafted;740 rather Everest told K&S that contact bounce was one theory that was being explored. 741 He also stated that he would not have referred to the power issue as an "anomaly.,,742 A few weeks later, on March 8, 2012, GM Legal Staff attorney Porter presented the _ matter to the Roundtable. 743 In his summary, Porter outlined how the SDM data was "confusing" and specificall y addressed the issue of how the power mode was recorded as being in Accessory: At this time we have identified 3 possible exp lanations for the accessory position data. I. _ turned the key to the acc position. This might have occurred if she was attempting to restart the engine after it stalled. 2. Ignition circuit "bounce." FPA has seen a few instances involving Cobalts operated on rough roads where Case Evaluation (Feb. 24, 2012), at 19 [DOC ID 000003895797; GMNHTSA000252849}. Case Evaluation (Feb. 24, 2012), at 19 [DOC ID 000003895797; GMNHTSA000252849}. ,\Cr.';';W or Brian Everest, April 29,2014. 741 J&B Interview or Brian Everest, March 27, 2014. 742 J&B Interview or Brian Everest, April 29, 2014. 743 _ Roundtable Case Summary (March 8, 2012), at 7~81DOC ID 000002818612 ; GMN HTSA000224895J. 163 Privileged and Confidenfial Protec1ed by AUorney~Client Privilege and as Allorney Work Product the ignition circuit goes to the ace position due to vibration. 3. Interference from the poorly designed and poorl y installed aftennarket al~tem. This mi ght have been the cause of the stalling condition for which _ took the car to [the dealer] several days before the crash.744 Among the Roundtable attendees were Jacl yn Pa lmer and Michael Gruskin. That same day, Palmer e-mailed Nowak-Vanderhoefand Kemp, along with GM engineering and PI employees, to schedule a "meeting so that Product In vestigations can update Legal, regarding the status of the in vestigation into Cobalt airbag nondeployments.,,745 On March 15, 20 12, Palmer met with Stouffer, Kemp, and others. According to Palmer, the other 746 participants included Nowak~Vanderhoef, Wachte l, Everest, Sprague, and Benavides. Palmer organized the meeting because she noticed the pa ssage of time and had heard from Sprague that there was minimal progress. 747 Palmer was "concerned that nothing seemed to have been figured OUt.,,748 At thi s point, neither Sprague nor the lawyers knew of the PRTS reports, the TSBs, the Wi sconsi n State Patrol report or the Indiana University study. Stouffer and Wachtel , on the other hand, had Elizabeth Kiihr' s fil e from 2005, which contained (a) 2005 news articl es from the New York Times , the Cleveland Plain Dealer, and the Sunbwy Times ; (b) several TREA D data reports regarding the Cobalt; (c) PowerPoint presentations, including presentations from the ISR and V APIR in 2005 ; (d) a cost estimate for changing the design o f the key; and (e) a copy of a PI Bulletin titled " Engine Stall s, Loss of Electrical Systems, and No DTCs." 744 _ 745 Roundtable Case Summary (March 8, 2012), at 7 rDOC ID 0000028 18612 ; GMNHTSA0002248951. E~mail [rom Jaclyn Palmer, GM, to Deb Nowak-Vanderhoef, GM , el al. (March 8, 2012) [DOC ID 00000 1745056; GMNHTSA000221020j. 746 J&B Interview o[ Jaclyn Palmer, April 24, 201 4 747 J&B Interview of Jaclyn Pa lmer, April 24, 2014. 748 J&B Interview of Jac\yn Pa lmer, April 24, 201 4. 164 Privileged and Confidenfia/ Protec1ed by AUorn ey~ Client Privilege and as Allorney Work Product According to Palmer, after the meeting Kemp no tifi ed Palmer that he would seek out an "executi ve champion" to oversee the FPE process 749 Kemp did not reca ll th is meeti ng, but recalled seeking out an exec utive champion in the com ing momhs. 2. 750 Ma rch 2012 : The Visit to the Daviso n Junkya rd and Re-Discove ry of the TSBs In his continuing support role, Sprague organ ize d a trip on March 28, 2012, to the Auto Sal vage Auction in Davison, Michi gan (the " Davison Junkyard") with electrical eng ineers who had been working with Stouffer on theories related to an electri ca l root cause. The group, includ ing John Do lan, Dolan's supervisor David Ca rey, Vipul Modi (the lead engineer responsibl e for airbag electronics) and Sprague, intended to extri cate t he S CM from the _ crash 2006 Cobalt.1S ! While moving cars into position to assess the S CM, one of th e engineers turned a key in a Cobalt igniti on and noticed that it turned extraordinarily easily.752 This discovery raised concerns. According to Do lan, the head of the Global Subsystem Leader Team on Passive Safety Control , the Ign iti on Switch turned so easily that the eng ineers decided that they sho uld measure the force required to turn the key. But they had not brought 749 750 J& B Interview of Jaclyn Pa lmer, April 24, 201 4. J& B Interview of Wi ll iam Kemp, May 9, 2014. J& B Interview of Jolln Dolan, Ma rch 21 , 2014; J&8 Interview of John Dolan, April 23, 20 14 Stouffer reponed that he found out about the trip after it occ urred. J&8 Interview of Brian Stouffer, May 13, 2014; J&B Interview of John Dolan, March 21 , 20 14. m J&B Interview of John Dolan, March 21 , 20 14; J&B Interview of John Dolan, April 23 , 2014; Cobalt Issue Activities Presentation, at 4 rDOC ID 000049367127; GMNHTSA0002826331, attached to e-mail from John Dolan, GM, to David Carey, GM, et al. (May 15, 2012) LDOC ID 000049367126; GMNHTSAOO0282632] ; J& B Interview ofVipul Modi, April 23. 2014; J&B Interview of David Carey, April 24, 20 14; 1&B Interview of Lisa Weber, April 24, 20 14; J&B Interview of James Churchwell, March 2 1, 201 4; 1& B Interview of James Churchwell, April 24, 20 14. 751 165 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product measuring tools with them. 7S3 They immediately dispatched an investigator to buy a fi sh scale from a local bait and tackle shop .754 Us ing the fish scale, the engineers measured the force necessary to move the key out of the Run position on a number of Cobalt vehicles at the junkyard.n ~ Dolan recorded the results and, after the visit, put them into a table. 7!16 He was surpri sed by the measurements: if they were correct, the measurements suggested a driver cou ld hit a pothole and the Ignition Switch might .. 757 move out ate f h Run position. According to Dolan, the next day he went to the office and pulled the Cobalt warranty data to determine whether there were customer complaints regardi ng the Ignition Switch moving out of the Run position.75~ To his surprise, he discovered many customer complaints and also the TSS that described both the low torque and the phenomenon of knocking the Ignition Switch out of Run and into Accessory. That day, he elevated hi s di scovery to Carey, who sai d that he presented the issue to Wachtel and to the Product Investigations group. According to Carey, he had a conversation with Wachtel in which he expressed that the torque issue could be significant to the ongoing Cobalt investigation;759 Wachtel did not recall this conversation. 76o The team also sent the infonnation to Stouffer, who started asking que stions of component engineers. 76 ! Although Stouffer still continued to question the probative value of the TSBs, they were now m J&B J&B m J&B 756 J&B m J&B h~ J&B 759 J&B 760 J&B 761 J&B 754 [ll!crvicw Interview Interview Interview Interview Interview Interview Interview Interview of JO[ln Do[an, March 21 ,2014; J&B of John Dolan, March 21 , 2014; J&B of John Dolan, March 21 , 2014; J&B of John Dolan, March 21 , 2014; J&B of John Dolan, March 21 , 2014; J&B of John Dolan, March 21 , 2014; J&B of David Carey, May 22, 2014. of Douglas Wachtel, May 22, 201 4. of John Dolan, April 23 , 2014. 11l!crvicw of John Interview of John Interview of John Interview of John Interview of John Interview of John 166 Dolan, April Dolan, April Dolan, April Dolan, April Dolan, April Dolan, April 23, 2014. 23 , 2014. 23 , 2014. 23 , 2014. 23 , 2014. 23 , 2014. Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product known for the first time by Sprague, and they were made available to in-house and outside counsel 3. April 2012: Punitive Da mages Eva lu ation a nd Another Wa rn ing Abo ut On Apri l 18, 2012, the law firm Eckert Seamans submitted a case evaluation to GM lawyer Jaclyn Palmer for the case. 762 The FPA engineer assigned to the case was Sprague, who had been present at the Davisonjunkyard trip . Thi s was the first evaluation for which outside counsel knew of the TSB - even though Stouffer had known of it since 20 I I and Wachtel since 2005. On December 13, 2009, was in the right front passenger seat in a 2005 Cobalt when the driver lost control of the vehicle on "black ice." The Cobalt ran off the roadway to the right, trave ling 60 feet off-road before striking two trees. The airbags failed to deploy and the SDM showed the vehicle in Accessory mode at the time of the accident. 763 The investigating officer concluded that _ was unbelted. _ sustained head injuri es and rib fractures. Case Evaluation (Apri l 18, 2012) [DOC ID 000001660990; GMNHTSAOO022 1074]. SDM Crash Data (Jan. 20, 2010), at 1, 3, 5 LDOC ID 000001660887; GMNHTSAOO0308798j. 167 Privileged and Confidenfia/ Protec1ed by AUorney~ Client Privilege and as Allorney Work Product CASE EVALUATION GENERAl.. MOTORS I..LC; GENERAL MOTORS COMPANY; o The SDM recorded a non-deployment event.(A copy of the CDR Rcport is atlacbed hereto as Exhibit " E" ). SinS'S' tbf Coba!! WI! in t he AsS'fUory Mode. in~tt9.d of Run Mode at thS' time of Ihe cra~h. the algoril hm thai the SDM runJ to dettnninc whether 10 dt ploy the . i rbag~ w.~ disabled. Thtrtfore the SDl\{ wu incapable of deploying the alrhus l'§ardll'S!< of the unrlty of the Impact. It recorded the e~'ent as a Non-Deployment Event, and would have recorded the event as such no matter how significant the crash event. In addition, John Spmgue advises that muc h of the typical pre=crash data, such as vehicle speed, yaw mIS, brake application. etc. is 1'«'orded as "invalid" on the CDR Report because the Cobalt ......as in AccesS()ry Mode. OM will be forced to explain that the airbags did not dcploy in this crash because the Cobalt was in Accessory Mode. While the S DM responded appropriately to the input it received from the Body Computer, it is not the expected result for the customer. It will be difficuh to explain why the ignition switch toggled to Accessory Mode simply from running off-ro ad . GM will also be forced to contend ....ith other incidents, some of which resulted in deaths, duc to the nondeployment of the frontal airbags in the 2005-2007 Cobalt. Those other incidents put GM at risk for imposition o f punitive damages in \Vest Virginia. Eckert & Seamans ' case evaluation spec ifically noted the TSB. 764 The evaluation explicitly drew a link between the airbag non-dep loyment and the vehicle' s power mode status of Accessory: GM issued [a Service Bulletin1 to address the problem of the dri ver inadvertently turning off the igniti on by bumping the key chai n with hislher knee while turning the steering wheel if the steering wheel was adj usted too low. This Bulletin addresses a similar problem as that seen in the field where the key in the igniti on switch in the 2005 Cobalt could togg le from the Run mode to the Accessory mode by traveli ng off-road or over rough terrain. 765 The eva luation explained that airbags cannot dep loy when the ignition is in the Accessory position: Case Evaluation (April 18, 2012), at 4 [DOC ID 000001660990; GMNHTSA00022 1074]. Case Evaluation (April 18, 2012), at 4 LDOC ID 000001660990; GM NHTSA00022i074 j. 168 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product Since the Cobalt was in the Accessory Mode, instead of Run Mode at the time of the crash, the algorithm that the SDM run s to determine whether to deploy the airbags was di sab led _Therefore, the SDM was in capabl e of de pl oying the airbags, regardless of the severity of the impact. 766 The evaluation also explained that Sprague believed that the Cobalt switched from Run to Accessory mode when the car ran over rough ground before the impact. 767 The evaluation furth er explained that there were other incidences of Cobalt airbag nondeployment in which "the system was not in Run Mode at the time of impact," thereby disabling the ai rhags. 768 The lawyer explained that " [i)t wi ll be difficult to explain why the ignition switch toggled to Accessory Mode simply from running off_road.,,769 The lawyer pointed out that prior non-dep loyment events in 2005 to 2007 Cobalts " had resulted in deaths," putting GM "at risk" for a punitive damage award in thi s case: GM will also be forced to contend with other inc idents, some of which resulted in deaths, due to the nondeployment of the fronta l airbags in the 200 5 ~2007 Cobalt. Those other incidents put GM at risk for imposition of punitive damages in West · .. 770 V Irglflla. A week later on April 25, 20 12, Palmer presented the _ case to the Roundtabl e. 771 Attendees included Larry Buonomo and Nabee l Peracha, a newly hired lawyer. 772 Palmer exp lained in the Roundtable summary that based on the SDM data, the Cobalt was in Accessory mode, which meant that the airbags could not deploy.773 Case Evaluation (April IS, 20 12), at 8 rOoe ID 000001660990; GMNHTSA0002210741 . Ca ~e Eva luation (A pril 18,2012), at 8 [DOC ID 00000 1660990; GMNHTSA0002210741_ Case Evaluation (Apri l IS, 2012), at S LDOe ID 000001660990; GMNHTSA00022 1074 J. Case Evaluation (April IS, 2012), at 19 LDOe ID 000001 660990; GMNHTSA000221074j. Case Evaluation (April IS, 2012), at 19 rOoe lD 000001660990; GMNHTSA000221074l Roundtable Case Summary (April 25, 2012) rOoe ID 000001 660996; GMNHTSA00022107S1. I Summary (April 25, 2012), at 2 rOOC ID 000001660996; GMNHTSA00022107Sl; Roundtable Case Summary (April 25, 2012) (bearing the handwritten notations of attendees), at 5 GMNHTSAOO02211511. Summary (April 25, 2012), at 2 LDOC ID 000001660996; GMN HTSA00022107SJ. """.m;· 169 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product Palmer also wrote that GM 's PI group had been investigating the Cobalt issue "for the past year," and that there was a relevant service hulletin about " low ignition torque/effort" : Product investigations has been investigating the Cobalt issue for the past year, but has not yet confirmed a root cause. There is a service bulletin related to the potential for the driver to inadvertently turn off the ignition (by contacting a large and/or heavy key chain with the knee) due to low ignition torque/effort, which is one theory that is currently being pursued. The possibility of a change in ignition status due to extreme off road events was previously explored. 774 The Roundtable granted settlement authority but the case did not settle at that point. 775 4. May 2012: Mec hanical vers us Electr ical Root Ca use At th is time, Stouffer was still working to determine the root cause of the Cobalt nondeployment issue and was investigating both electrical and mechanical components and subsystems, including the SDM and the Ignition Switch. 776 As part of Stouffer's ongoing in vestigation, he sought information about the Cobalt SDM from the engineers in the electrical department. 777 On May 2, 2012, Stouffer asked electrical engineers John Capp and Li sa Weber to meet with him and with Doug Wachtel. 778 Capp responded: " I don't have the hi story on thi s topic ... , but my understanding is that [we] have taken the position that engineering does not support any further steps.,,779 Capp was referring to a potential field solution that would modify Roundtable Case Summary (April 25, 201 2), at 2 [DOC ID 000001660996; GMNHTSA000221078]. Roundtable Case Summary (April 25, 2012), at 2 LDOC lD 000001660996; GMNHTSA00022i078 j. Douglas Wachtel, GM, to David Carey, GM, ef al. (March 29, 2012) [DOC lD 000000005003; GMNHTSA000221035]. m J&B Interview of John Capp, April 24 , 2014. m E~mail from Brian StoulTer, GM, to John Capp, GM, el al. (May 2, 20 12) rDOC 10 000003248745 ; GMNHTSA00061184Ij. 779 E-mail from Brian StoulTer, GM, to John Capp, GM, el al. (May 2, 20 12) [DOC 10 000003248745 ; GMNHTSA00061184Ij. 170 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product the SDM to allow airbag deployment while the car was in " Accessory. ,,78o However, this proposed fi eld soluti on did not gain traction within the e lectrical engi neeri ng group.7St The electrical engineers be li eved, correctl y, that the SDM was working as it had been designed and feared any alterations to the SDM, a corporate common component, might cause probl ems in vehicle lines that were not expe riencing the non~deployment issues. 782 Wachtel forwa rded Capp's e-mail response to Carmen Benavides. Benavi des, in turn, sent the e-mail chai n to Bill Kemp, and asked "[c]an we discuss the attac hed. I was unaware that a decision was made to not support. ,,783 At thi s time, Kemp took action and asked Terry Woychowski to "champi on" the Cobalt non~deployment investigation. 7 l!4 Woychowski was Vice President of Global Quality and 785 Veh icle Launch and a member of the committee in charge ofrecalls - the Executi ve Field Action Decision Committee (" EFADC,,). 786 According to Woychowski , a "c hampion" is an executive who helps a team remove roadb locks and obtain resources. 787 Shortly after enlistin g Woychowski, Kemp organized a meeting on May 15, 20 12, regarding the " Cobalt Airbag Issue." J&B Interview of John Capp, April 24, 2014. J&B Interview of John Capp, April 24 , 2014 ; J&B Interview of Kristin Siemen, April 17, 2014. 782 J&B Interview of John Dolan, April 23 , 201 4; J&B Interview orUsa Weber, April 24, 2014; J&6 Interview of David Carey, April 24, 20 14; J& B Interview of Kristin Siemen, April 17,20 14. 783 E~mail from Cannen Benavides, GM, 10 Wi lliam Kem p, GM (May 2, 2012) rDOC ID 000003248745 ; GMNHTSA000611841 ] (emphasis in original). 784 J&B Interview of Wi ll iam Kemp, May 9, 201 4. 785 J&B Interview of Terry Woyehowski, March 21 , 201 4. 786 J&B Interview ofCannen Benavides, March 25, 201 4; J&B Interview of Brian Everest, April 29, 2014. 787 When asked, Woychowski staled he did nOI view himsclfas the "champion" for Ihe Cobalt airbaglignition switch issue, but numerou s olher individuals involved with the investigation identified Woychowski as the "champion." J&B Interview of Terry Woychowski, March 21 , 2014; J&B Interview of William Kemp, May 9, 2014 ; J&B Interview ofCannen Benavides, March 25, 2014; J&B Interview of Brian Stouffer, May 9, 2014. Benavides and Kent have described the appointment ofa champion as "out of process" insofar as it was not proceeding through the ordinary course o f the FPE sleps. J&B Interview of Gay Kent, March 18, 2014; J&B Interview ofCannen Benavides, March 25, 201 4. It appears that almost all of the same individuals who would have participated in the FPE process were involved, however, wilh the addition ofa high~ra nking "champion," i. e., going "out of process" reflected more and highe r-level attenlion, not less. 780 781 171 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product Attendees included personnel from PI and engineering, including: Kemp, Carmen Benavides, Doug Wachtel, Woychowski , John Dolan, Vipul Modi and others_ 788 According to Modi , thi s was a meeting of high leve l managers and directors, and he and Dolan were not invited to present, but rather were asked to be there to lend their support and to answer any questions that . ht afl.se. 789 ITIlg In advance of the meeting, Dolan prepared a PowerPoint presentation for Dave Carey, his director. 790 The primary purpose of the Powe rPoint presentation was to convey Dolan ' s theory that the Igniti on Switch - not the SDM - ca used the Cobalt airbag non~deplo ym e nts . Dolan 's understanding was that due to the low torque of the Ignition Switch, a pre~crash bump or jolt caused the key to move out of the Run position before the crash, thus leading to the airbag nondeployment at the time of impact. Dolan arrived at thi s concl usion based on (1) hi s observations as to how easy it was to tum the keys in the Cobalts at the Davison Salvage Yard; (2) the Technica l Service Bulletin he di scovered the day afte r hi s trip to Davi son; and (3) his review of the portion of the SDM that records crash data and review of addi tional crash records. 791 Dolan 's PowerPoint included a poss ible expl anation for why the Ignition Switch was recorded in the " Run" position for so many of non-deployment incidents - a point that confused Stouffer. A slide in the Power Point stated that " [i]n the vehicl es where EDR recorded RUN power mode [ . . .] the crash sensing algorithm [is] recorded as bei ng inacti ve. The crash 788 E-mail from Dawn Wands,GM, to JeITrey Konchan, GM , el 01. (May 9, 2012) [DOC ID 000005178995 ; GMNHTSA00022I 079l; J&8 Interview of Lisa Weber, April 24, 2014. 7~~ J&B Interview ofVipul Modi, May 21 , 2014. 790 Cobal t Issue Activities Presentation [DOC ID 000049367127; G!VfNHTSA0002826331, attached to e-mai l from John Dolan, GM to David Carey, GM, ef 01. (May 15, 2012) [DOC ID 000049367 126; GMNHTSA000282632j. 791 J& B Interview of John Dolan, April 23 , 201 4; J&8 Interview of John Dolan, May 16, 20 14. 172 Privileged and Confidenfia/ Protec1ed by AUorney~ Client Privilege and as Allorney Work Product algorithm is inactive during the initial three seconds after transitioning from power mode OFF/ACC to RUN to run selfdi agnostics.,,792 Crash Records Review of EDR and Incident Reports - The SDM ind icat es tha t the vehicle power mod e is Ace o r OFF in the majority of report ed inci dents. In thl! vehicles wh ere EDR rei:orded RUN power mode holve the crash sensing ollgor lthm recorded as being inactive. The crash olIgor ithm is inactive dur ing Ihe initioll 3 seconds olfter trJnsi tioning from power mQcIe OFF/ACe 10 RUN 10 nm sell diaenoslics - All reviewed crash inciden ts had some form of pre-impact rough road input. - Some SDMs have no record ed history of an event - indicati nB tha t the module was asleep duri ng t he even L If a crash event has start ed w hile in power mode RUN, any transition from power mode RUN is igno red unti l t he crash event is completed - and the power mo de is recorded as RUN in EO R. Th e power mode behavior of t his SD M is th e same as all other SD Ms in GM vehicl es. At the time Dolan created this presentation, he believed that accident vehicles whose power mode status was recorded as " Run " but whose airbags failed to deploy likely experienced a change in status from Run to Acce ssory, and then back to Run . Because Dolan understood that it still took three seconds after a move from Accessory back to Run for the airbag to become functional , Dolan be li eved that the crash event occurred within three seconds of the power mode changing back to Run, but before the SDM had time to reset and trigger airbag deployment. Criti ca ll y, Dolan 's PowerPoint slide went on to say that "if a crash event has started wh ile in power mode RUN , any transiti on from power mode RUN [to ACCESSORY] is ignored Cobal t Issue Activities Presentation, at 2 rDOC ID 000049367127 ; GMNHTSA0002826331, attached to e-mail from John Dolan, GM to David Carey, GM, el at. (May 15, 2012) [DOC ID 000049367 126; GMNHTSA000282632J. 792 173 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product until the crash event is completed - and the power mode is recorded as RUN in EOR. ,,793 Neither the plain tex t of this bullet point nor its imp li cations are crystal clear. On its fa ce, the bullet point states that a transition from Run to Accessory would not be recorded if a crash began with the power mode in Run - the EOR wou ld still record the key as being in Run. Dolan explains now that the point of thi s statement was to convey that if the Ignition Switch changed position from Run to Accessory at the moment of the crash (that is, actua l impact with a large object), the SOM would ignore the change and still deploy the airbags. 794 According to Dolan, these two bu ll et points were intended to articulate his view at the time tbat a pre-impact event caused a change in the ign ition switch po sition from Run to Accessory and therefore airbag non-deployment. Wh ile the sw itch may have moved back to Run at some point prior to actual impact, it was too late for the SOM to finish resetting and trigger airbag deployment upon impact. Had it been otherwise, and even if the key moved out of Run at the time of impact, the airbags would still have dep loyed. It is unclear whether thi s topic was the subject of any discussio n at the May 15, 2012, meeting. Dolan recalls di scussion of hi s prese ntation generally and of slide 2 specifically.795 Stouffer does not recall any disc uss ion of the slide or bullet points, but he does recall Dolan Coba lt Issue Activities Presentation, at 2lDOC ID 000049367127; GMNHTSA000282633 J, attached to e-mail from John Dolan, GM to David Carey, GM, el al. (May 15 , 2012) [DOC 10 000049367 126; GMNHTSA000282632]. 194 The rcrerence to "crash event " in rhe bullet poim refers to the vchicle hilling a tree or another large object. It does not refer to the vehicle hitting rough, oIT-road terrain or shmbs. 1&8 Interview of John Dolan, May 16, 201 4. After the recall, GM developed a new theory as to why the SDMs recorded a mn power mode in some of the Cobalt non-deployment incidents. GM's current understanding is because the Cobalt SDM records vehicle data in I second ill\ervals or "snapshots," ifan ignition change from Run to Accessory occurs betwecn I second and 50 mi ll iseconds (the time it takes the SDM to detect a change in power mode) before the crash, (he S DM will record the vehicle as being in mn, but the airbag will not deploy because the ignition switch is actually in accessory. This theory is currently being validated with the assistance orColltinental. J&B Interview of John Dolan, May 16, 2014. 195 J&B Interview of John Dolan, May 16, 201 4. 793 174 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product presenting at the meeting. 796 By contrast, Carey and Mod i remember the meeting as being a general in troduction to th e Cobalt probl ems as opposed to a fo rmal presentation.797 Carey recalled that he did not present the slides at the meeting, despite his ro le as di rector of the SDM grouP. and he also did not remember Dolan making a presentation at the meeting. 798 Modi does not believe the PowerPoint presentation was given at the meeting, rather the onl y slide Modi thinks may have bee n used by Stouffer in the meeting was Dolan's data slide from the firs t Davison Salvage Yard visit. 799 That slide (when the measurements are co nverted to Newton~ centi meters) showed that the torque needed to turn the Igni tion Switch was above spec ifi cation on one vehicle, at specifi cation on two, and below specificati on on four. 8OO 796 1&8 Interview of Brian Stouffer, May 9, 2014. Stouffer's currcnt understanding of this bullct point is that it suggests that, if the power mode changcs to Accessory from Run after the crash event starts, but before the airbag deploys, the SDM will record the power mode as Run. Because an airbag makes a decision as to whether to deploy within .017 seconds, Stouffer believes now that the scenario described in the bullet point seems like a very remote possibi lity. He believes now that it wou ld be unlikely for the power mode to change nearly instantaneously at the time of the crash in each of the crash events in whieh the power mode was recorded as Run. 797 J&8 IllIerview of Vipul Modi, May 21 , 2014. Carey docs not recall a specific discussion about Dolan's present.ltion or slide 2 at the May 15, 2012 meeting. J&B Interview of David Carey, April 24, 2014. Specifically, Carey does not recall Dolan standing in the fro nt of the room and giving a PowerPoint presentation. J&6 Interview of David Carey, May 22, 2014. 798 1&8 Interview of David Carey, May 22, 2014. 799 1&8 Interview of Vipul Modi, May 21 , 2014. Kemp recalls nothing whatsoever about the mccti ng. J&B Interview of William Kemp, May 22, 2014. John Zuzclski recalled attending the May 15, 2012 meeting, and thought it was possible that Brian Stouffer gave a PowerPoint presentation, but was 1I0t certain. J&B Interview of 10hn Zuzelski, May 19, 2014. Kristin Siemen recalled anending one meeting on the Cobalt issue in the spring of 20 12, but did not remember the date of the meeting or any presentation being shown there. J&B Interview of Kristin Siemen, May 21 , 2014. Jeffrey KOllchan did 1I0t recall the exact meeting date and did not r"all the PowerPoint presentation at al\. J&B Interview of Jeff Konchan, May 22, 2014. Terry Connolly docs not know John Dolan and does nOl recal l 311ending any mccli ng on May 15, 2012. J&B i1l1erview of Terry Connol ly, May 22, 20 14. David DeFrain did not recall attending the May 15, 2012 meeting. J&B Interview of Dave DeFrain, May 22, 20 14. Lisa Weber attended the May 15, 2012 meeting, but did not recall Dolan's slides being presented there. J&B Interview of Lisa Weber, May 22, 2014. 800 Dolan explained that one can convert the fo rce data (measured in pounds) that arc renectcd in his presentation to torque (measured in N-cm) by multiplying fo rce by the distance from the center of the key to the edge of the key slol. l&B Interview of John Dolan, May 22, 2014. For original force measurements, see Cobalt Issue Activities Presentation, at 12 [DOC ID 000049367127; GMNHTSAOO0282633], attached to e-mail from John Dolan, GM to David Carey, GM, el al. (May 15, 2012) l DOC ID 000049367126; GMN HTSAOOO282632J. 175 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product Regardless of the content of the meeting, Stouffer, according to Dolan, agreed at the time that th e SDM was not the cause of the airbag non-deploymento; and that both he and Stouffer thought it did not make sense to look for solutions to the airbag non~deplo yme nts that were unrelated to the Ignition Switch. SOl Stouffer's in vesti gation indeed began to shift its focus to mechanical root causes and the action items developed at the meeting re flect as much.802 Doug Wachtel 's notes from the May 15, 20 12, meeting entitled "Cobalt Ignition Switch Assignments" were sent to Stouffer the following day, and reflect a recognition th at the Igniti on Switch could be the root cause ofthe non-dep loyment issue.803 At the same time it was looking for root cause, the investi gation also pursued soluti ons to the non-deployment issue. Namely, the SDM engineers were tasked with developing a means by which the SOM could remain active (i.e. , allow for airbag deployment) for some period of time after the power mode statu s moved out of Run . s. May 2012: The Second Trip to the Davison Junkyard On e of th e steps that emerged from the May IS, 2012, meeting was a trip to the Davison JunkYard on May 22, 2012, to measure the torque and pull down force associated with rotating the key in approximatel y 40 Cobalt, G5 , HHR, and Ion vehicles. Stouffe r and Sprague were J&B Interview of Jolm Dolan, May 16, 2014. E-mail from Douglas Wachtel ,G M, to Brian Stouffer, GM (May 16, 20 12) rDOC ID 000000005007; GMNHTSA00022 I0971, attaching Doug Wachtel's notes from the May 15,2012 meeting rDOC ID 000000005008; GMNHTSA00022I 098]. 803 E-mail from Douglas Wachtel, GM, to Brian StouITer, GM (May 16, 201 2) rDOC ID 000000005007; GMNHTSA000221097], attaching Doug Wachtel's notes from the May 15 , 2012 meeting [DOC ID 000000005008; GMNHTSA00022I 098 J. SOl 802 176 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product accompani ed by a summer intern, David Caples. 804 Before the trip, they obtained a list of availabl e Cobalt, HHR, and Ion vehicles. 80S At the salvage yard, Stouffer, Sprague, and Caples obtained measurements - this time using a genuine torq ue wrench rather than a fi sh scale - for the approx imate ly 40 Cobalt, G5, 807 HHR, and Ion vehicles. 806 Caples then graphed these measurements. In a May 23, 201 2 e~ mail , Caples wrote that there " is a noticeable dip in the torque for Model Years 2005 and 2006." Stouffer responded o n the same date, " I wasn' t sure we would see any trends in the data, but you've obvious ly found one. This will be helpful as we work to detennine root cause."S08 Stouffer observed that there was a slight upward trend in the data and that, to a certa in extent, earlier model years had lower associated torque values than later mode l years. R09 However, Stouffer al so observed fairly similar torque values for some vehicles in the 2007 and 2008 model years. 8 10 According to Stouffer, because the torque va lues in the 2007 and 2008 vehic les were si milar, he could not conclude that the Ign ition Switch c hanged in 2008. 811 Because in Stouffer's J&B Interview of Brian StolifTer, May 9, 201 4. J&B Interview of Brian StoufTe r, May 9, 201 4; e~mai l from John Sprague, GM, to Brian StoufTer, GM , el of. (May 17, 2012) [DOC ID 0000000460 II ; GMNHTSAOOO22I 0991. S06 J& B Interview of Bria n StoufTer, MlIY 9, 201 4; J&B Interview of David Caples, May 8, 2014. SI)7 J&B Interview of David Caples, May 8, 201 4; J&B Interview of Brian Stouffer, May 9, 2014. SI)8 E~ m ail from Brian SlOufTer, GM, to David ClIples, GM, el of. (May 23 , 20 12) [DOC ID 000000005033; GMNHTSA000221 122J , attaching grllph of ignition switch measurements lDOC ID 000000005034; GMNHTSA00022 I 124 ]; J&B Interview of Brian StolifTer, May 9, 2014. 8(t9 J&B Interview of Brian StolifTer, May 9, 201 4. SIO 1&B i1l1erview of Brian SlollfTcr, May 9, 201 4. In l.hc fall of 2013, aflcr recciving SlIhbaiah Ma ll adi ' ~ rcpon and thc Commodity Validation Form from Delphi , StoufTer reexamined his Davison Salvage Yard data and mapped the vehicle build dates onto the crashes to try to detennine whether there was a breakpoint in 2006 or 2007 aner which GM could be confident that the switches insta lled in Coba lts had higher torque values. J&B Interview of Brian StoufTer May 9, 2014. Because severa l of the n o n~dep l oy m en ts occurred in later-built MY 2007 vehicles, StoufTer could not identify a breakpoint before MY 2008. StoufTer speculated that within MY 2007 vehicles, there may have been "mixed stock," meaning that the defici ent switches may have been mixed with the good switches for the 2007 model years. J&B Interview of Brian Stouffer, May 9, 201 4. SII J& B Interview of Brian StoufTer, May 9, 201 4. 8(}.1 SI)5 177 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product view, the Davison data did not explain the absence of the non~deplo y ment condition in the 2008 model years, he was left continuing to look for the root cause.S12 6. June 2012: Kemp Persuades Federico to Help Two weeks after the May 15, 20 12 meeting, Wachtel e~mailed Kemp, copying Benavides and Stouffer, stating , "[t]he action items that were identified at your meeting on 5/ 15 have been completed, and we are ready for a follow-up meeting."S13 The next meeting would occur on June 21 , 2012. In earl y June, after Woyc howski retired, Bill Kemp asked executive director for Global Vehicle Integration Jim Federico (in hi s capac ity as Chief Engineer) to replace Woychowski.814 Although Kemp recall s asking Federico to step in as "champion" because he respected Federico and thought he would get things done, Federico did not remember Kemp using the term "champion," and understood hi s role to be "energiz[ing] the team " to find solutions for the Cobalt airbag non~deployment problem.8 1.5 Federico also understood the purpose of his team was to be a "solutions" team, while a different investigation sought the root cause. 8 16 Federico asserted that had he been leading the investigation , and understood the matter was urgent, he J&B Interview of Brian Stouffer, May 9, 2014. StoufTer was questioned about this sal vage yard trip during hi s May 1, 2013 deposition in _ Stouffer _ Deposition Transcript at33-39lDOC ID 00000177125IJ. Plainti IT's counsel showed~r the e-ma~ Caples and asked whether Stouffer found "a noticeable dip in the torque for ' 05 and ' 06." Stouffer . Deposit ion Transcript at 38-39. Stouffer said "if you do an average, Ihere's a dip from '03 10 ' 05 and ' 06, there sa slighlilpward trend, billnol comple!e separmion for the res! oflhe ~ears." StollfTer " Deposition Tmnscript at 33-39lDOC ID 00000177 1251j. 13 E-mail from Douglas Wachtel, GM, to William Kemp, GM (May 30, 20 12) lDOC lD 000002811396; GMNHTSA0004632071. See also e-mai l from Douglas Wachtel, GM, to Brian Stouffer, GM (May 16, 2012) rDOC ID 000000005007; GMNHTSA0002210971, attaching Doug Wachtel 's notes from the May 15, 2012 meeting rDOC ID 000000005008; GMNHTSA000221098j. Sl4 J&B Interview of Will iam Kemp, April 17, 2014; J&B Interview of William Kemp, May 9, 2014. Sl5 J&B Interview of Jim Federico, May 9, 201 4. Sl6 J&B Interview of Jim Federico, May 9, 201 4. Sl2 178 Privileged and Confidenfia/ Protec1ed by AUorn ey~ Client Privilege and as Allorney Work Product would have behaved differently - cl earing hi s schedule and working aggressively to solve the prohl em H I7 7. June 2012: The Stouffer Investigation Continues and He is Told There Are No Relevant Part Changes By the spring or summer of20l 2, Stouffe r had asked DeGi org io and Brian T hompson (DeGiorgi o' s supervisor) whether there had been any ha rdware changes to the switch. Both stated that they were confi dent that there had been no hardware changes to the switch. s lS At Stouffer's request, on June 25, 2012, Thompson and DeGiorgio sent Stouffer a spreadsheet which included comments on the effect of various work orders. s l9 Although in 2006 DeG iorgio had authorized the redesign of the Ignition Switch to inc rease the torque using a work order for the electrical system, Thompson and DeG iorgio described that very work order as having "NO AFFECT ON THE ROTATIONAL TORQUE OF THE SWITCH! OR LOCK HOUSrNG ASSEM BLY SYSTEM. ,,'20 Stouffer reli ed on their statement that no changes had been made. 82 1 He did not search for and review the Fonn 3660 that DeGi orgio signed that did specify the change in the detent and spring; this document, whi ch had been in GM ' s system since June 1, 2006, showed that there was a change that would have affected the torque required to turn the 19l1i tiol1 Switch . Stouffer did, however, personally review the work orders rel ated to the Ignition Switch on the GM J& B Interview of Jim Federico, May 9 , 2014 J& B Interview of Brian Stouffe r, April 4, 201 4. SI9 J& B Interview of Brian Stouffe r, April 4, 201 4; e-mail from Brian Thompson, GM to Brian Stouffer. GM (June 25, 20 12) rOOC 10000000005055 ; GMN HTSAOO0250057], and attached Annotated Spreadsheet of Ignition Sw itch Work Orders rOOC ID 000000005056; GMNHTSAOOO250OGOl ~20 Anno tated Spreadsheet of Ignition Swi tch Work Orders at rOOC 10 000000005056; GMNHTSA000250060.], attached to e-mai l from Brian Thompson, GM, to Brian Stou ffer, GM (June 25, 2012) rOOC ID 000000005055 ; GMNHTSA000250057]. 