- UNITED STATES DISTRICT COURT EASTERN DISTRICT OF MICHIGAN SOUTHERN DIVISION PATTI JO CAHOO, KRISTEN MENDYK, KHADIJA COLE, HYON PAK, and MICHELLE DAVISON, Plaintiffs, Case Number 17-10657 v. Honorable David M. Lawson FAST ENTERPRISES LLC, CSG GOVERNMENT SOLUTIONS, STEPHEN GESKEY, SHEMIN BLUNDELL, DORIS MITCHELL, DEBRA SINGLETON, and SHARON MOFFET-MASSEY, Defendants. __________________________________________/ OPINION AND ORDER DENYING MOTION TO CERTIFY CLASS The five named plaintiffs, all former claimants in Michigan’s unemployment compensation system, allege that their constitutional right to due process of law was infringed when the defendants designed, built, and implemented an automated system to detect and punish individuals who submitted fraudulent unemployment insurance claims. They seek to certify an opt-out class, the definition of which has evolved throughout the case, with the pace of evolution accelerating during the briefing on this motion. In 2012, the State of Michigan’s Unemployment Insurance Agency (UIA) began using its Michigan Integrated Data Automated System (MiDAS) to investigate and adjudicate fraud cases against claimants. It remained operative until August 2015, when the UIA discontinued use of that system for that purpose. The named plaintiffs each were adjudicated as having submitted fraudulent claims and were assessed penalties and interest, collected in some instances by seized income tax refunds and wage garnishments. They were denied due process, they say, because MiDAS failed to provide adequate notice of the fraud accusations and automatically adjudicated fraud claims through the rote operation of built-in decision trees that rigidly were applied either through automation or by UIA functionaries. They wish to represent a class of “[a]ll persons whose fraud determinations were made using the MiDAS system in any way from October 1, 2013 until August 7, 2015,” which, they believe, consists of about 67,000 individuals. They propose four subclasses in the alternative. The plaintiffs have identified a number of faults with MiDAS, several of which support a claim for denial of due process. However, the evidence does not sustain the idea that all members of the main proposed class, or even the proposed subclasses, experienced the same problems with their MiDAS encounters or suffered the same consequences. Put another way in class-action parlance, except for the defendants’ accountability, the plaintiffs have not identified a decisive common issue “that is central to the validity of each one of the claims” of each one of the class members. Wal-Mart Stores, Inc. v. Dukes, 564 U.S. 349-50 (2011). And they have not established that the named plaintiffs’ experiences were typical of the other absent class members whose fraud claims were adjudicated at various stages of the MiDAS process. There is one group of potential plaintiffs whose claims may benefit from class treatment: those individuals adjudicated guilty of fraud solely because they did not return their questionnaires. But the named plaintiffs who fall in that category are not suitable class representatives because their individual circumstances — particularly their failure to list the cause of action in their bankruptcy schedules — subject them to unique and individual defenses that threaten to become the focus of the litigation and consume a significant measure of their time and energy. Finally, although many class members’ claims raise legitimate grievances, the plaintiffs have not demonstrated that the class-action structure is superior to other methods of claim resolution. The motion to certify a class will be denied. I. Facts Under the Michigan Employment Security Act (MESA), an individual may be eligible for benefits if she is unemployed, registers for and actively seeks work, and is available to perform suitable full-time work. Mich. Comp. Laws § 421.28. A claimant also must show that he was neither fired for misconduct nor voluntarily left employment without good cause. Once the UIA determines that a claimant meets the basic requirements for benefits, the claimant must report biweekly to the Agency to maintain her eligibility. The claimant must answer a series of questions by phone through the Michigan Automated Response Voice Interactive Network (MARVIN) or by accessing the Michigan Web Account Manager (MiWAM). Failure to answer truthfully can expose claimants to penalties. MiDAS is the UIA’s fraud-detection software that records and searches for inconsistencies in the data that might indicate fraud. The data originate from several sources, including a claimant’s benefits application and the bi-weekly updates. Employers also independently submit information about employees, including descriptions on why an employee was discharged and quarterly wages paid. MiDAS went live on October 1, 2013. One main function was to detect, investigate, adjudicate, and issue determinations in unemployment insurance matters, including fraud cases, and then to assess and collect penalties. Efficiency was to be achieved by automating certain stages in the process. Defendant CSG recommended the automaton to “reverse the Agency culture where staff is reluctant to render a determination of fraud because it is viewed as too punitive on the claimant.” Defendant FAST designed the platform to search for discrepancies in the records of unemployment compensation recipients, automatically determine whether the claimants committed fraud, and execute collection proceedings, which included intercepting tax refunds and garnishing wages. MiDAS was intended to “increase timeliness and quality” of fraud determinations by using “system-assisted” adjudication. The parties use the term “auto- adjudication,” although that nomenclature is not used consistently in the briefs. The plaintiffs allege that FAST programmed MiDAS with an inherent bias to find fraud. Auto-adjudication is a process. This system starts with the automated generation of a flag, then leads to the automated generation of questionnaires, then to an automated determination based on logic trees, followed by an automated generation of a notice of fraud determination, then automated collection activity. A human could perform one or more of these stages, except for the generation of the fraud questionnaire. A. Investigating Fraud A finding that a claimant is guilty of fraud requires a determination that the claimant received an overpayment of benefits, which was caused by an intentional and material misrepresentation or omission. Mich. Comp. Laws § 421.54(b). For the first step in the fraud adjudication process, MiDAS used an electronic “cross-matching” mechanism to alert the UIA when income was reported for claimants or when some activity affected a claimant’s eligibility for benefits. MiDAS cross-referenced income received in a fiscal quarter under each claimant’s Social Security number with the corresponding quarterly report of the claimant’s employer. The UIA requested that employers itemize the claimant’s earnings for each of the 13 weeks in a quarter. If an employer did not itemize the claimant’s earnings, MiDAS used an “income- spreading” formula, which prorated the claimant’s quarterly earnings in an equal amount for each of the 13 weeks. The UIA then determined that claimants were guilty of fraud if they received benefits during any of the weeks in the quarter, despite lacking evidence of actual earnings. If MiDAS detected that an employer paid a claimant within a quarter in which the claimant received unemployment benefits, the issue was “flagged” as potentially fraudulent. MiDAS then sent a questionnaire to a claimant (Form 1713). If a claimant had a MiWAM account, the questionnaire would be posted to that account. For claimants who chose the paperless “Go Green” option, MiDAS sent a generic email to the claimant’s last email on file, notifying them to check their MiWAM account for correspondence. The emails did not inform claimants that the correspondence concerned eligibility or potential fraud. If a claimant