Proximate Cause and Independent Investigations

At some point, I'll tidy up prior posts in this blog that may or may not be helpful in light of the Supreme Court's decision in Staub v. Proctor Hospital.  There, the Supreme Court rejected the "hard and fast" rule that an "independent investigation" could automatically eliminate alleged bias by a non-decision-maker who provided information that was relied upon in the adverse employment action.  To quote the decision:

As we have already acknowledged, the requirement that the biased supervisor’s action be a causal factor of the ultimate employment action incorporates the traditional tort-law concept of proximate cause. See, e.g., Anza v. Ideal Steel Supply Corp., 547 U. S. 451, 457–458 (2006); Sosa, supra, at 703. Thus, if the employer’s investigation results in an adverse action for reasons unrelated to the supervisor’s original biased action (by the terms of USERRA it is the employer’s burden to establish that),then the employer will not be liable. But the supervisor’s biased report may remain a causal factor if the independent investigation takes it into account without determining that the adverse action was, apart from the supervisor’s recommendation, entirely justified.

But what is meant by "the traditional tort-law concept of proximate cause?"  To be honest, its use here seems out of place as it is a doctrine meant to relieve a remote tortfeasor (the person that is alleged to have caused harm to another) from liability.  Since employers, not biased supervisors, are liable for unlawful discrimination, using proximate cause to affect an employer' liability seems backwards.  The Court was not, we must assume, addressing the "traditional" effect of a proximate cause analysis but using it to illustrate the ultimate point, that the employer needs to show (and in some cases, such as USERRA, prove) that the supervisor's bias was not "a" cause of the employer's decision.

So what is proximate cause? This discussion of proximate cause is from a court of appeals decision (not in the employment discrimination field but a suit by shareholders against a corporations directors alleging a breach of the duty of loyalty) issued today:

The term “proximate cause” is pervasive in American tort law, but that doesn’t mean it’s well understood. A common definition is that there must be proof of “some direct relation between the injury asserted and the injurious conduct alleged.” Hemi Group, LLC v. City of New York, 130 S. Ct. 983, 989 (2010), quoting Holmes v. Securities Investor Protection Corp., 503 U.S. 258, 268 (1992). But “direct” is no more illuminating than “proximate.” Both are metaphors rather than definitions. What the courts are trying to do by intoning these words is to focus attention on whether the particular contribution that the defendant made to the injury for which the plaintiff has sued him resulted from conduct that we want to deter or punish by imposing liability, as in the famous case of Palsgraf v. Long Island R.R., 162 N.E. 99 (N.Y. 1928) (Cardozo, C.J.). The plaintiff was injured when a heavy metal scale collapsed on the railroad platform on which she was standing. The scale had buckled from damage caused by fireworks dropped by a passenger trying, with the aid of a conductor, to board a moving train at some distance from the scale. She sued the railroad; it would have been unthinkable for her to sue the scale’s manufacturer, even though if heavy metal scales did not exist she would not have been injured. No one would think the scale’s manufacturer should be liable, because no one would think that tort law should try to encourage manufacturers of scales to take steps to prevent the kind of accident that befell Mrs. Palsgraf. The railroad was a more plausible defendant; its conductor had tugged the passenger aboard while the train was already moving. But how could he have foreseen that his act would have triggered an explosion, as distinct from a possible injury to the boarder? If an accident is so freakish as to be unforeseeable, liability is unlikely to have a deterrent effect.

Coming closer to our case, the defendants cite our decision in Movitz v. First National Bank of Chicago, 148 F.3d 760 (7th Cir. 1998). The plaintiff had bought a building in Houston in reliance on what he claimed was the defendant’s misrepresentation of its value. Had it not been for the misrepresentation he would not have bought it. Shortly after the purchase the Houston real estate market collapsed and his investment was wiped out. The misrepresentation had not caused that collapse but it had been a cause of the plaintiff’s buying the building and thus had contributed to his loss. Yet we ruled, without using the term “proximate cause,” that he could not recover from the defendant because (among other reasons) that would produce overdeterrence by making the defendant an insurer of conditions that he could not control. Id. at 763. That would be as futile as making the manufacturer of the scale an insurer of Mrs. Palsgraf’s loss.