821 J& B Interview of Brian Stouffe r, April 4, 201 4. 817 SIS 179 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product system, which he had obtained from Sprague when Stouffer began working on th e investi gation, and did not observe any that indicated that hardware changes we re made to th e switc h_822 The mi sinformation provided by DeG iorgio caused Stouffer to maintain his beli ef that the Ignition Sw itch alone could not be the root cause of airbag non~d eploy ment. Stouffer furth er noted that hi s interpretation of the Davison Salvage Yard Study did not suggest to him that there had been a change to the switch between the 2007 and 2008 model years. 823 Over the cou rse of the summer and early fa ll , Stouffe r asked the engineers responsible for the steering column and the lock/igniti on cylinder whether there were any relevant hardware changes, including any changes that would have affected the torque required to turn the key in the Cobalt.1l24 Stouffer was told by those engineers that no rel evant changes were made to the steering co lumn, ignition cylinder, or key. Although at this point, the investigation was primarily focused on exploring possible mechani cal changes, in summer 20 12, Stouffer also asked the BCM engineers to notify him of any relevant electrical changes. 825 Stouffer was similarly told that no relevant changes were made to that component. 826 8. June 2012: The Shipp Expert Report Provides th e Correct Answer In June 2012 , the plaintiff's expert in the _ case, Erin Shipp, P. E. , issued her report, which was provided to GM during the courSe of litigation. J&B Interview of Brian Stouffer, April 4, 2014. J&B Interview of Brian Stouffe r, April 4, 2014 824 J& B Interview of Brian Stouffer, Apri l 4, 201 4; e-mail from Peter Judis, GM, to Brian Stouffer, GM (June 28, 20 12) lDOe 10 00000046058 ; GMNHTSA000320509j; e-mail from Brian Stouffer, GM, to Mark Beauregard, GM (Aug. 27, 2012) [DOC ID 000000005079; GMNHTSA00022 I 166]; e-mail from Mark Beauregard, GM, to Brian Stouffer, GM (Sept. 4, 2012) rDOe ID 000000005086; GMNHTSA000221169l ~2) E-mail from Alexander Ballios, GM, to Brian Stouffer, GM (June 14, 2012) rOOC ID 00000046051 ; GMNHTSA000273952j. 826 E-mail from Alexander Ballios, GM, to Brian Stouffer, GM, (June 14, 20 12) [DOC lD 00000046051 ; GMNHTSA000273952j. 821 823 180 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product Although outside counse l and Pa lmer were aware of the Ignition Switch TSB at this time, Shipp had not been provided with discovery material at this stage H27 Nonetheless, in preparing her report, Shipp located both the TSB and the 2007 Ind iana University study, whi ch was now known to GM lawyers and engineers working on the Igniti on Switch issue for the fi rst time. As to the 2005/2006 TSB, Shi pp wrote: General Motors knew that the design of the ignition switch was improper and could cause power interruption during [sic]. Th is would include the event as described in the bu lletin [a knee striking the key or keycha in], but also during events where the vehicle is subj ect to very rough terra in such as pre-crash events, it is likel y that the driver will move within the cab in and those movements would incl ude impact with interior components incl uding the igni tion key that is inserted 828 in the igni tion switch. She then po inted out that airbags will not deploy when the key is in the Accessory position, based on her review of wiring documents: I reviewed 2005 Chevrolet Cobalt wiring diagrams and have fo und the air bag system is active in the start/run and crank/run ignition positions but not active in 829 the accessory position. Shipp also referenced the fi ve-year old Indiana Uni versity study: [The data] in dicates that between 2 and I seconds before the start of the impact the key was turned to the accessories pos ition in the ignition switc h. These indicators are in agreement with the data from the crash report from the NHTSA E-mail rromJennirer Brooks, oulsidecounsel al Eckcn Scamans, 10 Jaci n Palmer, i GM, enclosing Shipp Report rDOe ID 000001660978]. (July 6, 2012) [DOC ID 00OO01662121 ~iscovery Tracking Log ror ~2~ Erin M. Shipp, Engineer's Report o ~ Crash,_ 1'. GM(Julle I , 012), at 7 rDOe ID 00000 1662 122 ; GMN H TSA000309665~ . 829 Erin M. Shipp, Engineer's Report 0 Crash,_ 1'. GM (June 12, 2012), at 7 [DOC ID 00000 1662 122 ; GMNHTSA000309665 . 827 18 1 Privileged and Confidenfia/ Protec1ed by AUorney~ Client Privilege and as Allorney Work Product database by Indiana Uni versity. 83o _......-- - -,- -_----__..-,-_....___ -,. . . ._. . . _.1*_____. ..._ ,_ _- ------"'_ . . .__ . . . . . __ . . _-ty-..-..,y 0.-.. -.0..-_ ......_ .............. ________ ... .. ... .... ""'... ...... ... _...,. ..... _"""'" , .. _10>50-.-..- ....... .....- .......... _ ........ _ . _ T... _ _ .......... _ ........ _ . . _ _ _ .... _ ..... ~ " .......... _awMuI _ _ ...... _ " ... ac Over the next few months, in anti cipation of the DFSS, Sto uffe r reached out to a number of different component engineers to disc uss possible so lutions. One so lution being contemplated in vol ved a modi fication to the Cobalt steering column shroud (bas ica lly, a cover to guard the ignition fro m contact with the knee).886 Specifically, one engineer suggested replacing the current Cobalt steering column shroud with a shroud that "provid[ed] more knee c1 earance.,,887 The engineer suggested that the Ion column shroud " ha[d] these design elements," and replacement of the shroud might be a solution,888 but others felt that the Ion might have the same Ignition Switch problem and changing the shroud to be simil ar to the Ion would not resol ve the probl em. 889 As noted above, thi s idea had been considered and rej ected in 2005 . By early 201 3, Sprague had identified two Ion n o n~d e pl oy m ent s that he thought might be comparable to the Cobalt non~deployments.89o However, because the Ion SDM did not record the signal positi on of the Igniti on Switch at the time of the acc ident, Sprague and Stouffer were unable to determine whether the Ion SDMs rece ived an "Off' or an "Accessory" signal at the time o fa crash.89 1 Neither Stouffer nor Sprague was co nfident that the two Ion incidents were J& B Interview of Brian Stouffer, Apri l 4, 201 4. J& B Interview of Bria n Stouffe r, April 4, 201 4 886 J& B Interview of Terry Connolly, April 28. 2014. 887 E-mail from Peter Judis, GM. to Mark A. Beauregard, GM,et a /. (Oct. GMNHTSA000221248l 888 E-mail from Peter Judis, GM, to Mark A. Beauregard, GM, efaJ. (Oct. GMNHTSA00022 I 248l; J&B Interview o f Brian Stouffer, April 4, 2014. 889 J& B Interview of Brian Stou ffe r, April 4, 201 4. 890 J& B Interview of John Sprague, May 9, 2014; J&B Interview of Brian 891 J& B Interview of John Sprague, May 9, 2014; J&B Interview of Brian 884 885 192 I I, 2012) [DOC lD 000000005108 ; I I, 2012) [DOC ID 000000005108 ; StoufTer, April 4. 2014. Stouffer, April 4. 20 14. Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product related to the Cobalt airbag non_deployments. s92 However, due to his concerns about the Ion , Stouffer did not agree with the suggestion to install an Ion-like shroud in the Coba lt.S93 On November 14, 2012, Federico hosted a meeting at which Stouffer gave an update on hi s investigation 894 and the electrical team explained why finding an electrical soluti on was difficult, specifically, that changes the team would need to make to the SDM to fix the problem would violate federal compliance requirements. s95 The team then shifted the discussion to the mechanical so lutions, and discussed various changes to the ignition cylinder and Ignition Switch to increase the torque required to turn the key as well as possible changes to the steering column.S96 To prepare for the DFSS, Stouffer hosted a brainstorming session on December 4, 2012, to identify potential mechanical so lutions. ~<)7 Following the meeting, Stouffer circulated a meeting invitation to Brian Thompson, Engineering Group Manager of Switches and Control s, Red X team member Dan Davis, DeGiorgio, and others for the Janua ry 8, 2013 DFSS meeting. 898 The meeting invite set forth a li st of potential so lutions that was generated at the Nove mber 14,2012 meeting hosted by Federico and discussed again at the team's December 4, 2012 brainstonning session.s99 J&8 Interview of John Sprague, May 9, 2014; J&B Interview of Brian Stouffer, April 4, 2014. J&8 Interview of John Sprague, May 9, 2014; J&B Interview of Brian Stouffer, April 4, 2014. g9~ J&8 Interview of Brian Stouffer, April 4, 2014 ; e-mail [rom Peter1udis, GM, to Terrence E. Connolly, GM, el af. (Nov. 1, 2012), fonvardin g calendar infonnmion from Cindy Mahan, GM, on hehalf of Brian Stouffer, GM lOOC ID 000000046119; GMNHTSAOO0343122j. 895 J&8 interview of Jim Federico, May 9, 2014. 8% 1&8 Interview of Brian Stouffer, April 4, 2014. 897 1&8 Interview of Brian Stouffer, April 4, 2014. ~9~ J&8 Interview of Brian Stouffer, April 4, 201 4; e-ma il from Brian Stouffer, GM, to Brian Thompson, GM, el af. (Dec. 14, 2012) rOOC iO 000000013059; GMNHTSAOO022 13931. 899 J&8 Interview of Brian StoufTer, April 4, 2014; e-mail from Brian Stouffer to Brian Thompson, el al. (Dcc. 14, 20 12) lDOC 10 000000013059; GMNHTSAOO0221393 j. 892 893 193 Privileged and Confidenfia/ Protec1ed by AUorn ey~ Client Privilege and as Allorney Work Product I. 2013: An Outsider Conclusively Demonstrates the Root Cause, But Still No RecaJi In early 201 3, Stouffer dec lared hi s two~yea r-old in vesti gation "stuck,,,900 even though he now had the Indiana Uni vers ity study correctl y identifyi ng the cause of the airbag nondeployments. 90 1 Even when the plainti ff's expert i n _ showed in April 201 3 that the Ignition Switch had in fact been redesigned in 2006902 - answering one of the key outstanding questions in Stouffer 's search for the "root cause" - GM hired an expert to reconfirm what it had already learned i n _ - an exe rcise that took another six months.90J When Stouffer received the written report from GM's expert in late October 201 3, Stouffer fin ally accepted the answer that Trooper Young and Indi ana University had arri ved at in 2007: that the Ignition Switch was inadvertently turning to Accessory, whi ch disabled the airbags. 904 By the end of 20 13, th e Cobalt Igniti on Switch had fi nally reached GM's committee that considers recalls (tJl e EFADC), but once there, more questi ons were raised about root ca use, and d ecisio n ~ma k ers were hamstrung by a lack of accurate data about what vehicles were affected and how many people may have been impacted by the defect. I. January 2013: DFSS Meeting The DFSS meeting was held on January 8, 20 13.905 Among others, Stouffer. DeG iorgio, and Dan Davis attended?06 The DFSS tea m reviewed a presentation that contained various J& B J& B 902 J& B 'IOJ J& B 9(J.t J& B 905 J& B 906 J& B 900 901 Interview Interview Interview Interview Interview Interview Interview of Brian Stouffe r, April 4, 20 14. of Brian Stouffer, April 4, 201 4. of Jaclyn Palmcr, May 8, 201 4. of Gay Kcnt, March 18, 2014. of Brian Stou ffe r, April 4, 201 4. of Brian Stouffe r, April 4, 201 4. of Brian Stouffe r, April 4, 201 4; J&B Interview of Dan Davis, M:'IY 8, 2014. 194 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product background information to assist them in working through viable optio ns. 907 In response to a question from Da vis, DeG iorgio sa id that he could duplicate turning the igniti on out o f Run with his own knee, and that he could do so with both a hol e and slot keyhead. 908 The recommendation from DFSS was that the best possible so]ution was adding a retooled Ion column shroud to the Cobalt.909 Chang ing the torque in the ignition sw itch was discussed but rejected due to the 1 8~24 months production estimate provided by DeGiorgio - agai n based on DeGiorgio 's faulty premise that the target torque was 100 N~cm .9 1 0 Agai n, DeGiorgio failed to di sclose that the torque had been increased - by a far smaller amount - in a 2006 redesign, which had addressed the prob lem. Davis recalls they a lso discussed changing the key by e liminating the fob entirely by making the key neither a slot nor a ho le so no keychain could be placed on it. However, the reaction of the group was that consumers would not like that design?ll After the January 8, 20 12 meeting, Stouffer and Sprague d iscussed the DFSS recommendation. 9 12 They agreed that adding the Ion co lunm shroud was not an acceptable solution because they were uncertain as to whether that the Ion had the same problem with airbag non~deployment. 9 13 J&B Interview of Brian Stouffer, Apri l 4, 2014; see generally DFSS Cobalt Steering Column Presentation [DOC ID 000000000150; GMNHTSA00022l395J. 908 J&B Interview of Dan Davis, May 23 , 2014. 909 J&B Interview of Brian Stouffer, Apri l 4, 2014. 910 J&B Ill1erview of Brian Stouffe r, Apri l 4, 2014 911 J&B Interview of Dan Davis, May 23 , 2014; J&B Interview ofOan Davis, May 8, 2014. 912 J&B Interview of Brian Stouffer, Apri l 4, 2014. 913 J&B Interview of Brian Stouffer, Apri l 4, 2014. They thought there were stalling complaints associated with the Ion and they suspccted that there might be no n~d ep l oyments associated with the Ion. TIley were unable to verify these suspected non-deployments, however, because the Ion uses a difTerent SDM than the Cobalt and the lon ' s SDM does not record crash data in the smne way that the Cobalt docs. Among other difTerences, the Ion SDM docs not record the power mode at the time of the crash incident. J&B Interview of John Sprague, May 9, 2014; J&B Interview of Brian Stouffer, April 4, 2014. 907 195 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product 2. February 2013: The FP[ Investigation is "Stuck" At thi s point in earl y 2013, Stouffer concluded that the investigation was "stuck. ,,91 4 DFSS had recommended a solution that did not appear likely to solve the problem, the Red X team had not moved the investigation forward, and none of hi s efforts to find solutions in the various component parts of the vehicle had borne fruit. Stumped and fru strated, Stouffer and Wachtel asked Federico and Kemp to retain a consulting finn named Exponent and, in particular, a consultant named Subbaiah Mall adi to assist the investigation. Multiple witnesses reported that Malladi is used only in the most difficult ofGM ' s cases. A February 20, 2013 e~ma il from Wachtel to Federico and Kemp, copying Benavides and Stouffer, stated, "we can not [sic] explain why there are non-deploy events on 2005 , 06 , 07 vehicles, whi le there are no known events on 2008 , 09 , 2012 vehicles .... I suggest that we ask Sabbiah [s ic] from Exponent to be part of thi s brainstorming effort.,,915 GM did not, however, then engage Malladi, and he did not begin work until May 20 13. J&B Interview or Bria n Stouffer, Apri l 4, 201 4. E-mail rrom Douglas Wachtel, GM, to Jim Federi co, GM , el al. (Fcb. 20, 2013) [DOC 100000032331281Wachtel claims Malladi was engaged shortly after Wachtel's February 20, 2013 e-mail and that he had a background call with Stouffer. J&B Interview or Douglas Wachtel, March 18, 2014 . \/14 915 196 Privileged and Confidenfia/ Protec1ed by AUorney~ Client Privilege and as Allorney Work Product From: Doug WlK:tI!el To : Jrn Fedeoc:o. V'JIDI8ITI J . Kemp CC: Carmen BenaVIOeS; Brian Stouffer acc: Sent o.te: Received Date: Subject: Attachmen tl: 201J.02·2O 22 32"00:000 2013.02·20 22 32"00:000 Coban artlag non-deploy "Privileged and Confidential" Prepared a\ the r~st of COU'1CeI 8 i1 merwlOl'lGd to you that I bllil8V8 that we are Shit mining a key peee of nIormalion Ifl this iI"1vfIstigation INa can not explain why there 8(e non~ events on 2005, 06, 07 vehicles, white thefe 8(e no known everws on 2008. 09. 2010 vehicles I thllk we need 10 gel OU' smartest lolts Irom I I involved SMTs back I~ther to brall'lStorm this rm ttmking 01 an off-stte meeong $Owe can focus everyone's anernlOfl. And it would help 10 injed nfIN blood 10 llO'ease our ability to get 10 the bottom of this I suggest that we ask S8boah "om Exponent 10 be part CJI this brttinslormrog effort. P lease let me know if you support th is IPI)t08Ch. The reason for this delay is unclear because witness accounts concerning how and when Malladi was hired vary and are somewhat inconsistent. Kemp recall s that he needed approval from senior management to hire Mall adi and that sometime after receiving Wachtel's February 20, 20 13 e~mail he sought approval from John Calabrese, Vice President ofG lohal Vehicle Engineering (GM's Chief Engineer), and a member of the EFADC, who was reluctant to incur several thousand dollars in expert fees. 9 16 According to Kemp, he suggested to Calabrese that Kemp could perhaps convince Porter and Phillip Holladay, outside counsel to GM on the _ matter, to hire Mall adi in connection with the _ litigation, so that Engineering would not have to pay Malladi's fees and his work wou ld be maintained as pri vi leged under the work product doctrine. 917 Porter recalled Kemp contacting him in April (when Porter was preparing Stouffer for his deposition in the _ 916 9L7 case), and telling Porter that FP E intended to hire J&B Interview or William Kemp, May 9, 2014. J&B Interview or William Kemp, May 9, 2014. 197 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product Malladi , but that Kemp and Porter ultimate ly agreed to hire Malladi under the auspices of the _ case to shield Ma lladi 's analyses under the work product doctrine_ 918 Holladay, howe ver, recalled that he recommended to Porter and Kemp shortly after DeGiorgio's April 29, 2013 _ deposition that GM hire Malladi?19 For hi s part, Malladi recall s being contacted by Holladay in Apri l or May 2013, and that hi s engagement was in connection with the _ litigation, not the FPE investigation. 92o In March 2013, Kemp asked Ca labrese to appoint someone to replace Federi co as "champion " of the Cobalt investi gation because Federico' s travel sched ule and other responsibilities rendered him unable to devote sufficient time to the rol e. Ca labrese assigned 92 1 Gay Kent, the General Director ofGM North America Vehicle Safety and Crashwonhiness. Kent, then in charge of the PI group, had been one of the engineers who had replicated the moving stall at the test track in 2005. Kent was the third "champion" in ten months. The same month that Kent became champ ion, Stouffer provided her with an ex planation of the investigation ' s progress. 922 On Apri l 4, 2013, Kent met with Stouffer and Wachtel to obtain additional infornlation about the investigation. 923 Stouffer did not conduct much work on the investigation in the spring of 20 13; he felt he had hit a dead end and beli eved that Malladi would be assuming a more active role. 924 J&B Interview of ROllald Porter, April 30, 201 4. J&B Interview of Phi lip Holladay, April 25, 201 4. 920 J&B Interview ofSubbaiah Mal1adi, May 7, 2014. 921 J&B Interview of Brian Stouffer, April 4, 201 4; J&B Interview of Cay Kent, Mareh 31 , 2014. 922 J&B Interview of Brian Stouffer, Apri l 4, 201 4; J&B Interview of Cay Kent, March 31 , 2014. 923 J&B Interview of Cay Kent, March 31 , 2014. Wachtel did not reca ll this meeting. J&B Interview with Douglas Wachtel, March 18, 2014. 924 J&B Interview of Brian Stouffer, April 4, 201 4. 918 919 198 Privileged and Confidenfia/ Protec1ed by AUorn ey~ Client Privilege and as Allorney Work Product 3. April 2013: GM Lawyers Review the"'Matter at the Roundtable While the FP E investi gati on fail ed to make progress, litigation continued. Th e Cobalt issue reappeared when, on April 10, 201 3, the Roundtable rev iewed the _ matter again. The Roundtable authorized settl ement, whi ch was approved by Buonomo, and th e _ matter settl ed. 4. April 2013: Depositions in the~Case Begin and There is a "Bombshell" at DeGiorgio's Deposition In Apri l 20 13, Philip Holladay of K& S began preparing to defe nd a series of GM deposi tions in th e _ case. Holladay reported that in preparing for the upcoming depositions, he rev iewed the FPE investigati on performed by Stouffer (and the work of Sprague before him). On April 29, Holladay defe nded DeGi orgio in a deposition at the Westin Hotel at the Detroit airport. According to an e-mail that Holladay sent to Porter from the deposition, plaintiffs counsel dropped a " bombshell ,,,925 namely evidence that the switch had, indeed, been changed from 2005 to 2008 - evid ence that had eluded GM engineers fo r years and that DeG iorgio testified, under oath, that he did not remember. Holladay reported that Ron Porter, the GM lawyer in charge of the _ litigati on, drove to the Westin on hi s way home to pick up photographs of the different sized plungers and caps that had been provided during the . . 91 6 depostUon. ~ J& B Interview of Philip Holladay, April 25, 20 14; e~ma il from Philip Holladay, K&S, 10 Ronald Poner, GM (Apc 29, 2013) [DOC 1D 000004640060], 926 J& B Interview of Philip Holladay, April 25, 201 4. For his pan , DeGiorgio said that after the deposition, he, indeed, took two switches apart and con finned the difference. J&B Interview of Raymond DeGiorgio, March 14, 20 14. He also told Brian Thompson, his supervisor, that he had been shown photographs of two different detent plungers of different lengths. DeGiorgio and Thompson then told their new group director, Craig Zinser. Zinser 925 199 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product Three days after the DeGiorgio deposition, Holladay participated in a telephone call with Porter and Bill Kemp _ Holl aday stated that GM uses Exponent and Subhaiah Ma lladi for its most important matters. During the call, Holl aday recommended hiring Exponent and Malladi for two reasons. First, Exponent would ass ist with the case. Second, GM needed to bring the FPE investigation to cl osure witho ut delay. Ma lladi became involved within days of the 927 DeGiorgio deposition. 5. May 2013 : Kent ' s " Kick orr' M eetin g In the meantime, in early May, Kent, the PI "champi on" since March, organi zed what she described as a " kick off' meeting that Kent, Ron Porte r, Bill Kemp, and Jenny Sevigny 928 attended. 929 Ma lladi and Holladay participated by phone . The purpose of this meeting was to disc uss the work conducted by Wachtel and Stouffer to date, and the infonnation presented at th is meeti ng was similar to what the investigators had presented to Kent at the April 4, 2012 meeting?30 Although Kent also invited Sprague, Stouffe r, and Wachtel to participate,931 they did not recall attending the meeting. Sprague never interacted with Mall adi directly, but came to be aware ofMa ll adi's invo lvement through Sevigny, who would gather infomlation from GM 932 employees for Ma ll adi's investigation. Stouffer reca ll ed that prior to the May 3, 20 12 told De Giorgio and Thompson to let lega l handle it. J&B Interview of Raymond DeGiorgio, March 19, 2014. Zinser had no recollection of this discuss ion with DeGiorgio and TIlOmpson. J&B Intcrview of Craig Zinser, May 27, 2014. 927 J& B Interview ofSubbaiah Malladi, May 7, 2014. 928 E-mail from William Kemp, GM, to Ronald Porler, GM (May 2, 20 13) IGMNHTSA000445329; KSGMOOOl69121 (K&S May 20) 3 Invoicc); J&B Intcrview of Gay Kent , March 3 1, 2014. 929 J&B Interview of Gay Kcnt, March 31 , 2014. \/30 J&B Interview of Gay Kcnt, March 31 , 2014. 931 E-mail from William Kcmp, GM, 10 Ronald Porter, GM (May 2. 20 13) TGMNHTSA000445329; KSGMOOOI6912] (K&S May 2013 Invoice); J&B Interview of Gay Kent. March 3 1. 2014. 932 J& B Interview of John Sprague, May 15, 2014. 200 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product meeting, he and Wachtel were asked not to participate. 933 Porter confirmed that he asked Stouffer and Wac hte l not to participate in the call with Malladi because he tho ught their presence would make it more difficult to assert pri vilege. 934 A lthough he did not attend the meeting, Stouffer walked by the conference room while it was taking place, and heard Ho lladay providing Malladi with a "dow nload" on Sto uffer's investigation. 935 On May 10,20 13, Kent, Sevigny, Porter, Ho ll aday, and Kemp met again on the Cobalt project, with Malladi participating by phone. 936 Stouffer and Wachtel agai n did not participate, and Kent understood that Stouffer did not attend because he was go ing to be deposed in the _ case. 937 There is disagreement about what Malladi was asked at this meeting to do. Kent says she tasked Malladi with determining w hy there were differences in the Coba lt 's Ignition Switch torque across model years?38 Kent also remembered requesting that Malladi conduct an investigation to determine the connection between the Igniti on Sw itch and the non~d ep l oym e nt of the ai rbags,939 but Malladi did not recall being asked to undertake s uch an investi gation, and he did not act on such a request, if made; hi s report was narrower and focused on the issue of difference in torque across model years. 940 Ma ll adi stated that he was not asked to and did not perform work to determine the root ca use of the airbag n o n~d ep l oy me n t, and that he did not draw J&B Interview of Brian Stouffer, April 4, 2014, J&B Interview of Ron aid Porter, April 30, 201 4. 935 J&B Ill1erview of Brian Stouffer, April 4, 201 4 936 J& B Interview of Gay Kent, March 31 , 2014; K&S May 2013 _ Invoice (Jun. 28, 2013), at4 LDOC ID 20 1405160003712J (showing participat ion by Kent, Sevigny, Malladi , Porter, and Holladay). 937 1&B Interview of Gay Kent, March 31 , 2014. Stouffer described that he and Wachtel were "no longer invited" 10 the Malladi meeting and he thought it was "odd," thinking that he would have beller briefed Malladi than the lawyers. J&B Interview ofBriall Stouffe r, April 4, 2014. 938 J&B Interview of Gay Kent, March 31 , 2014. 939 J&B Interview of Gay Kent, March 31 , 2014. 9~O J& B Interview ofSubbaiah Malladi, May 7, 2014;_ v. GM presentation LGMNHTSA000466745j. 933 934 201 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product any conclusion regarding whether the change in the Ignition Switch or the lower torque caused the SDM to read the power mode as Accessory or the non~deployment of the airbags 9 41 There is also di sagreement about how Malladi 's work on _ intersected with work on the open FPE investigation. Mall adi said he did not recall knowing that his work related to an open FPE investigation until after the _ case settled in the fall of 20 13. 942 Kent said that she communicated the urgency of the situation, but that she knew Malladi needed to fin ish some work on the _ li tigation before he cou ld tum his full attention to the FPE project?4J Holladay said that whil e part of his reason for hiring Mallad i was to assist on the _ Ma lladi 's work on th e _ case, litigation was structured to ensure that the open FPE investigation moved towards closure without del ay.944 Whatever the instruction, after thi s meetin g, Malladi focused solely on the issue that had already been considered by _ expert witness - the question of differences in the Ignition Switch in different model years.945 Sevigny was designated to be the point of communication between Kent, the FPE investigators, and Malladi ?46 Going forward, Sevigny routinely provided updates 011 Malladi ' s progress and provided Mall adi with data and informati on.947 J&B Interview ofSubbaiah Malladi, May 7, 2014. J&B Interview ofSubbaiah Malladi, May 7, 2014 9H J& B Interview of Gay Kent, March 31 , 2014. 944 J& B Interview of Philip Holladay, April 25, 20[4. 9-15 J&B Interview ofSubbaiah Malladi, May 7, 2014;" v. GMpresentation[GMNHTSA000466745l ').16 J&B Interview of Gay Kent, March 31 , 2014. \I~1 J&B Interview of Gay Kent, March 31 , 2014. Malladi recalls Holladay telling him that DeGiorgio stated that no changes were made to the ignition swi tch plunger. Holladay told Malladi that the plaintiffs expert had said that a 2005 switch had a different plunger and spring than a service part switch obtained from a dealer. J&B Interview of Subbaiah Malladi, May 28, 2014. 9-11 ').11 202 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product 6. May 2013: GM Lawyers and Outside Co unsel Review The~ Matter Again On May 15, the week after Kent's "kick-off' meeting and abo ut two weeks after the DeGiorgio depositi on, GM lawyers reviewed t h e _ matter at another Roundtable. 948 Porter 's summary of the Roundtable included: NEW INFORMA TION: Plaintiff claims the engine turned off because the switch Ignition switch moved from the run to the accessory position. The _ requires about 8 N /cm to tum from Run to Accessory. The torque c urve on the drawing works out to about 20 N/cm. 949 The Roundtable gave settlement authority, but the case did not settle and di scovery continued. K&S submitted a second case eva luation for t h e _ matter on July 22, 2013. 950 Holladay's evaluation included a conclusion that a jury would almost "certainly" find the Ignition Switch unreasonably dangerous, that low torque would lead to inadvertent shutting off of the engine: This case is a very poor trial candidate. A jury here will almost certainly conclude that the Cobalt's ignition switch is defective and unreasonably dangerous because the torque effort required to move the key from run to accessory is too low, which leads to inadvertent key move ment and the engine shutting off with little or no warning. 951 K&S warned that this was known virtuall y from the vehicle launch: This phenomena was identified almost immediately after the 2005 Cobalt went into production and there were several newspaper and trade publication articles shortly after the car's launch that fla gged the issue. As discussed in more detail below, the iss ue was assessed internally in a series of investigations conducted as part of the Product Resolution Tracking System and ultimately addressed by 9~8 Ronald Porter. _ Sett lement Review Committee Roundtable Agen~C lD 000162270688 J. 9-19 Ronald Poner, _ Settlement Review Committee [DOC ID 0000028186121. \/)0 Ronald poner, _ Sett lement Review Committee 0000028 18612]. 951 Ronald Porter, _ Sett lement Review Committee 100000028 186 12. r - Case Summary (Aug. 8, 2013) [DOC ID 000002818612J ; Case Summary (Aug. 8, 2013), at 4-5 (emphasis omitted) Case Summary (Aug. 8, 2013), at 11 rDOC ID Case Summary (Aug. 8, 2013), at 12-13 [DOC 203 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product issuing an Informati on Service Bulletin in the Fall of2005 that provided a fi eld service fix for customers who experienced an incident invo lving inadvertent key movement_ 952 The evaluation then catalogued the airbag non~deplo y ment issue, even though airbag non~deployment was not an all egation i n _ More recently, this issue has surfaced again as part of an on~going FPE investigation into why frontal air bags have not deployed in ce rtain high~s peed multiple~impact frontal collisions invo lving 2005~2007 Chevrolet Cobalts. In more than half of those incidents, it appears the reason that the air bag did not deploy was because the car's ignition was in than the run position. While there is no allegation here frontal air bags should have deployed, the on~going . into plaintiffs' expected theme that the original Information Service Bulletin was an inadequate " ba nd~aid fix" for a significant safety issue that should have been addressed through a recall and des ign change. Given these facts, we believe that GM 's defense here necessarily wi ll have to focu s almost exclusively on causation.9SJ Holl aday warned of "a substantial adverse verdict.,,954 Among the reasons he listed: "the ignition switch and key cylinder used on the 2005 Cobalt were problematic from the outset and plaintiffs will have little problem convincing most jurors that these components were substandard and defective.,,955 In concluding, Holladay wrote, "This case needs to be settled. ,,956 He warned that plaintiff would argue that GM had known of the defect from the time that Coba lts first "rolled off the assembly line" and yet it has "essentially done nothing to correct the problem for the last nine years": Ronald poner, _ IDOOOO02818612L - 95) Ronald Porter, _ IDOOOO028 186 12 954 Ronald Poner, _ IDOOOO02818612L - \/)) Ronald Poner, _ 952 L- IDOOOO028 1 86 1 2 ~ Ronald Porter, IDOOOO028 186 12 . 956 Settlement Review Committee Case Summary (Aug. 8, 20 13), at 12 ~ 13 rDOC Sett lement Review Committee Case Summary (Aug. 8, 2013), at 13 LDOe Settlement Review Committee Case Summary (Aug. 8, 2013), at 13 rDOC Settlement Review Committee Case Summary (Aug. 8,2013), at 13 rDOC Sett lement Review Committee Case Summary (Aug. 8, 2013), at 36 [DOC 204 Privileged and Confidenfia/ Protec1ed by AUorney~ Client Privilege and as Allorney Work Product is littl e doubt that a jury here will find that the ignition switch used on 2005 Cobalt was defective and unreasonab ly dangerous, and that it own torque specifications_ In addition , the PRTS docume nts referenced above and the on-going FPE in vestigation have enabled plaintiffs ' counsel to develop a record from whi ch he can compellingly argue that GM has known about thi s safety defect from the time the first 2005 Cobalts rolled off the assembly line and essentially has done nothing to correct the probl em for the last nine years. He specifically will criticize GM for not doing more than implementing the field service campai gn back in 2005, and point to GM's failure to take any action in the on-going FPE investigation that has now been dragging on fo r almost two years as proof positi ve of GM 's consc ience indiffere nce and willful mi scondu ct when it comes to the sa fety of its vehicles' occupants?57 EVALUATION AND RECOMMENDATION there is little doubt thai ajury here will Cobalt was defective and unreasonably danaero1.lS, and that j, did no, mffi OM's own torque specifications. In 8(idition, the PRTS documents referenced above and the on~going FPE investigation have enabled plaintiffs' counsel 10 develop 8 record from which he C8rt compellingly argue thaI OM has known aboullhis !l8fety defect from the time the (irst 2005 Cobalts rolled ofT the assembly line and essentially has done This case needs to be find that the ignition switch u."ed nothing to eonect the problem for the last ninc years. He specifically will criticize GM fo r not dOing mon: than implcmcnti,ng tbe field service campaign back in 200S, and point 10 GM's fpjlure to take lillY action in the on-going FPE investigation that tw now been dragging on for almost two years as ptoofpositive of GM' s conscience indifference and willful misconduct when it comes 10 the safety of its vehicles' occupants. 7. July 2013: MaUadi Presents His Findings OraIJy On July 30, 2013, the week after GM rece ived K&S's second case evaluation and about ten weeks after GM 's May 10, 20 13 meeting with Ma lladi , Malladi orally briefed Porter, Sevigny, Kemp, Kent, and a K&S lawyer. According to Kent, Malladi reported that the Ignition Switches install ed in ' 05 to ' 07 MY Cobalt vehicles did not meet GM's torque specifi cations. 958 Porter recalled that the call focused on the GM car dealer's expert witness, who thought the S CM was responsible for the crash, and 9S7 Ronald poner,_ Settlement Review Committee Case Summary (Aug. 8, 2013), at 36 rOOC ID 0000028 18612]. 9S8 J&B Interview of Gay Kent, March 18, 2014. 205 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product comm unications with co-counsel in the case. 959 Kemp recounted that Malladi reponed on differences in th e sizes of springs and switches uncovered through x-rays, and asked GM to 960 obtain records from Delphi to confinn that part changes had been made. Ma lladi has no recollection of the July 30, 20 13 phone ca l1. 8. 961 August 2013: ~SRC Meeting On August 7, 2013, a week after hearing Malladi 's results, Porter presented _ to the Settlement Review Committee.962 Buonomo, Gruskin, Kemp, and Nowak-Vanderhoefare listed as attending?63 The Settlement Review Committee received the previo us Roundtable summaries and Ho ll aday's case evaluation . It al so recei ved a summary drafted by Porter that contained a list of factors impacting exposure. Even though _ was not an airbag case, Porter 's exposure summary incl uded a paragraph on the ongoing airbag investigation in volving Cobalts produced before 2008: There is a co ntinuing GM FPE investigation into airbag non-deployments in Cobalts whic h may be, at least partially, linked to the ignition switch issue. We know of approximately 20 non-deployments of frontal ai rbags in Coba lts where the crash forces are sufficient to have properly deployed the fronta l SIR. In half of these crashes, the SDM printout indi cates the igniti on was in the " Acc" position. In the Cobalt, the airbag will not dep loy with the key in "Acc." In the other cases, the SDM indicates " Run" and there is no known rea son for the failure J&B Interview or Ronald Porter, May 13,20 14. J&B Interview or Wi ll iam Kemp, May 9, 2014. 961 J&B Intcrview orSubbaiah Malladi, May 27, 2014. Documcntary cvidcncc rcnccts, howevcr, that Malladi told Porter he received a dial-in and planned 10 ca ll in on July 30, 20 13, [DOC ID 000004634434], and Ihal on July 31 , 20 13, Susan Clare or K&S requested, in an e-mail to Malladi, "copies or the TSBs you mentioned yesterday related to the BCM." [DOC ID 000004634389J. 962 Ronald Porter, _ Sculcmcnt Review Committcc Casc Summary (Aug. 8,2013) [DOC ID 000002818612]. 