did not “Go Green,” the questionnaire was mailed to the claimant’s last address on file. The UIA made no effort to ascertain whether the email or physical addresses of former claimants were still valid; the claimants were responsible for keeping their contact information current. Moreover, regular mail was frequently returned to the UIA as undeliverable, but the UIA took no steps to update MiDAS to indicate that the residential addresses were no longer current. MiDAS continued to send correspondence to the same undeliverable addresses. The questionnaire posed two questions: “Did you intentionally provide false information to obtain benefits you were not entitled to receive?”, and “Why do you believe you were entitled to benefits?” The second question included the following multiple choice options: 1. I needed the money 2. I had not received my payment when I reported for benefits 3. I reported the net dollar amount instead of the gross dollar amount paid 4. I did not understand how to report my earnings or separation reason 5. I thought my employer reported my earnings for me 6. Someone else certified (reported) for me 7. Someone else filed my claim for me 8. Other Because MiDAS was set up to review claims from the six preceding years, questionnaires were sent to claimants whose benefits had expired already. The system did not provide any other means of notifying claimants of the questionnaire’s existence. Unless a questionnaire actually was sent to a claimant or that claimant regularly checked her MiWAM account, that claimant received no notice of the alleged fraud flag. B. Fraud Adjudication, Notices, and Collection MiDAS’s automation implemented a series of “logic trees,” which engaged binary if-then paths that led to certain adjudicatory results. A path was determined by a claimant’s response to the questionnaire or by a failure to respond at all. Failure to respond timely to a questionnaire (10 days for “green” claimants and 14 days for those corresponding via paper) resulted in a default determination that the claimant knowingly and intentionally misrepresented or concealed information to receive benefits unlawfully. According to a Department of Labor monitoring report, “Most of the intentional misrepresentation cases reviewed were auto-adjudicated due to the claimant’s failure to return the questionnaire.” ECF No. 433-17. When a claimant did respond to the questionnaire, MiDAS’s programming, based on its logic trees, found intentional fraud whenever a claimant chose answers 1, 6, or 7 to question two on the form, even if she indicated that she did not intentionally provide false information. MiDAS had no programming that permitted it to determine whether a claimant actually intentionally misled the UIA or concealed information. Once a default fraud determination was made, MiDAS automatically issued three notices. First, it issued a primary “Notice of Determination” (Form 1302), which explained why the UIA believed it overpaid (i.e., a claimant quitting a job voluntarily, failure to report wages earned, etc.) but not its reasoning for why it believed the plaintiff’s alleged misrepresentation was intentional. For example, the primary notice issued to plaintiff Patty Jo Cahoo read: Issues and Sections of Michigan Employment Security Act involved: Voluntary Quit and 29(1)(a). You quit your job with RANDSTAND EMPLOYMENT SOLUTIONS LP on January 11, 2013 due to other personal reasons. Your leaving was voluntary and not attributable to the employer. You are disqualified for benefits under the MES Act, Sec. 201(a). Rework begins with week ending January 19, 2013. You will not receive benefits until you satisfy the rework requirements. Rework Requirements: Claimant is disqualified until completion of a $4,344.00 earnings rework requirements which has been satisfied. Calculation of interest and penalty amount is shown later on this form. If you disagree with this determination, refer to “Protest Rights on the reverse side of this form. Cahoo Primary Determination Notice, ECF No. 461-18, PageID.28597-600. MiDAS then automatically generated a second “Notice of Determination” (also designated as Form 1302), which generally informed the claimant that his actions misled or concealed information to obtain benefits. It also announced that benefits were terminated on any active claims. Cahoo’s second notice read: Issues and Sections of Michigan Employment Security Act involved: Misrepresentation and 62(b). Your actions indicate you intentionally misled and/or concealed information to obtain benefits you were not entitled to receive. Benefits will be terminated on any claims active on January 05, 2013 [(the mail date was May 27, 2015)]. You are disqualified for benefits under MES Act, Sec. 62(b). Restitution is due under MES Act, Sec. 62(A). The wages used to establish your claim are cancelled and no further benefits will be paid based on those wages. In addition, you are required to pay the penalty assessed based on this determination under MES Act, Sec. 54(b). If the amount of restitution due is less than $500, the penalty is double the restitution due, except that for a subsequent intentional representation the penalty amount is four times the restitution due. If the amount of restitution due is $500 or more, the penalty is four times the restitution due. Calculation of interest and penalty amount is shown later on this form. If you disagree with this determination, refer to “Protest Rights on the reverse side of this form. Cahoo Second Notice of Determination, ECF No. 399-29, PageID.17636. After that, MiDAS sent a “List of Overpayments,” which demanded payment of actual benefits overpaid as well as a statutory penalty for fraudulent misrepresentation of two to four times that amount, depending on whether the value of improper payments equaled or exceeded $500. It also alerted the claimant that the charges would accrue interest at a rate of one percent per month. The first two notices have the same title and same form numbers, but different contents. Further, the notices did not reference each other and often had different case numbers. The reverse side of the primary determination notice indicated that the claimant was entitled to appeal a determination of fraud to an Administrative Law Judge (ALJ) within 30 days. The notices did not inform claimants about their abilities to file late appeals for good cause. Because of errors notifying claimants, many claimants did not become aware of their assessments until after the appeal deadlines passed. If a claimant did not appeal the determination within 30 days, MiDAS automatically issued a letter (Form 1088) demanding that the claimant pay restitution, a penalty, and accrued interest. The delinquency statements explain that the UIA may collect the charges by intercepting state and federal income tax refunds, garnishing wages, or litigation. If a claimant failed to pay the charges voluntarily, MiDAS issued tax refund intercepts to the IRS and State of Michigan. And if intercepting a claimant’s tax refunds did not satisfy the claimant’s debt, the UIA then garnished her wages. The plaintiffs allege that tax intercept and wage garnishment notices were automatically sent by MiDAS in the same manner as all other correspondence from the UIA, with the same absence of measures to ensure that claimants received actual notice. C. Human Involvement The plaintiffs originally alleged that the UIA hardly ever employed any human review when making these automated determinations between October 2013 and August 2015. The plaintiffs have since abandoned that position but maintain that the human oversight was cursory at best. According to MiDAS’s Implementation Specifications, “[i]ssues may be configured to be auto-adjudicated, have system-led adjudication, or be adjudicated solely based on staff input.” MiDAS Nonmonetary Determinations Implementation Specification, ECF No. 385-1, PageID.16806 UIA employee Kristine Kratz testified that MiDAS adjudicated fraud in “stages,” which were identified as “created,” “opened,” “pending fact finding,” “determination issued,” and “issue closed.” Kratz Dep., ECF No. 399-28, PageID.17628. If the word “batch” appeared next to any stage, it signified that MiDAS had changed the status automatically. A human could perform any of these stages, except for the generation of the fraud questionnaire. Human involvement could only be established if a staff member edited the file, which could be seen in the system. MiDAS could open cases by using the “cross-matching” system discussed above. Staff members also could open a fraud investigation based on a hard copy objection mailed in by an employer. MiDAS or the staff member would then update the case from “opened” to “pending fact finding,” and the claimant’s name would appear next to that stage on MiDAS’s “Nonmonetary issue” screen. MiDAS then automatically generated the fraud questionnaires, which the system processed for those who chose to correspond online. For claimants who chose mail correspondence, staff members could click on a screen to indicate that the UIA never received a response or they could enter the response manually. In both instances, staff members or MiDAS would update the case from “pending fact finding” to “pending adjudication,” and their name would appear next to that stage. MiDAS then would make an automated fraud determination, usually at night, based on the UIA’s logic trees. Staff members could also make the fraud determinations based on the same logic trees. In either instance, the case would be staged from “pending adjudication” to “determination issued” to “issue closed,” and the name of whoever prompted the decision would appear beside the stages on the “Nonmonetary Issue” screen. MiDAS then would automatically issue a notice of determination. D. The End of Auto-Adjudication In April 2015, a putative class of plaintiffs sued state officials in their official capacities alleging constitutional violations relating to UIA’s fraud adjudication practices. See Zynda v. Arwood, 175 F. Supp. 3d 791 (E.D. Mich. 2016). In August 2015, the UIA turned off MiDAS’s auto-adjudication processes for claimant fraud. On February 2, 2017, the Zynda parties entered into a settlement agreement in which the UIA agreed to review determinations of claimant fraud that had been reached while it was using the auto-adjudication functionality. The UIA decided to review all fraud determinations from October 2013 through August 2015, even though it estimated that only about 14.8% of those determinations potentially resulted from what the Zynda plaintiffs referred to as “auto- adjudication.” The examiners during the Zynda review overturned all fraud adjudications that were based on a claimant’s failure to respond to fact finding. E. Governmental Review 1. 2015 DOL Monitoring Report and 2016 Follow Up Letter On November 13, 2015, the United States Department of Labor issued a Monitoring Report of the UIA’s adjudication practices. It found that the State’s practices violated section 303 of the Social Security Act, 42 U.S.C. § 503(a), in six ways, five of which are relevant to this case: (1) the UIA determined willful misrepresentation without staff intervention to assess credibility and intent, often based on a claimant’s failure to respond, (2) overpayments were sometimes established based on estimated wages (via prorated “income spreading”), (3) fraud questionnaires provided limited choices for claimants to explain themselves, (4) questionnaires and fraud determination notices did not clearly state the issue or reason for the UIA’s suspicion, and (5) the UIA did not monitor the system for undeliverable email and did not systematically check on whether a claimant’s account was opened. 2015 DOL Monitoring Report, ECF No. 461-8. 2. Auditor General July 2015 Memorandum and April 2016 Audit Report The Michigan Office of the Auditor General also reviewed the UIA’s MiDAS procedures. In July 2015, it issued a performance memorandum in which it found that the “UIA did not effectively and efficiently process claimant and employer mail that was returned undeliverable.” 2015 Auditor General Memorandum, ECF No. 461-13, PageID.28466. The next year, the Auditor General issued a performance audit in April 2016 in which it found that the UIA “needs to improve its efforts to obtain and/or consider supporting information and provide claimants with the facts and rationale when it determines that claimants provided false or misleading information” to “help claimants better understand the allegations against them to make informed decisions on their next course of action.” Auditor General 2016 Performance Audit, ECF No. 461-13, PageID.28472. The Auditor General considered this a “material condition,” — the highest category of importance. Id. at PageID.28516. It also found a “reportable condition” (of lesser importance) that the “UIA did not effectively and efficiently process claimant and employer mail that was returned undeliverable and without a forwarding address.” Id. at 28472. F. The Plaintiffs 1. Patti Jo Cahoo Patti Jo Cahoo collected unemployment benefits from January 12 through June 29, 2013. She chose to go paperless, directing the UIA to provide her with email alerts. Shortly after Cahoo went “green,” her former employer protested her eligibility, alleging that she voluntarily quit. On May 2, 2014, a UIA examiner, “robinson12,” reviewed Cahoo’s file and opened a fraud case. That caused the MiDAS system automatically to post a fraud questionnaire on Cahoo’s MiWAM account and send her an email notifying her to check her account. Cahoo alleges that she was unaware of the posting and did not respond. More than one year later, the UIA determined that Cahoo committed fraud based on her failure to respond, and MiDAS issued two notices of determination, both dated May 27, 2015. Her Primary Notice of Determination explained that she “quit [her] job with Randstad Employment Solutions on January 11, 2013 due to other personal reasons,” and that her “leaving was voluntary and not attributable to the employer.” Her Secondary Notice of Determination accused her of “intentionally misle[adding] and/or conceal[ing] information to obtain benefits [she] was not entitled to receive.” Attached to the secondary notice was a list of overpayments charging Cahoo $7,095.00 for overpayment and $28,380.00 as a penalty. Cahoo alleges that she was unaware of the fraud determination and did not hear about the determination until she filed for bankruptcy “months later.” Her deposition testimony contradicts that position; she testified that she may have had at least some notice. She also conceded that the fraud questionnaire and notices of determination were posted on her MiWAM account, which the record reflects someone checked 22 times between the time the questionnaire was sent in 2014 and when the fraud determinations were issued in May 2015. She also admitted to deleting regularly UIA notifications sent to her email address and neglecting to read UIA letters posted on her MIWAM account. On September 22, 2016, Cahoo filed for Chapter 7 bankruptcy; she received a discharge on January 15, 2017. She did not list her claim in this case in her bankruptcy petition, but she included her UIA debts in her schedule of liabilities. On January 4, 2017, the UIA filed an adversary complaint, opposing the discharge of the UIA debt based on Cahoo’s alleged fraud. But around February 3, 2017, Cahoo received a redetermination of her fraud case under Zynda, which overturned her fraud determination. The parties stipulated to dismiss the adversary proceeding on February 16, 2017. The UIA’s designated witness, Kevin Smith, reviewed Cahoo’s file and opined that UIA examiners were involved in Cahoo’s overpayment determination and that a UIA claims examiner took the actions that led to the determination. However, he “didn’t see [the] staging screen” for Cahoo’s intentional misrepresentation case, which indicated that some stages proceeded by “batch.” 