Proximate cause does not mean "but for."  Each of the examples the court of appeals gave could be "but for" causes of the perils described.  It means too remote

There is a lot to be said about the decision in Staub but for now, employers should continue to investigate allegations of employee misconduct only they should be sure to establish that information from an allegedly  biased (assuming that fact is asserted to begin with) should be independently corroborated so that the supervisor's alleged bias is as remote to the decision as possible.

When Investigations Go Wrong

Just as a housekeeping matter, on July 27, 2010, the Home Depot decision I discussed in this post was vacated by the court of appeals and the lawsuit was dismissed.  (They even took down the reported decision so the link below doesn't work.)  This happens, most likely, when the parties reach a settlement, as one legal journal has reported.  It means that as legal precedent, the decision is no longer binding. Of course, the cost of taking the case all the way to court of appeals before settling it was significant so I hope employers will learn from Home Depot's experience.

I am returning to one of my frequent topics, how an employer can avoid liability by conducting an independent investigation.

I've emphasized that employers need to make sure to conduct a fair investigation. This means, as one post emphasized, that employers need to listen to the employee and form an opinion only after hearing the employee's version of events. It also means, as a minimal standard, looking into how similarly situated employees have acted without being disciplined.

An appeals court decision involving Home Depot managers from Florida and Alabama provides additional fodder for employer investigatory mistakes. Two Home Depot store managers were fired after complaining about sexual harassment by a regional HR Manager. The store managers lost the harassment claim but persuaded the appeals court to reverse the retaliation claim.

Home Depot deserves some criticism for the manner in which it terminated the store managers. They didn't produce a written policy (about personal use of a company cell phone) and did not have very strong evidence that the store managers gave out discounts to vendors beyond what other store managers had done. Worse still, there was some evidence that some of the "illegal" discounts were actually authorized by district managers; the discount policy was not put into evidence as an exhibit.

It is never good when the court's discussion of the investigation starts by saying that it was "riddled with inconsistencies and contradictions." Perhaps no adequate investigation could have rescued Home Depot - but that is an issue for a trial. For employers that want to avoid a jury trial in a termination case, however, the decision points to several problems to avoid in conducting an investigation into employee misconduct.

Timing is everything. Home Depot maintained it fired the store managers because it began investigating improper discounts in "late Spring of 2005." The earliest documentation in the investigatory file, however, was dated late September 2005. Also some of the documents that were purportedly relied to make the termination decisions were dated after the store managers were fired. (Even so, one decision-maker testified she reviewed these documents before she decided to fire the store managers.)

Conduct an even-handed investigation. Home Depot maintained that the investigation into improper discounts was district wide. Yet, the investigator's notes about the two store managers who were fired was "far more detailed" than for any other store manager. The investigator maintained he was "never particularly focused" on the two fired store managers. No doubt, it is natural to see an increase in detail when the violation might lead to firing an employee. The important point here is that in any investigation, the same type of information should be retrieved for everyone, not just the ones who are ultimately fired.

Firewall the accused. The whole point of an investigation is to provide the employer with a non-discriminatory motive. That point is lost where, as here, there was evidence that the person accused of harassment is not obviously excluded from the investigation (in any investigatory capacity). Home Depot maintained otherwise (it appears the decision-makers were never told about the allegation of harassment) but when the termination notice is signed by the accused harasser (among others), and the accused harasser delivers the bad news, the employer is not making it easy for the court to see that the investigation is fair and above board. (It didn't help that the accused harasser bragged, allegedly, about getting the store managers fired.)

Follow company policy. The court of appeals roundly criticized Home Depot because "no specific policy violations" were listed on the termination notices even though the forms said to specify the policy that had been violated. The problem was particularly acute here because the store managers maintained that there was no policy. Strictly speaking, this last point wasn't a failing of the investigation itself but the court found this failure called into question the legitimacy of the investigation. It is a good practice point. Any investigation should, at a minimum, identify the exact policy that governs the situation.

Retaliation, Honest Beliefs and Employer Investigation

Taking a risk that I am harping on it too much, today's Sixth Circuit dismissal of a retaliation claim again confirms the value to an employer of conducting an investigation.