963 _ Settlcmcnt Review Committee Mecting (Aug. 7, 2013) [DOC ID 000 120897828]. Evidcncc shows that, although she is listed on the settlement authority document as atlending, Clark Dougherty, who was on maternity leave, did not attend in person or by phone, or access the meeting materials. J&B Interview or Lucy Clark Doc# 4216521 Document History (May 22, 2014) I DOC ID Dougherty, April 16, 2014; SRC Packagc - _ 24 I 263003J; Buonomo Meet Me Con rerenc~og ror Aug. 7, 20 13 (May 20, 2014) LDOC ID 241263002]. 959 960 206 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product of the airbag to deploy. All of these events involve Cobalts produced before 200S 964 The SRC authorized settlement, which was approved by Buonomo, and the case settled in September 201 3 for $5 million, the maximum the SRC could authorize without the approval of the General Counsel. 965 The evide nce shows that neither Kemp, Buonomo, Gruskin , Porter nor any other GM lawyer elevated Holladay's letter or specific issues related to t h e _ case to General Counsel Michael Millikin prior to sett lement. 966 9. October 2013: Malladi's Written Presentation Malladi presented hi s findings ora ll y on a September 20, 201 3 conference call. The meeting invitation to Benavides, Wachtel, Stouffer, Kent, Porter, Malladi, and Kemp noted: " FPA, Legal Staff and Subbaiah" had investigated "ignition sw itches on 2005~20 I 0 Cobalt vehicles as a resu lt of issues that arose in a product liability lawsuit. T hat case has resolved and we would like to review with you the work that was done and identify any remaining items that need to be completed. ,,967 After recei ving the oral download ofMalladi 's report on September 20, 2013 , which demonstrated that the ignition switch had been changed between MY 2006 and MY 2007, Stouffer started trying to find a contact at Delphi, the supplier of the Ignition Switch, to obtain additional information. Stouffer asked DeGiorgio and Porter for Delphi contacts before Ronald Porter, _ Settlement Review Committec Casc Summary (Aug. 8, 2013), at 2~3 [DOC 10 0000028 186121. 965 _ Scttlcmcnt Review Commil\CC Mccting (Aug. 7, 2013) [DOC ro 0001208978281 ; Mielmcl Gmskin, GM, Legal $taffSettlement Review Processes (March 23, 2009) LDOC ID 0 _000 I 6584J. 966 J&B Interview ofLawrcncc Buonomo, April 16, 2014; J&B Interview of William Kemp, Apri l 17, 201 4; J&B Interview of Jac1yn Palmer, April 24, 2014. 967 E-mail fromJennifer Sevigny. GM . toBrianStouffer. GM. el al. (Scpt. 17, 2013) LDOC ID 000000027269J. 964 207 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product recalling that GM had a Delphi contact on government investigations (Gary Greib) whom he could ask_ Stouffer contacted Gre ib on October 23, 20 13 968 About two weeks after the September 20, 2013 call with Malladi, on October 8, 2013 , Wachtel e~mai l ed Benavides again, this time to tell her that he and Stouffer needed three main pieces of information to " bring the [ignition switch1 issue forward": ( I) " List of confirmed/suspected cases where non-deploy is suspected to have occurred from ignition being shut off - Brian has this"; (2) " Delphi information that shows dates that a new switch was substituted into production" ; and (3) " A copy of the information Subaiah [sic] presented to us during our recent conference call. Jenny Sevigny and Bill Kemp have requested this from Ron Poner, but we do not have the information.,,<)09 About three weeks later on October 28, 2013, Malladi sent hi s written Power Point presentation to Porter and Holladay, copying Wachtel , and apologizing for the delay.97o In an interview, Malladi explained that he did not respond with the slides ri ght away because he realized that there were x~rays he wanted to comp lete first; he had the impression at that time that GM wanted to take " some field action" with respect to the vehic les at issue. 971 Wachtel forwarded the e~mai] to Stouffer the fo llowing day.972 Stouffer said he did not know why it took so long to get Malladi's results, and said Wachtel had asked Kemp for them. 973 Malladi 's 2 I-page slide deck reported two findings: J&B Interview of Brian Stouffer, Apri l 4, 201 4; e-mail from Brian Stouffer, GM, to Ronald Poner, GM (OCL 14, 2013) [DOC ID 000000108574] 969 E-mail from Douglas Wachtel, GM, 10 Brian SloufTer, GM (Oct. 9, 2013) LDOe ID 000000108837 ; GMNHTSA000283360J. 970 E-mail fromSubbaiah Malladi, Exponent, 10 Ronald Poner, GM , Phil Holladay, K&S , and Douglas Wachtel, GM (OCL 28, 2013) [DOC ID 000004623644; GMNHTSAOO0466744l; see rDOe ID 000004623645 ; GMNHTSA0004667451 for attachment 10 e-mail. 971 J&B Interview ofSubbaiah Malladi, May 7, 201 4. 972 E-mail from Douglas Wachtel, GM, lO Brian SloufTer, GM (Oct. 29, 201 3) [DOC ID 0000001034661973 J&B Inlerview of Brian Stouffer, April 4, 201 4. %8 208 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product • One, there had been "a design change" in Cobalt ignition switch part # I 0392423 between MY 2006 and MY 2007 MY Cobalts (as t h e _ expert had said some fi ve months before). • Two, "certain [Cobalt] ignition switches did not meet specifications.,,974 Malladi reported in an interview that he had not been asked to opine on whether that low torque was the cause of the Cobalt airbag non-deployments, which he described as a complicated issue because of multiple potential causes. 9 75 10. November 2013: Stouffer Arrives at and Presents His Conclusions, Which Are the Same as Trooper Young ' s, the IV Investigators ', and Erin Shipp' s With both t h e _ plaintiff's expert and Malladi now having established that the Ignition Switch had changed, on October 29, 201 3, Delphi provided documents to GM confirming that DeGiorgio had approved a change to the Ignition Switch spring and plunger in April 2006. 976 Wachtel reported that these documents helped GM identify the Ignition Switch as the so urce of the airbag non~de plo yment issue. Wachtel, who had overseen Stouffer's in vestigation for over a year by thi s point, believed that GM's investigation had been impeded by previous statements that the Ignition Switch had not changed.977 Stouffer agrees. In November 20 13, Stouffer fina ll y reached the same conclusi on that Trooper Young had in 2007. An Investigation Status Review ("ISR") meeting took place on November 5, 201 3. 978 974 _ I'. GM prescntation IDOC ID 000004623645; GMNHTSA000466745j. 975 !r'B"Tt'ncrview ofSubbaiah Malladi , May 7. 201 4. 976 E-mail from Douglas Wachtel, GM , loJamie Morri son, GM, el al. (OCL 3, 2013) [DOC 10 000000117060; GMNHTSA000466736J; e-mail from John Murawa, GM, to Mark Johnson, GM (Dec. 9, 201 3) [DOC ID 0000 13664326; GMNHTSA00022372IJ ; e-mail fromGaryGreib, Delphi, to Brian Stouffer, GM (Oct. 29, 2013) rDOC ID 000000131995 ; GMNHTSAOO00028891 ; see Prcscntation cntitled "Enginccring Change Support Infonnmion" rDOC ID 000000131996; GMNHTSAOOOO028961 ; Rcport entit led "Spring Dimensiona l rDOC ID 000000131997 ; GMNHTSAOOOO02899 J; Form LDOC ID 000000131998; GMNHTSAOOOOO2900 J. 977 J&B Interview of Douglas Wachtel, March 18, 201 4. 978 Cannen Benavides was in allendance a t the ISR on the Cobalt a nd believed at that time that there should be a recall. J&B Interview of CarnIen Benavides, May 19, 2014. Two days later, she and Dale Furney presented to 209 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product Stouffer presented findin gs that indicated a relationship between the low torque in the Cobalt Ign ition Switch and incidents of air bag n o n ~ dep l oy me n t. In his PowerPoi nt presentation on th e 2005 to 2007 Cobalt, he reported: "A review of se lected Cobalt & G5 fro ntal crash events indicates some airbag non deploys have occurred where the ign ition switc h was in accessory or off. The noted field events involve vehicles going off the road and/or hi tting small er objects before a more significant impact." He hypothes ized that "during the off road event the driver's knee is interacting with the keys and/or the mass of the keys is causing the ign ition to rotate.,,979 2005-7 Cobalt, G5, Pursuit, 2003-2007 lon, 2006-2007 HHR Condition: A review of selected Cobah & GS frontal crash events indicates some airba, non deploys have occurred where the Ignition $witch wa$ in aCCe5$ory or off. The co n d ition ap pears t o be limited to 200S·07 Cobalt & G5 vehicles. The noted fie ld events involve vehicles going off the road and/or hitting smaller objects shortly before a more signifICant impact. Questions: - W hy no incid{!nt~ on Ion or HHR Ion I!o Oan 2 an:httKtur~1I!. GM LAN on Cobalt - loth diubIt 5.OM wilh by off. but Cobllt v.1I stol • St81.. wi crash ,..:ord ...... IonWlt root Ion has different (olumn s~roud whkh could olffeet POII!nt~ for key lntet'oKtlon Ion (uSlom~may be len nlcel'( to h;we the type of cfiJSJl n~ for the condition Ion hasdiffertf'lt SOM and supplier II'I~ Cobalt HHR hn more d.ar,nte 10 I~e d,i¥v(s knee ..,fIIon - Why no incidents on 2008·10 Cobalt? ""ltlon swttchwil'l r~ to h_1on£fl pIIlnt~r and sprlineto Inc,~.~ ~ffort (confirmed 10/29}1)) - Ihrtn"""~~$Gq~lItion~.lI'Itnown(Vll.".tiont~t.4/25/Oi1. Se,"... YJfdwmpltJ ~.d 'lfX11 ~'" .... bnttt 1IIu...... (1I'I1aIowrI1f.., ... rtpltc..:l *" IrUCM;e). Root Cause: - The hypothesis is that durinc the off road ~nt tt'ie dri ver's knee is interacting with the keys and/or tl'le mass of the keys is causina the ignition to rotate NHTSA on a number of pre-selecled lopics, inchldi ng the condilions GM would analyze in conneclion wilh a " loss of propulsion. " J&B Interview of Carmen Benavides, May 19,20 14; e-mail from Dale Fumey, GM, 10 Cannen Benavides. GM (Feb. 1&,20 14) LOOC ID 000005127381; GMN HTSA000442785]; Presenlalion LOOC ID 000005127383; GMNHTSA0004427911. In her May 19,2014 interview, Benavides stated her bclief that the Cobalt ignition switch problem was not relevant 10 the framework for assessing stalls in her November 7, 2013 NHTSA presentation; she reasoned that in the Coba lt situat ion, where the ignition was moved out of Run, the engine would be expected to shut ofT, and therefore the Cobalt condition was not a stall . 1&B Interview of Cannen Benavides, May 19, 2014. 979 See Presenlalion to ISR, al I LDOC lD 000003100829; GMNHTSA000250865 ; GMHEC00025 1168 J. 2 10 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product On November 19, 20 13, Ma ureen Foley-Gardner e-mai ledA licia Boler-Davis, GM ' s Senior Vice- President of Global Q uali ty and Customer Experience, to inform her of the iss ue with tbe Cobalt igni tion switch. She noted in her e-mail that she wanted to provide a "[h]eads up on high dollar FP E issues," estimating that the issue wo uld cost $35 million.98o At this point, GM fi nally began its fonnal internal processes to commence a reca ll . 11. Conclusions: 2011 to 2013 The failures from 20 II to 20 13 demonstrate a lack of urgency in the face of a pattern of airbag fa il ures, an unwillingness by GM personnel to re-evaluate their concl usions, a lack of accountability or leadership in dri ving the investigations to a conclus ion, and a continued reluctance to elevate issues. The 2011 failures include: • In ea rly January 20 II , ~s - including sa fety lawyers and product litigation lawyers - attended the _ Roundtable and decided that it was important to meet to "see some additional information" about the a irbag "anoma ly." It took until July to schedule the meeting, a delay no witness could explain. • In July 20 II >GM lawyers called a meeting to make sure that senior engineering management had "eyeballs" on the issue and "not let it flow th.rough the normal process." Yet, Wachtel opened the investigation and did little else. The investigation would remain open and unresolved for two-and-a-halfyears. Neither the lawyers nor the engineers elevated the issue to the top levels of manageme nt. • When he was assigned the investigation in the summer of 20 I I , Stouffer obtained an old PI file (Kiihr ' s) with background on the 2005 moving stall in vestigation , and obtained the TSBs. Stouffer did not share thi s information w ith GM product liability lawyers or with Sprague. Stouffer now knew (I) that the Cobalt had a history o f the Igniti on Switch being turned accidentally to Accessory because of low torque, and (2) that airbags do not dep loy when the Ignition Switch is in Accessory. Regard less o f whether there were other issues affecting other model years, any driver who turned the Cobalt Igniti on Switch off with a knee (a problem documented in the TSB) was disabling airbags. Instead of addressing thi s safety iss ue, he worked to find a solution that would sol ve the problem for all years and all positi ons of the Igniti on Switch. E-mail from Maureen Foley-Gardner, GM , to Alicia Boler-Davis, GM (Nov. 20, 2013) [DOC 1D 000003036088; GMNHTSA000223499] ; J&8 Interview of Alicia Boler-Davis, May [6, 20 [4. 980 2 11 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product The consequence of the failures of2011 was delay. The unexplained gap between the _ Roundtabl e and the July meeting (a meeting ostensibl y to convey urgency) constiruted a delay of a half a year. And the investigation that proceeded from that meeting moved forward without any sense of urgency, ultimately taking two~and-a-halfyears. Failures in the Cobalt investigation continued in 20 12: • In March 2012, an electri cal engineer found the TSB after a vis it to ajunkyard where he noticed that the keys turned extremely easily and doing hi s ow n search ofGM records to see whether thi s was a documented problem. The TSB was forwarded to FPA engineers, in-house GM product li abi lity lawyers and outside counsel. Based on the TSB, outside counsel reac hed the same conclusion that Trooper Young and Indi ana University had five years earli er, and warned of the risk of puniti ve damages because "GM will be forced to contend with other incidents, some of which resu lted in deaths, due to the nondeployment of the frontal airbags in the 2005-200 7 Cobalt.-- No lawyer elevated the issue to the General Counsel. • In May 2012 , with the search for root cause making little progress, Kemp persuaded Terry Woyc howski (a Vice President) to "champion" the issue, but Woychowski soon retired. Then Kemp recruited Federico (an Executi ve Director) to replace Woychowski as the second "champion," but Federico viewed hirn selfas doing nothing but "energiz[ing] the team," not bringing the investigation to a conclusion. Neither Woychowski nor Federico elevated the issue up their chain of command. • In June 2012 , DeGiorgio once again told Stouffer, this time in writing, that there was no effect on the torque of the switch due to changes made in the Ignition Switch in 2006, again diverting investigative efforts. • In June 20 12, plaintiffs expert Erin Shipp located both the TSB and the Indiana Uni versity study and came to the same correct conclusion as the Indiana University team. Having been made aware of the Indiana Uni ve rsity study by the expert, Palmer transmitted it to Wachtel, Stouffer, Sprague and Kemp in July. But still GM engineers did not accept the correct theory. Rather, th ere were more meetings and referral to yet another group, the Red X team (which made no progress whatsoever). 2 12 Privileged and Confidenfia/ Protec1ed by AUorn ey~ Client Privilege and as Allorney Work Product Finall y, in 20 13, now that anoth er plaintiff' s expert had provided GM engineers with photographs of how th e MY 2008 igniti on switch differed from earl ier Cobalts, th e response was referral to an outs ide ex pert and yet more meetings: • In February 201 3, Stouffer - after almost two years - declared himsel f "stuck," and he and Wachtel recommended hiring an outside expert, which did not happen unti l the "bombshell" at DeG iorgio' s depositi on in April 2013 . • In April 20 13, GM lawyers and engineers received photographs from expert proving that DeGiorgio's statements had misled his colleagues fo r years; contrary to his repeated denials, there had been a part change years before that affec ted ignition switch torque. Th e plaintiff's expert had done what none of the GM e ngineers had done; he took apart two sw itches and compared them. • The response to the revelati on at the DeGiorgio deposition was to hire an expert. Gay Kent, who recentl y had become the third "champion," and who had personaLly replicated a moving sta ll eight years earlier in 2005, soon held what she called a " kick off meeting" that included the expert. It took six months after the DeGiorgio deposition for the expert to provide his written report concurring that the Ignition Switch had changed - just as the photographs had shown. • In July 20 13, outside counsel warned of a "compelling[]" case th at GM had known about the defect since 2005 "and essentia ll y has done noth ing to correct th e probl em for the last nine years." Still, no GM lawyer appri sed the General Co unsel. The in vesti gation under Kent awaited the completion of the outside expert' s report. Delphi, GM's suppli er which built the igniti on switch, was not contacted until October. • In November 20 13, after receiving the written report from the outside expert verify ing that the ignition switch had, indeed, changed in MY 2008 Cobalts, Stouffer fi nally reached the same conclusion that Trooper Young and Indiana Uni versity had in 2007: that the Ignition Switch was inadvertent ly being turned to Accessory and that, consequently, the airbags would not deploy. Stouffer's conclusion was reported to yet furth er committees, and it would not be until January 3 1, 2014 that the recall was ordered. a_ By this period, most of the infonnation silos relating to the root cause had collapsed. A new set of fai lures became more predominant. One was the failure to rev isit one's own conclusions. The GM investigator had been so focused on the SDM solution th at he failed to act on what was right before hi s eyes - regardl ess of whether the Coba lt was in Run or Accessory 2 13 Privileged and Confidenfia/ Protec1ed by AUorn ey~ Client Privilege and as Allorney Work Product from MY 2008, dri vers were kneeing off the power in Cobalts, as they had been since 2005, and in those instances airbags did not deploy by GM's own des ign_ Stouffer failed to revisit his first conclusions, albeit afte r having been mi sled on multip le occasions by DeGiorgio ' s repeated deni als. After conclusion of the long investi gation, the next step was referral to more comm ittees. J. The 2013- 2014 Recall Process In mid-D ecember 201 3, the Cobalt Ignition Switch fi nally eme rged from the FP E process and reached the Executi ve Field Action Decision Committee ("E FADC"), which is responsibl e for iss uing recalls. The first EFA DC meeting, however, ended without a recall dec ision and with a request for follow-up analysis to detennine the root cause of the airbag non-deployments. The failure to issue a recall at thi s EFADC meeting was another mi ssed opportunity, stemming from the continued, singular focus on findin g the root cause, and from a lac k of urgency due to a lack of awa reness of fataliti es among key decisionmakers. The EF AD C did not meet again until six weeks later - January 3 1, 2014 - and at that point issued a recall of MY 2005 - 2007 Chevrolet Cobalts and Pontiac G5s. The Saturn Ion and Chevrolet HHR were bri efl y discussed at the January 31, 20 14 meetin g. But EFADC members were incorrectly told, however, that although those models used the same igniti on switch as the Cobalt, they were different in several materia l respects and - based on Stouffe r's incomplete TREA D data pull in 20 11 , whi ch had not been updated - had not experienced crashes involving airbag non-deployments. It was not unti l the latter part of February, when EFADC member John Calabrese requested a fresh data pull, that the GM employees involved in the ign ition swi tch issue 2 14 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product discovered numerous accidents in volving the Ion and HHR in which the airbags did not deploy, incl uding fata l Ion cTashes in late 2003 and 2004. They also discovered numerous PRTS reports for the Ion related to moving stalls , some of which dated back to 2004. Calabrese called an emergency EFADC meeting for February 24, 201 4, and the EFADC illunediately expanded the recall to MY 20 0 3~2 0 07 Ion, HHR, Solstice and Sky vehicles. In the weeks that fo llowed, it came to GM's attention that housing ki ts containing the defective Ignition Switch were used to repair some MY 2008- 201 1 Coba lt, G5 , Ion, HH R, Solstice and Sky vehicl es, necessitating yet another reca ll. I. December 2013 ""Mini"-FPERC Meetin g In earl y December 20 13, the Cobalt reached the Field Performance Evaluation Recommendation Committee ("F PERC"), an executive comm ittee that reviews the details of an FPE investigation and makes a recommendation to the EFADC.98 t An informa l "small group" of FPERe members met on December 2, 20 13; Stouffer presented at the meeti ng, and attendees discussed prior crashes, fata li ties, and avail able warranty infomlation.98 2 Among the attendees were Benavides, Foley~Ga rd n er, Kemp, Kent, Stouffer, Wac htel , and John Murawa, an FPE investigator who woul d be replacing Stouffer upon hi s retirement. 983 Camlen Benavides, Product Invest igations Topics presentation (July 25 , 2013), at 2 LDOC ID 000003722007 J. allached \0 e-mail fromCanne n Benavides \oWill iamJ. Kem p, GM , er al. (July 25. 201 3) [DOC ID 000003722006]. 982 J&B Interview ofBriall Stouffe r, April 4, 201 4; J&B Interview of Maureen Foley-Gardner, March 26, 2014_ 98) J& B Interview ofCannen Benavides, March 25, 2014; J&B IllIerview of Maureen Foley-Gardner, March 26, 20 14; J&B Interview of Mark Johnson, April 23, 2014. There may have been other attendees as well, including Greg Hall. Kent could not recall whether she allended this meeting; others recalled that Kent did in fact allend. 1&B Interview of Gay Kent, March 18, 2014; J&B Interview o f Maureen Foley-Gardner, March 26, 2014. T here is some difTerence of opinion as to whether the fi rst meeting on December 2, 20 13 was a fonnal FPERC mccting or an infonnal meeting. Benavides and Mark Johnson reca lled that this was not a fonnal FPERC meeting . J&B Interview ofCannen Benavides, March 25, 201 4; J&B Interview o f Mark Johnson, April 23, 2014. Foley-Gardner recalled it as a "small group" rather than the usual FPERe allendees. J&B Interview o f Maureen Foley-Gardner, March 26, 981 2 15 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product At that time, Stouffer claimed to know of only two fatal accidents that might be associated with the safety defect: the fatality Sprague had described to Stouffer in 20 II , and the _ fatali ty described in th e Indiana Uni versity study that Palmer sent Stouffer in 2012. 984 This FP ERe meeting was the first time Foley-Gardner learned of fataliti es potentially associated with the Cobalt Igniti on Switch. 98s Accord ing to Foley~Gardner, the Cobalt Igniti on Switch issue went straight from the December 2, 2013 "mini" FPERe to the EF ADC because there was a desire by the legal department and Benavides to get the issue to the EF ADC. 986 It is unclear whether the FPERC considered recommending a recall of Delta and Kappa platform vehi cles other than the Cobalt, even though those ve hicl es shared the same ignition switch and had been included in the 2005/2006 TSBs. 9R 7 At that po int, the FPE investigation was still relying on Stouffer's incomplete data collection efforts in 2011 and was not aware of numerous ai rbag n o n~d eployment incidents involving the Ion and HHR, including several fatalities in Ion crashes. Stouffer retired on December 4, 20 13, and Wachtel left GM on December 20, 20 13 (Wachtel had formally retired in A ugust, but remained at GM as a contract empl oyee into December). The Cobalt matter was transitioned to Murawa, who had less than two weeks to delve into the details of the lengthy FPE investigation before presenting to the EFADC on 20 14; e-mail fromM,llIrcen Foley-Gardner, GM, toWilliam Kemp, GM , ef 01. (Nov. 20, 2013) IGMNHTSA0002235 I I ; DOC lD 000005057057]. 984 J&B Interview of Brian Stouffe r, April 4, 2014. Stouffer was also aware of the . aeeideIII , but he did not believe it was a ease in w hich the airbags should have deployed, regardless of the pOSition of the ignition switch, because it was a side impact crash and there were 11 0 side impact airbags on the vehicle. J&B Interview of Brian StoufTer, April 4, 2014. 985 J&B Interview of Maureen Foley-Gardner, March 26, 2014. n6 J&B Interview of Maureen Foley-Gardner, May 19, 2014. 987 At (ile December 2, 2013 meeting, Benavides asked Stouffer to (rack down the PPAP for the ignition switch from Delphi, and StoufTer contacted Gricbe at Delphi later that day 10 request (he PPAP. J&B Interview of Brian StoufTer, April 4, 2014. 2 16 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product December 17, 201 3. Murawa consulted with Stouffe r before and after hi s departure and met with DeG iorgio, Sprague, and Brian Thompson to understand the facts 988 The day before the EFADC, the Director of Global FPE e~ma il ed two coll eagues saying that she was " not clear on the ' rush ' for tomorrow 's meeti ng," adding "[w]e have skipped the step in the process w here we talk execution detai ls" and that she hoped to have a disc ussion rather than a decision "so we can understand the service procedure and the part availabili ty, etc.,,9&9 2. Dece mber 17, 2013 EFADe Meeting On December 17, 201 3, the Cobalt Ignition Switch fi nally reac hed the EFA DC and its three d ec i s i o n ~ makers - John Calabrese, VP of Engineering, Ali cia B o l er~ Da v i s , SVP, Global Quality & Customer Experience, and Gerald John son, VP of Manufacturi ng - who must be unanimous in order for a recall to be issued? )!) There are no minutes of EF ADC meetings and notes are not taken, so we do not have any written record of the discu5sion?91 J& 8 Interview of8rian Stouffer, Apri l 4, 2014; J&8 Interview of John Murawa, April 23 , 2014. Benavides stated that Murawa becamc an cxpcn in a vcry shon period of tilne and that Stouffcr' s depanurc did 110t slow the pace of the FPE process. J&B Imcrvicw ofCannen Benavides, March 25, 2014. Both Benavidcs and Mark Johnson claim to be the one who chose Murawa to replace Stouffer. J&B Interview of Mark Johnson, April 23 , 2014; J&B Interview of Camlen Benavides, M,lrch 25, 201 4. 989 J&B Interview of Maureen Folcy~Ga rdn er, May 19, 2014; c-mail from Maureen Foley-Gardncr, GM , to Jeffrcy Wrona, GM (Dec. 16, 2013); J&8 Ill1erview of Jeffrey Wrona, Apri l 23, 2014 990 J&B Interview of Gay Kent, March 18, 2014; FPE Overview presentation (Feb. 2014), at 6-7 LDOC ID 000005 I 29665J, attached to e-mail from Deborah Nowak-Vanderhoef to Cannen Benavides (Jan. 30 , 2014) LDOC ID 0000051296641. Boler-Davis joincd the EF ADC in 2012 and Johnson in July 2013. J&B Intcrview of Alicia Boler-Davis, March 21 , 2014; J&B Intcrview of Gerald Johnson, May 15, 2014. Neither oftheln rcceivcd any training regarding their EFADC duties. J&8 Interview of Alicia Boler-Davis, March 21 , 2014; J&8 Intcrview of Gerald Johnson, March 20, 2014. 991 J&B Interview of Alicia Boler-Davis, March 21 , 2014. According to Boler-Davis, in April 2014, the EF ADC began taking meeting minutes. J&B Interview of Al icia Boler-Davis, May 16, 2014. 988 2 17 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product EFA DC members rece ived a single PowerPoint slide on the day before the meeting, which did not include an y information on fatalities or the length of the FPE in vestigati on 9 92 Ca labrese attended in person, Boler-Davis sent Jeffrey Wrona as her proxy,993 and it is unlikely that Johnso n attended. 994 Other attendees included Benavides, Foley~Gardner, Kemp, Kent (by phone), Wachtel, and possibly others. 995 No one at the meeting questioned why th e investigation had taken as long as it did ? J6 Murawa presented at the meeting, and went through a number of PowerPoint sli des. Ca labrese was dissatisfied with the evidentiary support M urawa and his team provided and rai sed a number of factual questions that he said required further analysi s before a recall decision could be made. In particular, he believed that the FPE team had not yet identifi ed a root cause for the ai rbag non~deployment.997 He thought the FPE team should have been more prepared at the meeting. 998 In contrast, Murawa thought the EFADC had all the information necessary to John Murawa, 2005 ~ 2007 CobaltlPursuitlG5 presentation [GMNHTSAI00097398; DOC ID G MC8~ 0000OO7129 11J, attached to e-mail fro m Jolm Murawa, GM, to Maureen Foley-Gardner, GM (Dec. 16, 2013) IGMNHTSAI00097397; DOC 10 GMCB-00000071 29IOI; c-mai l from Chrystal Holmes, GM, to Foley-Gardner, GM , el al. (OCI. 4, 20 13) rGMNHTSA000223969; DOC ID 0000051158401; J&B Interview of Maureen FoleyGardner, May 19, 2014. 993 Wrona could not recall any other time when he served as Boler-Davis 's proxy at an EFADC meeting prior to December 17, 201 3. J&B Interview of Jeffrey Wrona, April 23, 2014. 994 J&B Interview ofCannen Benavides, March 25, 20 14; J&B Interview of Maurecn Foley-Gardner, March 26, 20 14; 1&B Interview of John Calabrese, March 28, 20 14; J&8 Interview of John Murawa, April 23 , 2014; J&B Interview of Jeffrey Wrona, Apri l 23, 20 14. In one of his interviews, Johnson stated that he participated by phone and WebEx, but he later claimed he did not participate al all. J&B Interview of Gerald Johnson, March 20, 2014. Accord ing 10 records of the teleconference line used to host this meeting, Gerald 10hnson did not dial in to the Deccmber 17, 201 3 EFA DC mccti ng. J& B Ill1 ervicw of Gerald Johnson, May 15, 20 14 995 J& B Interview of Maureen Foley-Gardner, March 26, 201 4; J&B Interview of John Calabrese, March 28, 20 14. Wachtel recalled in his interview that Stouffer presented at the December 17, 20 \3 EF ADC, but that is contradicted by other attendees and by the fac t thm Stouffer rctired earlier in December. J&B Interview of Douglas Wachtel, March 18, 2014. Wachtel himself retired shon ly aft er the Dccember 17, 2013 EF ADC, and had no funher involvement in the recall process. J&B Interview of Douglas Wac htel, March 18, 2014. 996 J&B Interview of John Murawa, April 23, 2014. 997 J&B Interview of John Calabrese, March 28, 2014. 998 J& B Interview of John Calabrese, March 28, 20 14. 992 2 18 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product issue a recal1. 999 Because answering Ca labrese ' s questions about the root cause required additional work, th e EF ADC meeting ended without a recall deci sion. Although it is possible that fatalities were mentioned at the meeting, they were not a focus of the discussion and they may not have been mentioned at all. Backup slides to Murawa 's presentation referenced fi ve fatalities and other serious injuries, 1000 but Murawa may not have presented those slides at the meeting. (Notably, Stouffer recalled that when he retired two weeks before the EF ADC meeting he was aware of only two fatalities). Murawa stated that he did present that slide,loo l while Ca labrese, Wrona, and Foley-Gardner did not recall fatalities being di scussed at the meeting. 1002 However, Wrona was aware at the time that fatalities were invol ved. l °O.' Other attendees said they did not learn about fatalities until sometime after the meeting. 1OO4 Boler~Davis explained that in EFADC meetings generally, backup slides to presentations are not di stributed and are not presented unless a particular backup slide includes infonnation responsive to a question posed by an EF ADC member. 100.5 When interviewed, Boler-Davis stated that death and injury data should always be included in EFADC presentation materials provided to members and should not be considered backup.l 006 It is al so unclear whether the Ion or HHR were discussed at the December 17 EFADC. ]007 These model s are referenced in some of the slide s that Murawa prepared, but whether those m J&B Il1\ervicw of John Mlirawa, Apri l 23,2014. 1000 Presentation from December 17, 2013 EF ADC Meeting, a1 28 [DOC ID 000001788022; GMHECOOOO02899]. 1001 J&B Interview of John Murawa, Apri l 14, 201 4. 1002 J&B Interview of Jolm Calabrese, March 28 , 201 4; J&B Interview of Jeffrey Wrona, April 24, 20 14; J&B Interview of Mallreen Foley-Gardner, March 26, 20 14. 1003 J&B Interview of Jeffrey Wrona, April 24, 2014. llJO.l J&B Interview of Alicia Boler-Davis, May 16, 2014. lOOS J&B Interview of Alicia Boler-Davis, May 16, 2014. 1006 J&B Interview of Alicia Boler-Davis. May 16, 2014. 1007 Neither the Sky nor the Solstice was d iscussed at the meeting. 2 19 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product slides were actua ll y shown is not c1ear.l OOS One witness recalled that the Cobalt was the primary vehi cle disc ussed, but that there were questi ons about the Ion and HH R related to potential differences in the positioning, orientation, and operati on of the igni tion switches in those cars. lOO9 Even if the Ion and HHR were di scussed, the EFADC was not provided with accurate data regardi ng airbag non~d ep l oyme nt incidents in those ve hicl es. Murawa's presentation was based on Stouffer's incomplete data pull from 2011 , in which Stouffe r had erroneously omi tted MY 2003-2004 Ions. IOIO Conseq uently, no informat ion about the fata l Ion crashes involving airbag non-dep loyments was provided in the EF ADC ' s materials. When asked why he did not support an immediate reca ll given the fact th at fatali ties had occurred, Calabrese explained that, at thi s point, he did not see a reason to rush to a recall because the backup slides to the EF ADC materi als showed a strong downward trend in the occurrence of airbag non-deployments and he wa nted the investigation to be done correctly so that the solution arrived at would actually fix the probl em. 1011 3. Developments Followin g Decemb er 17, 2013 E FADC Meeting Boler-Davis learned about the meeting' s outcome in a subsequent call on unrelated matters. I012 Boler- Davis stated that had she known at the ti me of the December 17, 20 13 EF ADC meeti ng that fatalities were invo lved, she would have treated the issue with more urgency.I Ol3 1008 Kent did not recall a discussion as to whether other Dclta Platfonn vehicles, including the Ion and the HHR, might also present a safety issue. J&B Interview of Gay KCllt, March 18, 20 14. Benavides recalls that the Ion and HHR were discussed. J&B Interview ofCannen Benavides, March 25 , 2014. 1009 J&B Interview of Jeffrey Wrona, March 20, 2014. JUJU J&B Interview of Brian StoufTer, April 4, 2014; J&B Interview ofCannen Benavides, March 25 , 2014. lOll J&B Interview of John Calabrese, March 28, 2014. 1012 J&B Interview of Alicia Boler-Davis. March 21 , 2014. lOll J&B Interview of Alicia Boler-Davis. March 21 , 20 14. 220 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product In late December 2013, Calabrese informed Mary Barra that the EF ADC was considering a possible recall , but dlat more anal ys is was needed hefore a recall decision could be made. The conversation was brief, and Calabrese did not provide Barra with details about the issue. 10l4 Barra told him to "get the right data; then do the right thing." IOI ~ This is the first time Barra learned of the Cobalt Ignition Switch issue. Kemp says that he rai sed with Clark Dougherty hi s disappointment that GM had not yet decided to recall the Cobalt and told her that GM would possibly face civi l penalties from NHTSA. 10I6 The first time that Clark Dougherty spoke to General Counsel Michael Millikin about the Cobalt ignition switch issue was in December 2013. 1017 She told interviewers that at that time she did not understand the seriousness of the issue because she did not have the key facts. 101 R As a result, she merely told Millikin that there was an issue with the Cobalt and that Engineering was looking into it. 101 9 Around Christmas, she to ld Mi ll ikin she would give him a fu ll update after tile holidays . 1020 4. J anuary 31, 2014 [FADe Meeting The EF ADC met a second time on January 3 1, 2014, and considered infonnation prepared in response to Calabrese's questions. 1021 Calabrese and Boler~ Davi s attended in person; J&B interview of Mary B,lrra, April 14, 201 4. J&B Interview of Mary Barra , Apri l 14, 201 4. 10 16 J&B Illl crvicw of William Kemp, Apri l 17, 2014 101 7 J&B Interview of Lucy Clark Dougherty, April 16, 201 4. lOI S J&B Interview of Lucy Clark Dougherty, April 16, 201 4. 10 19 J&B Interview of Lucy Clark Dougherty, April 16, 2014. lOW J&B Interview of Lucy Clark Dougherty, April 16, 2014. lin! Presentation from January 31 , 2014 EFADC Meeting rDOC ID 0000 133 12002; GMNHTSA000324650l; J&B Interview ofCannen Benavides, March 25, 201 4; J&B Interview of Gerald Johnson, March 10, 201 4; J&B Interview of Gary Altman, March 14, 20 14; J&B Interview of Gay Kent, March 18, 20 14; J&B interview of William Kemp, May 9, 20 14. l0 14 lO IS 22 1 Privileged and Confidenfia/ Protec1ed by AUorn ey~ Client Privilege and as Allorney Work Product Johnson likely did not attend. 1022 Murawa presented once again , and other attendees included Altman, Benavides, Foley-Gardner, Federico, Kemp , Kent, Thompso n, and possibl y others_lOB According to Boler-Davis, six weeks was not an unusuall y long time to wai t for an FP E investl gative team to complete additi onal research requested by the EFA DC. 1024 However, Foley-Gardner stated that it was not typical for that much time to elapse between EF ADC meetings. 1025 There had been an EFADC meeting earl ier in January, but Murawa was not ready to present at that meeting. 1026 At the January 31, 2014 EF ADC meeting, Murawa presented ( 1) evidence that there was a statistically signifi cant difference between the torque on pre-MY 2008 and post-M Y 2008 CobaJts, consistent w ith the pan change by Delphi; (2) ev idence that during rough dr iving conditions, the mass of the keys could exert enough fo rce to tum the Ignition Switch from Run to Accessory or Off; and (3) the results of hi s comparison of knee cl earance and key cylinder position across GM's vehicl es, showing that the Cobalt was not an outlier among GM's fl eet, and that therefore kn ee cl earance and key cylinder pos ition were not the root cause. 