2. Kristen Mendyk Kristen Mendyk received unemployment benefits from July 2, 2011 to March 31, 2012. She chose to receive correspondence from the UIA by mail. Based on an employer protest, the UIA investigated her for fraud in 2012 and again in 2013. The UIA sent her two fraud questionnaires, one on July 20, 2012 (pre-MiDAS) and the other on November 4, 2013. The UIA sent the questionnaires to an address Mendyk provided in Saint Charles, Michigan, where her ex- husband lived. However, she testified that she moved out in 2011, briefly returned from March to June 2014, then left again after she could not reconcile with him. Mendyk alleges that she never received the questionnaires and therefore never responded. In November 2013 and January 2014, the UIA adjudicated three claims of fraud against her: one dated November 1, opened by “wisemans 1” and adjudicated by “batch”; one dated November 8, 2013; and another adjudication on January 14, 2014. The fraud determinations also went to the Saint Charles, Michigan address where Mendyk no longer resided. On February 25, 2014, the UIA sent a monthly delinquency statement charging Mendyk $9,355.00 for overpayment, penalty charges of $22,702.00, and $691.15 for interest. The UIA intercepted Mendyk’s federal tax refund and garnished her wages between May 2014 and July 2016, collecting a total of $5,983.71 from her. Mendyk alleges that she did not learn about the fraud determinations until she attempted to file another claim in 2017. She acknowledged, however, that she had to keep her address current with the UIA. Moreover, her claim files reflect that she began checking her MiWAM account regularly since January 10, 2014. The UIA eventually reconsidered Mendyk’s determination in response to a DOL review of MiDAS auto-adjudications. On October 12, 2016, a UIA employee mailed her a redetermination notice, but it was returned as undeliverable. Between December 30, 2014 and October 12, 2016, several letters that the UIA sent to Mendyk were returned and marked “return to sender,” or “not deliverable as addressed,” including a 2014 letter attempting to collect $30,080 and the 2016 redetermination notice. Mendyk filed for Chapter 13 bankruptcy on December 16, 2016; the UIA filed an adversary proceeding contesting dischargeability. Although Mendyk included her UIA debts on her schedule of liabilities, she did not list the cause of action she asserts in this case. She was represented by counsel in her bankruptcy and discussed this action with him. On February 28, 2017, Mendyk informed her Chapter 13 trustee of her potential due process claim, which the trustee authorized her to pursue. She later entered into a consent judgment with the UIA on April 12, 2017, in which she admitted to the alleged fraud and agreed to pay $6,793.70. She received a discharge on July 24, 2018. The UIA’s designated witness, Katherine Potter, reviewed Mendyk’s file and opined that a UIA examiner made Mendyk’s fraud determinations. The file indicated that stages opened, pending factfinding, and pending adjudication were made by “wisemans 1;” and pending adjudication, redetermination issued, and case closed by “batch.” 3. Khadija Cole Khadija Cole received unemployment benefits from March 1 through June 28, 2014 and opted for email correspondence with the UIA. After suspecting that she failed to disclose vacation pay for the week of March 1, 2014, despite repeatedly certifying that she was not receiving such payments, the UIA posted a fraud questionnaire to her account on October 3, 2014. The UIA also detected earning discrepancies for the weeks of April 5 to June 28, 2014, and on February 12, 2015, MiDAS issued another questionnaire inquiring about unreported wages. On both occasions, MiDAS generated an email instructing Cole to check her MiWAM account. Cole never responded to the questionnaires, and on October 16, 2014 and February 26, 2015, the UIA determined that Cole committed fraud. MiDAS issued corresponding redetermination notices to her MiWAM account. Cole insists that she never received any emails, fraud questionnaires, fraud determinations, or notices of the fraud charges against her whatsoever. Cole testified that she first learned about the fraud determinations when she received a statement of debt by mail from the UIA around the summer of 2016. However, the record reflects that someone frequently logged into her MiWAM account, including before the UIA sent the first fraud questionnaire in October 2014, and after it issued the second determination notice in February 2015. The UIA ultimately assessed $4,950 for overpayment and $19,346 as a penalty. Cole’s claim file reveals extensive adjudication activity by “batch.” The first case concerned vacation pay during one week in March 2014 (received, created, opened, and pending fact finding by “puskarm”; pending fact finding, pending adjudication, redetermination issued, and issue closed by “batch.”). Cole alleges that the UIA improperly flagged the second case based on MiDAS’s “income spreading” programming. As evidence, she points to the UIA’s “Notice of Redetermination” issued on February 27, 2015, which found that she received exactly $733.45 in wages every week between April 5 and June 28, 2014. MiDAS also issued a list of overpayments, stating that it overpaid exactly $362.00 each week between April 5 and June 28, 2014, and charged her $4,706 for overpayment and $18,824.00 as a penalty. On October 3, 2016, Cole filed for Chapter 13 bankruptcy. In her petition, she included her UIA debts on her schedule of liabilities, but she did not list her cause of action in this case as an asset. She was represented by counsel during her bankruptcy proceedings and discussed this lawsuit with him. While her bankruptcy was pending, Cole admitted to committing fraud and entered into a consent judgment with the UIA on March 29, 2017. Her bankruptcy proceedings are ongoing. 4. Michelle Davison Michelle Davison received unemployment benefits from May 25 through November 23, 2013, and from March 22 through July 19, 2014. She applied for her second round of unemployment benefits on March 17, 2014, contending that her employer laid her off. She opted for regular mail correspondence and directed the UIA to contact her through an address on Leerda Street in Flint, Michigan. Davison’s former employer protested her eligibility around March 2014, alleging that she voluntarily quit. A UIA employee opened an investigation on October 7, 2014, and MiDAS sent a fraud questionnaire to Davison’s designated address the following day. She never responded. Davison testified that she likely was no longer living at that address, had never received the questionnaire, and was unable to respond. Her recollection, though, was quite foggy; she testified that she could not remember exactly when she moved from the listed address. On October 22, 2014, the UIA mailed a notice of determination to the Flint address stating that she was ineligible for benefits. The notice was returned as undeliverable on November 11, 2014. On December 29, 2014, Davison called the UIA to unlock her MiWAM account, and during that call, she updated her address to reflect that she was residing at Resource Genesee, a homeless shelter on Saginaw Street in Flint, Michigan. The UIA never resent the fraud questionnaire, nor did it do anything to verify that Davison received it. On January 15, 2015, the UIA issued a fraud determination to Davison’s updated address and assessed $6,222 for overpayment and $740 as a penalty. Davison alleges that she never received any correspondence and first learned about the fraud determination from the IRS when the UIA intercepted her 2015 federal tax refund. However, she admitted that she did not check her mail consistently, and the record is devoid of evidence that the 2015 fraud determination was ever returned to the UIA as undeliverable. Although she opted for mail correspondence, Davison’s claim files indicate that she frequently checked her MiWAM account in 2014 and 2015. She eventually filed a late appeal of her fraud determination after her Zynda review and prevailed on May 8, 2017. The UIA’s designated witness, Jessica Hart, reviewed Davison’s file and opined that a UIA examiner made Davison’s fraud determinations. “Batch” appears beside only one entry in the relevant time period, which changed the stage of Davison’s case from “pending fact finding to pending adjudication.” The names of UIA employees appear beside every other entry. Additionally, claims examiner Chris Corey swore in an affidavit that he adjudicated Davison’s case “based upon the information [he] would have had at the time the determination was issued,” and that it was approved by his manager. 5. Hyon Pak Hyon Pak received unemployment benefits from September 24, 2011 through March 3, 2012. Before the UIA implemented MiDAS, it opened a fraud investigation and sent Pak three pre-MiDAS questionnaires by mail on September 24, 2012. The first two concerned employer Express Services, and the third addressed employer Eagle Industries. He responded to all three on September 26, 2012. On October 8, 2012, Pak received two more pre-MiDAS fraud questionnaires, one regarding Express Services, the other regarding Eagle Industries. He did not respond to them, apparently believing they were duplicates. About a year later (and after MiDAS rolled out), Pak chose to “go green” and provided the UIA with his email address. About one year after “going green,” on December 29, 2014, the UIA posted four fraud determinations on Pak’s MiWAM account, all of which were adjudicated by MiDAS. The System had gone back over two years to review Pak’s September 26, 2012 responses to the UIA’s questionnaires. However, a staff member had to enter his responses into MiDAS and open the fraud investigation because the old questionnaires prompted handwritten responses. About four hours after the employee opened Pak’s cases, “batch” changed the stages from “pending adjudication” to “issue closed.” The UIA issued determination notices that day, December 29, 2014. UIA employee Mandy Brickel, remarkably, testified that she saw no evidence of auto-adjudication on Pak’s case, and Pamela Sagady swore that she adjudicated Pak’s claims “based upon the information [she] would have had at the time . . . , including the fact that he had a prior 2012 determination of fraud.” She explained that “Pak did not respond to the factfinding he was sent; had he responded, I would have reviewed and considered his responses in adjudicating the issues in his case.” Pak alleges that he was unaware of the fraud determinations until 2016, when the UIA intercepted his federal tax refund. But during his deposition, he conceded that the UIA posted the correspondence on his MiWAM account, and that someone (possibly his wife) accessed his account on December 29, 2014 and January 2, 2015 — right after the UIA emailed him to check his account. He filed an untimely appeal in 2017, but the UIA eventually reversed Pak’s fraud determinations after it permitted him to appeal again. II. Discussion The plaintiffs’ remaining claim in this case is that the defendants deprived them of their property without due process when MiDAS determined that they had committed fraud when applying for and receiving unemployment compensation benefits. They seek to recover the consequential damages that flowed from those determinations; they already have received equitable relief in the form of re-adjudications of the fraud accusations as a result of the Zynda settlement. It appears that during the time it was operational, MiDAS was involved in some way in adjudicating the fraud cases of 67,740 individuals, according to an excel spreadsheet that the State of Michigan produced. The plaintiffs seek to sweep all of those people into a class under Federal Rule of Civil Procedure 23(b)(3) now defined as: All persons whose fraud determinations were made using the MiDAS system in any way from October 1, 2013 until August 7, 2015. Alternatively, the plaintiffs seek to certify the following four subclasses: a. All claimants who received a fraud determination and received fraud questionnaire Form 1713. b. All claimants who received a fraud determination and received fraud determination Form 1302. c. All claimants who received an automated default determination of fraud based on prorating earnings weekly over a quarterly earning period. d. All claimants who received an automated default determination of fraud based on selecting multiple choice answers 1, 6, or 7 to the fact finding questionnaire. A class action is an “exception to the usual rule that litigation is conducted by and on behalf of the individual named parties only”; it “provides a procedure by which the court may exercise . . . jurisdiction over the various individual claims in a single proceeding.” Califano v. Yamasaki, 442 U.S. 682, 700-01 (1979). “[T]o justify a departure from that rule, ‘a class representative must be part of the class and “possess the same interest and suffer the same injury” as the class members.’” Dukes, 564 U.S. at 348-49 (quoting East Tex. Motor Freight Sys., Inc. v. Rodriguez, 431 U.S. 395, 403 (1977). To ensure that is the case, a party seeking to certify a class must satisfy Rule 23(a)’s requirement of numerosity, commonality, typicality, and adequacy of representation. Clemons v. Norton Healthcare Inc. Ret. Plan, 890 F.3d 254, 278 (6th Cir. 2018). And the party must demonstrate that the class action fits within at least one of the three categories identified in Rule 23(b). Ibid. A. Numerosity Numerosity is shown when the number of potential plaintiffs is too large to make joinder practicable but not so large as to make administration impossible. Senter v. Gen. Motors Corp., 532 F.2d 511, 523 (6th Cir. 1976). There is no strict numerical test to determine when the class is large enough or its members too numerous to be joined under the Federal Rules of Civil Procedure. Id. at 523 n.24. “When class size reaches substantial proportions, however, the numerosity requirement is usually satisfied by the numbers alone.” Ross v. Abercrombie & Fitch Co., 257 F.R.D. 435, 442 (S.D. Ohio 2009). A class of 67,740 individuals certainly is large enough to satisfy this requirement. Some of the defendants, however, contend that the list that yielded that number is misleading because it includes both staff-adjudicated and auto-adjudicated fraud determinations. It may be true that some of the individuals’ claims were adjudicated with human input. But that is irrelevant, since the plaintiffs made clear that when they say that cases were “auto-adjudicated,” they mean they were subject to the automated process of the MiDAS system regardless of human involvement. And there apparently is no dispute that all the people named in the list furnished by the State were subjected to MiDAS’s decision trees in some way. The numerosity requirement is satisfied. B. Commonality Commonality means that “there are questions of law or fact common to the class.” Fed. R. Civ. P. 23(a)(2). Not all questions of law and fact raised in the complaint need be common. Olden v. LaFarge Corp., 203 F.R.D. 254, 269 (E.D. Mich. 2001), aff’d, 383 F.3d 495 (6th Cir. 2004). But “[c]ommonality requires the plaintiff to demonstrate that the class members ‘have suffered the same injury.’” Dukes, 564 U.S. at 349-50 (quoting Gen. Tel. Co. of Sw. v. Falcon, 457 U.S. 147, 157 (1982)). Here, the plaintiffs all contend that they suffered a violation of their right to due process of law when they were adjudicated as having committed fraud and their property was taken from them. And federal law requires that claimants be provided an “[o]pportunity for a fair hearing, before an impartial tribunal, for all individuals whose claims for unemployment compensation are denied.” 