1027 Ca labrese concluded that the FPE team had suffic ientl y established a root cause to warrant a ll)22 In one of his interviews, Jolmson stated that he partici pated by phone and WebEx, but he later claimed he did not participate at all. J& B Interview of Gerald Johnson, March 10, 2014; J& B Interview of Gerald Johnson, March 20, 20 14. See also J&B Interview ofC;mnen Benavides, March 25, 20 14 (stating that Johnson did participate); J&B Interview of Maureen Foley-Gardner, Ma rch 26, 201 4 (noling that Johnson did not attend and that she reached out 10 Johnson later lhat day 10 oblain his vOle in suppon of lhe recall) 1023 J&B Interview of Gay Kent, March 18, 201 4; J&B Interview of Maureen Foley-Gardner, March 26, 2014. lI)24 J&B Interview of Al icia Boler-Davis. March 21 , 2014. She noted that Kemp sometimes called her to convey a sense of urgency regardi ng matters before the EFADC, but that he did not do so in connection with the Ignition Switch. /d. at 6. Jim J&B Interview of Maureen Foley-Gardner, May 19, 2014. lI)26 Foley-Gardner said lhat she had reached out to Murawa in early January to schedule a follow -up EFADe , but that he said he was not ready yet. J&B Interview of Maureen Fo l ey~Gardner, March 26, 20 14. lI)27 J&B Interview of Jolm Calabrese, March 28 , 201 4; J&B Interv iew of John Murawa, April 23, 2014. 222 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product recall .1028 Boler~ Davis agreed, and Johnson gave his approval by phone later that day.1 029 The recall covered MY 2005 to MY 2007 Chevy Cobalt and Pontiac G5 vehi cles It is not clear whether fa tali ties were di scussed at this EFADC meeting. 10JO T he EF ADC debated whether to recall any model years of the Ion or HHR, and decided not to because the available data did not indi cate that an airbag non-dep loyme nt problem existed in those vehicles. 1031 That information was erroneous; it was based on Stouffer's incomplete data collection efforts in 20 I I. Based on that fau lty premi se, the EF ADC was also told that vari ous factors could account for the lack of air bag no n ~dep l oyme n t events in the Ion and HH R, including the vehicl e suspension system and the angle of the steering column. 1032 Bo l er~D av i s informed Mary Barra of the recall on the same day the EFADC issued its decision.10]] She later observed that had she known about fa talities, her communication to Barra would have been " totally di fferent," and she woul d have incl uded the Ion and HHR in the recall .1034 J&B Interview of Gay Kent, March 18, 2014; J&B Interview o f John Calabrese, March 28, 2014. J&B Interview of Alicia Boler-Davis, March 21 , 2014; J&B Interview of Maureen Foley-Gardner, May 19, 20 14. lOW By this time, Davidson had pulled death and injury data relating to the Cobalt and provided it 10 Murawa. J&B Interview of Dwayne Davidson, April 22, 201 4. Benavides recalled that fatality data was presented 3t the mccting, while Foley-Gardner stated that it was not. J&B Intervicw of Carmen Benavides, March 25 , 2014; J&B Interview of Maureen Foley-Gardner, May 19, 2014. !OJI J&B Interview of John Ca labrese, March 28, 2014. TIle Solstice and Sky were not addressed at Ihe January 31 , 20 14 EFADC. IOn J&B Interview of Alicia Boler-Davis. March 21 , 2014; J&B Intcrvicw of Alicia Boler-Davis, May 16, 2014; J&B Interview of John Calabrese, March 28, 2014. Kent statcd tha t it was her understanding based on the January 31 , 20 14 meeting that "the Ion and HHR had a ' different electrieal architecture,' such that the deficient ignition switch did not present a safety issue in these vehicles." J&B Interview of Gay Kellt , March 18, 2014 . She funher noted that on January 31 , 2014, she did not have accurate data on the crashes and claims associated with those models. Id. 1033 J&B Interview of Alicia Boler-Davis. March 21 , 20 14. 1034 J&B Interview of Alicia Boler-Davis. March 21 , 20 14. 1028 !O29 223 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product On February 7, 2014, GM informed NHTSA ofa safety recall for MY 2005-2007 Cobalts and 2007 Pontiac G5s, approximatel y 780,000 vehicles at an estimated cost of $32.3 million . 103S The recall included all MY 2007 Cobalts because the company could not identify the "break point" during MY 2007 when the redesigned switch began to be installed in vehicles. 5. Boler~Da v i s 10J6 Developments Following January 31, 2014 [FADe MeetingfFebruary 7,2014 Recall Letter said that she first learned of fatalities only in early February during a leadership call including Barra, Millikin, and Reuss. 1037 In the same period, Kemp disclosed the Ignition Switch issue to Millikin for the first time. He could not expla in why he did not raise it with Mill ikin earlier and in hindsig ht says he probably should have. 1038 Kemp also met with Calabrese and Kent in early Febmary, and they discussed the Cobalt in vestigation and the need to identify process failures at GM and what could have been done differe ntl y.l039 Kent and Kemp mentioned that it was difficult to conduct an investigation when there is ongoing litigation regarding the same issue (i.e. , t h e . case). 1040 On February 17, 2014 , ajoumali st sent a series of questions to Alan Adler asking why the other model vehicles cited in the TSBs were not recalled as part of the MY 2005-2007 Cobalt and Pontiac G5 recall. 1041 On February 19, 2014, during the course of preparing a chronol ogy for submi ssion to NHTSA , Calabrese was asked about the Solsti ce and Sky and why they had GM Recall 573 Letter (Feb. 7, 2014) IDOe 10 000002796457; GMNHTSA000439255]. E-mail from Alicia Boler-Davis, GM. to Maureen Foley-Gardner, GM, el al. (Feb. 6, 2014) [DOC ID 000000080008; GMNHTSA000373278]; e -m~i l from M~ ureell Foley-Gamer, GM, 10 Alicia Bo ler-Da vi~ , GM , and Gerald Johnson, GM (Feb. 5, 20 14) LDOC ID 000001738002/ GMNHTSA0002246 17J; J&B Interview of Alicia Boler-Davis. March 21 , 20 14. 1037 J&B Interview of Alicia Boler-Davis, March 21 , 2014. 1038 J&B Interview of William Kemp, Apri l 17, 2014. JOSY J&B Interview of Gay Kent, March 18,2014. 104\l J&B Interview of Gay Kent, March 18,2014. 1041 E-mail from James Healey. USA Today, to Alan Adler, GM (Feb. 17, 2014) [DOC 10 000002819087; GMNHTSA00032527IJ. 103S 1033 224 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product not been included in the initial recall. He wa s troub led and requested a data sweep for those and other De lta and Kappa Platform vehicles that shared the same Ignition Switch_I042 The next day, the New York Times publi shed an article criticizing GM for not recall ing all of the Delta and Kappa platform vehicl es included in the 200512006 TS8s. I043 Calabrese and others met on Friday, February 2 1, 20 14, and over the weekend to review the data. They reviewed a full TREAD data pu ll for Delta and Kappa vehicles, as well as the results of a PRTS search . lo44 Although they found no incidents in volvlng non~deployments in the Solstice or Sky, they identifi ed 22 incidents in Ions, and six for the HHR.I045 This is the first time that anyone involved in the investigation realized that Stouffer's 2011 data collection was incomplete and that Stouffer had missed key information about Ion and HHR airbag deployment inc idents. non~ lo4 6 Calabrese called an emergency EFADC meeting for February 24, 2014, to consider expand ing the recall. 6. February 24, 2014 EFADe Meeting The EFADC met a third time on February 24, 2014, and expanded the recall to the Chevrolet HHR and Pontiac Solstice for MY 2006-2007, the Saturn Ion fo r MY 2003- 2007, and J&B Interview of John Calabrese, March 28, 2014. E~ mail from Alan Adler, GM, 10 Lucy Clark Dougheny, GM, el aI. , forwarding a New York Times article entitled G.M Recalls SOlli e Cars, Bill Not All, Willt Ignilion Swi/clt Problem (Feb. 20, 2014) lDOC ID lO·n lO·t) 000005082754; GMNHTSA0002906431. In response 10 Calabrese' s request for a data search, Brian TIlOmpson "obtained special pri vilcdgcs rsicl in Ihe PRTS s.yslem 10 a cce~~ ' legal' incidel1l~ s.uch a~ FPR~ _ Th c~e are not available to the typical ~y~tem mer" Mllrawa explained in a February 23 , 201 4 e-mai l that running an ordinary search without these privileges '\vould potentially not get all documents, but we would assume avery thing LsicJ available was in our possession. This could have been what happened 10 Stouffer." E-mail from John Murawa, GM, to Mark Johnson, GM (Feb. 23 , 2014) rDOC ID GMCB-000000713224; GMNHTSA 100 t 053961JI).t, J&B Interview of John Calabrese, March 28, 2014. 11)46 J&B Interview of John Calabrese, March 28, 2014; J&B Interview of Brian TIlOmpson, April 23 , 2014; e-mail from John Calabrese, GM, to Mark Reuss, GM, el al. (Feb. 23 , 2014) [DOC ID 000002693074; GMNHTSA000225149J. )1)4,1 225 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product the Saturn Sky for MY 2007. 1047 Attendees included Calabrese, Boler~Davis , and Johnson (by phone), as well as Foley~Gardner.1 048 The EF ADC included the Solstice and Sky in the recall because the EF ADC did not have any infonnation indi cati ng that the 19nition Switch did not pose a similar ri sk in the Solstice and Sky.l o49 A presentation prepared for the meeting noted that adding all of these model s, approximately 842,000 vehicles, to the recall would increase the cost of the recall by an estimated $34.3 million. 7. t050 March 2014 Service Parts Recall In the weeks that followed, GM personnel di scovered that the original Ignition Switch may have been used to repair certain MY 2008-20 11 Cobalt, GS , lon , HHR, Solstice and Sky veh icles. 1051 As a result, on March 28, 20 14, GM extended its January 3 1 and February 28 recall s to include replacing all Ignition Switches in the MY 2008- 20 II population of model s covered by those recalls, adding approximately 970,808 additional vehicles, at a proj ected cost of $39.7 million. VII. 1052 GM'S LEADERSHIP, THE TONE AT THE TOP, AND GM CULTU RE Thi s section addresses a number of issue s that do not fit neatl y into a chronology, but that broadly affected the entire company during the relevant period. Below we address the rol e of the current senior leadership at GM and the Board; the extent of their knowledge of the issues di sc ussed above; the messages sent by senior leadership and the environment that they created Prcscntation fmm EFAOC Mccling (Feh. 24, 201 4) [DOC ID 00000173 1002; GMHEC0000027351. J&B Interview or Maureen Foley-Gardner, March 26, 2014; J&B Interview of Gerald Jolmson, March 20, 201 4. 11)49 J&B Interview of Alicia Boler-Davis. March 21 , 2014; J&B Interview of John Calabrese, March 28, 2014. )(}5() Prcscntation from EFAOC Mceting (Feb. 24, 2014) rODe ID 00000173 1002; GMHEC0000027351. )M) J&B Intcrvicw of Cindy Hasscin, April 1, 2014 ; J&B Intervicw of Nancy Mclean, April 2, 2014; J&B Interview of Gary Smits, April 2, 2014; J&B Interview of Deb Nowak-Vanderhocf, April 3, 20 14; J&B Interview ofCanncn Bcnavides, April 4, 2014; J&B Interview of Greg Hall, April 3,2014; J&B Intcrview of Manhcw Schroedcr, Apri l )1)47 1048 3. 2014. l(}S2 GM 573 Letter regarding recall N 14092 (March 28, 2014) LDOC ID 000050895592]. 226 Privileged and Confidenfia/ Protec1ed by AUorn ey~ Client Privilege and as Allorney Work Product for employees ("tone at the top"); and GM' s cu lture and the impact that it may have had on how GM ' s employees responded to the Cobalt' s problems_ A. Current Se nior Leaders hip The current C EO Mary Barra, the current Executi ve Vice Pres ident, Global Product Development, Purchasing & Supply Chain Mark Reuss, and the current General Counsel Mi chael Millikin did not learn about the Ignition Switch 's safety issues and the lengthy delay in addressing th em until after the EFA DC dec ided to issue the recall on January 3 1, 2014. A few weeks earlier, in December 201 3, Ms. Barra and Mr. Millikin had bee n provided limited informati on that data was being reviewed that could lead to a reca ll. As part of Jenner' s engagement, we were asked to prioritize our review of th e in volvement, if any, of these three current senior leaders in the events that led to the belated Ignition Switch recall. We conducted a thorough investigati on that included (a) review of a large number of documents collected from numerous custodia ns, including potentially relevant e-mails any of the three seni or leaders sent or received at pertine nt times; (b) interviews of GM empl oyees in th e three leaders' respecti ve chains of reporting who might have di scussed Cob a lt~ related issues with them; and (c) interviews of each of Mary Barra, Ma rk Reuss and Michael Millikin. All of th e evidence we reviewed corroborated the concl usion that none of the three current leaders had knowledge of the prob lems with the Cobalt 's Ignition Switch or nondeployment of airbags in the Cobalt until Dece mber 20 13 at the earl iest. Before becoming GM ' s CEO on January 15, 20 14, Mary Barra had served for the preceding nearly three years as Senior Vice Pres ident and then Exec uti ve Vice President for Global Product Development, Purchasing & Supply Chain. Before becoming responsible for 227 Privileged and Confidenfia/ Protec1ed by AUorn ey~ Client Privilege and as Allorney Work Product Global Product Development in February, 2011 , Barra had served in other capaciti es that did not typi call y in volve recall s or the recall dec isi on~making process_ 10S3 As head of the Product Development organi zation, Barra had responsibility for the global vehicle engineering fun ction, headed by John Calabrese, as well as the global quality function headed by Terry Woychowski and th en Alicia Boler_Davis. los4 The engineering and quality fun ctions encompassed many of the perso nnel involved in the FP E process as we ll as the members of the EFADC. Ba rra became well acquainted with the recall process when the issues in volving the Chevrolet Volt' s lithium~ ion battery arose in 2011 . Based on that experi ence and others she bel ieved th at recall iss ues were addressed with appropriate urgency and that the recall dec ision-makin g process worked well. 1055 Barra first began to learn of some aspect of the Cobalt ' s Ignition Switch issue in December 20 13, when she was told by Calabrese that the company was working on an analysis that might lead to a decision to recall the Cobalt. She did not learn more about the matter until shortly after the EF ADC made the reca ll decision on January 3 1, 2014.1056 Our review of contemporaneous e-mails and oth er documents corroborated tJl ese reco ll ections. In that connection, we investi gated certain e ~mail s that rai sed issues with the Cobalt or with mov ing stalls with Barra before year~ e nd 201 3, but tho se e ~mail s all concerned other issues, not the Ignition Switch. First, in Ma rch, 201 0, Barra was among many rec ipi ents of an e~ma il invitation to an FP E team meeting attaching information about a power steering J&B Intcrvicw of Mary Barra , April 14, 201 4. In mid-201 3 thcn-CEO Daniel Akerson shifted the reponing line for Boler-Davis as Senior Vice Presidcnt of Global Quality and Customcr Experiencc directly to the CEO. 1055 J&B Interview of Mary Barra, Apri l 14, 201 4. 1056 J&B Interview of Mary Barra, Apri l 14, 201 4; J&B Interview of Alicia Boler-Davis, March 25 , 20 14. 1053 11)>4 228 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product prob lem in the Coba lt.1 057 By that date Barra had become Vice President of Global Human Resources and had no role in such a meeting or its subj ect matter ; she does not recall the e~ mail or its attachment, which concerned a separate power steeri ng issue, not the Igni tion Switch. 1058 Second, on October 3, 20 11, after Barra had become Senior Vice President for G lobal Product Development, the then-Vice President fo r G loba l Q uality sent her a press report about a power steeri ng problem in the Ion and questioning w hether it should have been included in GM's recall of the Cobalt a year earlier for a similar issue. 1059 Again, nothing in this report menti oned any prob lem with the Igni tion Sw itch, moving stall s, or a irbag n o n ~dep l oy m e n t in either the Ion or the Cobalt. Third, on April 22 , 20 12, Barra was one of two recipients o f an e ~mai l fro m a former GM emp loyee reporting a moving stall in hi s Bui ck that he d escribed as an " isolated/freak" incident attribu table to the key design and s uggesting that the company investigate and perhaps issue a service bulletin. Barra forwarded the e-ma il to the VP for Global Q ua lity, Terry Woychowski , and asked him to investigate the matter. 1060 Barra 's e~mai l was then forwarded within the company and an e~mail discussion ens ued ultimately invo lving eight people, but no longer incl ud ing Barra . In o ne s uch subsequent entry in the e~ m ail string, Wachtel mentioned that early Coba lts had a simil ar issue but a design change to the key had addressed the issue. I061 Barra was not co pied on any of the subsequent e-mai ls and did not learn of the Cobalt issue at that time. 1062 J&B Interview of Mary Barra, April 14, 2014. J&B Interview of Mary Barra, April 14, 2014. 1059 J&B Intcrvicw of Mary Barra , April 14, 2014. 1060 J&B Intervicw of Mary Barra , April 14, 2014. 11J61 J&B Intcrview of Mary Barra , April 14, 2014. Barra statcd that none ofthc subscquent c-mails in thc chain, following her April 22, 2012 c-mail, had e ver comc to hcr attcntion. ld. Barra did not rccall anyone communicati ng to her a resolution of the concern she forwa rded to Woychowski 011 April 22, 20 12. ld. She also stated that neither Cannen Benavides nor Maureen Foley-Gardner - both of whom arc copied on Doug Wachtel ' s Apri l 23 , 2012 e!(IS7 lOSS 229 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product Finally, on November 19, 2013, Ali cia Boler-Davis responded Maureen Fole y~ Gardn er to an e-mai l from that had alerted Boler-Davis to a likely Cobalt field action involving its Ignition Switch and airbag non-deployment, by saying that Foley-Gardner should se nd her a quick summary that she would then share with Barra. 106., Barra was not copied on that e-mail, and Boler-Davis reports that she did not share any infonnation with Barra about the Cobalt issue until January 31 , 20 14. None of these e-ma il s is evidence that Barra was informed about the Cobalt Ignition Switch issues before Calabrese alerted her in December to the anal ys is that was being undertaken on that subject. Mark Reuss succeeded Barra as EVP for Global Product Deve lopment, Purchasing & Supply Chain, when she became CEO. Before that he had been Presidem ofGMNA from 2009 through 20 13, and before heading GMNA he had served in various engineering positions within GMNA except for brief periods when he led GM 's business in Australia and New Zea land in 2008-09 and a short stint in 2009 as Global Vice President of Engineering. 1064 He joined GM in 1986. He did not have engineering responsibility for the Cohalt in any of hi s positi ons.1 06 5 As President ofGMNA, Re uss was typicall y infonned of recalls after th e EF ADC had made lts decisions. He learned of the Cobalt recall after the EF ADC made its deci sion on mai l - brought the Cobalt complaints referenced in Wachtel's email 10 her allention, or ever spoke to her about issues with the Cobalt during the time period of this email.ld. We confinned ill imerviews or through counsel that neither Wachtel 's April 23, 20 12 e-mail descri bing 2005 Cobalt complaints nor the historical complaints themselves were brought to Barra 's attention. Wachtel confirmed that he did not speak to Barra about his email or the topic of the 2005 Cobalt complaints, of whieh - per hi s email - hc heard nonc after 2005. Terrencc Connolly also confinncd lha! hc ncver discllssed lhe ema il with Barra. J&B Imerview of Terrence Connolly, April 28, 2014_ Cannen Benavides did not recall the e-mail and did not recall having any discussions with Barra about it. Maureen FoleyGardner said she would 110t have paid close attention to the e-mail because she was just coming into her role, and she did not recall any follow-up to thc c-mnil. J&B Interview of Maurecn Foley-Gardner, May 19, 2014. Jcffrcy Konchan did lIot rceallthc e-mn il. J&B Intcrvicw of Jcff Konchan , May 7, 2014. JIJ6! J&B Intcrvicw of Mary BnITa , April 14, 201 4; J&B Intervicw of Alicia Bolcr-Davis, March 25 , 2014. 1063 J&B Intervicw of Mary Bnrra , April 14, 201 4. 106-1 J&B Interview of Mark Reuss, April 15, 2014. I06S J&B Interview of Mark Reuss, April 15, 2014. 230 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product January 3 1,201 4. 1066 None of the documents we reviewed or interv iews we conducted identifi ed any evidence that is in consistent with Mr. Reuss ' recollection Michael Millikin became General Counsel in Jul y 2009, after more than three decades in various positions in GM ' s Lega l Department. Between 2000 and 2009 he had responsibili ty in various positions for coordinating GM's global lega l processes and did not have any role in recall s, product li ability matters or safety-related issues in the United States. Until the recall decision was made in 20 14, he had not been in formed of the lengthy review of the Cobalt ' s Ignition Switch issue in whi ch GM lawyer Bill Kemp had participated for years. He also had been unaware of the litigation involving fatal accidents by Cobalt drivers, such as the _ case. IO t';7 Mi ll iki n's direct and indirect subordinates, who at various times before 20 14 became aware of th e problems with the Cobalt, confirmed in inte rviews that they did not bring th em to Mi ll ikin 's attention. 1068 The _ case had been settled for $5 million, the highest amount for which his subordi nates had authority to reso lve a li tigation without obtaini ng his approval. In December, 20 13, the General Counsel ofGMNA, Lucy Clark Dougherty, mentioned to Mi ll ikin that a possible recall was under consideration. 1069 Kemp and Clark Dougherty reported the recall decision to Millikin and the facts underlying it for the fi rst time on February 6, 2014.1070 Kemp acknowledged that he coul d not explain why he had not raised the Cobalt safety issues with J&B Interview of Mark Reuss, April 15, 2014. J&B Inlerview of Michael Millikin, April 4, 2014; J&B Interview of Lawrence Buonomo, April 16, 2014. 1068 J&B Inlerview of William Kemp, Apri l 17, 2014; J&B Interview of Lawrence Buonomo, April 16, 2014; J&B Interview of Lucy Clark Dougherty, Apri l 16, 2014; J&B Interview of Jaclyn Palmer, April 24, 20 14 ; J&B Interview of Ronald Porter, April 30, 2014. 1069 J&8 Interview of Michael Millikin, April 4, 2014; J&B Interview of LllCY Clark Dougherty, April 16, 2014. lI)70 J&B Interview of Michael Millikin, April 4, 2014; J&B Interview of LllCY Clark Dougherty, April 16, 2014. 1066 1067 23 1 Privileged and Confidenfia/ Protec1ed by AUorn ey~ Client Privilege and as Allorney Work Product Millikin before then. 107 1 Our investi gation did not identify any documents that were inconsistent with these recollecti ons of Millikin or the lawyers who reported to him B. The Board To investigate the Board's role in the oversight of safety issues and the information that the Board rece ived, we interviewed several current and former outside directors and chief executive offi cers, as well as many of the GM executi ves who made pertinent reports to the Board during the II-year peri od from 2003 to early 201 4. We also reviewed written materials prov ided to the Board during that period. The Board ' s oversight of vehicle safety had multiple face ts; no single committee of the Board was responsible for all vehicle sa fety~ related issues, and the Board and its committees received a wide variety of reports that related at least in part to vehicle safety. Th ose reports included informati on in an aggregate fonn and did not di scuss individual safety issues or indi vidual recalls except in rare circumstances. Typically, neither the Board nor any of its commi ttees reviewed the Company ' s response to spec ifi c product de fects, which were the subject of hundreds of product investi gations globally each year, culm inating in an average of approx imately 30 recall s each year. \072 As di sc ussed in more detail below, the Board and various committees rev iewed infonnation fro m management on a vari ety of different s afety~re\ ated issues: a) evaluations of the quality and safety of GM' s vehicles by thi rd parties such as Consumer Reports and JD Power (the full Board) ; l(}7) J072 J&B Intcrvicw of William Kemp, Apri l 17, 2014. See Recall Data from NHTSA, 001 website, ami/able at http://www-odi.nhtsa.dot.gov/downloadslflatfiles.cfm. 232 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product b) the Company' s interactions with NH TSA (the full Board); c) recalls and warranty costs (the full Board); d) the annual internal audit plan, which on two occasions during the relevant period included audits that encompassed the reca ll decision~making process (the Audit Conunittee); e) risk management matters generally, whi ch potentially could have raised issues relating to vehicle safety (the Aud it Committee); and f) the Company' s public position on potential legislation and other public policy matters, which also could have touched on vehicle sa fety (Public Policy Committee). As described elsewhere in this repon, the slow pace of the Cobalt investigation and the emerg ing pattern of accidents potentially related to the safety defect were not escalated to the Company's most seni or executives who from time-to~time met with the Board. The Board of Directors was not informed of any problem posed by the Cohalt ignition switch until February 20 14. 1. Board Oversight Process a. The Full Board During the period in question, the Board as a whole received three different types of periodic reports that related to safety issues. 1. Peri od ic Reports Quality Reports. One means by which the Board exerc ised overs ight of safety-related issues was its review of informati on from management concerning vehicle "quali ty" iss ues. GM had a V ice President of Quality who oversaw a group that interacted w ith the engineers on 233 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product quality issues; worked to reduce GM 's warranty costs; tracked how GM's vehicl es were eval uated by publi cations like COl/sumer Reports and.lD Power; identified areas for improvement, including potential solutions or best practices used by other OEMs; and participated in evalu ations of potential defects and recall dec isions. I073 The Company considered safety issues to be partly the responsibility of thi s quality function , along with the engineering and manufacturing functions, because proper design and protection against production defects should ensure safe vehicle operation, and minimize safety issues resulting from product failures. I074 The Board recognized that "qua li ty" encompassed safety .1075 During the period 2003-20 14, the Board regularly received reports from management on the quality ofGM 's vehicl es. 107o [n the earlier part of thi s period, the reports focused on speci fi c regions,I077 and in the later years, after the Company had g lobalized its quality function, the reports were typically more detailed and oft en included a focu s on GMNA.I 078 These reports were often accompanied by an oral presentation by the Vice President o f Quality or another J&B IllIerview of Kevin Williams, May 7, 2014; J&B Interview of James Hresk.o, May 7, 2014. See. e.g., J&B Interview of Mary Barra, April 14, 201 4; see also J&B Interview of Phil Laska\vy, May 20, 2014; 1&B Interview of Daniel Akerson, May 15, 201 4; J&B Interview of Kevin Williams, May 7, 2014. 1075 See. e.g. , J&B Interview of Ph il L1skawy, May 20, 2014; J&B Interview of Erroll Davis, May 20, 2014; J&B IllIervicw of Daniel Akerson , May 15, 2014. 1076 See. e.g. , Quality Update Preselllation to General Motors Board of Directors (Del. 10, 2013) [DOC TO 000240787014]; Minutes of the Regular Meet ing of the Board of Directors of General Motors Company (Oct. 9~IO, 20 13), at 10 [DOC TO 00024078701 31; 2011 Qual ity Perfonnanee Summary Presentation from Boardbook for the General Motors Board of Directors Meeting (Jan. 10- 11 , 2012) [DOC TO 000240752013] ; Consumer Reports Presentation from Boardbook for the General Motors Board of Directors Meeting (Feb. 2-3 , 2(09) LODe ID 0002407570471 ; Minute s of the Regular Meet ing of the Board of Directors of General Motors Corporation (Feb. 23, 2009), at 3 rOOC ID 0002407870101; GM Non h America Quality Review Presentation from Boardbook for the Genera l Motors Board of Directors Mect ing (June 1-2, 2004) [DOC ID 00024075701 0]; Minutes of the Regular Meeting of the Board of Directors of General Motors Corporation (June 1-2, 2004), at 2 [DOC lD 000240787006]. 1077 See. e.g. , GM Asia Pacific Quality Review Presentation from Boardbook for the General Motors Board of Directors Mecting (Feb. 2-3, 2004) rODe ID 0002407570071; Minutes of the Regular Meeting of the Board of Directors of General Motors Corporation (Feb. 2-3, 2004), at 2 rDOC ID 0002407870051J 1)7~ See. e.g. , 2011 Quality Perfonnancc Summary Presentation from Boardbook for the General MOlors Bo.1rd of Directors Meeting (Jan. 10- 11 , 20 12) rDOe ID 0002407520131; Quality Review Presentation from Boardbook for the General Motors Board of Directors Meeting (Dec. 6-7, 2010) [DOC ID 000240757074]; Minutes of tile Regular Meeting of the Board of Directors of General Motors Company (Dec. 6-7, 20 10), at 3-4 LODe lD 000240787012]. 1073 lInl 234 Privileged and Confidenfia/ Protected by AUorney~Client Privilege and as Allorney Work Product knowledgeable executive. 1079 Simil ar presentations were made at the regular meetings held among the CEO and hislher direct reports. 1080 These GMNA quality reports typically included in formation on three broad topics, all of which re lated in part to safety. First, these reports foc used on GM ' s rankings in buyers ' guides li ke Consumer Reports and JD Power. IOS I Because these publ ications are highl y respected by consumers and often playa part in buying decis ions, GM personnel believed it was important to compete with other manufacture rs in thi s foru m, and the Board was particul arly focused on these issues. I082 Consumer Reports and JD Power both include important safety~related information. For example, Consumer Reports includes eva luations of each car's "accident avoi dance" (based on testing of braking and steering systems) and crash test performance (based on NHTSA and Insura nce Institute of Highway Safety testing).I083 JD Power reported on ini tial quality (how 1079 See. e.g, Quality Update Presentation to General Motors Board of Directors (Oct. 10, 2013) LDOC ID 0002407870141; Minutcs of the Rcgular Mcct ing of thc Board of Dircctors of General Motors Company (Oct. 9~10, 20 13), at 10 [DOC ID 0002407870131; G lobal Quality Proccss from Boardbook for lhc Gencral Motors Board of Dircctors Meeting (April 16-17, 2012) [DOC ID 000240752012] ; Minutes of the Regular Mccting oCthe Board of Directors of General Motors Company (April 16-17, 2012), at 2-3 LDOC ID 000238341259J; Quality Update from Boardbook for the General Motors Board of Directors Meeting (Aug. 20-21 , 20 12) 1DOC 10 0002407520111; Minutcs of the Regular Mecting of the Board ofOircctors ofGcncral Motors Company (Aug. 20-21 , 2012), at 5 [DOC 100002383412621 ; Prcsident's Repon Quality Ovcrvicw Prcscntation from Boardbook for the Gencral Motors Board of Dircctors Mccti ng (OCI. 6, 2009) [DOC ID 0002407570541; Minules of the Rcgular Mecting of the Board of Directors of General MOlors Company (Oel. 5-6, 20(9), al2 [DOC ID 000240787011]. lOW See. e.g., Global Quality Update 10 GM Executive Opcrations Committee (Dcc. 18, 2012) [DOC ID 000240756004]; J&B In terview of D;mie l Akerson, May 15 , 20 14. IOSI See. e.g., 2011 Quality Perfonnance Summary Presentation from Boardbook for the General MOlors Board of Directors Meeting (Jan. 10-11 , 2012) 1DOC lD 00024075201 3 1; Consumer Repons Presentation from Boardbook for thc Gencral Motors Board ofOircctors Mcct ing (Fcb. 2-3, 2009) [DOC ID 0002407570471 ; Minutes ofthc Rcgular Mccting of the Board of D irec l o r~ ofGcncral MOIor~ Corporation (Feb. 2-3 , 2009), aI 3 [DOC ID 000240787010]; GM North America Quality Review Presentation from Boardbook for the General Motors Board of Directors Meeting (June 1-2, 2004) LDOC ID 000240757010J; Minutes of the Regular Meeting of the Board of Dircctors of Gencral Motors Corporation (Junc 1-2, 2004), at 2 [DOC ID 00024078700611082 See. e.g., J&B Intervicw of Gcorge Fishcr, May 21 , 2014; J&B Intervicw of Danicl Akerson, May 15, 2014; J&B Interview of Thomas Stephens, May 7, 2014; J&B Interview of Erroll Davis, May 20, 2014; J&B Interview of James Hresko, May 7, 2014. 1083 See Consumer Reports, "Compare Safety Ratings" (last updated Feb. 20 14), ami/able al http://www.consumerrepons.org!cro/2012/ 12/compare-safety-ratings/index htm; see also General Motors, 235 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product many defects were reported in the fi rst three months after purchase/084 as well as dependability over time_ I08S The reports to the Board concerning the ranki ngs ofGM's cars in COl/slimer Reports and JD Power made it clear that GM had ample room for improve ment in those eva luati ons, but they did not identify that GM 's vehicles were unsa fe or less safe than those manufactured by GM 's competitors. I086 Second, the quality reports incl uded in fo rmation about reca lls. That information typ ica ll y did not reference specific vehicl es or ind ivid ual recall dec isions, but rather represented aggregate data about GM 's recall s over time. More specifica lly, management typicall y provided the Board with data concerning the costs it incurred annuall y fo r "field actions," which included recall s, as well as the actual number of recall s GM issued each year and the number of vehicles Consumer Reports Presentation from the GM Board of Directors Director's February 2009 Session Boardbook (Feb. 2-3, 2009), at 3-5 [DOC 10 000240757047[. 1084 See. e.g. , J. D. Powcr, " 2013 In itial Quality Study - Car Ratings & Rankings," amilable al hnp://autos.jdpowcLcom/ratingsJqua lity hun; see also GCllcral Molors Quality Review Prcscntation from thc GM Board of Director's June 2004 Session Boardbook (June 1-2, 2004), at 3-5 LDOC ID 00024075701Oj. lOSS See. e.g., J. D. Power, "2013 U.S. Vehiclc Depcndability Study" (Fcb. 13,20 13), available at hup://autos.j dpowcLcomJcolltcntJprcss-rc Icasc/b TiXtc5/20 13-u-s-vchie1c-dcpcndabi Ii ty-st udy. hun rOOC ID 0002407870391; see also Gcncral Motors Qua lity Updatc Prescntationto thc Gcncral Motors Board of Dircctors (Oct. 10,20 13), at 12 rDOC ID 00024078701 41. ]086 See. e.g., Gcncral Motors, Quality Updlltc Presentation \0 Gcneral Motors Board of Dircctors (Oct. 10,2013) LDOC ID 000240787014J ; Minutcs of the Regu lar Meeting of the Board of Directors of General Motors Company (Oct. 9- 10, 2013), at 10 IDOC ID 000240787013]; 2011 Quality Perfonnance Summary Presentation fTom Boardbook for the General Motors Board of Directors Meeting (Jan. 10-1 1, 20 12) lDOC ID 000240752013 J; Quality Review Presentation from Boardbook for the General MOlOrs Board of Directors Meeting (Dec. 6-7, 20 I0) fDOC ID 0002407570741; J.D. Power Survcy Rcview/Quality Updatc Prescntation from Boardbook for thc Gcncra l M010rs Board of D ircelor~ Mcc1i ng (June 2-3, 2008) [DOC ID 000240757043]; Minu1es Oflhc Regular Mecting of the Board of Directors of General Motors. Corporation (June 2-3, 2008), at 3 1DOC lD 000240787009J ; Consumer Reports Review from Boardbook for the General Motors Board of Directors Meeting (Dec. 3-4, 2007) lDOC ID 0002407570381; Minutcs of thc Rcgular Mcet ing of thc Board of OircclOrs of Gcncral Motors Corporation (Dcc. 34, 2007) fDOC ID 0002407870081; J. D. Power Survcy Rcview Prcscntation for thc Gcncral Motors Board of Directors Meeting (June 4-5, 2(07) LDOC ID 000240757035J; Minutes of the Regular Meeting of the Board of Directors of General Motors Corporat ion (June 4-5, 2007), at2 LDOC ID 000240787007J; GM North America Quality Rcvicw Prescntation from Boardbook for the Gcncral Motors Bomd of Dircctors Mccting (Junc 1-2, 2(04) [DOC ID 0002407570tO[. 236 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product encompassed by those recall s. lOS7 Those numbers varied from year-to-year, sometimes significantly. For example, in December, 201 0, the Board received a presentation from the Vice President for Quali ty showing that GM ' s fi eld actions had increased that year~ management committed to " tum th is around while protecting the customer. "I088 The Board also reviewed the same annual recall-related statistics for GM's principal competitors. Based on the recall -related info nnation they received over the years , the Board members and executives we interviewed believed that GM ' s recall process was fu nctioning appropriately.1089 Third, the quality reports foc used on GM's warranty cost experience, whi ch was also an ind icator of whether its vehicles were safe to drive. 1090 Warranty claims originate with customers experiencing real-world iss ues with their veh icl es. And while warranty cl aims are by no means limited to problems with safety equipment or that otherwise reflect a vehic le' s safe operation, a lower warranty cost genera lly indicates that a vehicle is dependable and reliable, wh ich tends to mean the vehicl e is safer as we ll . l091 As described earlier in thi s report, GM's engineers also viewed warranty claims as one source of information about possible safety issues. Litigation Reports. In addition to the quality reports, the full Board received a second regular report concerning safety iss ues: The Board received, in connection with essentially every See. e.g., 2011 Quality Perfonnance Summary Presentation from Boardbook for tile General Motors Board of Directors Meeting (Jan. 10-11 , 2012) LDOC ID 00024075201 3J; Quality Review Presentation from Boardbook for the General Motors Board of Directors Meeting (Dcc. 6-7, 2010) [DOC ID 0002407570741; Minutes of the Regular Meeting of the Board of Directors of General Motors Company (Dec. 6-7, 20 10), at 3-4 [DOC ID 0002407870121. 1088 Quality Review Pre~entatio n from Boa rdbook for the General Motors Board of Directors Mccting (Dec. 6-7, 20 10), at 15. LDOC ID 000240757074j. lOS') See, e.g., J&B Interview of George Fisher, May 21 , 2014 ; J&B Interview of Erron Davis, May 20, 2014; J&B Interview of Phil Laskawy, May 20, 2014; J&B Interview of George ("Rick") Wagoner, May 12, 20 14; J&B Interview of Mary Barra , Apri l 14, 2014. IIJ90 See. e.g., President ' s Repon Quality Overview Presentation from Boardbook for [he General Motors Board of Directors Mccting (Oct. 6, 2009) [DOC 100002407570541; Minutes of the Regular Meeting of the Board of Directors of General Motors Company (Oc\. 5-6. 2009), at 2 [DOC ID 00024078701 I}, 1(191 J&8 Interview of George ("Rick") Wagoner, May 12, 2014. 1087 237 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product meeti ng, a written legal update on pending or recent lega l matters, encompass ing not onl y product li abili ty li tigati on but also other types of di sputes incl uding antitrust, environmental, and empl oyment matters. Ion Prior to GM 's bankruptcy, the Board also received an Annual Report on Product Liability Litigation that described both acc ident cases and asbestos matters. 1093 In general, these report s summarized the number, severity, complexity and subj ect matter of product liability claims that GM had experienced in the most recent period. Both types of reports highlighted claims that were noteworthy for some reason, for example, cla ims involving un usua lly large verd icls. 1094 These reports incl uded safety~related in formation to the extent that safety issues were all eged and at issue in certain litigated matters, such as ro llover risks or seat belt system fai lures. 1095 No cl aim, settlement, or other aspect of any li tigation involving the !On See, e.g., Lcgal Revicw and Updatc, from Boardbook for thc Gencral Motors Board ofDireelOrs Mceting (Fcb. 3·4, 2003) [DOC ID 0002407570021; Lcgal Rcvicw and Updatc, from Boardbook for the General Motors Board of Directors Meeting (Feb. 6, 2006) LDOe ID 000240757022J ; Legal Report, from Boardbook for the General Motors Board of Directors Mceting (Sept. 8·9, 2009). [DOC JD 0002407570531; Legal Rcport to GM Board of Dircetors, from Boardbook for the Gcneral Motors Board of Dircctors Mecting (OCI. 10- 12,20 12) [DOC ID 0002407870401. 1093 See, e.g., Lcgal Revicw and Update, from Boardbook for thc General Motors Board of DireclOrs Mceting (Fcb. 3·4,2003) lOOe LD 000240757002J; Lcgal Review and Updme, from Boardbook for the General Motors Board of Directors Meeting (Feb. 2-3, 2004) rooe ID 0002407870471; Legal Review and Update, from Boardbook for the Gcncral Motors Board of Dircctors Mect ing (Jan. 31 - Feb 1,2005) [DOC ID 0002407870481; Legal Revicw and Updatc, from Boardbook for the Gencral Motors Board of DircclOrs Meeting (Fcb. 6, 2006) [DOC ID 0002407570221. 1094 See. e.g. , Lcgal Revicw and Updllte from Boardbook for the General Motors Board ofDireetors Mccting (Feb. 2-3, 2004) rDOe ID 000240787047]; ILegal Rcview and Update from Boardbook for thc Gcneral Motors Board of Dircctors Meeting (Jan. 3 I-Feb. 1, 2005) [DOC ID 000240787048 ]; Lcgal Revicw and Updatc from Boardbook for the General Motors Board of Directors Meeting (March 1, 2005) LOOe ID 0002407870481; Legal Review and Update from Boardbook for the General Motors Board of Directors Meeting (May 9, 2005) [DOC 10 000240787045]; Legal Revicw and Updatc from Boardbook for the Gencral Motors Board ofDireetors Mecting (May 2, 2006) [DOC ID 000240787046]. ](195 See, e.g., Legal Review and Update from Boardbook for the General Motors Board of Directors Meeting (Feb. 3-4,2003), at 131DOC lD 0002407570021; Legal Review and Update from Boardbook for the General Motors Board of Directors Mceting (Dcc. 2, 2003) rDOe ID 0002407870431; Lcgal Rcvicw and Updatc from Boardbook for thc Gencral Motors Board ofDireetors Mect ing (Feb. 2-3, 2004) rDOe ID 0002407870471 ; Lcgal Rcview and Updatc from Boardbook for the Gcncral Motors Board of Directors Mceting (Jan. 31 - Fcb 1, 2005) [DOC ID 0002407870481; Legal Revicw and Updatc from Boardbook for the Gencral Motors Board ofDireetors Mecting (Feb. 6. 2006), at 8 [DOC lD 000240757022]; Legal Review and Update from Boardbook for the General Motors Board of Directors Meeting (May 2, 2006) LDOe ID 000240787046j. 238 Privileged and Confidenfial Protec1ed by AUorney~Client Privilege and as Allorney Work Product Ignition Switch, airbags, or moving stall s in the Cobalt vehicles was specificall y referenced in an y of the litigation reports provided to the Board. Regulatory Reports. The third type of s afety~re l ated report received by the fu ll Board concerned the regulatory environment, including matters such as responses to the passage of the TREAD Act, GM's NHTSA obligations, and various sa fety~ re lated initiatives that had been proposed. As requirements under the TREAD Act and GM 's obligations to NHTSA evolved, at times the Board rece ived additional repo rts about these matters,IO'J6 including a 2004 report that the Company had agreed to pay a $ 1 milli on fin e to reso lve a disagreement with NHTSA. 1O'J 7 Thi s fine was reported to the Board as follows: Over the past two years, GM has disagreed with the federa l traffic safety agency (NHTSA) over whether certain windshield wiper problems in 2002 and 2003 SUVs and certain engine stall ing issues invol ving recent mode l Saabs posed an unreasonable risk to safety and should be recalled for that reason. To apply pressure on GM , the agency iss ued a civil pena lty demand letter to GM alleg ing that GM had violated the Traffic Safety Act by not providing timely notification of and remedy for these alleged defects. Th is led to hig h ~level di scussions which toned down the rhetoric and resolved a number of other open issues on a basis satisfactory to GM. However, as part of thi s reso lution, GM agreed to pay a one million doll ar penalty relating on ly to the windshield wiper iss ues. The agenc y agreed to downplay the penalty and forego any press release. Through a continuation of the h i gh~leve l dialogue between agency officials and GM ' s top engineers , GM will endeavor to minimize and avoid such situations in the future. 1098 Overview of Gcnera l MOlor~ North America Lc gall~~lIe~ and Activity, from Boardhook for the Gcncral Motors Board of Directors Meeting (Feb. 29, 2004), at 9 LDOC ID 000240787055J ; U.S. Legislative and Regu latory Issues from Boardbook for the General Motors Board of Directors Meeting (April 5-6, 2009), at 8 LDOC 10 0002407870641; Legislative and Regulatory Issues Update and Regulatory Issues Primer from Board book for the General Motors Board ofDircctors Meet ing (Nov. 2-3, 20(9), at 6-7 rDOC ID 0002407870491. 11)97 Legal Review and Update from Boardbook fo r the General Motors Board of Directors Meeting (Aug. 3, 2004), at 10 [DOC ID 000240787042]. 1098 Legal Review and Update from Boardbook fo r the General Motors Board of Directors Meeting (Aug. 3, 2004), at 10 LDOC ID 000240787042]. 1096 239 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product In contrast, in 2013, the Board, including the then-CEO. was not infonned that NHTSA had recently criticized GM's responsiveness regarding several spec ified recalls, service bulletins, and investi gations. 1099 11. Specific Reports In unusual circumstances, management also reported about speci fic safety~relate d events. In one such case, in June, 2011 a Chevrolet Volt, GM's first production vehicle with an electric propulsion system, su ffered what was described as a "thermal event" three weeks after being subjected to severe NHTSA crash testing. I100 Management reported to the Board on the status of GM ' S response to the Volt event,llOI and kept the Board apprised thereafter as di scussions with NHTSA over the appropriate resolution proceeded. The Board also recei ved a report on the Toyota unintended acceleration problems in 20 10. 1102 GM's management had conducted a "deep dive" to determine whether any GM vehicles (other than the Pontiac Vibe, which was developed and produced with Toyota) were susceptible to a similar ri sk of unintended acceleration, and determined that they were not. lJ03 Management's report to the Board in March 2010 indicated that Toyota's problem was not shared by GM or other major manufacturers. 1104 E-mail from Michael J. Robinson, GM, 10 Gay Kem, GM (July 24, 201 3) LDOC 10 000240755008j. See NHTSA "Chevrolet Volt Battery Incident Overview Report" (Jan. 20 12), a vailable al www nhtsa.gov/staticfiles/nvs/pdflFinal_ Reports.pdf [DOC 10 000240787024l 1101 See Minute~ oflhe Regular Meeling of I.he Board ofDireelor~ ofGenernl MOlor~ Company (Dec. 13, 20(1). LDOC ID 000240752007J; J&B Interview orrhil Laskawy, May 20, 1014; J&B Interview of Daniel Akerson, May 15, 2014. 1102 Legislative and Regulatory Issues Update from Boardbook for the General Motors Board of Directors Mccti ng (March 1-2, 2010), at I [DOC ID 0002407570641 ; see also J&B Interview of Phil Laskawy, May 20, 2014. IIOj E-mail from Michael P. Millikin, GM, to Deborah Nowak-Vanderhoef, GM (Feb. 17, 20(0) [DOC ID 1099 1100 000006293898]. 11(1.] See Legislative and Regulatory Issues Update from Boardbook for the General Motors Board of Directors Meeting (March 1-2, 2010), at I lDOe ID 000240757064j. 240 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product b. Board Committees In addition to the reports to the full Board described above, two Board committees - the Public Policy Committee and the A udit Committee - had specific responsibiliti es that related to so me aspects of vehicl e sa fety. I. Public Poli cy Com mittee The Public Poli cy Committee, as its name suggests, focused on the positions GM adopted on public policy issues, and on GM's acti viti es with respect to pending legislation aDd regul ation that affected GM. The scope of th e Public Po licy Committee's responsibility in th e peri od from 2003 through mid-20l3 encompassed the "soc ial and political " issues that might affect the Company's business and reputation; the li st of such issues incl uded "automotive safety" but also encompassed other issues such as corporate socia l responsibility, diversity, environmental matters, education, employee health and sa fety, trade, and philanthropic activiti es.IIOs In mid20 13, the Public Policy Committee revised its charter to make clearer that its focus was on issues, including vehicle safety, onl y to the extent they affected the Company's reputation in the legislative and regulatory arenas. I 106 In clarify ing its charter, the Public Policy Committee noted General Motors Company Public Policy Commitlee Charter (Nov, 15, 201 1) [DOC ID M_OO206123]; General Motors Company Public Policy Commince Chaner (Oct. 6,2009) rOOC ro 0002407570551; see also General MOlors Corporal ion Proxy SlaiCmelll (Apri l 17, 2003), aI 7 [DOC ID 000240787056]; General MOlors Corporation Proxy Statement (April 29, 2005), at 7 LOOC ID oo0240787057J; General Motors Corporation Proxy Statement (April 25, 2(08), at 12 LOOC 10 000240787058J; General MOlors Company Proxy Statement (Apri l 26, 2012), at 22 [DOC 100002407870591. 1106 See Proposed Public Policy Commincc Responsibilities, rrom Materials ror Ihe General MOlors Company Publ ic Policy COlluninee Meeling (June 5, 2013), at 1 rOOC ID 0002407550061; see also General Motors Company Public Policy COllunillee Chaner (June 6, 2013) rOOC 10 0002407550071; Olher GM Board Committees - Ri sk Management Responsibilities, rrom Materials ror the General MOlors Company Fin:'lIlce and Risk Policy Comminee (Oct. 4, 2010), at 1-4 LOOC ID 0000000 181721. 1I0S 24 1 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product that with respect to «Product Safety," the oversight responsibility already resided with the " Full Board .,,1107 Consistent w ith these iterations of its charter, the Public Policy Comminee djd not have oversight responsibility for the recall deci sion-making process, nor did it undertake to review management 's processes for ens uring compliance with NHTSA's regulatory requirements concerning recalls. 11. Audit Committee The Audit Comminee, in addition to its core function overseeing GM's financial reporting process and systems of di sclosure and internal controls, was al so responsible for oversight of GM 's external and internal auditors, and its ri sk management process. I 108 These aspects of the Audit Committee's work provided two ove rsight mechanisms for the Audit Committee to receive information about processes that related to vehicle safety, among many other subjects. First, each year the Audit Committee reviewed the annual internal audit plan that had been prepared by the internal audit function in conjunction with management. 11 09 Prior to that review, the chairman of the Audit Committee typically personally reviewed the proposed plan in J&B Interview of Errol! Davis, May 20, 201 4; Proposed Public Policy Committee Risk Assignments, from Materials for the General Motors Company Public Policy Committee Meeting (June 5, 2013), at 2 [DOC 10 0002407550061; e-mail from Michael Mi llikin, GM , to Daniel Akerson, GM (Feb. 4, 2013) (noting that the Quality risk received regular allCI11 ion by the Board of Dircelors) [DOC ID 000240752014] ; Minules Of lhc Public Policy Committee (June 12, 2012), at 2 [1405150oo11905J; Memorandum from Bob Ferguson to Members of the Public Policy Committee and Daniel Akerson (Oc\. 11 , 2011) [DOC lD 000240757081J; Finance and Risk Committee Meeting Presentation (March 14,2011), 3t4-5 rDOC lD 0000000183431))08 Audit Committee Charter (Nov. 15, 2011) rDOC ID 00024075202 11I IIJ'} See. e.g. , 2010 Global Audit Plan, from Boardbook for the General Motors Board of Directors Meeting (Nov. 30-Dec. 1, 2009) roOC lD 000240787061 t 2009 Global Audit Plan, from Materials for the General Motors Audit Committee Meeting (Dec. 1, 2008) [DOC ID 000240787060]; Minutes of Meeting of the Audit Conmlittee (Dec. 3, 2007), at 4 [DOC 10 000240787053 J. 1107 242 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product detail at a spec ial ha lf~day meeting with seni or internal a udit staff II 10 The internal audi t plan covered a broad array of topics affecting the globa l enterprise; the internal aud itors have conducted on average more than 100 internal audits each year over the past decade. I III During that period, two internal audi ts incl uded a review ofGM's process concerning fie ld actions and recalls or Field Performance Eva luations. An audit of the Globa l Quality organization in 20 13 identifi ed modest opportunities for process enhancement, but did not find any prob lem with the process by whi ch safety issues were identified or addressed (and di d not refer to any spec ifi c vehi cle issue).1112 In 2006, internal aud itors conducted a similar audit of GM~Europe's quality fun ction, which it found satisfactory.IIB Thus, the interna l audit fu nction rev iewed GM's process for evaluating possibl e defects or for making recall dec isions, and that rev iew was subj ect to the Audit Commi ttee's oversight. Second, the Audit Committee oversaw GM 's risk management process, including rev iewi ng the "ri sk factors" described in GM 's public disclosures, 1114 and meeting regularl y with the Chief Risk Officer in the years hi s work was overseen by the Committee. 11 15 In recent years, See. e.g. , Gencral Auditor's Rcport (March 3, 2008) roOC ID 000240787027]; J&B Intcrview of Phil Laskllwy, May 20, 2014. Illl See. e.g., Audit Status Rcport - All Closed Audits 2003~2014 roOC ID 000240794002]; Minutcs of the Audit Committce (Dec. 1,2008), at 21DOC 10 0002407870521. 1112 Audit Report on Globa l Quality (Nov . 25, 2013) rDOC ro 00024075501ll Illl Repon on Audit of General MOIors Europe Quality (Planning and Operations) (April 12,2006) [DOC ID 000240757023]. 1114 See, e.g., Enterprise Risk Management, from Boardbook for the General Motors Board of Directors Meeti ng (Nov. 30-Dec. I, 2009), at3~4 rOOC ID 000240787026]; Minutes of Meeting No. 443 of the Audit Commil\ee (OCI . 1,2007), al 4 [DOC ID 000240787051 1IllS Agenda of the Regular Meeting of the Board of Directors of General Motors Corporation (Nov. 18, 2013), at 2 LDOC ID 000240787017 j; Agenda of the Regular Mceting of the Board of Directors of General Motors Corporation (March 18,20 13), at 2 rDOC ID 0002407870161; Agenda of the Regu lar Meeting of the Board ofDireclOrs of General Motors Corporation (Aug. 20~21 , 2012), at 2 rDOC ID 000240787015]; J&B Interview of Brian TIlclen, May 2, 2014. For a two~yearperiod between mid~2010 and mid-20 12, the Board had a "Finance and Risk" committee 10 which oversight of general risk management was allocated. Finance and Risk Poliey Commil\ee Charter, from Materials for Meeting of the Finance and Risk Policy Committee (Oct. 1,2010), at 14-16 [DOC ID 0000000 18172]; Finance and Risk Commi llee Meeting Presentation (May 16,20 11), at 3 LDOC ID 000000018399]. 1110 243 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product the Company 's interna l " Enterprise Ri sk Management" ("E RM ") process has identified "quality" as an important ri sk and defin ed it as: "[m]aj or or chroni c product problems could occur, resulting in negative public image, large product recall campai gns andlor significant, unexpected increases in warranty expenses.,,1116 Management and the Board have addressed thi s risk through their emphasis on quality improvement, 111 7 and no specific vehi cle safety issue was brought to the Audit Committee's attention as part of its oversight of risk management other than a risk assoc iated with lithium ion batteries. lll s 2. The Cobalt Ignition Switch a. The Board's Lack of Knowledge Before 20 14, none of the written reports to the Board included any information concerning the Ignition Switch. We have not identifi ed any evidence that the Board was otherwise informed of this problem, of GM 's delay in address ing it, or of the fatalities or accidents associated with it. Moreover, none of the executi ves we inte rviewed who presented on quality or recall issues at Board meetings in the pertinent period were aware of the Ignition Switch problem. Bcforc and aftcr that pcriod, the Audit Committec had gcneral ovcrsight responsibility for risk managcment. Gcncral Motors Company, 20 13 Proxy Statcmcnt (April 25, 2013). at 23 rDOC ID 0002407870291; Audit Committce Chartcr within Gcneral Motors Corporation, 2006 Proxy Statement (April 28, 2006), at A~ l - A-3 [DOC ID 00024078702811116 20 13 Key Risk Classification (De<:. 18, 2012), at 16 [DOC ID 0002407520 161; Audit Committee Meeting Documents (Aug. 19, 2013), at 155-159 rDOC ID 0000000285821: see also Gcncra l Motors Company, 20 13 Annual Rcpon (Fonn IO-K) (Fcb. 6, 20 14), at 23 [DOC ID 000240787033 ]; Gcneral MOlors Company, 201 2 Annual Repon (Fonn IO- K)(Feb. 15 , 201 3), at 25 LDOC ID 000240787031J; General Motors Company, 2011 Annual Repon (Fonn IO- K) (Feb. 27, 201 2), at 25-26 [DOC ID 000240787032J; General Motors Company, 2010 Registration Statement (Fonn S-I ) (Aug. 18, 2010), at 24 rDOC ID 0002407870621; Gencral Motors Company, 2010 Annua l Rcpon (Fonn 10-K) (March 1, 2011), at 38 rDOC ID 0001607432881; General Motors Company, 2009 Annual Rcpon (Fonn 10-K) (Apri l 7, 2010), at 31 rDOC ID 00024078703 1l 1117 J&B Intcrview of Phil Laskawy, May 20, 2014; J&8 Interview of George ("Rick") Wagoner, May 12, 2014. 1118 Materials for the General Motors Company Audit Committee Meeting (March 18,20 13), at 117 [DOC ID 000240787039j. 244 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product As noted above, the Board did not typically receive reports on the subj ect of individual defects or indi vidual recalls; th e Board also did not receive data about traffic fataliti es associated with GM's vehicl es, except in the context of some of the regular liti gation update reports that referenced a small number of cases unrelated to the Coba lt. b. The March S, 2009 Body/Exterior SMT VPR Update Slide Deck In March 2009, CEO Rick Wagoner may have viewed a slide deck that incl uded, in a "back-up" slide, a reference to the Cobalt's inadvertent c ut~ offiss u e. That slide, in a 72-slide presentation, described the then~recent change in the Cobalt's key design from a slot to a hole. 11 19 The back~up slide focused solely on warranty cost reduction and did not characterize the matter as a safety issue or mention airbag non-dep loyment, accidents or fatalities. 1120 Wagoner does not recollect reviewing any part of the slide deck. Jl2 1 Because the slide deck was fo und in the data collected fro m Wagoner 's computer, and the potential importance of Wagoner 's possible review of the back~up slide, we conducted an extensive investigati on of its origin and of the meetings in which it was discussed to determine whether Wagoner was informed of the change or of the unmentioned safety impli cations. The background to th e slide deck follows. As discussed above, in early 2009, GM personnel implemented a change to the design of the Cobalt 's igniti on key to a hole from a slot beginning with production of the Cobalt ' s 20 10 1119 llW 1121 BodyfExtcrior SMT, VPR Update Plan/Perfonnancc (March 5, 2009), at 34 [DOC 10 W_006073501. BodyfExterior SMT, VPR Update Plan/Pcrfonnancc (March 5, 2009), at 31, 34 [DOC ID W 00607350]. J&B Interview of George ("Rick") Wagoner, May 12, 2014. - 245 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product model year in mid-2009. Wagoner did not attend the meeting in January 2009 at which the keyhole design change to save warranty cost was discussed. I 122 About six weeks later, certain slides that had been used in the January meeting were included among 57 "back-up" slides that appeared in a 72-slide deck that in turn was used for one of nine separate presentations made at a meeting on March 5, 2009, of the Vehic le Program Review ("VPR") team, which met on a close-to-weekly basis. 1123 The presentation that included the back-up slides focused on both material and warranty cost reductions achieved by the "Body/Exterior SMT," one of the "system management teams" that focused on cost. quali ty and related issues with respect to eight separate systems across all GM vehicle platforms. I 124 The presentation's primary slides addressed such actions as «supplier workshops" and a new " week of claim tool" devised to expedite iss ue identification; some of the many back-up slides in the presentation focused on "side closure" issues, and in turn the back-up slides to the "side closure" back-up slides included the January 2009 slides about changing the Coba lt 's key head design from a slot to a hol e. 1125 Wagoner did not attend the March 5, 2009 meeting or, at least typically, any other of the VPR meetings. 1126 He and more than 20 other executives were routinely e-mai led summaries of the di scussion after each week's V PR meeting ended, and they may have had access to the slide J&B Interview of Gary Altman, May 20, 2014. See Gary Altman, Program Quality Sufficiency Rcview (Jan. 29, 2009), at 38-40 IDOC ID 000007551904; GMNHTSA000281704l; Body/ Extcrior SMT, VPR Updatc Plan/Pcrfonnallcc (March 5, 2009), at 30-34 rDOC 1D W_00607350]; e-mail from RaIlCharbY, GM, loAmhony Brown, GM , c1. al. (Feb. 27,2009) [DOC ID 000106626436]; VPR Meeting Notes (March 5, 2009) IDOC ID H_ 00262247], attached to e-mail from Richard Ceppos , GM, to George ("Rick") Wagoner, GM, et. al. (March 5, 2009) IDOC ID H_00262244]; VPR Meeting Notes (March 5, 2009) [DOC 1D H_00262247l; e-ma il from Richard Ccppos,GM, to Edward Whitacre, GM (Dcc. 17, 2009)fDOC 10 W 00000480]. 11 24 J&B Intcrvicw of Gary Altman, May 20, 2014; J&B Interview of Jamcs Hcntschel, May 27, 2014. 112S BodylExtcrior SMT, VPR Update Plan/Perfonnancc (March 5, 2009) rOOC 10 W_0060735011126 J&8 Interview of George ("Rick") Wagoner, May 12, 2014; J&B Interview of Richard Ceppos, May 21 , 2014; J&B Interview of James Hentschel, May 27, 2014; J&B Interview of Laurie Huizar, May 23, 2014. 1122 1123 246 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product deck presentations on a shared inte rnal website. 1127 The engi nee r who believes he presented the Body/Exterior SMT's information at the VPR meeting states that the back-up slides concerning the Cobalt key design change were not di scussed at the V PR meeting, and the summary of the March 5, 2009 meeting does not menti on that subj ect. 1128 Forensic analysis indicates that the BodylExterior SMT presentation was opened on the evening of March 5, 2009, on Wagoner's computer, located in two files in a temporary folder bearing generic names. The presentation was not found in the data of the other custodians whose electronic files were coll ected in the investigation. There is no forensic evidence that Wagoner reviewed any specifi c slide within the presentation. As noted, Wagoner does not recollect viewing the presentation or the back-up slides; about three weeks later, on March 29, 2009, Wagoner agreed to resign as CEO at the request of the U.S. governme nt 's Auto Task Force. 1129 Contemporaneous e-mails he exchanged with the person who provided the summa ry notes of the meeting do not mention the Cobalt issue or any other specific topic. 1l3O 3. Other Board Activities One way in w hi ch the Board expressed its commitment to improving the quality ofGM's vehi cles was through the design of the short-term incentive compensation - bonus - plans that app li ed not only to corporate officers but also to employees at the executi ve, director, manager, and supervisor levels. With a sing le exception in the past decade, the annual targets GM had to 1127 See e.g. , E-mail from Ran Charby, GM , lo Amhony Brown, GM , CL al. (Fcb. 27 , 2009) [DOC ID 000106626436]; e-mail fromRichardCeppos. GM . toGeorge(..Rick..)Wagoner. GM. et.al. (March 5, 2009) LDOC ID H 00262244]; e-mail from Richard CepPos, GM , loEd Whitacre, GM (Dec. 17, 2009) [DOC ID W 00000480J. 1128 J&B Intcrvicw of James Hentschel, May 27, 2014; VPR Meeting NOles (March 5, 2009) rDOC ID H_002G22471, attached to e-ma il fromRichardCeppos. GM.toGeorge(.. Rick ..)Wagoner. GM .ct. al. (March 5, 2009) [DOC ID H 00262244]. 1129 J&B Interview- of George ("Rick") Wagoner, May 12, 2014. llW E-mail from George ("Rick") Wagoner, GM, to Richard Ceppos, GM (March 6, 2009) [DOC ID W_00037388]; e-mail from George (" Rick") Wagoner, GM , toRichard Ceppos, GM (March 8, 2009) lDOC ill W_00041634]. 247 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product achieve before any bonus would be payable included improvement of the quality of GM ' s vehi cles_ The quality component in the annual target constituted 10% of the overall calcu lation, and three finan cial and sales components accounted for the balance for each year in the past decade except 2010, when the target's components were limited to the three finan cia l and sales metrics. 'U' In about mid~2013 , Al icia Boler~Davis, the head of Global Quality, became a direct report to then~CEO Dan Akerson (having previously reported to Mary Barra as EVP of Global Product Deve lopment, Purchasing & Suppl y Chain) and proposed an increase in the relative weighting of the "quali ty" factor in the incenti ve pl an to 25%. The Board's Executive Compensation Committee approved the increased importance of quality improvement in determining the extent of bonuses, effective thi s year. I m Si milarly, the annual performance obj ectives for the CEO set by the Board's Executi ve Compensation Committee included quality improvement as among the many factors by which the CEO's performance would be assessed by the Board. 1l33 C. Tone at the Top on Safety It is impossible to catalog all possible directives and management actions that might generall y have influenced how GM emp loyees viewed their roles and responsibilities. It is even more difficult to ascertain how the general tone set by sen ior leadership affected specific decisions made by individuals. Where indi vidual s refe renced specific management directives as the ca use of their actions, we have identified them. In most circumsta nces, however, we could J&B Interview of Janice Uhlig, May 12,2014 & May 16,2014. J&B Interview of Janice Uhlig, May 12, 20[4 & May 16, 20[4. 1133 J&B Interview of Janice Uhlig, May 12, 2014 & May 16, 2014. 1131 1m 248 Privileged and Confidenfia/ Protec1ed by AUorn ey~ Client Privilege and as Allorney Work Product not ascribe a particular management action or policy directi ve from a senior executive as the reason for an y specifi c acti on But the tone set at the top is relevant background for assess ing GM's approach to the issues di sc ussed in thi s report. Repeated throughout the interview process we heard from GM personnel two somewhat different directives - "when sa fety is at issue, cost is irrelevant" and "cost is everything." It is worth examin ing how those two messages co llided. GM personnel were quite consistent in say ing that they understood that safety was a critical priority and that, if they identifi ed a safety probl em, cost should not be a factor in deciding whether and how to address the safety prob lem. For example, a senior manager in Accessories Engineering said that safety is the top pill ar at GM. I I .l 4 Training materi al and di rectives that have been located for FPE process, the key engineering process for reviewing sa fety issues, make no references to cost~benefit analyses. lm Indeed, an "FPE Sensitivity and TREA D Training" presentation fro m May 2004 by Kev in Will ia ms, G M North America Vice Presi dent of Quality, notes: The harsh real ity is ~ we are competing in a new world , one that demands a culture where there is no tolerance for defects at any point during in [sic] the vehi cle deve lopment and manufacturing process. Beca use the marketplace has zero tolerance for defects, thi s organization will have no to lerance for defects. If I so und alanned, I am. . . . You must also become gatekeepers o f quality. Consider every issue a potential defect and ri sk to the customer. I like to say 'stand in front of th e train '. Stop the problems from flowing downstream! If, despite your best efforts, you cannot stop a problem, your next action is to pull your andon [sic] and escalate the issue up to someone who can. We have esca lati on systems in place ,lor th·IS very reason SO don ' t heSltate . to use them. 1136 Messages such as this are consistent with what we heard from many employees. l lJ~ J&B Interview of Khris Lcc, April 3, 2014. See. e.g., New GMPT FPE Director's Orientat ion (Jan. 10, 2008) rDOe ID GMHECOOO384265 ; GMNHTSA000383962 ; DOC ID 000001345940]. 1136 Kevin Williams, FPE Sensitivity and TREAD Training (May 2004), at 68 LDOe 10 000087206552]. 1135 249 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product That sa id, the 2000s was a time of extraordinary cost~cutting at GM. The messages from top leadership at GM - both to employees and to the outside world - as well as their actions were focused on the need to control costs. We heard repeatedl y from GM personnel about the focus on cost~cutting and the problems it caused. For example, an engineer stated that an emphasis on cost control at GM «permeates the fabric of the whole culture. ,,1137 Cos t~c utting impacted all aspects of the business. Keeping projects on time - because of the impact on cost - became a paramount concern. One witness expressed concern that the cost~ and time~cutting principles known as the "Big 4" emphasized timing over quality.IIJ8 Those principles were introduced to GM in the early 2000s. I139 Those responsible for a vehicle were responsible for its cost, but if they wanted to make a change that incurred cost and affected other vehicle s, they also became responsible for the costs incurred in the other vehicles. I 140 For example, if the Cobalt team wanted an ignition switch replaced, the other vehicle lines that used the ignition switch would request that the cost for their new switches be paid for by the Coba lt team because the Cobalt team requested the change. I J4 1 Reductions in staff, especially in Engineering, meant that employees were forced to do more with less. In the time leading up to the bankruptcy, one cost~cutting measure was to decrease the Engineering headcount by adding to the responsibilities of the Design Release Engineer (the position held by Ray DeGiorgio). J 142 Witnesses stated that the reduction in force J&B J&B IISY J&B ))4(1 J&B 1141 J&B 1142 J&B IJ37 1138 Interview of Philip Honon, March 3 [, 20 [4. Interview of Mark Reuss, April 15, 2014. Interview of Mark Reuss, April 15, 2014. Interview of$tcve Oakley, April 23 , 20[4. Interview of$teve Oakley. April 23 , 2014; J&B Interview of Sarah DeVries. May 13 & May 16, 2014. Interview of Mark Reuss, April 15. 2014. 250 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product created a difficult en vironment in w hi ch people were overworked and the quality of work suffered _ The cost~c u tt in g naturally flowed through to suppliers. O ne cost-cutting measure in the time leading up to GM ' s bankruptcy was to source parts routinely to the lowest bidder, even if they were not the hi ghest quality parts. I 143 We have uncovered no evidence that any employee made an explicit trade-off between safety and cost in the investigation of the Cobalt. To be sure, the Coba lt engineers working in the 2004-2006 timeframe rejected various fixes to the moving stall issue because there was "no acceptable business case ," but those engineers' error was that they fa il ed to understand the connection to airbags and the safety issue that they were facing. Having wrongly identifi ed the issue as a customer convenience issue , cost considerations that would otherwise have been immaterial became part of their calculus. That noted, we cannot conc lude that the atmosphere of cost-cutting had no impact on the failure of GM to resolve th ese issues earli er. As discussed above, GM was under tremendous cost pressure, and it was imposing tremendous pressure on its suppl iers to cut costs. Engineers did not believe that t hey had extra funds to spend on product improvements. Staff was cut dramatically. It is not feasible for three to do ajob as effectively as eight (in the case of the team charged with pulling TREAD data), and there were specific impacts. As noted above, the responsibility of the OREs was increased. The group charged with the TREAD database "pared down" their affirmative searches to look for trends because of their reduced headcou nt. 1I 44 J&B Intcrvicw of Mark Reuss, April 15,2014. Interview of Dwayne Davidson, May 6, 2014. The TREAD database team was originally approximately sixteen or seventeen employees as GM worked to develop its TREAD database, but that nLUnber was reduced to 11 43 11 4,1 J&B 25 1 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product When belts are tightened, most function s are impacted in some way, a.nd we cannot assume that safety was immune_ D. Culture 1. Resistance to Raising Issues Some witnesses said that there was resistance or rel uctance to raise issues or concerns in the GM culture. For example, a Red X Manager said that, ifan employee tried to raise a safety issue five years ago, th e employee would get pushback. 1145 Mary Barra explained that problems occurred during a prior vehicle launch as a result of engineers being unwilling to identify issues out of concem that it would delay the launch. 1146 Similarly, an employee survey refl ected an issue related to speaking up. In a Corporate Executi ve Board Company survey admini stered at GM in 20 13, GM participants' rate of reporting mi sconduct they observed was below the benchmark rate developed by the Compliance and Ethics Leadership Counci l based on the responses and experi ences of participating companies. 1147 A small munber of participants also suggested a fear of retaliation. Whi le the survey comments were unconnected to safety questions, issues of culture cannot be eas il y confined. Some witn esses provided examples where culture, atmosphere, and the response of supervisors may have di scouraged indi vidual s from rai sing safety concerns, including, in a different context than the Cobalt, supervi sors warning e mployees to " never put approximately cighl artcr thi ~ stan ~lIp pcriod. J&B Inl crvicw of Dwaync Davidson, May 6, 20 14. Thc subscqllclll reduction to three employees was due to cost cutting. J&B Interview of Dwayne Davidson, May 6, 2014. 114S J&B Interview of Bill Merrill, March 18, 2014. 1146 J&B Interview of Mary Barra , Apri l 14, 201 4. 1147 The Corporate Executive Board Company, 2013 RiskClarity Report - Findings and Next Steps (March 20, 20 14) [DOC ID 000145003510.1. In a more recelll workforce survey administered by GM ' s Human Resources fun ction to all salaried employees, conducted only weeks ago, there was signi fi cant improvement in the responses conceming willingness to "speak up" and trust in management over a similar survey conducted in 2012. GM "Workplace of Choice" Survey Results, May 20, 2014. 252 Privileged and Confidenfia/ Protec1ed by AUorn ey~ Client Privilege and as Allorney Work Product anything above th e company,,1 148 and " never put the company at risk.,,1149 The form er Cobalt Brand Quality Manager said that he felt that GM "pushed hack" on describing something as a safety issue during a relevant time period. llso Wheth er general "cultural" issues are to blame is diffi cult to ascertain, but th e story of the Cobalt is one in whi ch GM personnel fail ed to raise signifi cant issues to key d ec i s i o n~mak e rs . Senior attorneys did not elevate the issue within the Legal chain of command to the General Counsel - even after receiving the _ evaluation in the summer of 20 13 th at warned of the ri sk of puniti ve damages because ofa "compelling[]" argument that G M had "essentially. done nothing to correct the problem for the last nine years." Il5l Engineers, too, fai led to elevate the issue. Starting in mid ~ 2 0l 2 , there were three high-level managers brought in as "champions" - Woychowski , Fede ri co, and Kent. The very reason they were brought in was to help reso lve an unex.plained pattern of airbag non-deployments in an expeditious manner. But they d id not el evate the issue to their superiors, and the common thread was to hold more meetings and refer the matter to additional groups or committees. Si milarly on the issue of culture, GM employees received formal training as to how to write about safety iss ues. Im A PowerPoim presentation from 2008 wa rned employees to write J&B Interview of Dan Davis, March 14, 2014. J&B Interview of Christina Connor, March 19, 201 4. 