42 U.S.C. § 503(a)(3). To succeed on their procedural due process claims, the plaintiffs must prove “three elements: (1) that they have a property interest protected by the Due Process Clause; (2) that they were deprived of this property interest [by a state actor]; and (3) that the state did not afford them adequate pre-deprivation procedural rights.” Cahoo v. SAS Analytics Inc., 912 F.3d 887, 900 (6th Cir. 2019) (citing Hahn v. Star Bank, 190 F.3d 708, 716 (6th Cir. 1999)). It is not hard to find commonality among all putative class members on the first two elements. Each had a property interest in receiving unemployment compensation benefits, and each apparently was deprived of some measure of that benefit when MiDAS labeled them a fraudster. It is the third element that is the crux of this lawsuit, and it is the one that is the most problematic. A fundamental aspect of adequate pre-deprivation procedural rights that the Due Process Clause demands is notice and the opportunity to be heard. Goldberg v. Kelly, 397 U.S. 254, 267 (1970). The notice must be “reasonably calculated, under all of the circumstances, to apprise interested parties of the pendency of the action and afford them an opportunity to present their objections,” and “must afford a reasonable time for those interested to make their appearance.” Mullane v. Central Hanover Bank and Trust Co., 339 U.S. 306, 314 (1950). Whatever device the local government chooses for giving notice, it must after all be “through means that ‘one desirous of actually informing the absentee might reasonably adopt.’” Ming Kuo Yang v. City of Wyoming, Michigan, 793 F.3d 599, 602 (6th Cir. 2015) (quoting Mullane, 339 U.S. at 315). And when the chosen method fails, “it must take ‘additional reasonable steps’ to notify the interested party.” Id. at 603 (citing Jones v. Flowers, 547 U.S. 220, 234 (2006)). The record in this case contains strong evidence that MiDAS’s built-in procedures failed to satisfy rudimentary due process requirements. For instance, the MiDAS-issued fraud questionnaires were not reasonably calculated to apprise claimants of the fraud charge against them because they did not provide adequate notice of the plaintiffs’ alleged misconduct and prevented them from intelligently objecting to the possibility of losing their benefits and being accused of fraud. The same fault can be found with the fraud determination notices. Although the primary determination notices (re: overpayment) provided a basic explanation, the secondary determination notice (re: fraud) did not explain why the UIA believed that claimants’ representations were knowingly or intentionally false. Moreover, the notices did not provide an opportunity for pre- deprivation review. Similarly, MiDAS’s manner of notifying claimants of the proceedings against them by email was not reasonably calculated to apprise claimants of the pendency of the action and afford them an opportunity to present their objections. The emails did not include copies of the UIA correspondence, did not inform claimants that the correspondence concerned benefits eligibility or potential fraud, and did not include any warnings or any sense of urgency. The UIA’s use of “auto-adjudication,” that is, determining fraud (which required a finding of intentionality) by the rote application of MiDAS’s decision trees, likely deprived some claimants of the opportunity for a fair hearing before an impartial decisionmaker. That fault was particularly apparent when automated fraud adjudication was based on a failure to respond. Likewise, MiDAS’s use of income-spreading deprived some claimants of the opportunity for a fair hearing before an impartial decisionmaker. Although prorating earnings is not inherently problematic when used to investigate overpayment, a fraud adjudication based on nothing but a conflict with prorated earnings is neither fair nor impartial. To satisfy the commonality requirement, the claims “must depend on a common contention . . . of such a nature that is capable of classwide resolution — which means that determination of its truth or its falsity will resolve an issue that is central to the validity of each one of the claims in one stroke.” Dukes, 564 U.S. at 349-50. “It is not every common question that will suffice, however; at a sufficiently abstract level of generalization, almost any set of claims can be said to display commonality. What we are looking for is a common issue the resolution of which will advance the litigation.” Sprague v. Gen. Motors Corp., 133 F.3d 388, 397 (6th Cir. 1998) (en banc). Not all of MiDAS’s systemic faults affected each of the putative class members in the same way. For some, the notices were confusing. Others did not even receive a notice (and therefore could not have been confused by them). Some putative class members suffered fraud adjudications for failing to respond, some out of neglect and others due to not receiving notification that the questionnaires had been sent or because they did not receive them at all. Some, but not all, were victimized by the income-spreading formula that may have attributed earnings incorrectly to weeks when no work was performed. And the level of human input into the decision-making process varied from case to case. Whether the UI furnished adequate process through MiDAS at all of these stages is necessary to determine liability for various class members. But there is no single answer to these questions that can advance the litigation for the class as a whole within a level of specificity that is acceptable. And an overriding question as to all of the putative class members is whether they received actual notice of the charges and proceedings. See Krueger v. City of Eastpointe, 452 F. Supp. 3d 679, 690 (E.D. Mich. 2020) (“If an individual did not receive procedurally adequate notice but was notified in fact that the government intended to deprive the individual of property or a liberty interest, no claim under procedural due process will lie.”). That determination requires claimant-specific answers. And all of this undercuts a finding of commonality. Nonetheless, there is one area that fits with Dukes’s rubric: the accountability of the separate defendants for MiDAS’s systemic flaws and their impact on the class members. Although the Supreme Court in Dukes “elevate[d] the [Rule 23](a)(2) inquiry so that it is no longer ‘easily satisfied,” Dukes, 546 U.S. at 375 (Ginsburg, J., concurring in part), the plaintiffs have met the commonality requirement on that question. A decision on the defendants’ responsibility for creating and launching a fraud adjudication system that foreseeably would deny claimants notice and a proper determination of intent-based liability without an opportunity to be heard in their defense is “central to the validity of each one of the claims” and could be decided “in one stroke.” Dukes, 564 U.S. at 349-50. Deciding that question in the plaintiffs’ favor would not end the liability inquiry for each member of the class, for the reasons discussed above. But for the purposes of class certification, Rule 23(a) does not require a single answer to the ultimate question of liability; it simply requires “a common issue the resolution of which will advance the litigation.” Sprague, 133 F.3d at 397. The plaintiffs met the commonality requirement. C. Typicality The plaintiffs also must show that their claims, as “representative parties[,] are typical of the claims . . . of the class.” Fed. R. Civ. P. 23(a)(3). The named plaintiffs’ claims are typical if they “arise[] from the same event or practice or course of conduct that gives rise to the claims of other class members, and if [their] claims are based on the same legal theory.” In re Am. Med. Sys., Inc., 75 F.3d 1069, 1082 (6th Cir. 1996) (internal quotation marks omitted); see also Sprague, 133 F.3d at 399 (stating that the essence of typicality boils down to the notion that “[a]s goes the claim of the named plaintiff, so go the claims of the class”). Typicality requires that a “‘sufficient relationship exist[ ] between the injury to the named plaintiff and conduct affecting the class, so that the court may properly attribute a collective nature to the challenged conduct.’” Stout v. Byrider, 228 F.3d 709, 717 (6th Cir. 2000) (quoting Sprague, 133 F.3d at 399). It has been said that the commonality and typicality requirements “‘tend to