115(1 J&B Interview ofStcvc Oakley, May 7, 20 14. 11)1 _ Selllcment Rcview Committee Roundtable Case Summary (J uly 22, 2013), at 36 rOOC ro 0000028 18612]; J&B In terview of Ronald POrler, Apri l 30, 2014; J& B Interview of William Kemp, April 17, 2014; 1&B Illlcrview of Lawrence Buonomo, April 16, 2014; see also J& B Interview of Michael Milli kin, Apri l 4, 201 4. 1152 J&B Interview of Mickey Sabol, April 24, 201 4. 1148 ll49 253 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product "smart," and not to use "j udgmental adj ectives and speculation."IISJ Employees were given a number of words to avoid, with suggested replacements· • " Problem = Issue, Condi tion, Matter" • "Safety = Has Potential Safety Implications" • " Defect = Does not perform to design,,1154 Employees were also given examples of sente nces not to use, including "Dangerous. almost caused accident" and "This is a safety and security issue .... ,,1155 And they were told, in what the author descri bed as an attempt at humor, not to use phrases such as " Kevorki anesque," "tombli ke," or " mani acal," or "rolling sarcophagus.,,1156 The "actual examples" provided in the presentation described how a plaintiffs lawyer had used a memo from a senior manager at another automaker warning that a risk of conducting a survey about a problem was that it could provide "product liabili ty credence to a hypothes is we have long ignored.,,1157 In addition to being trained on how to write, a number of GM employees reported that they d id not take notes at all at critical safety meetings because they bel ieved GM lawyers did not want such notes taken. No witness was able to identify a lawyer who gave such an instruction, no lawyer reported having given such an in struction, and we have found no documents or e~ m a il s reflecting such an in struction. Th e no-notes direction, however, reached 2008 QJ Interior TeclmiC,l1 Learni ng Symposium Soft Trim Breakout Session (2008), at 34 IGMNHTSA000258573l. 11 54 2008 QI Interior Technical Learni ng Symposium Soft Trim Breakoul Session (2008), 3142 IGMNHTSA000258573 J. 1155 2008 QI Interior Teclmica l Learning Symposium Soft Trim BreakOlll Session (2008), at 39 IGMNHTSA000258573]. 1l.S6 2008 Q 1 lnterior Technica l Learni ng Symposium Soft Trim Breakout Session (2008), at42 fGMN HTSA0002585731. The employee who presented the training was later told by a lawyer who saw a version of this training to remove the slide listing words never to be used. J&B Intervicw of Mickcy Sabol, April 24, 2014. ll57 2008 QI interior Teclmica l Learning Symposium Soft Trim BreakOlll Session (2008), at 37 IGMNHTSA000258573 J. 1153 254 Privileged and Confidenfia/ Protec1ed by AUorn ey~ Client Privilege and as Allorney Work Product the status of an urban myth that was foll owed, an instruction passed fro m GM employee to GM employee over the years_ Thus, as we learned in our in vestigation, for many meetings - of GM' s many committees - there are no cl ear records of attenda nce or of what was disc ussed or decided. Leadership at GM has tried to counter thi s culture with cl ear messages that employees should rai se issues. "Winning With Integrity" (the code of conduct) instructs employees to raise probl ems (although it does not explicitly reference vehicle safety) and ensure they rece ive proper attenti on, and to conduct themselves with the hi ghest ethical standards. Recommendations as to how to do thi s better are found below. 2. The '"GM Salute" and the '"GM Nod" A cultural issue repeatedly described to us and borne out by the evidence is a proliferation of committees and a lack of accountability. The Cobalt Igniti on Switch iss ue passed through an astoni shing number of committees. We repeatedly heard from witnesses that th ey fl agged the issue, proposed a solution, and the so lution died in a committee or with some other ad hoc group ex pl oring th e issue. But determining the identity of any actual dec ision-maker was impenetrabl e. No single person owned any dec ision. Indeed, it was often difficult to determine who sat on the committees or what they considered, as there are rarely minutes of meetings. On e witness described the G M phenomenon of avoiding responsib ility as the "GM sa lute." a crossi ng of the arms and pointing outwa rd towards others, ind icating that th e responsib ility belongs to someone else, not me. 1158 It is th is same cab ining of responsibility, the sense that someone else is responsible, that permeated the Cobalt in vestigation for years. 1158 J&B Interview of Andrew Srenz, April 22. 2014; J&S Interview of Alberto Manzor. May 27, 2014. 255 Privileged and Confidenfia/ Protec1ed by AUorn ey~ Client Privilege and as Allorney Work Product Similarly, Mary Barra described a phenomenon known as the ·'G M nod.,,11S9 The GM nod, Barra described, is when everyone nods in agreeme nt to a propo ~ed plan of acti on, but then leaves the room with no intention to follow through, and the nod is an empty gesture . ll 60 It is an idiomatic recognition ofa culture of that does not move issues forward quickly, as the story of the Cobalt demonstrates. 3. Failure to Share or Gather Knowledge/Information Silos Repeatedly, over a decade, GM personnel fail ed to search fo r, share or gather knowledge, and th at failure had seri ous consequences. There are multiple components to these failures, in volving individual mistakes, organizational dysfunction, and systems inaccessibl e to some and impenetrable to many. In 2004 and 2005, when complaints of moving stall s came in, the engin eers who considered the issue did not know that the vehicl e was designed so that the ai rbags would not deploy when the igniti on switch was in Accessory. As a consequence, the engineers fail ed to recognize the stalls as a safety issue and resolve the problem qu ic kl y. Even the committees (VAPIR and CPI T) that were designed to have c ross ~d isc iplin ary members did not connect the dots. Other ex amples of failure to share informati on and gather information include: 1159 1160 • In 2006, DeG iorgio did not change the part number, a system that by its very design is intended to share informati on with engineers - namely the critical piece of informati on that a part 's fit, form or fun ction had changed. DeGiorgio compounded the matter by later telling hi s colleagues that there was no change. As a consequence, for years engineers investigating the matter thought that the Ignition Sw itch was not to blame for airbag non-depl oyments. • In 2007, when Sprague was directed to start tracki ng airbag non-deployments , no one shared with him (and he did not find ) all of GM's prior work on the Cobalt moving stall J&B Interview of Mary Barra, Apri l 14, 201 4. J&B Interview of Mary Barra, Apri l 14, 201 4. 25 6 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product issue - not the PRTSs and not the TSS. As a consequence, he did not have an opportunity to connect the non-dep loyments with the problem of low torque in the Ignition Switch • From 2007 on, GM employees failed to find publicly avai labl e materials. This incl uded the Indiana Uni versity study, which a pl aintiffs expert reportedly found on the NHTSA website. 11 6 1 It also included GM's own TS8s, w hi ch were available both internally and publicly. • In 2011, Stouffer failed to check the TREAD database correctly. and thus did not obtain all of the info rmation available on it. As a consequence, he missed the Ion and HHR fataliti es, which were not discovered until February 2014 - after the fi rst recall. • From 2011 on, the fact that the Cobalt acc idents led to fatalities was not shared with all relevant decision-makers. These indi vidual s say that, as a consequence, they lacked a sense of urgency. In contrast, some of the mechanisms to share information did work, but still did not yield res ults. One reason for thi s was the use of "root cause" as a trigger for action. 4. " Root Cause" There are two dates at which criti ca l infonnation resided in a single place, and yet action was not taken. Witnesses state that the reason for this lack of action at either time was that the "root cause" was not known with certainty or that a full sol ution had not been devised. • In 2011, shortly after Stouffer was assigned the investigation, he knew: (1) that the MY 2005 and 2006 Cobalts had a hi story of the Ignition Switch being turned accidentall y to Accessory because of low igniti on torque (he had the TS8s); and (2) that airbags do not dep loy when the Ignition Switch is in Accessory. In other words, regardless of whether other later model year Cobalts perform ed better, any time a MY 2005 or 2006 Cobalt Ignition Switch was inadvertentl y turned to Accessory, the airbags were turned off Instead of addressing this repeating problem, the investigators worked to find an ultimate solution that would solve the problem for all years and for every permutation of the Ignition Switch. The consequence was a two~and~a-halfyear delay. 1161 Erin M. Shipp, P.E., Engineer 's Report of the _ Crash, Robson Forensic (June 12, 2012), at 5 rDOC ro 0000016621221. We have been unable to detenni~hciher the Indiana University study was available on the NHTSA website in 2007. 257 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product • In late April 2013, after the DeGiorgio deposition , engineers and GM Legal Staff knew: ( I) that the Coba lt had a hi story of the Igniti on Switch being turned acc identally to Accessory because oflow torque (they had the TSB, the Indiana University study, and two reports from plaintiffs' experts); (2) that airbags do not deploy when the Ignition Switch is in Accessory; and (3) that the Ignition Switch had changed for MY 2008 Coba lts (they had seen photographs during DeGiorgio ' s deposition). In other words, even the difference between early Cobalts and MY 2008 Coba lts was now explained. The reca ll did not occur until February 20 14, approximately nine months later. In the interim, GM employees confirmed what pl aintiff's expert had to ld them, researched additional iss ues, and worked on a field sol uti on. At both of the times described above, there was a view that no action should be taken until the "root cause" of the problem was full y understood and a so lution developed. To be sure, GM needed to investigate to understand the problem in the Coba lt. But the search for root cause became a basis for doing nothing to reso lve the problem for years. The lengthy search for root cause thus diverted GM from its obligations and fa iled to produce the required urgency to bring the matter to fast closure. Indeed, it still took nine months after DeG iorgio' s deposition for GM to order a recall and. even then, not all affected model years were included. VIII. RECOMMENDATIONS As described throughout thi s report, there were multiple, interrelated factors that led to GM's d ecad e~ l o n g failure to recognize the safety defect in the Cobalt and initiate a recall. It is critical that GM minimize the risk of this happening again. Because of the multiplicity of causa l factors, there is no simple soluti on, but the lessons learned from this failure can lead to a reexamination of the Company' s poli cies, procedures, and culture, and provide the Company with an opportunity not onl y to improve but to innovate and become an industry leader in the processes used to ensure consumer safety. We made preliminary recommendations to GM as the investigation was proceeding. With the benefit of further investigation and ana lysis, thi s section elaborates 0 11 and revises those 258 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product recommendations. It describes a variety of measures that, worki ng together, can improve the speed and manner in whi ch GM addresses safety issues and help bring GM to the forefront of the industry in its approach to safe ty. A. Organizational Structure A successful overhaul of processes begins with examination of the organizational structure, which influences the degree of coordination between groups with interrelated responsibilities and the extent of emphasis the Company places on key prioriti es, including safety. Recommendations as to organi zational structure include: I. Ensure that the responsibilities of the Vice Pres ident of Global Vehic le Safety I 162 are appropriately de fin ed to comprehensively cover safety and compliance issues, and ensure that suffi cient resources are made available to a llow this executive to comprehensively address safety and compliance issues. Ensure that this executive has direct access to the CEO and the Board and is required to make reports to the Board at least quarterl y. 2. Ensure that all departments , divisions, or groups that have substantial responsibiliti es concerning the identification, investigation, or remediati on of safety issues have a direct or indirect reporting line that leads up to th e Vice President of Global Vehicl e Safety. These groups include, for example, FPA, Quality, the CPITs and the VAPIR s. Ensure that Global Vehicle Safety also coordinates with the Legal Staff and is aware of safety issues arising through matters handl ed by the Legal Staff. 3. Review the activities of all orga nizational departments, divisions or groups that have safety-related responsibilities (such as FPA, PI , and Quality) for the purpose of identifying any areas where multipl e groups have similar or overlapping functions such as investigating or reso lving safety issues. Where overl ap is identified, consolidate or coordinate those functions to ensure, for example, that Quali ty and PI do not have independent a nd parallel responsibility for identi fying and resolving a safety defect but rather that each group is aware of, and not duplicating, the other's activities. To the extent that areas of overlap remain, ensure that a clear owner is identified. 1162 The Vice President of Global Vehicle Safety is a new position created ill the wake of the Ignition Switch recall. Jeff Boyer has been appointed to this position. 259 Privileged and Confidenfia/ Protec1ed by AUorn ey~ Client Privilege and as Allorney Work Product B. Cultural Emphasis on Safety An organizational emphasis on, and commitment to, sa fety is a criti cal comp onent to mitigating the risk ofa situation like thi s recurring. To e nsure that a commitment to consumer safety is a prominent part of the Company' s culture and is embedded in the fabri c of the organization, we recommend the foll owing: 1163 I. Implement regular communications with empl oyees about safety to rai se awareness and rein force the tone at the top, for exampl e by iss uing periodi c bulletins from the Vice President of Global Vehicle Safety and/or the CEO that include updates or reminders on safety issues, or including a column on safety in an employee newsletter. Ensure that employees understand that they have an obl igati on to ra ise any concerns they have about safety or compl iance, and to continue to rai se those conce rns if they do not believe those concerns have been resol ved. 2. We understand that GM has created a new "Speak Up For Safety" program to encourage employees to rai se safety issues, and we recomm end that GM promote that program through visibl e communi ca tions, such as posters on employee bulletin boards. Bulletins or newsletters could include fea tures recognizing employees who have raised safety issues and highl ight the significance of the potenti al safety probl em averted by the esca lati on of th e issue. 1163 3. Consistent with , 19 of the May 16, 2014 NHTSA Consent Order, visi bl y promote and rigorously enforce the n on~re ta l i ation po licy, including for employees who report concerns regard ing actual or potenti al safety ~re l a t e d defec ts or potenti al n on~ co mplian c e with the Federal Motor Vehicle Safety Standards. 4. Regularly communicate to suppli ers the importance of safety and GM 's ex pectati on that suppli ers will promptly and accurately identi fy any potenti al safety issues. Provide suppli ers with access to the Speak Up for Safety hotline or a simil ar mechanism for rai sing such safety issues. 5. Consistent with , 20 of the May 16, 2014 NHTSA Consent Order, explicitly communi cate to empl oyees that they should not be reluctant to classify issues as safety issues or potenti al safety issues, includ ing in written work, and eliminate any language in any guidelines, training decks, or poli cies that suggest otherwise or that cauti on against using spec ific words or phrases that mi ght be deemed too We understand that this is in progress with feature stories on GM's intemal website. 260 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product sensitive or inflammatory. Communi cations to employees on safety policy issues like thi s should come from the level of the Vice Pres ident of Global Vehicle Safety or hi gher. 6. Deve lop protocols for escalating potential safety issues to appropriate levels of management. Such protocols should include identification of the types of issues to be escalated (based for example on their severity, frequency, or the amount of time they have been open) and the level to which they should be escalated. The protocols should make cl ear that employees who are aware of unresolved safety issues should escalate them, and should defi ne circumstances in whi ch the concerns should be raised to the level of the Vice President of Global Vehicle Safety and, ultimately, the CEO. The protoco ls should also define other circumstances in which the CEO should be made aware of safety issues, including instances in which safety issues ri se to a certain level of severity or in whi ch an investigation has not been resolved within a defined period of time. 7. As agreed to in ' 19 of the May 16, 20 14 NHTSA Consent Order, continue to review and strengthen the process for expeditious reporting by employees of potential or actual safety issues and non~co mpli ance. This includes the development and roll~out of training programs instructing employees on how to identify safety issues and how to raise them with supervisors and executives (see more discussion on training below). 8. Ensure that supplier quality issues that may raise safety concerns for vehicl es in production are reported as safety issues and are not addressed solel y as suppli er quality issues. Impl ement specifi c procedures for personnel dealing with quali ty issues to track and report to the safety organization those quality issues that raise safety concerns, and train quality personnel on those procedures. C. Individual Accountability All empl oyees have responsib ility for raising safety issues, and many employees have responsibility for addressing safety issues once they have been identified. The Company should take steps to ensure that employees are aware of their safety-related responsibilities and that individuals are accountable for addressing the safety issues for which they are given responsibility. Specific recommendations are as follows: I. Identi fy clear owners of safety and compliance issues to increase accountability. Clarify internal roles by defining responsibilities and accountability for those involved in processes relating to the eva luation and resolution of safety issues, 26 1 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product incl uding responsibili ties for feed ing items into the FPE process, such as from VA PIR and CPI T meetings. Cla ri fy the role of co mmittees such as C PITs and VA PIRs in identi fy ing safety issues and referring them to the FPE process Des ignate one person on each such committee as a safety li aison with specific responsibility for elevating safety issues. 2. Include appropriate identification, elevation, and resolution of safety and compli ance iss ues as a factor in employee performance evaluations. 3. Require employees to certify annually th at they have reported any safety issues of whi ch they are aware and to identify, as part of that certi fication, any safety issues they are aware of that have not been reso lved. D. Communications Between and Within Groups Breakdowns in communication between and withi n groups were a critica l part of the fa ilures described in thi s report. Consistent with GM's comm itment in ~ 21 of the May 16, 20 14 NHTSA Consent Order to improve processes for the purpose of encouraging and improving infonnation~s h ari n g across functio nal areas and disciplines, many of tile recommendations throughout this section are intended to address inadeq uate communication. Specific additional recommendations are: I. Provide regul ar written or oral updates by the Legal Staff for relevant engineering groups on all eged or potential defects. 2. Formalize coordination between and among PI, Legal Staff, FPA, FPE, the new "emerging issues" group, and rel evant engi neeri ng teams with accountabi lity for sa fety and compliance issues, for example by imp lementi ng a cross~ fun ct io n a l committee to track and review potential safety issues and set clear action items. This coordination should include the sharing across groups of information concerning signi fi cant FPA matte rs, FPE investigations, serious or hig h ~vo lume warranty claims, tre nds or pattern s identified, and NHTSA interactions. 3. Conduct an assessment of the adequacy of mechanisms for ensuring coordination between groups handling diffe rent subsystems of the same vehicle regarding safety~ re l ated information and items that may affect safety performance, and implement improvements as necessary. For example, e nsure those working on th e ignition switch understand how it interacts with airbag deployment. Simil arly, ensure that committees such as the VA PIR and CPIT are structured with suffic ient 262 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product cross-functional expertise to ensure that safety issues are properly identified and addressed. 4. Assign th e Global Ethi cs and Compliance Ce nter ("GECC") to oversee the review of isslies rai sed through the new Speak Up For Safety program. This will have the benefit of ensuring that review of issues rai sed through the Speak Up For Safety program is coordinated with review of issues ra ised to the compliance organization, and to ensure that interna l inquiries and investigations prompted by th e Speak Up For Safety program are appropriately tracked and handled consistent with GECC procedures. E. Commun ications with NHTSA NHTSA should be viewed not only as a regul ator but also an ally in the effort to ensure that the Company's vehicles are as safe as th ey can be. Interactions with NHTSA should be consistent with that type of relationship. Specifi c recommendations for ma intaining an appropriate relationship are: I. As required by ~ 21 of the May 16, 20 14 NHTSA Consent Order, improve communi cation with NHTSA regarding safety~ re l ated defects. Consent Order ~~ 25~28 require the Company to have monthly and quarterly meetings with NHTSA and to be prepared to address va rious issues at those meetings. As part of thi s effort, the Company should ensure that both the quality and the frequency ofN HTSA communications are sufficient to keep NHTSA fully informed of sa fety ~re l ated iss ues. The individual responsible for implementing this item should report directl y or indirectly to the Vice President of Global Vehicl e Safety. 2. Create a centralized database for all communicat ions with NHTSA and train indivi duals who communicate with NHTSA to file thei r communications in this database. 3. Ensure timely notification to NHTSA of safety~related defects. Evaluate and clearly define, consistent with federal law, the trigger for the NHTSA five~day reporting req uirement. 263 Privileged and Confidenfia/ Protec1ed by AUorn ey~ Client Privilege and as Allorney Work Product F. Role of Lawyers The Legal Staff can and should play a criti ca l and unique role in assisting with the identification, anal ys is, and resolution of safety issues that have given rise to customer claims. To ensure that Legal Staff play this criti cal role effectivel y, we recommend the following steps: I. Hold regul ar di scussions between each product liti gation attorney and the Practi ce Area Manager and Gl obal Process Leader for Liti gation regarding whether the attorney has observed trends or potential safety issues in lawsuits or n ot~in~ s uit matters ("NISMs"). 2. Des ignate a member of the Legal Staff as a liaison to provide regular reports (and reports where matters are partic ularly significant or urgent) to the Global Vehicle Safe ty organization on safety ~re l a t ed issues identi fied in matters handl ed by Legal Staff. Hold regular meetings between the Global Vehicle Safety group and the Legal Staffliaison. During these meetings , the Global Vehicle Safety organization will present to the Legal Staff signifi cant or unreso lved sa fety issues, and the Lega l Staff liaison will report on any issues identifi ed in matters handled by Legal Staff that the liai son has not already communi cated. real~lime 3. Hold monthly meetings between the GM Lega l Staff and FPA engineers responsibl e for each specialty (e.g. , n o n~de pl oym e nt li tigation, rollover/roof crush, power steering, etc.) to di scuss observed trends a nd potenti al safety issues in that specialty area. Include in these meetings the ES IS claim administrator with the most experience in that speci alty area. If any trends or sa fety issues are identifi ed, ensure that the Global Process Leader for Litigation is informed, and that the trends or safety issues are di scussed at the next scheduled Ro undtable meeting and elevated as appropri ate. 4. Ensure that at the onset ofliti gation, the Lega l Staff and FPA generate a li st of all PRTSs, FPRs, TSBs, recalls, product investigations, and issues in the FPE process for the subj ect vehicl e make and model. Include thi s infonnation in the new suit package prov ided to outside counsel at the onset of litigation. Ensure that the earl y techni cal assessment (" ETA") refl ects considerati on of whether there are relevant PR TSs, FPRs, TSBs, reca ll s, product investiga tions, and issues in the FP E process that affect the integrity of the make and model invol ved. 5. Include in the Roundtable case summary fonn a category for any safety issues, trends, lawsuits and NISMs simil ar to the case discussed at the Roundtable. 264 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product 6. Institute a process to elevate unreso lved techn ical issues expeditiously. Ifan FPA engineer is unable to determine any reason for a component failure following a vehi cle inspection and technical analys is, require the FPA engin eer to initiate a conference call with in~h o u se and outside coun sel within a defined period to di scuss the iss ue and appropriate follo w~up . Depending on the circumstances, appropriate follow~up mi ght include referring the issue to the PI or FPE processes. 7. Provide specific guidance concerning the types of issues that should be elevated to the General Counsel. Those issues should include serious safety issues and safety issues that are not being resolved expeditiously, and more broadly, any concerns that could have a signifi cant impact on the Company, its customers, or its shareholders. Explain in this guidance that lawyers should request the General Coun sel's assistance when important processes are not progressing with appropriate speed. Include such guidance in the orientation program for all new hires to the Legal Staff. 8. Consistent with "Winning with Integrity" and other prior internal guidance, ensure that in~house counsel are aware of the expectation that they wi ll respond appropriately if they become aware of any threatened, on~go in g, or past violation ofa federal , state or local law or regulation, a breach of fiduciary duty, or violation ofGM policy, including the expectation that if they raise such an issue and believe it has not been addressed appropriately, they will bring the situation to the attention of their supervisors, and if they believe their supervisors have not addressed it appropriately, to higher levels including the General Counsel if necessary. Ensure that thi s expectation extends to issues of safety. Communi cate similar expectations to outside counsel in written guidelines. 9. Consistent with the "Winning with Integrity" guidelines for employee conduct, specifi ca lly memorialize in writing for in~house and outside counsel the expectation that they will report to appropriate Legal Staff any observed violation of law or GM policy, including possible violations of policies pertaining to th e recall decision-making process. Circulate thi s policy to Legal Staff on a regul ar basis, and circu late the policy to any outs ide attorneys to whom the poli cy applies. Include this policy in ori entation programs provided to in~house lawyers when they join the Company. 9. Provide guidance for product liability attorneys on how to recognize and communi cate safety issues to ensure that they are properly addressed notwithstanding ongoing liti gation or claims acti vity, while fulfilling their obligations to defend the Company in litigation and appropriately protect attorney-client privilege. 265 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product G. Interactions with Suppliers The Company ' s interactions with suppliers should reflect the s ignificance that supplier quality control has on vehicle safety. Specific recommendations for enhancing suppli er relationships are as fo ll ows: I. Ensure that procedures are in place so that supplier quality issues that may impact safety in production vehicles are conununicated to appropriate personnel with vehicle safety responsibility. 2. Identi fy all responsible units and personnel who are accountable for monitoring and resolving concerns with suppli er quality, and ensure that they have a direct or indirect reporting line up to the Vice President for Global Vehicle Safety. 3. Cond uct an assessment of the adequacy of measures taken to ensure identification of any supplier parts that do not meet GM specifications, both by the time of launch and periodically during production, including independent testing in appropriate circumstances. 4. Ensure that appropriate procedures are in place for addressing parts, including parts provided by suppliers, that do not sati sfy GM specificat ions. These procedures should include a specific protocol for reviewing authorizations of out~ of~specification parts, tracking out~of~specification parts, identifyi ng who should be notified of them , and identifying and elevating any potential safety issues that might be associated with the use of out~of~specification parts. Require high~l eve l review for approval to use the part. 5. Ensure that the Company receives and maintains, as part of its own files, all supplier design change and work order documents relating to supplier parts. 6. Work with suppliers to increase GM's access to stored and recorded data on SDMs, CDRs, and EDRs used in GM products. H. Data Storage, Retrieval, and Ana lysis Several of the problems discussed in thi s report could have been avoided or resolved much more quickly if the individuals involved had timely searc hed for or received accurate and complete data that in many cases was already in GM ' s possession. Accordingly, the Company has committed, in '1/22 of the May 16, 2014 NHTSA Consent Order, to revise its product quality 266 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product ana lytics to improve its ability to identify safety consequences and the severity of those consequences, as we ll as to assess the number or rate of all egations, complaints, inc idents, reports and/or warranty claims relating to potential safety~related defects. Specific recommendations to fulfill these commitments and to improve the Company's ability to store, retrieve, and analyze data in a way that promotes identification of safety issues are as follows: I. Ensure that the Global Vehicl e Safety organization monitors all safety~ re lat e d databases (including TREAD, CYE P, 1164 GART, PRTS, ES IS, CTF, etc.) as well as publi c sources, such as NHTSA and other crash investigation reports, for trends or patterns of iss ues that may raise safety concerns. This entai ls two sub~ i ss u es. First, monitor the databases active ly so that there is a mechanism for identifying patterns of defects or incidents that rai se potential safety concerns, as opposed to searching the databases only when a perce ived need arises. Second, establish mechani sms for coordinating the mon itori ng of these databases to ensure that patterns of simil ar defects and incidents across different databases will be vis ible and receive attention. 2. Establi sh a standardized process for the regular monitoring of safety recalls by other manufacturers, and for the mapping of these recalls to existing internal and external investigations by PI. 3. Ensure that staffing levels in the TREAD group and elsewhere are sufficient to allow proactive monitoring of data for safety issues. Similarly, ensure that sta ffin g levels of personnel responsib le for other company databases, for infonnation technology, and elsewhere are sufficient to ensure accuracy and completeness of the information in Company databases. 4. Improve the search functions in th e databases to make it easier to locate and review data. Determine whether improvements can also be made to increase the accuracy of searches, by minimizing irrelevant data caphlred by searc hes whil e maximizing the ability of searches to retrieve data that is relevant even though it may have been captured from different sources usi ng different language and different fields. In addition, address diffi culties in searching across several databases, for example by conso lidating the databases or improving the search mechanism, and ens ure that re levant personnel are trained on search methods. llM The Company Vehicle Evaluation Program ("CVEP") database collects infonnation provided by employees who drive company-owned vehicles. 267 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product 5. Work with data analytics suppliers to identify and implement availab le automated tools, or commiss ion new ones if appropriate, that will enhance the Company's abi li ty to monitor, retrieve and analyze safety~ related data 6. Provide product investigators and related personnel direct access to relevant databases where appropriate, and provide indirect access to any databases that require special expertise, such as TREAD. 7. Ensure that relevant analytics, procedures, and policies focus on severity and identification of potential safety-related consequences of a defect, even where the defect is infrequent. 8. Conduct an assessment of the adequacy of processes for dealers to identify and report sa fety~re l ated iss ues that do not otherwise get reported through the warranty data coll ection process. Ensure that GM's expectations for dealers to report safety issues are clearly defined. 9. Accept NHTSA 's offer to consult with GM about industry best practices for identifying safety ~re l ated defects through the sc reening of data. 10. Require that ESIS monitor its data concerning alleged safety defects for patterns and trends and bring such patterns and trends to the Co mpany 's attention. I. Engineering Processes and Databases Functions such as the PRTS and EWO processes are critical to execution ofGM's engineering functions, and also to tracing prior engineering decisions when product investigations are conducted. Recommendations to improve these processes are as follows: I. Ensure adequacy of current processes for transparency, tracking, and, where appropriate, elevation of iss ues being dealt with in the PRTS system. Ensure that th e processes focus on quality and safety. 2. Ensure adequacy of process for determining the correct severity classification, including review by supervi sory personne l of initial severity determinations. In addition , when a PRTS has been open for a set amount of time (such as 60 days), implement a review of its severity level based on any newly avai lable information or analysis. Establish a process through which PRTSs categorized as level I (safety issues) are automatically elevated fo r additional review, in coordination with a member of the Global Vehicle Safety organization. 268 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product 3. Ensure that the standard for cl os ing PRTSs without act ion is clearly defined and sufficientl y ri gorous. Ensure that a PRTS cannot be closed without action absent clear s i gn~offb y named indi viduals and appropriate levels of review. Re~affirm that the lack of an «acceptable business case" is not an acceptable reason for closing out a PRTS if the PRTS invo lves a safety issue. 4. Require s i g n~ o ffb y a validation engineer on all EWOs, with the va lidati on engineer to consider any impact on safety and compliance as to both the design change itself and other affected subsystems in the vehicle. 5. In any processes and databases (such as PRTS and GART) where severity levels are assigned, review the criteria used to determine whether problems are identifi ed as safety issues to ensure that those criteria adequate ly encompass all defects that have the potential to adversely affect consumer safety. In addition, GM should ensure that indi vidual decisions about severi ty levels are reviewed, so th at a single engineer does not make the decision. J. Product Investigation Process The product investigation of the Coba lt Ignition Switch began nearl y a decade before any recall was initiated. Multi ple failures contributed to that extraordinary delay. We recommend the following spec ific steps to improve the Product Investigations process: 1. Standardize the process for collecting and reviewing infonnation that gives rise to product investigations. As part of this standardi zation, identify the categories of source infonnation that should be reviewed during any investigation. Train investi gators on the revised process. 2. Establish clear criteria for the process for assigni ng Pis and for documenting those assignments (or the reasons for not making an assignment) to ensure that Pis are assigned in a manner that ensures that the req uisite expertise and resources are applied to each matter. 3. Formalize other procedures governing th e PI process, including how issues are escalated within the process and who is responsible and accountable for monitoring the progress and bringing the investigation to a conclusion. Ensure expedited review and escalation of investigations involving injuri es and fatalities, and any other investigations where at any point a safety or compliance issue is implicated. 269 Privileged and Confidenfia/ Protec1ed by AUorn ey~ Client Privilege and as Allorney Work Product 4. Ensure that there are adequate mechani sms to enable product investi gators to identify and review any prior Company investigations regarding the same condition, in the same or other models. 5. Establish clear deadlines for each element of the PI process such that any mi ssed deadline results in increased attenti on. For exampl e, for any in vesti gation that is not resolved within a defined time period , such as 60 days, ensure that the progress of the investi gation is monitored and documented, and in appropriate circumstances, escalated. 6. Standardize the processes for creating documentation reflecting the progress and results of investi gations. Documentation might be on a tracking form and/or on a separate progress reporting form , and it should include such information as maj or investi gative activities conducted since the last report, current conclus ions or hypotheses, open issues, and next steps. 7. Ensure that the PI group is staffed suffi ciently to address these recommendati ons. K. FPE Process GM has agreed in ' 21 of the May 16, 2014 NHTSA Consent Order to increase the speed with which recall dec isions are made, including by c1 ari fying the reca ll decision-making process to decrease the number of steps pri or to making the final decision of whether to conduct a recall. We recommend th at GM review its processes and implement changes consistent with that agreement. Specifi c recommendations are as foll ows: I. Creat e a specifi c form for initi ating the FP E process that clearly identifies the issue(s) to be investigated, lists the parts that could be implicated, identi fies the suppliers of those parts, lists the potenti ally impacted vehicle models and years, identifi es the source of the information that triggered the investigation, identifi es the investigator who will have responsibility fo r the investigation , and identifi es the indi vidual who will be responsib le for monitoring the matter and bringing it to conclusion. 2. The Company has committed, in '1r 24 of the May 16, 20 14 NHTSA Consent Order, that it shall not delay holding any meeting to decide wheth er or not to recommend or conduct a safety reca ll because the Company has not yet identified the prec ise cause of a defect, detennined a remedy for the defect, or prepared a plan for remedying the defect. In order to comply with that commitment, develop 270 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product and adopt policies a nd procedures that implement this commitment and train relevant employees on those policies and procedures. 11 6S 3. The Company has committed, in '1)24 of the May 16, 20 14 NHT SA Consent Order, that it shall ensure that the deci sio n-makers for recall s are informed of safety-related concerns in a reasonably exped iti ous manner, including by ensuring that GM ' s corporate structure enables its safety organization to promptly bring sa fety-related iss ues to the attention of committees and individual s w ith authority to make sa fety recall dec isions. Accordingly, in conjunction with its review of the FP E process, the Company should assess whether the current organizational structure - incl uding the appointment of the Vice Presi dent for Global Vehicle Safety and the restructuring of the groups that report to him - fulfill s thi s commitment. 4. In '1)21 of the May 16, 20 14 NHTSA Consent Order, GM agreed to increase the speed with whi ch recall deci sions are made, including by clarifying the recall decision-making process to decrease the number of steps pri or to making the final decision of whether to conduct a reca ll. l1 65 As GM revi ses thi s process, formali ze other procedures governing the process, including how the process is initiated, who may initiate the process, how issues are escalated w ithin the process , who is responsible and accountable for monitoring the progress and bringing the matter to a conclusion, and the chain of command throughout the process. Mai ntain a written policy governing thi s process that clearly defi nes the procedural steps in the process and the role of each participant. Ensure that part of the FPE process includes identification and review of any prior FPE processes regarding the same condi tion in the sam e or other model s. 5. Establi sh clear deadlines for each element of the FPE process such th at any mi ssed deadline res ults in increased attention. For example, for any issue that is not resolved within a short, defined time period, such as 60 days, ens ure that the progress of the issue is monitored and, in appropriate ci rcumstances, escalated. 6. Standardi ze and improve documentation of dec ision points during the FPE process, including documentation of deci sions made and attendees present at those decision points. In particular, require that FPE participants document poten tial defects and remed ies that are rejected during or as the result of the FPE process , such as root causes that are considered and rejected and model years that are examined but not included in a fin al recall . 7. For any investigation that is not resol ved within a defin ed time period, such as 60 days , ensure that the progress of the in vestigation is documented. Documentation might be on a tracking form and/or on a separate progress reporting form , and it We understand that this is in progress. 27 1 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product should include such information as major investigati ve activities conducted since the last report, current conclusions or hypotheses, open issues, and next steps. Document the cone lusion of an in vestigation on a form that records the conclusi ons reached, the action recommended, and the reasons supporting the conclusions and recommended action. 8. Ensure that systems are in place for maintaining and organizing documentation of agendas, presentations, summaries, decision memoranda, and other documents reflecting the steps in the FPE process, and implement improvements as necessary to ensure that those material s are organized, maintained, and accessible. 9. Establish guidelines governing the ability of an investigator to remove an issue from the FPE process because PI determines it is unrelated to safety. L. Policies and Training In many instances in this Report, personnel were not aware of a relevant policy to govern their decisions or did not sufficientl y understand an existing policy, and many employees reported receiving little to no training on policies and substantive issues relevant to their roles. As GM fulfills its training~related obligations in the May 16, 2014 NHTSA Consent Order, we recommend that it undertake the following specific step s to ensure that appropriate policies are in place and that suffi cient training is provided: 1. Review policies and procedures that bear on safety issues to ensure that they are eas ily understood and readily accessible. In particular, revise the GRC 3.1.1 policy on determining when a new part number must be assigned. Clarify the policy to require assignment of a new part number whenever a design change alters the expected performance ofa supplied part, and remove the " legitimate business case" lang uage. 2. Review GM's compliance policies and procedures to identify revisions to ensure adequate, appropriate, and consistent treatment of safety issues. For example, make clear that Winning With Integrity covers the identification and raising of safety issues , or iss ue a companion policy covering those issues. 3. Improve or create standards for elevating safety issues and ensure that training covers these standards. The standards should cover, for example, situations where GM engineers are unable to detemline why a product failed to perform appropriately. 272 Privileged and Confidenfia/ Protec1ed by AUorn ey~ Client Privilege and as Allorney Work Product 4. Ensure that personnel, including legal personnel, personnel responsible for IT systems and databases, and personnel who are assigned to monitor data streams or identify emerging issues and trends, have appropriate backgrounds and training. For ex ampl e, hi gh~l ev elle ga l personnel may benefit from background or training in product litigation issues and recall s. Similarly, emp loyees with responsibility for IT systems and databases would benefit from background or training in the relevant systems or databases, and employees who monitor and digest warranty data may benefit from engineering backgrounds. 5. Train employees to recognize and elevate sa fety issues. Emphasize that safe ty issues should be actively identifi ed and addressed regardless of whether vehicl es are in the design phase or the production phase. 6. Provide employees with appropri ate functi on-specific training to recognize safety issues and respond appropriatel y. This training should include, among other things, what constitutes a safety issue and the role of various organizational units and processes in handling safety issues. For engineers, training should include the safety and compliance ramifications of des ign changes. 7. Train PIs and FPA engineers, and related personnel , on sources of data that are publicl y available and may have information relevant to particul ar investi gations and poss ible safety iss ues. 8. Train Legal Staff who deal with safety issues on the NHTSA recall process. 9. Extend training on safety issues to supplier empl oyees embedded within GM where appropriate. 10. Train engineers on policies that expl ain the circumstances under which an engineer may dev iate fro m a Component Techni cal Speci fication when approving a part for production, and how such a deviation must be documented. II . Train employees on lessons learned from the Cobalt investigation. 12. Communicate to empl oyees that a low frequency of occurrence should not prevent an issue fro m being escalated for appropriate investi gation and review ifit otherwise presents a potenti al safety issue . 13. Educate employees that shielding seni or exec utives or other employees from infonnation in order to allow them to deny knowledge is not acceptable; on the contrary, employees should err on the side of elevating potenti al safety issues. 273 Privileged and Confidenfia/ Protec1ed by AUorn ey~ Client Privilege and as Allorney Work Product M. Compliance, Auditing, and Oversight To ensure proper implementation of remedi al measures discussed above, we recommend the following: I. After Speak Up For Safety has bee n in pl ace for a yea r, survey employees to assess their attitudes toward safety and their willingness to rai se safety concerns. In addition, within a reasonable time (pe rhaps one year) after the date of this Report, have a source outside of the safety organi zation (such as GM Audit Services) conduct an audit to assess the progress and effectiveness of the remedi al measures implemented. Also, implement regul ar aud its of the rec all decisionmaking process to ensure that it is functioning as intended and identi fy improvements. For exampl e, audits should address whether issues that are safety ~ related are being appropri ately designated as such. 2. Ensure that the Global Ethi cs and Compliance Center monitors relevant aspects of th e safety program, including investi gation of Speak Up For Safety complaints, the recall decision-making process, traini ng on safety and recall issues, and periodic evaluati on of safety procedures and metri cs, in conjunction with the regular audits described in item I above. 3. Require the Global Vehicle Safety organi zation to document the improvements it is making related to safety, and its impl ementation of these recommendations, to facilitate review of those improvements. 4. Ensure that the Board receives regul ar reports fro m the Vice President of Global Vehicl e Safety on at least a quarterly basis and, after one year, evaluates the Company's progress in implementing the comm itments made in the May 16, 2014 NHTS A Consent Order and any recommendations in thi s Report the Company decides to implement; and take such other steps as the Board may find useful to monitor management ' s compliance with global vehicle sa fety~re lated obli gations and Company vehicle safety-related policies. N. Recordkeeping Adequate recordkeeping will enhance th e ability of product investigators and others with safety responsibilities to review past engineering and investigati ve measures that may impact current investigations, and will also fac ilitate sub sequent reconstruction of investigati ve steps when necessary . In addi tion to the recommendations identi fied in connec tion with specific 274 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product processes, as discussed above, recommendations to improve the Company's record~keep in g relating to sa fety issues are the following· I. Document the di scussions during Settlement Roundtables and Settlement Review Committee meetings, including follow-up from those meetings, and not just the final decisions. 2. Maintain all significant engineering records documenting design decisions, including documentation of all design changes within a model year on a particular platform. 3. Document FPE decision points, as well as decisions made at CP ITs and VAPIRs, specifi cally incl uding decisions not to take suggested field actions or to downgrade the type of action taken. 4. Document, on the EWO foml, whether a proposed EWO involves a change to fit, form, or function. 5. Ensure that all results of safe ty and compliance testi ng are adequately documented. 6. Ensure that all electronic documents relating to safety issues that are maintained in GM's records (including records maintai ned by Global Vehicle Safety and the Lega l Staff) are kept in a searchable format to the extent technologically possible. For exampl e, perfonn optical character recognition ("OCR") on PDF documents before storing them. O. Conclusion GM has begun undertaking some of the measures discussed above. Under the direction of Jeff Boyer, who has ass umed the recently created role of Vice President for Global Vehicl e Safety, the Company has instituted a Speak Up For Safety program designed to encourage employees to identify, report, and elevate safety issues, and train them how to do so. Staffing has been increased subslanliall y lo allow more issues lo be invesligaled and expedite lhe investigation process. A new emerging issues group has been created to monitor and analyze data streams, and that group is working with data ana lytics suppliers to assess the suppliers' capabilities and develop recommendations for appropriate systems to employ. We recommend 275 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product that G M continue to implement these measures, implement the add iti onal steps identified above , and use the self-examination prompted by the Igniti on Switch failures , so that the Company is an industry leader in the way it handles consumer safety issues. 276 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product APPENDIX A: NHTSA Reporting Requirements This Appendix summarizes GM 's various reporting obligations to NHTSA under the Safety Act and TREA D Act during the relevant time period. Reporting a " Defect . .. Related to Motor Veh icle Safety" The Safety Act requires that a motor vehicle manufacturer notify NHTSA when it " learns" one of its vehicl es contains a " defect" that is " related to motor vehicle safety.,,1l66 NHTSA 's implementing regulation likewise provides: «[ eJach manufacturer shall furni sh a report to the NHTSA for each defect in hi s vehicles or in hi s items of original or rep lacement eq uipment that he or the Admin istrator detennines to be re lated to motor vehicl e safety."lt67 NHTSA has furt her spec ified that the Defect Informati on Report must be submitted " not more than 5 working days after a defect in a vehicle or item of equipment has been determined to be safety related."t I68 NHTSA has not promulgated a comprehensive definition of "defect," choos ing instead a "common sense" approach that looks to the facts of each particular case. 1169 The statute defin es a defect "related to motor vehicle safety" as a defect that creates an "unreasonable ri sk" of an 49 U.S.C § 301l8(c)(I). 49 CF. R. § 573.6(a). 11 68 49 CF.R. § 573.6(b). 11 69 A case entitled United States 1'. General Motors Corp., 518 F.2d 420 (D.C CiT. 1975) ("Wheels"), established the standard for determining whether a malrlmction in a motor vehicle constitutes a "defect" under § 30lI8(cXl): 11 66 1167 [A] vehicle or componelll 'colllai m; a defect' if it i~ ~Hbject 10 a ~ignificallt number of failnrcs in lIonnal operation, including fail ures either occurring during specified use or resulting from owner abuse (including inadequate maintenance) that is reasonably foreseeable (ordinary abuse), but excluding fail ures attributable to nonnal deterioration of a component as a result of age and wear. [l in all cases the manufacturer may prove, as an affinnative defense, thai the failures resulted from unforeseeable owner abuse (gross abuse) or unforeseeable neglect ofvchicle mai ntenance. Id. at 427. A "significant number" offai lures according to thc court is "non-de mil/imis," but also "normally will not be a substantial percentage of the total number of componcnts produccd." Jd. at 438 n.84. 277 Privileged and Confidenfia/ Protec1ed by AUorney~ Client Privilege and as Allorney Work Product accident, or of a risk of injury or death as the result of an acc ident. l17O As with the tenn "defect," NHTSA has not defi ned " motor vehicl e safety," believing "any attempt to defin e safety related defect would be ill-advised ... . The fact that [safety] determinations may encompass a wide variety of factual situations, and may consequentl y be difficult to make, does not mean that it is necessary, desirable, or even possible to rep lace the decision with a simple formula. ,,1171 Some defects such as total loss of contro l of a vehicle or an engi ne fire create a per se unreasonable risk. ll 72 For other defects, however, courts look to statistical and other evidence to determine whether the circumstances in which the defect occurs (or could occur) actually pose an unreasonable ri sk of accident or injury.l 173 Courts have rej ected the notion that any risk to vehi cle safety - no matter the cost required to eliminate the ri sk - is automaticall y an "unreasonable" risk.11 74 Finally, regarding what it means for a manufacturer to "learn" ofa safety-related defect to trigger the reporting and recall requirement, neither the statute nor N HTSA regulations provide an explicit standard. However, in interpreting the predecessor statutory provision courts 49 U. S.c. § 30102(a)(8). NHTSA, Resp. to Pe t. for Reconsideration of Defect Noti fi cation Rcgs., 38 Fed. Rcg. 9509, 9510 (April 17, 1973). 1172 See. e.g., Uniled Slales 1'. General Molors Corp ., 561 F.2d 92 3, 92 8 (D.C. CiT. 1977) (holding that an unreasonable safety risk occurred when part of the steering mechanism of the car fa iled causing sudden and complete loss of steering control); United Slates 1'. General MOlors Corp., 565 F.l d 754,757 (D.C. Cir. 1977) ("[WJe can see no question but that engine fires .. . are extremely dangerous for all involved and should be considered an unreasonable ri sk to safety."). 1171 See. e.g., Smith v. Ford Motor Co., 749 F. Supp. 2d 980, 99 1 (N. D. Cal. 2010) (finding that an ign ition failure lha! could prevenl a vehicle from being staned did nOl conslilule a lhrea! 10 mOlor vehicle safely because ''lpJlaintiffs offer no evidence that the ignition-lock defect causes engines to shut ofT unexpectedly or causes individuals to stop their vehicles wlder dangerous conditions."). 1174 See. e.g., Cenler for A ulD Safely v. Peck, 751 F.2d 1336, 134411 .5 (D.C. Cir. 1985) ("The principle that an ' unreasonable risk' provision requires eve n insignificant risks 10 be eliminated if that can be done at (presumably) insignificant cost would tum many areas of regulation into unending pursuit of the trivial."); United S tates v. General MOfOrs Corp., 656 F. Supp. 1555, 1579 (D. D.C. 1987) ("Pl f the only ' remedies' are ineffect ive, prohibitively expensive, or affimlatively detrimental to public safety, even a significant risk may nevertheless be ' reasonable' as a matter of law."). 1170 11 71 278 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product have made clear that constructive rather than actual knowledge is effectively required, and this knowledge may be inferred when a manufacturer rece ives suffic ient customer compl aints or . concernm . g the delect !-". • Il7S warranty ca I lms at Issue. EWRs and Quarterly Reviews The TREAD Act requires automobi le manufacturers to submit quarterly reports, known as Early Warning Reports ("EW Rs"), to NHTSA. 1I76 For car manufacturers, the data must include counts of property damage claims; warranty reports; consumer comp laints; and field reports broken down by make, model, and model year and by system category (e.g., steering, braking) .I177 Manufacturers must al so submit summaries of eac h claim against the company for death or injury allegedl y related to a defect. 117s NHTSA mines these reports for possible safety defects that may warrant a fonna l investigation. I 179 We do not understand that GM is all eged to have violated its obli gation to submit these EWRs, and such routine re porting is not the focus of thi s in vestigation. 1m See. e.g., Ulliled Slales v. General MOlors Corp., 656 F. Supp. 1555, 1559 n.5 (D.D.C. 1987) (finding that manufacturer cannot evade its reponi ng requi rement by intentionally failing 10 reach detennination that a defect is safety-related); Uniled Slales v. General MOlors Corp. , 574 F. Supp. 1047, 1050 (D. D.C. 1983) (manufacturer incurs reponing obligation when it "actually dctcnnined or should have determined" that a safety-related defect exists). See also Doll v. Ford Molar Co., 81 4 F. Supp. 2d 526, 537 -38 (D. Md. 20 11)(rcquiring duty to rcpon safety defect where Ford had received "customer complaints filcd dircctly with Ford, Ford 's authorized dealerships, NHTSA, internet websites, and other public venues"); III re Porsche Ca rs North America, IIlC., 880 F. Supp. 2d 801, 816- 17 (S. D. Ohio 2012) (same); Greal Wesleru Cas ually Co. I '. Volvo Trucks Norfh America, No. 08-CV-2872, 20 10 WL 422294, at ·2 (N. D. 111. O CI. 20, 2010) (finding a plaint iff had alleged sufficiem facls hy claiming "Volvo knew that the tmck was defective because it had received several fi re-related liability claims and warranty claims since August 2003"). 11 76 49 C.F. R. § 573.7. u n 49 USC § 30 I66(m)(3)(A)(i); 49 C.F.R. § 573.6(c)(2)-(8). 1 17~ 49 V.S.C § 301 66(m)(3)(A)(i). 1179 See Statement of the Honorable David Friedman, Acting Administrator, NHTSA, Before the U.S. Senate Comminee on Commerce, Science, and Transportation Subcommittee 011 Consumer Protection, Product Safety, and Insurance (April 2, 2014), at 5, ami/able at hllp://www nhtsa.govrrestimony [DOC lD 000162977022 J. 279 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product Recalls A recall can arise either from a manufacnlrer' s determination that a safety or compli ance defect ex ists (i.e. , "self-generated" ) or from an order by NHTSA requiring a recall (i.e., " influe nced") during or after a NHTSA investigation. As noted, when a manufacturer learns that a sa fety defect ex ists that requires a recall, it must file a defect information report within fi ve days with NHTSA under 49 C.F.R. Part 573, known in the industry as a "573 Report."t ]80 The 573 Re port must identi fy the line of vehi cles or equipment under recall , the number of affected veh icles or pieces of equipment, a description of the defect, and a description of the remedy. The manufacturer must a lso include a chronology of its own investigation and decision-making process. lt 8t If NHTSA has concerns that the manufacturer' s action was untimely, it may open a Timel iness Query ("TQ") to learn more about that process. It 82 Under 49 C.F .R. Part 577, manufacturers are required to noti fy owners of vehicles and vehi cle equipment under recall. 1183 The "577 Letter" must state that the manufacturer has determined that there is a safety defect in a vehicl e or piece of equipment; where and when a remedy to the defect can be obtained; and that the remedy will be provided free of charge. ]t84 49 C.F.R. § 573.6(b) . 49 C.F.R. § 573 .6(c)(6). ]]82 S]alcmelll of thc Honorable David Friedman, Acting Administrator, NHTSA, Before the U.S. Scnate Comminee on Commerce, Science, and Transportation Subcommittee on Consumer Protection, Product Safety, and Insurance (April 2, 2014), at 7, available al http://www.nhtsa.gov!Testimony LDOC ID 000162977022 J. 11 83 49 C.F.R. § 577 el. s eq. 11 M 49 C.F.R. § 577.5(e), (gXI); Statement of tile Honorable David Friedman, Acting Administrator, NHTSA, Before the U.S. Senate Comminee on Commerce, Science, and Transportation Subcommittee on Consumer Protect ion, Product Safety, and Insurance (April 2, 2014), at 7, amilable a/ http://www.nhtsa.govfTestimony [DOC ID 000162977022]. !l80 11 8 ] 280 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product The manufacturer must then track how many of the items under recall receive the remedy ("recall com pl etion" ) and report th e numbers to NH TSA fo r six quarters_ 11 8S NHTSA's Office of Defects Investigation administers TREAD Act requirements, and investlgates defects brought to its attention by manufacturers or by members of the pUblic. 1l 86 Jl85 Sec Statement orthe Honomb1c David Friedman, Acting Administmtor, NHTSA, Berore the U.S. Senate Committee on Commerce, Science, and Tmnsponation Subcommittee on Consumer Protection, Product Sarety, and Insurance (April 2, 2014), at 7, ami/able at http://www nhtsa.govn-estimony[DOC ID 000162977022]. 1186 See Statement orThc Honorable David Friedman, Acting Administmtor, NHTSA, Berorc the U.S. Senate Committee on Commerce, Science, and Tmnspon ation Subcommittee on Consumer Protection, Product Sarety, and Insurance (April 2, 2014), at 4, ami/able at http://www nhtsa.govn-estimony[DOC ID 000162977022]. 28 1 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product APPENDIX B: GM ' s Vehicle Investigations Processes Problems with the Cobalt were identifi ed and in vestigated through three different establi shed GM processes : one within the engineering and program deve lopment area ; One under a dedicated investigative function; and one in the context of addressing product liability claims against GM . Thi s Appendix describes how those processes worked during th e relevant time period. A. Investigations in Engineering and Program Development Process Customer satisfaction and sa fety issues were tracked and addressed through a variety of systems and committees di scussed in this section. t. Problem Resolution Tracking System (" PRTS") PRTS is a database used by GM to document and track engineering problems identifi ed in testi ng, in manufacturing, or through warranty data and customer feedback. PRTS issues could originate from GM employees in a number of areas: Engineering, Industrial Engineering, Service, Worldwide Purchasing, and Assembl y Manufacturing. 11 87 A PRTS issue is generally addressed by engineers in the functional area(s) in volve d - for example, with regard to the Coba lt Igniti on Switch, by engineers responsibl e for the steering column or for the ignition sw itch. 11 88 A large number of PRTS issues are opened for every vehicle. According to one witness, it is typical for a vehicle to have had 500 to 600 PRTS issues opened before launch . 1189 )[87 GM, CP [P and PRTS+ Process Overview (last modi fi ed March 30, 2006), at 3 [nOC ID GM CB~ 00000097727S1. 1I~~ J&B Interview of Joseph Manson, March 19, 20 [4; J&8 Interview of Doug Parks, March IS, 2014. 1189 J&B Interview of John Hendler, April 25, 20 [4 ; see also, e.g., GMXOO I VAPIR Meeting Agenda (Jan. IS, 2005) attaching a spread sheet ofsixty~fi ve open PRTS issues for the Cobalt [DOC 10 14021301695327; GMNHTSA000227763 J. 282 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product After identifying the issue, the originator ofa PRTS report selects a severity level for the prohl em H90 The severi ty leve l is a significant factor in the priority given to a PRTS report, with more severe issues addressed more urgently.1191 The originator selec ts the severity level from a drop~down menu that includes brie f descriptions of what each level signifies. Prior to 20 I 0, there were four seve rity levels that could include probl ems impacting the customer: 1192 Code I : Poss ibly Safety / Regulatory Issues / Walk Home / No Build Code 2: Major Issues - an issue that would cause the customer to immediate ly return the veh icl e to the dealershi p or cause excessive cost or labor impact at the assembl y plant Code 3: Moderate Iss ues - fi x on the next trip to dealership or cause moderate cost or labor impact at the assembl y plant Code 4: Annoyance / Continuous Improvement l1 93 A fifth option, "Code $," was used for Cost Reduction Issues. l )94 In March 20 I 0, Code 3 was revised to "Moderate Build Concern (A ssembly producti vity, Part substitution request, etc.) with NO CUSTOMER IMPACT.,,11 95 The memo announcing that change explained that all issues affecti ng customers should now be rated either level I or level 2 "to focus the organization on fixing every customer issue." II96 Prior to 2010, an issue could be seen as "moderate" or an "annoya nce," and thus given a low priority, even ifit directly affected customers. 1190 GM, CPIP and PRTS+ Process Overview (lasl modified March 30, 2006), at 10~14 rDOC 10 GMC8000000977278] 1191 Dennis Korinek, GM, Priority Drivers and the Prioritization Fonnulas Used in PRTS+ (revised Nov. 11 , 2005), at 2, 32 lDOC ID GMCB-000000977300]; J&8 Interview of Yvollne Cummings, April 3, 2014; J&B Interview of Albcno Mainor, May I, 20 14; J&8 Interview of Gregory Schone, April 2, 2014. 1192GM , Memo Definition # 20100330 (March 2010) rDOC ID 000240765039l 1I9J GM, CPIP and PRTS+ Process Overview (last modifi ed March 30, 20(6), at 14 [GMCB-000000977278l 119~ GM, CPIP and PRTS+ Process Overview (last modifi ed March 30, 20(6), at 14 [GMCB-000000977278l 1195 GM, Memo Definition # 20 100330 (March 2010) [DOC ID 000 240765039]. 1196 E-mail from Kathy Doher, GM, toWW_YBE Staff, GM, et a/. (March 30, 2010) lDOC lD 000007716693]. 283 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product Safety-related issues may be elevated out of the PRTS process for immediate review by the Fie ld Performance Evaluation ("FPE") process, discussed below.lI 97 In the 2005 timefram e, when an engineer was assigned to serve as the initial "Champion" for a PRTS report, the e-mail announcing the assignment directed the engineer to the PPE procedure for potential sa fety issues. 11 98 The CPIT was responsible for officiall y openi ng and reviewing the PRTS reports, includ ing deciding the appropriate severity designation. 1199 Some witnesses have indicated that the severi ty level of a PRTS report was re viewed in committee meetings or by program 1200 . engmeers. There are fi ve steps in the PRTS process: identification of the issue; identification of the root cause; identifi cation of a solution; implementation of the solution; and feedback . mil At each step, a "Champion" is assigned to resolve and close that step. 1202 The Champion may assign a "Champion Designee" to document the required information for that step. 1203 The Champion is typically someone with functional experience with the parts involved, such as the Design Release Engineer or Engineering Group Manager. 1204 The 2006 training module included a process for assigning a new Champ ion if the Champion has been incorrectly assigned, J&8 Interview of Joseph Manson, April 22, 201 4; J&8 Interview ofUlri Queen, May 12, 2014. See. e.g.. E-mail fromPRTSRestriclcd@GMAPP . GM . loKcvinGannon. GM (May 25, 2005) [DOC 10 000011020033 ; GMNHTSA000218751l 1199 J&8 Il11 erview of Joseph Mamon, Apri l 22, 2014 ]200 J&8 Interview of Steve Longman, March 24, 201 4; J&B Inlerview of Joseph Manson, April 22, 2014. 1201 GM, CPIP and PRTS+ Process Overview (last modified March 30, 2006), al 15 [DOC ID GMC8000000977278]. 1202 GM, CPIP and PRTS+ Process Overview (lasl modified March 30, 2006), at 15 rOOC ID GMC80000009772781; PRTS+ 2.0 Overview Trifold (Nov. 2, 2006) rOOC ID GMCB-000000977276\. 1203 GM, CPIP and PRTS+ Process Overview (lasl modifi ed March 30, 2006), al IS rOOC ID GMCB000000977278]. 120-l J&8 Interview of Joseph Manson, March 19, 201 4. 1197 1198 284 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product although it did not include a process for changing the severity level assigned to the PRTS reporL 120~ There was a timetable for taking a PR TS report from initiation to conclusion, but witness reports differ as to what the precise timetable was. One witness said the timing guidelines varied over time, and recalled that the goal in 2009 was to reso lve a PRTS issue within 90 days.1206 Other witnesses reported that a PRTS inquiry was usuall y completed in 3 0~35 days. 1207 Engineering meetings allotted time to PRTS issues that were more than 35 days old and still in the early stages of investigation. 1208 Typically, a PRTS report would be closed at the end of the Feedback step, after a solution had been implemented and it had been confirmed to have addressed the concern. 12 0') However, a PRTS report could also be closed at an earlier stage, "Without Action," with the approval of the Current Production Improvement Team ("CPIT") (discussed below). l21O The PRTS forms did not necessarily reflect that approval process when the PRTS was closed without action. 121l Witnesses differed as to whether a Program Engineering Manager (" PEM") could close a PRTS report on his ow n; most witnesses sa id that the approval of either the CP IT or the Vehicle Chief Engineer ("VCE") was required, depending on the severity level of the issue in the PRTS. 1212 1205 See GM, CPI P and PRT$+ Process Overview (last modified March 30, 2006), at 18 [DOC ID GMCB~ 000000977278J. 1206 J&B Interview of Yvonne Cummings, April 3, 2014. 1207 J&B Interview of Albeno Manzor, May 1, 2014; J&B Interview ofRajiv Mehta, April 3, 2014. 1208 See. e.g. , GMXOOI VAPJR Mccting Agenda (Jan. IS, 2005) [DOC ID 14021301695327}. 1209 GM, CPIP and PRT$ + Process Overview (last modified March 30, 2006), at 22 IOOC ID GMCB000000977278J. 12 10 GM, CPIP and PRT$ + Process Overview (last modified March 30, 2006), at 23 rOOC ID GMCB000000977278]. I ~ I I See. e.g. , PRTS NI72404 (Nov. 19, 2004) rOOC ID 000001771143; GMNHTSAOOOI42658] ; PRTS N182276 (May 16, 2005) [DOC lD 14021301731575]. 1212 See. e.g. , J&B Interview of Joseph Manson, March 19, 2014 (ePIT approval was required to close a PRTS Tepon); J&B Interview of Albeno Manzor, May 1,2014 (c losing a PRTS without action required Chief Engineer's 285 Privileged and Confidenfia/ Protec1ed by AUorn ey~ Client Privilege and as Allorney Work Product Under the GM policy in effect at the time, the PEM did not have authori ty to close the PRTS report; the CPIT had to approve the closure ofa non-safety related PRTS when no action was taken . 12l3 2. Vehicle and Process Integration Review ("VAPIR") Meetings Poss ibl e PRTS solutions would be di scussed at weekly Vehicle and Process lntegration Review ("VA PIR") meetings, attended by program engineers responsibl e for the vehicle. VA PIR meetings would include the Vehicle Program Manager, the Vehicle Architecture Manager, the Business Manager, and Vehicl e System Engineers ("VSEs") for various components of the vehicle. 1214 The PEM and sometimes the VCE would also attend. The VSEs acted as li aisons between the program engineers, who would be responsible for a parti cular vehi cle program (lik e the Cobalt), and system engineers, who would be responsibl e for a system (like Electri cal or Chassis).1215 Safety engineers typica ll y would attend VAPIR meetings during the development phase of a vehicle, but would attend meetings after production began only if a sa fety-related topi c were on the agenda. 1216 The purpose of VAPIR meetings was to di scuss engineering solutions, to keep issues movi ng in accordance with PRTS timelines, and to overcome engineering roa dblocks. 1217 Over the life of a vehicle program, the content of VAPIR meetings would change, focusing first on approval and would trigger scmtiny at the CPIT and Vehicle and Process Integration Review ("V AP IR") meetings); l&B Interview of Doug Parks, March IS, 2014 (PEM could close a PRT5 report himself). 1213"Closed Without Ac(ion" Approval Guidelines By Issuc Typc & Closurc Code [DOC ID 5006S97_000000629]. 12 14 See, e.g. , GMXOOI VAPJR Mccti ng Agcnda (Jan. IS, 2(05) rDOC ID 1402 1301695327]. 12 ]) 1&8 Intcrvicw of John Hcndler, April 25, 201 4. 