merge.’” Ball v. Union Carbide Corp., 385 F.3d 713, 728 (6th Cir. 2004) (quoting Rutherford v. City of Cleveland, 137 F.3d 905, 909 (6th Cir. 1998)). Together they “serve as guideposts for determining whether under the particular circumstances maintenance of a class action is economical and whether the named plaintiffs’ claim and the class claims are so interrelated that the interests of the class members will be fairly and adequately protected in their absence.” Rutherford, 137 F.3d at 909. It is here that the plaintiffs stumble. “[C]lass certification is inappropriate where a putative class representative is subject to unique defenses which threaten to become the focus of the litigation.” Baffa v. Donaldson, Lufkin & Jenrette Secs. Corp., 222 F.3d 52, 59 (2d Cir. 2000) (internal quotation marks and citation omitted). Patti Jo Cahoo, Khadija Cole, and Kristen Mendyk each filed for bankruptcy before filing this action, and the defendants argue forcefully that they either are not the real parties in interest or they are judicially estopped from bringing their claims because they failed to disclose their property interests in this litigation to their bankruptcy trustees. The plaintiffs insist that this issue is a red herring and does not destroy typicality. But “unique defenses will destroy typicality . . . ‘where the defenses against the named representatives are likely to usurp a significant portion of the litigant’s time and energy, and there is a danger that the absent class members will suffer if their representative is preoccupied with defenses unique to it.’” Willis v. Big Lots, Inc., 242 F. Supp. 3d 634, 646 (S. D. Ohio 2017) (quoting Bentley v. Honeywell Int'l, Inc., 223 F.R.D. 471, 484 (S.D. Ohio 2004)). The real-party-in-interest and judicial estoppel defenses already have dominated the merits motion practice in this case. They have taken center stage in a later-filed motion to dismiss raising a standing question, and they are one of the main focuses of the summary judgment briefing, not to mention the present motion. The issue is not trivial and will continue to distract the three plaintiffs from the merits of their due process claims. At first blush, Michelle Davison and Hyon Pak appear to have claims typical of many members of the class: that MiDAS made fraud determinations based only on the failure to respond to the questionnaires due to the rote application of the logic trees the defendants built into the system. However, their individual circumstances vary significantly. Davison never received the fraud questionnaire because she moved to a different address than the one the UIA had on file. Even though she subsequently updated her address, MiDAS never re-issued the fraud questionnaire to her. Pak, on the other hand, received three fraud questionnaires in 2012 (before MiDAS became operational and before the UIA updated the questionnaires). He replied to those questionnaires but failed to reply to follow-up questionnaires, mistakenly believing they were duplicates. Pak eventually opted to “go green,” and the UIA posted four MiDAS fraud determinations on Pak’s MiWAM account two years later, in December 2014. However, a staff member entered his 2012 responses into MiDAS and opened the fraud issues because the old questionnaires called for handwritten responses. Pak alleged that he was unaware of the fraud determinations until 2016, when the UIA intercepted his federal tax refund. But during his deposition, he conceded that the UIA sent notices online as he requested, and that someone (possibly his wife) accessed his MiWAM account on the same day the UIA emailed him to check his account. The plaintiffs take great pains to show that the crux of their class action is both the substance and manner of notice. But Davison and Pak offer nothing concerning the substance of the notice provided by the UIA, as they contend that they never actually received the MiDAS- generated questionnaires or fraud determinations. That leaves the question of the sufficiency of delivering notice, but Davison and Pak have divergent circumstances about the manner in which they were notified about their fraud adjudications, and Pak faces a significant challenge proving that he was not actually aware of the fraud determinations. And the use of subclasses does not fix the adequacy problem. Sprague v. General Motors Corporation explains why. In that case, the plaintiffs brought contract and estoppel claims on behalf of a class of early retirees who alleged that GM promised them free medical coverage in retirement. Sprague, 133 F.3d at 395. The en banc court held that the plaintiffs fell short of establishing typicality because “[s]ome class members may have signed the same form, some may have received the same documents, or some may have attended the same meetings about the early retirement program, but taken as a whole the class claims were based on widely divergent facts.” Id. at 399. The plaintiffs’ solution to those differing circumstances was to propose four subclasses that depended on which forms (if any) class members had signed. Id. at 397. But the subclasses “did not solve the problem” as they “are not a substitute for compliance with Rule 23.” Id. at 399. n.9; see Rombiero v. Unumprovident Corp., 385 F. App’x 423, 432 (finding no typicality because even though “all of the plaintiffs might have been subjected to some or all of [the defendant’s] alleged wrongful practices” that “does not eliminate need for individualized assessment as to the ultimate propriety of benefits decisions affecting each and every class member”). The named plaintiffs all attack various aspects of the MiDAS adjudication processes and seek to hold the defendants accountable for their respective roles in designing and implementing the system. But each appears to have been harmed by different flaws in the system, which may have had due process violations built in at several stages of its fraud detection and adjudication procedure. Many members of the putative class may share their respective complaints. But the named plaintiffs’ claims are not typical of the class as a whole, as they have defined it, or any of the proposed subclasses. D. Adequacy of Representation This finding dovetails into the fourth prerequisite for class certification: that “the representative parties will fairly and adequately protect the interests of the class.” Fed. R. Civ. P. 23(a)(4); see also In re Am. Med. Sys., 75 F.3d at 1083. “Adequate representation” invokes two inquiries: (1) whether the class counsel are “qualified, experienced and generally able to conduct the litigation” and (2) whether the named plaintiffs have interests that are “antagonistic” to the other class members. Stout, 228 F.3d at 717. “Interests are antagonistic when there is evidence that the representative plaintiffs appear unable to vigorously prosecute the interests of the class.” Id. at 717. As the Sixth Circuit has noted, “[t]he adequate representation requirement overlaps with the typicality requirement because in the absence of typical claims, the class representative has no incentives to pursue the claims of the other class members.” In re Am. Med. Sys., 75 F.3d at 1083. There is no dispute on the first point. None of the defendants question the qualifications of plaintiffs’ counsel or argue that they cannot vigorously prosecute the interests of the class. Class counsel are experienced litigators who have done extensive work in identifying, investigating, and litigating the potential claims in this action. However, because the claims of the named plaintiffs are not typical of the proposed class as a whole, there is good reason to question whether they will “adequately protect the interests of the class.” Fed. R. Civ. P. 23(a)(4). Each of these plaintiffs will try to make his or her case by exposing a particular flaw in a system with many faults. That effort would not necessarily conflict with the claims of absent class members who were harmed by other aspects of MiDAS that trenched upon due process rights. But