12 16 J&8 Intcrvicw of Albcrto Manzor, May 1, 2014; see also J&B 11lIcrvicw of Joseph Manson, April 22, 2014 (explaining that safcty enginccr would attend a CPIT meeting, if there was a safety issue before the CPIT). 12 17 J&8 Interview of Joscph Manson, April 22, 201 4. 286 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product development issues, then on va lidation iss ues, then on pl ant or build iss ues. 12l 8 Although VA PIR meetings would focus on e ngineering solutions, they still would consider cost when deciding which solutions to recommend, if the issue presented did not relate to safety.1219 The notes from VAPIR meetings are minimal and not consistently recorded. To the exte nt we have them , the notes are limited to a sentence or two entered next to an item on the agenda. 3. 1220 Cross-VAP[R Meetings Engineering leaders for different vehicle programs wo uld come together in less frequent Cross-V APIR meetings to di sc uss issues affecting multiple vehicle programs, such as a proposed part change that would impact vehicles on more than one vehicle pl atform. 122 1 One witness said Cross-VAPIR meetings would occur once or twice a month . 1222 The Cross~ V APIR attendees , would vary depending on the parts and vehicles on the agenda for di scussion.m . 4. Current Production Improvement Team (" CP IT") Meetings For vehicles in production, PRTS issues and solutions were presented to the Current Production Improvement Team ("CPIT"), which managed business plan objectives for vehicl es in production. 1224 Each vehicl e line would have its own CPIT, respon sible for dec iding whi ch J&8 Interview of John Hendler, April 25, 201 4. J&8 Interview of Gary Altman, March 14, 2014; see also, e.g., GMXOO I VAPIR Meeting Agenda (Sept. 20, 2005) rDOe ID 15032000612752; GMNHTSA0003295321 ("Trush to advise Joe on possible cost rcduction otherwise will SlOp") 122(l See. e.g. ,GMXOOI V APIR Meeting Agenda (Jan. 18, 2005) IDOe ID 14021301695327; GMNHT$A000227763 J; GMXOO1 VAPIR Meeting Agenda (Apri I 12, 2005) IDOe 10 14032600001274J. 1221 J&8 Interview of John Hendler, April 25, 2014; J&8 Interview of Joseph Manson, April 22, 20 14; J&8 Interview of Albeno Manzor, May 1, 2014; J&8 Interview of Steven Oakley, April 23 , 2014. 1m J&8 Interview of Joseph Manson, April 22, 2014. 1m J&8 Interview of John Hendler, April 25, 201 4; J&8 Interview of Alberto Manzor, May 1,2014. 1224 GM, CPIP and PRTS+ Process Overview (last modified March 30, 2006), at 3 [DOC ID GMC8~ 000000977278J. 1218 1219 287 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product improvements to ma ke to that vehicle, and approving PRTS sol utions proposed by engineers. 1225 The C PI T is chaired by the Vehicle Line Director ("V LD") on behalf of the Vehicl e Lin e Execu ti ve ("V LE") fun ction.l226 The VLE would review issues left unresolved by the CPIT. l227 CPIT meetings would typically be attended by the Program Engineering Manager for the vehicle progra m, the Plant Manager, the Proj ect Work Engineer, the Supplier Quali ty Engin eer, the Brand Quality Manager, the Quality Assurance Manager, and the Quality Engineeri ng Manager. 1228 There are no consistent records of who attended each e PIT meeting. 5. Program Execution Tea m (" PE.T") The Program Execution Team (" PET") is above the CPI T in the vehicle program hierarchy.l229 The PET would consider pl atform~wide business decisions and incl uded representati ves fro m marketing and planning, as well as leadership from the CPIT and VAPlR processes. 1230 The PET made commercial decisions and brought a marketing perspective to engineering issues. 1231 For example, it may be advantageous from an engineering standpoint to standardize a component across vehicle models, but the PET coul d make a decision based on its knowledge of customer preferences for diffe rent vehi cles. I 232 The VLE, who was described by one witness as the " top of the food chain," has ultimate authority for cost decisions made at the PET.l233 122S J&8 Interview of John Hendler, April 25, 2014; J&6 Interview of Eric Kenar, March 25 , 2014; J&B Interview of Joseph Manson, April 22, 20 14. 1216 Glohal CIiITenl Product Improvemclll. Process (lIpdalcd April 17, 20(9), at I I [DOC ID GMCB-OOOOO09773071. 1227 J&8 Interview of Joseph Manson, Apri l 22, 2014; J&8 Interview ofStevell Oakley, April 23, 20 14. 1228 Cobalt CPIT Weekly Agenda (June 3, 2005) (DOC ID S007 I 16_000008738J. 1229 J&8 Interview of Chris Bembe, Apri l 22, 2014; J&8 Interview of Albeno Manzor, May 1, 2014. 1230 J&8 Interview of Chris Bembe, April 22, 2014; J&8 Interview of John Hendler, April 25, 2014; J&B Interview of Albeno Manzor, May 1, 2014. 1231 J&8 Interview of John Hendler, April 25, 2014. 1232 J&8 Interview of John Hendler, April 25, 2014. 1m J&8 Interview of Joseph Manson, Apri l 22, 2014. 288 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product 6. Field Performance Evaluation ("FP[") and Product Investigations The Field Performance Eva luation (" FPE") process was a dedicated investi gation process used to evaluate safety, compliance, emission, and customer satisfaction issues, and to determine whether a field action is necessary . FPE is a separate process from the PRTS process, although an FP E issue may (but does not necessarily) ari se out of the PRTS system. 1234 The FP E process is run by the FPE Director in GM's Globa l Quality and Customer Experience division, separate from the Engineering function. Issues referred to the FPE process could arise from virtually any part of the company and be based on a problem identified by any source. 1 23~ One avenue to elevate an iss ue through the FPE process is through an internal investigation conducted by the Product Investigations ("P I") grouP. a designated group of engineers devoted to such investigations. m 6 After the PI group' s in vestigation, its findings could be elevated through the FPE process to determine whether a field action was appropriate. 1237 Investigators had a target goal of bringing an issue through the FPE process within 40 days of the matter being assigned to PI, but there was no speci fi c consequence to missi ng that deadline.1238 For example, as long as an investigator kept the supervisor apprised of an investigation ' s progress, it was acceptable for an investigation to last more than 40 days.1239 There was also a perception that an issue cou ld not proceed to the deci sion~making J&8 Interview of Joseph Manson, April 22 , 201 4; J&B Ill\erview of Alberto Manzor, March 2 1, 2014. Gcneral Motors Global FPE Process (Feb. 2, 2009), at 3 [DOC ID 00000 15562061, attached to e~ma il from DOllglas Wach1el, GM , 10 Canncn 8cna v id c~ , GM (Jan. 13 , 2013) lDOC ID 00000 1556205]. 12)6 Maureen Foley-Gardner, Reinvent ion of FPE (Aug. 2012), at 13, 16 [DOC ID 00005 1944665 J (showing that Product Investigations handles the initial analysis/investigation in a ''Traditional FPE"); J&8 Interview of Maureen Foley-Gardner, March 26, 2014. 1237 Maurccn Foley-Gardner, Reinvention ofFPE (Aug. 20 12), at 13, 16 [DOC ID 0000519446651 (showing that Product Investigations handles the initial analysis/investigation in a "Traditional FPE"); J&8 Interview of Maureen Foley-Gardner, March 26, 2014. 1238 J&8 Interview of Brian Stouffer, May 9, 201 4. 1239 J&8 Interview of Brian Stouffer, May 9, 201 4. )234 1235 289 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product levels of FPE until a root cause was determined. 1240 A flow chart from 2005 shows " Detennine Root Cause" as a step before deciding to conduct a safety field action. 1241 The first place where the Pl group would present its findin gs was at one or more Information Status Review ("ISR" ) meetings, held weekly and attended by the FP E Director and the Product Investigations Director, as we ll as by representatives from the Legal Department, Customer Care and After Sales, FPE, and Product Investigations. I242 On occasion, a PI engineer might bring an issue to ISR to draw attention to the case or to present diffic ult issues raised by the investigation , before determining the root cause. 1243 However, issues presented at an ISR meeting were usuall y referred to the next level for further review and recommendation suggesting that issues did not typically go to ISR until the investigation was fairl y far along. 1244 For most issues, the next stage after the ISR would be a meeting of the Field Perfonnance Evaluation Team ("FPET"), which could recommend possible field actions (including recalls) to address the problem. 1245 After FPET, the issue would move to the Field Performance Evaluation Review Committee ("FPERC") to detennine whether to reconunend a field action to the Executi ve Field Action Decision Committee (" EF ADC,,).1246 The EF ADC would make the fina l deci sion regarding recall s and other field actions for GM. The deci s i on~makers on the EFADC are the Senior Vice President for Global Quality and Customer Experience, the Global Vice J&B [Ill crvicw of Brian Slollffcr, May 9, 201 4. ]24] FPE Process (Junc 22, 2005), at 2 l DOC 1D S 0 [3189 [6J. ]242 J&B Intcrvicw of Maureen Folcy-Gardncr, March 26, 2014. 1243 J&B Intcrvicw of Brian Stouffcr, May 9, 20 [4. 124·1 J&B Intcrvicw of Maurccn Folcy~Gardncr, March 26, 2014. 1 ~4' Maurccn Foley-Gardncr, Reinvention ofFPE (Aug. 20[2), at 20 rDOC 1D 0000519446651; J&B Intcrviewof Maurcen Folcy-Gardncr, March 26, 2014. 12~6 Maurccn Folcy-Gardncr, Reinvention ofFPE (A ug. 2012), at 20 [DOC 1D 000051944665] ; J&B Intcrviewof Maurcen Folcy-Gardncr, March 26, 2014. 1240 290 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product President for Engineering, and the Global Vice Pres ident for Man ufacturin g. 1247 In order to issue a recall , the three EFA DC members needed to he unanimous. 1247 J&B Interview of Maureen Foley-Gardner, March 26, 2014. 29 1 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product APPENDIX C: GM's Approach to Potential Safety Problems Through Vehicle Development Process In the 2000s, GM utilized a number of processes to assure safety and quality in the development of a vehicle and its component parts. 1. The Vehicle and Component Development Process Since at least the early 2000s, new GM vehicles go through GM's Global Veh icl e Deve lopment Process ("GVDP" ). I 248 The GVDP guide s pre-production vehicles through a series of " milestones," described by GM employees as "gates" or "checkpoints," along the path from vehicle design, through component and vehicle validation, to vehicle launch and production. I249 At each step of the GYDP, GM engineers work to ensure that a vehicle and all of its parts satisfy all applicable requirements and specifications. mo 2. Technical Specifications In the earl y 2000s, as they do now, technical specifications served as a foundation for the vehicle and component development process. GM maintained both Vehicle Technical Specifications ("VTS") and Component Technical Specifications ("CTS,,).ml VTS contained standard requirements for all GM vehicles, including requirements mandated by the Federal Motor Vehicle Safety Standards ("FMVSS,,).1252 CTS contai ned detailed descriptions of the Plans in Motion: The GVPO Workshops (2001), aI 9 [DOC ID 000140632991] Plans in Motion: The GVPO Workshops (2001) LOOC ID 000 14063299 1]; J&B Interview OfOOllg Parks, May 1-2, 20 J 4; 1&B Interview of Lori Queen, May 12, 2014. GM has employed varying names and ternlinology for the steps and features of this process since the early 2000s. See 1&B Interview of Chris Berube, Apri l 22, 2014. Wherever possible, this Report employs the tenninology used eontcmpomneously with the events it describes. I ~W J&8 Interview of Chris Bembe, Apri l 22, 201 4. ]251 J&8 Interview of Doug Parks, May 1-2,20 14. 1252 J&8 Interview of Chris Berube, Apri l 22, 2014; J&B Interview of Ernie McCutchen, May 16, 2014; J&B Interview ofOoug Parks, May 1-2, 201 4. 1248 ]249 292 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product requirements to build the part, as we ll as a Validation Testi ng Plan for the component.12S3 3. Design Release Engineers and the Product Development Team GM 's Design Release Engineers ("OREs") had responsibility for working with GM 's suppli ers to develop speci fi c vehicle components for use in particular GM vehicles - their "design " responsibiliti es - and to ensure that those components satisfied GM's requi rements and specifications before ultimately approving the part for use in a GM vehicle - "releasing" the part. 12S4 The ORE was the primary point of contact with a supplier during the part development process. 12SS A suppli er selected to develop a vehicle component received from GM the relevant CTS, along with any other vehicle-speci fic requirements for the part. 12S6 The suppli er then worked to design and develop a pan that satisfied all GM's requireme nts and specifications. l2S7 OREs held regular Product Development Team (" PDT") meetings wi th the supplier. 1258 Other GM engineers, such as Validation Engineers ("VEs") and Supplier Quality Engineers ("SQEs") occasionally participated in these meetings. 12S9 At PDT meetings, a DRE provided feedback on work done by the supplier, including the results of any tests the suppli er performed on the part. 1260 If either th e ORE or the s liPpli er wanted to make a change to a part, the proposed change would be disc lissed at a PDT meeting. 1261 4. Prototype Parts and Validation Testing Validation testing, the process through whi ch GM verified that its vehicles and their J&B Intcrvicw of Ernic McCutchcn, May 16,2014; J&B Intcrvicw of Doug Parks, May 1~2 , 2014. J&B i1l1crvicw of Raymond DcGiorgio, M~ rc h 14,2014; J&B Ill1crvicw of Thomas lJllcr, April 24, 2014; J&B Interview of Doug Parks, May 1-2, 2014. ms J&B Interview of Thomas Viler, April 24, 2014. 1256 J&B Intcrvicw of Thomas Vllcr, April 24, 2014. 1257 J&B Intcrvicw of Raymond DcGiorgio, March 14, 2014; J&B Intervicw of Thomas Vllcr, April 24, 20 14. 1m J&B Intcrview of Raymond DcGiorgio, May 7-8, 20 14; J&B Interview ofTIlornas Vllcr, April 24, 2014. 1259 J&B Intcrview of Raymond DcGiorgio, May 7-8, 20 14; J&B Interview of Emie McCutchen, May 16, 2014. 1260 J&B Interview of Raymond DeGiorgio, May 7-8, 20 14; J&B Intcrvicw of Emie McCutchen, May 16, 2014. 1261 J&B Interview of Andrew Brenz, April 23, 2014. 1253 1254 293 Privileged and Confidenfia/ Protec1ed by AUorney~ Client Privilege and as Allorney Work Product component parts met GM requirements and specifi cations, was a key component of the GV DP.1262 At multipl e stages in the vehicle developme nt process, GM and its suppli ers conducted validation testing on both individual vehicle components and on integrated pre~ " veh"IC Ies. P~ 63 productlOn Validation testing was conducted in two general stages: Design Validation (" DV") and Production Validation (" PV,,). 1264 During the DV stage, GM' s suppliers produced prototype parts and prototype vehicl es for use in conducting validation testing. 1265 Earl y prototype parts and vehicles, commonl y referred to as " Beta-level" parts or ve hicles, were assembl ed by hand using prototype tools made of aluminum, referred to as "'soft tooIS.,,1266 Conducting DV testing on prototype pans allowed GM to learn about any problems with the pans - either as standalone parts or as components of an integrated vehicl e - that required redesign as early in the process as possible. 1267 Later in the GVDP, GM and its suppliers conducted PV testing on components and vehi cles manufactured using production tools, in the faci lities where the vehicl e would eventually be produced for sa le, in order to prove that the individual components, as well as the vehicle as a whole, met requirements and spec ifications. 1268 GM employed Validation Engineers responsible for overseeing the va lidation process and approving the results. 1269 As explained further below, the role and scope of involvement of J&8 Interview of Raymond DeGiorgio, March 14, 2014; J&8 ]nterview of Dou g Parks, May 1-2, 20 14. J&B il11erview of TIl omas 11l1er, April 24, 2014. m·l J&8 Interview of Thomas Viler, April 24, 2014. 1265 J&8 Interview of Raymond DeGiorgio, May 7-8. 20 14; J&B Interview of Emie McCutchen, May 16, 20 14: J&8 Illlerview ofTIlOmas Viler, April 24, 2014. 1266 J&8 Interview of Raymond DcG iorgio, May 7-8, 2014; J&6 Interview of Ernie McCutchen, May 16, 2014; J&8 Illlerview of Thomas Viler, April24, 2014. 1267 J&8 Interview of John Hendler, May 20, 2014 1268 J&8 Interview of John Hendler, May 20, 2014; J&6 Interview of Thomas Viler, April24, 2014. 1269 J&8 Interview of Eugene Carnago, May 15. 20[4; J&8 Interview of Emie McCutchen, May [6, 20[4. 1262 1263 294 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product Validation Engineers varied between 2000 and the present. 1270 5. C o mpone nt- Level Validatio n Testi ng Since the early 2000s, GM ' s suppliers have been primarily responsible for conducting component~level validation testing .1271 Supp liers communicated the results of component~level validation testing to GM in an Analysis/DevelopmentlValidation Plan & Report ("ADVP&R,,).1272 ADVP&Rs tracked the spec ific va lidation plan laid out in the CTS, and provided test results on sample parts that showed whether the part met the spec ifications required by the CTS.1273 DV stage ADVP&Rs also contained the results of the suppl ier' s component~ level Design Failure Mode and Effects Ana lysis (" DFMEA"), a process intended to identify potential fa ilures that might occur in a pan, eva luate the risk of such an occurrence, and determ ine whether th e part could be designed to compensate for such a fa ilure, as well as of the supplier's Process Fail ure Mode and Effects Analysis (" PFMEA"), an analysis like a OFMEA l' '.' lor manullactun ng processes.1274 In the earl y 2000s, ADVP&Rs containing the results of co mpon ent~ l eve l validation testing would have been reviewed by the ORE responsib le for the part. 1275 In order to approve a part for production, the ORE was required to sign a Commodity Validation Sign~offform , also mo J&8 Interview of Eugene Camago, May 15, 201 4; J& B Interview of Ernie McCutchen, May 16, 2014. J&8 Interview of Raymond DcGiorgio, March 14,2014 ; J&B ]nterview of Thomas Viler, April 24, 2014. 127l J&8 Illl erview of Doug Parks, May 1~2 , 201 4 ]27) J&8 Interview of Thomas Viler, April 24, 201 4; see, e.g., AnalysislDevelopmentNalidation Plan & Report ("ADVP&R") for Suppliers for Delta Z Ignit ion Swi tch (Jan. 10, 2002) [DOC ID 00005 1786004; GMNHTSA00025778 11 ; Analysis/DevelopmentNa lidatioll Plan & Report ("ADVP&R") for Suppliers for Delta Z Ipllition Switch (May 21 , 2002) rDOC ID 000051786005 ; GMNHTSA000257795l J" 74 J&8 Interview of James Federico, March 20, 2014; J&B Interview of Emie McCutchen, May 16, 2014; J&B Interview of Andrew Brenz, Apri l 23, 2014. 1275 J&8 Interview of Raymond DeGiorgio, March 14, 201 4; J&B ]nterview of Dou g Parks. May 1-2, 20 14; J&8 Interview of Mallhew Schroeder, April 2, 201 4. 1271 295 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product known as Fonn 3660. 1276 It is unclear whether, in the early 2000s, a GM Validation Engineer with responsibi lity for an indi vidual component would also have been assigned to review and approve ADVP&R results; witnesses interviewed diffe red as to thi s fact. 1277 By 2006, however, GM undoubtedly employed Validation Engineers with component~level responsibilities, who also were required to sign off on Foml 3660 approva ls. 1278 If the ADVP&R showed that a part did not meet specifications, the DRE and VE would generally work w ith the supplier at PDT . · ue. 1"'79 meetings to a dd ress t he Iss ~ Witnesses simi larly provided differing accounts of whether, in the early 2000s, a DRE could have unilaterall y approved a part for production that was out of specification, and how this approval would have been documented. Several GM personnel interviewed thought that, even in 2001 or 2002, a DRE would have required the approval of at least an Engineering Group Manager ("EGM") or Validation Engineer to approve an out~of~spec part. 1280 Two validation engineers, however, thought that, in the earl y 2000s, a DRE had the ability to approve a part that did not pass component~leve l vali dation testing. 1281 While some witnesses thought that a deviation from specification could have been documented on GM Fonn 14 11, others thought that, in the early 20005, Form 141 1 was used as an interim approva l, to give the suppli er additional time to satisfactori ly complete vali dation testi ng. 1282 J&B [l11 erview of Eugene Camago, May 15,2014; J&B Imerview of Emie McCulchcn, May 16, 201 4 Compare J&B Interview of Emie McCutchen, May 16, 201 4, with J&B Interview of Eugene Carnago, May 15, 20 14. 1278 J&8 Intcrview of Ernic McCutchcn, May 16, 2014; J&B Intervicw of Eugcne Camago, May 15, 2014. 1279 J&8 Intcrview of Ernic McCutchcn, May 16, 2014. 121:W J&8 Intcrvicw of Ernic McCutchcn, May 16, 201 4; J&B Intervicw of Doug Parks, May 1~2 , 2014. ]281 J&8 Intcrview of Eugcne Camago, May 15, 20[4; J&B Intcrview ofMichacl Bcrdcn, May 27, 2014. 1282 Compare J&B Intervicw of Doug Parks, May 1-2, 201 4; J&B Intcrview of Gary Altman, May 8, 2014, wilh J&B Intervicw ofRalfNickcl, March 26, 201 4 ; J&B Intcrvicw of William Skclton, April 25 , 2014. 1276 ]277 296 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product 7. TALC and Human Factors Review Before receiv ing final approval , some component parts were also reviewed b y GM's Touch, A ppearance, Li ghting, and Control ("TALC") team and/or Human Factors group. 1283 The TALC team assigned to the vehicl e worked with the vehicle's Chie f Engineer to make sure that the vehicle and its components had a con sistent feel and appearance. 12M TALC testing would occur at both the prototype phase ofa car's development and also the production phase. 1285 The Human Factors engineers evaluated a vehicle from a customer's perspecti ve with a focu s on four criteri a: ( 1) sense, (2) reach, (3) understand, and (4) manipulate. 1286 Their work focused on, for example , reachability of items on the dashboard from the driver's seat, or making sure there was adequate cl earance around the ign ition switch for a driver 's hand to turn . 1287 When evaluating a program and providing input, typically two Human Factors engineers would evaluate a vehi cle together. 1288 They used specific forms, called Human Factor Criteria I.nventory ("HFC I") forms. 1289 The form s contained a li st of criteria to evaluate the vehicle, or a specific component within the vehicl e, and the engineers gave it a letter grade from A to F.1290 These evaluations were presented at vehicl e program team meeti ngs w ith comments explaining J&B Intcrvicw of Joan Hcnley, May 14,2014; J&B Intcrvicw of Manhcw Schroedcr, April 2, 201 4; J&B IlIIcrvicw of William Skcllon, Apri l 25 , 201 4; J&B IlIIcrvicw of Thomas LJllcr, April 24,20 14 128~ J&B Interview of Matthew Schroeder, April 2, 201 4. Some employecs interviewed thought the T ALe team was focused most on appearance and lighting rather than feel. See J&B Interview of Joan Hertley. May 14,2014. 1285 J&B Intcrvicw of William Skelton, April 25, 2014. 1286 J&B Intcrvicw of Joan Henley, May 14, 2014. 1 2~7 J&B Intcrvicw of Joan Henley, May 14,2014. 1288 J&B Intcrvicw of Joan Henley, May 14,2014. 1289 J&B Interview of Joan Hertley, May 14, 2014. 1290 J&B Interview of Joan Hertley, May 14, 2014. 1283 297 Privileged and Confidenfia/ Protec1ed by AUorn ey~ Client Privilege and as Allorney Work Product the scores.129 1 If a low grade was given, there was typically di scussion with the program team about how to address th e Human Factors engineers' concerns. 1292 8. Vehicle-Level Validation Testing GM al so conducted system~ and vehicle~le ve l validation testing to ensure vehicle components fun ctioned properl y when integrated, and that the vehicle as a whole satisfied durab ility standards and appli cable regulations. A vehicle's PEM was responsible for preparing an Analysis/DevelopmentlValidati on ("ADV") Plan for his or her vehicle that incl uded all of the require ments contained in GM ' s VTS, as well as a va lidation testing plan for the vehicl e, and for ensuring that the vehicle as a whole was in fact full y va lidated. 1293 Vehicle-level validati on engineers became involved in the vehicle development process as soon as there was a prototype car to validate and possibly earlier, to help develop the va lidation testing plan. 1294 To coordinate and oversee vehicle-l evel va lidation testing, eac h vehicl e had a lead Validation Engineer responsible for the validation of the vehicle as a whole, as well as systemlevel Va lidati on Engineers representing and responsible fo r the various subsystems of th e vehic le, such as elec trical, chassis, powertrain, and interi ors. 1295 Vehic le-l evel va lid ati on engineers coordinated with the engineers responsibl e for vehicl e technical speci fications to enSure that tests required by the vehicl e-level technical specifications relevant to the electri cal system were performed on the vehicles, and to furth er ensure that all requ ired certi fication testing - such as for FM VSS compliance - was performed. 1296 J&B J&B J29j J&B m~ J&B 1295 J&B 1296 J&B 1291 1292 Interview of Joan Henley, May 14, 2014. Interview of Joan Henley, May 14, 2014. Interview of Doug Parks, May 1-2,201 4. Interview of Ernie McCutchen, May 16, 201 4. Interview of Ernie McCutchen, May 16, 201 4. Interview of Ernie McCutchen, May 16, 201 4. 298 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product Deve lopment Engineers who worked at the Milford Provi ng Grounds actuall y performed validati on test dri ving. 1297 Vehicl e~l evel Va lidation Engineers superv ised the Developm ent Engineers performing the tests that related to their particular subsystern. ' 298 In addition to s ub system~speci fi c tests , each vehicl e had Validation Engineers with responsibility for performing durability testing. ' 299 As part of durability testing, in addition to putting a vehicl e through driving tests at Milford, Validation and Development Engi neers, along with OREs, supported cross-country ride trips for 8- 10 cars to test the vehicle during hard driving over long distances and ensure that the performance of all ve hicl e features could be tested. B OO Vehicle issues identified during validation testing were documented in th e PRTS system. UOI Pre-production issues captured in a PRTS repon would be reso lved throughout the PV stage and monitored by the program team exec uti ves, including the VAPIR committee and the Vehicl e Line Executi ve, as the vehicle moved closer to launch. 1302 Approximately 10 weeks before a vehicl e went into production, any PRTS reports that remained unresolved were presented during laun ch meetings. 1303 All PRTS reports classified as severity level 1 had to be resolved before launch. l304 J&B J&B 1299 J&B 1300 J&B lJOI J&B ]302 J&B ]303 J&B ]3O.J J&B 1297 1298 Interview of Ernie McCutchen, May 16, 2014. Interview of Ernie McCutchen, May 16, 2014. Interview of Ernie McCutchen, May 16, 2014. Interview of Ernie McCutchen, May 16, 2014; J&B Interview of Chris Skaggs, May 20, 20 14. Interview of Ernie McCutchen, May 16, 2014; J&B Interview of Doug Parks, May 1-2, 2014. Interview of Lori Queen, May 12, 2014. Interview of Lori Queen, March 17, 20 I4. Interview of Lori Queen, May 12, 2014; J&B Interview of Doug Parks, March 18, 20 14. 299 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product At the end of the va lidation process, the Chief Engineer or Program Engineering Manager must sign a Product Validation Complete Letter (" PVC L") certifying that the vehi cle compli es with a ll applicable specifications and is ready for producti on. l3OS 9. Captured Test Fleets As launch approached, l ate~stage p re~prod u cti o n veh icles also underwent Captured Test Fleet ("CTF") testing, in whi ch vehicles were given to progra m team executives to drive. 1306 The cars were built at the assembly plant, given VIN numbers, test run for three months, and then sold as used cars. 1307 Problems that arose during Captured Fleet Testi ng were reported in field reports submitted by the employees and recorded in a separate database maintai ned by the Quality Group.130S The CTF reports could contain reports of malfunctions, but the reports were intended to contain anything that the driver did not like about the vehicle (dri ver preference). The reports woul d be reviewed, summarized and grouped into recurring categories by the Qual ity Group. These groupings and summaries would be entered into a spreadsheet, which was e-mailed to the chi ef engineer, the Program Engineering Manage r ("PEM"), a nd the program team. The program team held week ly meetings at which thi s informat ion was discussed. l309 10. Th e C ompan y Vehicle Evaluation Program ("CVE P") Later, as part of the Company Vehi cle Eva luation Program ("eVEP"), several hundred additional salable vehicles would be made avai lab le for GM employees to drive in order to help identify potenti al problems. i31 O eVE P programs are diffe rent fro m CTF programs, but they can J&8 Intcrvicw of Emic McCutchcn, May 16, 2014; J&8 Intcrvicw of Doug Parks, May 1-2, 2014. J&8 Intcrvicw of Lori Quccn, May 12, 2014. lJ07 1&8 Intcrview of Joscph Taylor, May 16, 2014. 1308 1&8 Intcrvicw of Lori Quccn, May 12, 2014; J&8 Intcrvicw of Joscph Taylor, May 16, 2014. 1309 1&8 Interview of Joseph Taylor, May 16, 2014. 131 0 1&8 Interview of Gary Dowd, May 8, 2014; J&B Interview of Joseph Taylor, May 16, 2014. 1305 1306 300 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product be run concurrentl y. eTF was pre-production , and eVEP was post~production. eVEP cars were onl y provided to indi viduals w ith the program team. Any incident reports w ith a eVEP vehicl e would go to GM ' s service engineering. The eVEP cars were treated like cars sold to real customers in that a c ustomer would take his or her car to a dealer, but eVEP vehicles were taken to GM's service eng ineering garage. Warranty reports from dea lers based on eVEP reports generated fro m the service engineering garage wo uld go into a master warranty database. The actua l eVEP reports (from the drivers), as well as the e TF reports, were only seen by the program team.1J11 1311 J&B Interview of Joseph Taylor, May 16, 2014. 301 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product APPENDIX D: Production Part Approval Process ("PPAP" ) Before any component part (new or modified) can be used in a production vehicle, it is required to pass through the Production Part Approva l Process (" PPAP"). The part approval staffs at Chrysler, Ford, and General Motors, working under the auspices of the Automotive Division of the American Society for Quality Contro l (" ASQC") and the Automotive Industry Action Group ("AIAG") , developed standardized PPAP guidelines. 13l2 The purpose of the PPAP is "to determine ifall customer des ign record and specification requirements are properly understood by the suppli er and that the process has the potential to produce product consistently meeti ng these requirements during an actua l production run at the quoted production rate.,,13i3 A PPAP " package" consists of 19 required categories of documents/inputs. These include: (I) design records of saleable product; (2) engineering change document; (3) General Motors Commodity Validation Sign-Off(Form 3660); (4) design failure mode and effects analysis ("DFMEA"); (5) process flow diagrams; (6) process failure mode and effects analysis (" process FMEA"); (7) control plan (pre-launch control plan GP~I2 , plus production control plan); (8) measurement systems ana lys is studies; (9) dimensional results ; (10) material, perfonnance test results ; (11) initial process study; ( 12) qualified laboratory documentation; (13) appearance approval report (if app li cable); (14) sample product; (15) mastersample; (16) checking aids; (17) records of compliance to customer specifics; (18) part submission warrant; and ( 19) bulk material requirements checkli st. 1314 1312 13ll 1314 AIAG PPAP Manual (3d cd. Aug. 2000), at v rDOC ro 000123252680]. AlAG PPAP Manual (3d ed. Aug. 2000), at I [DOC ro 000123252680]. GM Global Supplier Quality Manual (Sept. 2008), at 22 LOOe lD 000049740003 J. 302 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product The Part Submission Warrant ("PSW") represents the supplier 's confirmation that the . . an d tests 0 f pro dllctlon . parts conlorm 'G '· 13 1> GM ' s "Cc mmo d·Ity inspection to M s reqUirements_ J Validation Sign Off' (Form 3660) serves as GM's engineering approval for the part. 13 ]6 Under current GM practice, both the ORE responsible for the part and a component level Validation Engineer must sign off on the Form 3660 to authorize the change. 13 17 The PPAP Manual provides: Any results that are outside specification are cause for the supplier not to submit the parts, documentation and/or records. Every effort shall be made to correct the process so that all design record requirements are met. If the supplier is unable to meet any of these requirements, the customer 1J 18 shall be contacted for determination of appropriate corrective action. In certain cases, specifications may be changed to resemble the actual perfonnance of the part (i.e., the part provides the desired feel) , and those changes can be memorialized in the component technical specifications for the part. 1319 However, some GM employees described a separate deviation process that allows a component to be used in production vehicles notwithstanding noncompliance with GM 's speci fications. 132o Under this process, to approve a nonconforming part for production, the ORE must sign off on the deviation on th e validation paperwork and obtain approval from at least the Validation Engineer, and, potentially, the ORE's AlAG PPAP Manual (3d cd. Aug. 2000), at 90 [DOC ID 000 I23252680j. J&8 Interview of Ralf Nickcl, March 26, 2014. 13 17 See Gcneral Motors Commodity Validation Sign~OfT(Fonn 3660) (May 3, 2007) [DOC 10000004253531 ; GMNHTSA000223928]. 1318 AIAG PPAP Manual (3d cd. Aug. 2000), at 90 [DOC ID 000 123252680l 13 19 J&8 Interview of Doug Parks, May 1-2, 201 4; 1&B Interview o f Thomas Utter, April 24, 2014. 132<1 J&B Interview of Thomas Utter, April 24, 2014; J&B Interview of Raymond DeGiorgio, May 7-8, 2014. 13IS 13 16 303 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product supervi sor, the EGM. \32 1 It is unclear whether such a process for the approval of deviati ons from specifications existed in 2002 because GM protocols previously were not as well defin ed. 1322 Ifa supplier is unable to satisfy the PPAP requirements and must submit an action plan to achieve full PPAP approval , the suppli er is responsible for providing a GM Form 1411 to the GM Suppli er Quality Engineer for sign off and, depending on the issue, may have to obtain additional approvals, including from the DRE. 1323 Form 1411 contains a section entitled "Supplier Performance and Validation Requirements.',1 324 Thi s sec tion sets forth three choices to describe the PPAP status: (I) " Performance/validation requirements met, signed copy of GM 3660 submitted in package"; (2) " PerformanceN alidation requirement items I & 5 on GM 3660 completed satisfactorily"; and (3) «Performance requirements NOT fu lly met; status acceptable to move to a Saleable status" (but " SMT Director and/or SMT Validation Director signature required,,). '325 The supplier also must set forth in the Form 1411 an action plan to obtain full PPAP approval. 1326 The supplier. not GM , is responsible for maintaining the complete PPAP package. J32 7 Specifically, the AIAG PPAP Guidelines provide that the supplier "shall " maintain a PPAP part file for the documentation associated with the PPAP package and make " readil y available for 1321 J&B Interview of Raymond DeGiorgio, May 7-8, 201 4; J&B Intcrview of Doug Parks, May 1-2, 2014; AnalysislDevelopmentN alidation Plan & Rcport (May 3, 2007) I DOe ID 000009486444; GMNHTSAOOO2238011. 1322 J&B Interview of Raymond DeGiorgio, May 7-8, 201 4. 1323 GM Global Supplier Qllalil y Manual (SCpL 2008), at 23 [DOC ID 000049740003]; J&8 hucrvicw ofDolIg Parks, May 1-2, 2014. 132~ GM, PPAP Worksheet (GM 1411) (March 6, 2005) LDOC ID 000116935005]. 1325 GM, PPAP Worksheet (GM 1411) (March 6, 2005) rOOC ID 000116935005]. 1326 GM, PPAP Worksheet (GM 1411) (March 6, 2005) rOOC ID 0001 I6935005l 1m AIAG PPAP Manual (3d cd. Aug. 2000), at 3 rOOC ID 000123252680l GM Global Supplier Quality Manual provides that under the current system, GM maintains in GQTS copies of the Part Submission Warrant, Oimensional Report, Material and Perfonnanee Test Report, AAR, Fonn 1411 and Fonn 3660. GM's Global Supplier Quality Manual (Sept. 2008), at 22 LDOe ID 000049740003 j. 304 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product customer use" sample production parts, ma ster samples, or checking aids. 1328 The AIAG requires suppli ers to retain the PPAP package for "the length of time that the part is active . plus one calendar year" and requires suppliers to ensure that any superseded PPAP packages are included or referenced in the new PPAP packages. 1329 1328 )329 AlAG PPAP Manual (3d cd. Aug. 2000), at 3 [DOC ID 000123252680]. AlAG PPAP Manual (3d cd. Aug. 2000), at 17 LDOC lD 000 I23252680J. 305 Privileged and Confidenfia/ Protec1ed by AUorney~Client Privilege and as Allorney Work Product APPENDIX E: GM's TREAD Datahase After the TREAD Act (the "Act") became law in 2000, GM developed a database, known as "TREAD," to house the data required to be reported quarterly to NHTSA under the Act. I BO In addition, GM investigators can pull data from TREAD on an ad hoc basis to support GM investigations. 1331 Since 2003 , TREAD has been the principal database used by GM to track incidents related to its vehicles. 1332 The TREA D database is organized to track and report data in categories created by NHTSA covering each of24 different systems in a ve hicle (e.g. , steering or airbags) as well as two additional data categories: fire and rollover. 1333 The database draws information primarily from seven sources o f data from within GM: • • • • • • Service Requests ('