neither would it promote (or protect) those interests. And none of these plaintiffs has been harmed by the procedures identified in three of the four proposed subclasses. In addition, the ongoing issues regarding Cahoo’s, Cole’s, and Mendyk’s bankruptcy filings render them inadequate class representatives for the same reason it renders their claims atypical. That remains true, even if their respective bankruptcy trustee could take over their claim. See Griffith v. Javitch, Block & Rathbone, LLP, 358 B.R. 338, 341-42 (S.D. Ohio 2007) (“except in unusual circumstances that do not exist here, a Chapter 7 trustee is not an adequate representative for a class of non-debtors, even if the class includes the debtor.”); Dechert v. Cadle Co., 333 F.3d 801 (7th Cir. 2003) (vacating district court’s class certification order appointing a Chapter 7 trustee as the class representative in a Fair Debt Collection Practices Action.). E. Predominance and Superiority “Rule 23(b)(3) classes . . . must meet predominance and superiority requirements, that is, ‘questions of law or fact common to class members [must] predominate over any questions affecting only individual members’ and class treatment must be ‘superior to other available methods.’” Sandusky Wellness Ctr., LLC v. ASD Specialty Healthcare, Inc., 863 F.3d 460, 466 (6th Cir. 2017). A proposed class generally will fail the predominance test when “‘the legal or factual questions that qualify each class member’s case as a genuine controversy,’” are not “‘subject to generalized proof, and thus applicable to the class as a whole.’” Id. at 468 (quoting Amchem Prods., Inc. v. Windsor, 521 U.S. 591, 623 (1997), and Bridging Cmtys., Inc. v. Top Flite Fin. Inc., 843 F.3d 1119, 1124 (6th Cir. 2016)). The test is “more demanding” than the commonality determination. Comcast v. Behrend, 569 U.S. 27, 34 (2013). The analysis “begins, of course, with the elements of the underlying cause of action.” Erica P. John Fund, Inc. v. Halliburton Co., 563 U.S. 804, 809 (2011). But beyond simply “advance[ing] the litigation,” Sprague, 133 F.3d at 397, the “key [to finding preponderance] is to identify[] the substantive issues that will control the outcome” of the case. Sandusky Wellness Ctr., 863 F.3d at 468 (citation omitted, alteration in original). “[C]ourts should ‘consider how a trial on the merits would be conducted if a class were certified.’” Id. (quoting Gene & Gene, LLC v. BioPay, LLC, 541 F.3d 318, 326 (5th Cir. 2008)). The defendants insist that individualized questions and circumstances permeate this action. In addition to the likelihood that each class member’s claim will focus on a discrete flaw in MiDAS, the defendants point out that each individual also may be subject to critical defenses, like actual notice of proceedings. See Krueger, 452 F. Supp. 3d at 690 (“If an individual did not receive procedurally adequate notice but was notified in fact that the government intended to deprive the individual of property or a liberty interest, no claim under procedural due process will lie.”); Hroch v. Omaha, 4 F.3d 693, 696 (8th Cir. 1993) (“Because [the plaintiff] had actual notice that the City intended to condemn all buildings on the premises, there was no procedural due process violation”); Roberts v. Girder, 237 F.Supp.3d 548, 555 (E.D. Ky. 2017) (“due process may be satisfied by actual notice of the decision to demolish the property and the consequences for failure to timely challenge the determination”). The determination of whether each claimant received actual notice “control[s] the outcome” of the claim as it relates to the manner of notice. Sandusky Wellness Ctr., 863 F.3d at 468. The plaintiffs argue that the manner of the notice was constitutionally deficient, but they also insist that the substance was deficient as well. Cf. Krueger, 452 F. Supp. 3d at 689 (“Plaintiff does not contest that the Forfeiture Notice meets this standard. It provides a description of the property seized, includes notice that the government intends to proceed with forfeiture, and details the procedures established”). However, the plaintiffs cite no authority suggesting that the deficient substance of notice cannot be cured by actual notice. This case is complex and, like MiDAS itself, involves many moving pieces and individualized considerations that are outcome determinative. Zinermon v. Burch, 494 U.S. 113, 127 (1990) (due process “is a flexible concept that varies with the particular situation.”). The plaintiffs, therefore, have not demonstrated that the common questions presented predominate over the panoply of individualized considerations. “Analyzing superiority entails balancing ‘the desirability’ of class treatment with ‘the likely difficulties in managing a class action.’” Sandusky Wellness Ctr, 863 F.3d at 471 (quoting Fed. R. Civ. P. 23(b)(3)). This case is fraught with individualized outcome-determinative considerations, and the plaintiffs have not identified a single course of wrongful conduct that ties every claim together. That renders the management of the class impracticable. Young v. Nationwide Mut. Ins. Co., 693 F.3d 532, 545 (6th Cir. 2012) (“[w]here many individual inquiries are necessary, a class action is not a superior form of adjudication.”). F. Issue Class The plaintiffs present a fallback argument that if their proposed class or subclasses are not certified, then the case could proceed as an issue class. “When appropriate, an action may be brought or maintained as a class action with respect to particular issues.” Fed. R. Civ. P. 23(c)(4). “Rule 23(c)(4) contemplates using issue certification to retain a case’s class character where common questions predominate within certain issues and where class treatment of those issues is the superior method of resolution.” Martin v. Behr Dayton Thermal Prod. LLC, 896 F.3d 405, 413 (6th Cir. 2018). Under the “broad view” of Rule 23(c)(4) adopted by the Sixth Circuit, “courts apply [Rule 23(b)(3)’s] predominance and superiority prongs after common issues have been identified for class treatment under Rule 23(c)(4). The broad view permits utilizing Rule 23(c)(4) even where predominance has not been satisfied for the cause of action as a whole.” Id. at 411 (citations omitted). The plaintiffs contend that the following common questions that would be suitable for issue certification: “method of notice, content of notice, failure to provide for a meaningful hearing and failure to provide an impartial process.” Mot. Class Cert., ECF No. 278, PageID.8010. But these issues are not “common issues” among the class, that is, “one where the same evidence will suffice for each member to make a prima facie showing or the issue is susceptible to generalized, class- wide proof.” Tyson Foods, Inc. v. Bouaphakeo, --- U.S. ---, 136 S. Ct. 1036, 1045 (2016). “The predominance inquiry asks whether the common, aggregation-enabling, issues in the case are more prevalent or important than the non-common, aggregation-defeating, individual issues.” Tyson Foods, 136 S. Ct. at 1045. As discussed above, the manner of notice and the UIA’s failure to provide a meaningful hearing and impartial process are fact-intensive inquiries that still would fail the preponderance and commonality requirements of Rule 23(a)(2) and (b)(3), as the resolution of those issues would depend on whether the individuals received actual notice. And the certification of the plaintiffs’ proposed issues would not be superior to individualized adjudication. III. Conclusion The plaintiffs have satisfied the numerosity requirement of Rule 23(a), and they have identified some questions common to their proposed class. But they have not satisfied the other requirements of Rule 23(a) or (b)(3). Accordingly, it is ORDERED that the plaintiffs’ motion to certify this case as a class action (ECF No. 278) is DENIED. s/David M. Lawson DAVID M. LAWSON United States District Judge Date: December 22, 2020
Document Info
Docket Number: 2:17-cv-10657
Filed Date: 12/22/2020
Precedential Status: Precedential
Modified Date: 6/22/2024