Understand my point of view: when my partners and I defend a lawsuit for an employer, we work hard to get the claims dismissed prior to a trial. (It isn't that we don't like trials, some of us do, but it is much more efficient and less risky for the client to get the claims dismissed early on.) With few exceptions, we have to show that even if the evidence were as the plaintiff says it is, our client should still win.

The decision today concerned a lady who alleged she had been harassed and then fired because she complained of harassment. Understand the timing, the employee complained about racial and sexual remarks by a co-worker, filed an EEOC charge, and was then terminated because, the employer asserted, she had filed a false injury report.

The employee claimed she had injured her back when a co-worker (the person she had accused of making harassing remarks) moved a truck while she was standing in its flatbed. The employer investigated even to the point of conducting a hearing where it received testimony (while the decision doesn't explain why a hearing was held, it was likely to satisfy some provision in the union contract). The co-worker and four other employees said they didn't see her in the bed of the truck and that she was, in fact, standing beside the truck when it was started. At the hearing, and at trial, the employee adhered to her story, saying the four witnesses were lying to protect the co-worker.

In dismissing her claim, the court avoided addressing the temporal proximity issue (or the more interesting question about the filing of an EEOC charge over the harassment claim after the employee had been suspended pending an investigation into the making of a false report. Rather, the court held that even if the employee had not filed a false report (as she continued to assert) and even if the four witnesses were lying to support their co-worker, the employer's investigation was sufficient justification for firing the employee. As I said earlier, when an employer is faced with competing versions of events, "there is probably no practical step an employer can take beyond independently investigating the misconduct charges that will reduce the chances of an employee's racism influencing its behavior." Brewer v. Bd. of Trs. of Univ. of Ill., 479 F.3d 908, 920 (7th Cir.), cert. denied, 128 S. Ct. 357 (2007).

Why is this so important? At the start I said employers benefit from early dismissals of lawsuits and to get them dismissed we have to show the material facts are not disputed. Here, there was a factual dispute - was the employee in the bed of the truck when it was started - but the employer's investigation resolved that fully and fairly. The investigation made the factual dispute immaterial because the focus was then on the honesty of the employer's decision, not on whether or not the employee had in fact filed a false report. Had the investigation not occurred (such as had the employer promptly fired the employee based solely upon the co-worker's story), or had it been slipshod, the court could have held a jury needed to determine whether the employer retaliated against the employee.

Conducting a Fair Investigation Into Employee Misconduct

Two decisions from the Sixth Circuit issued around Thanksgiving serve to emphasize the importance of an employer conducting a quality investigation before it fires an employee for misconduct. While I'll try to simplify the facts somewhat, understanding some facts are important to understanding why the Sixth Circuit held, in both cases, why a jury could reasonably find the employer discriminated against the employees.

Madden v. City of Chattanooga, the court of appeals affirmed a bench trial liability finding in a race discrimination claim. The black plaintiff worked in the service department and was fired for setting off a firecracker at work. His supervisor reported him. Plaintiff never disputed doing the deed. White employees had also set off firecrackers at work without being disciplined and there was testimony that the same white supervisor who reported the plaintiff had been present when white employees set off firecrackers and he had not reported them. Another supervisor who saw white employees use firecrackers at work merely told the employees to "knock off" the horseplay.

Madden would be a routine disparate treatment case except for the fact that the decision to fire Madden was made by senior managers who, the facts showed, did not know that white employees had set off firecrackers and for whom there was no evidence of a biased motive. (The Seventh Circuit calls this a "cat's paw" type of case. Shager v. Upjohn Co., 913 F.2d 398, 405 (7^th Cir. 1990)). In a crucial footnote, the Sixth Circuit pointed out that the "scope and nature" of the City's investigation before firing Madden was "unclear" but appeared to have been "limited" to interviewing the biased supervisor and the plaintiff.

To explain, the federal courts of appeal widely diverge on when the bias of a subordinate is imputed to the decisionmaker and the Supreme Court has twice expressed an interest in resolving the dispute. But one recognized way to avoid the cat's paw problem altogether is to conduct an independent investigation because that investigation should break the causal chain between the biased report of the subordinate and the final employment decision. Wilson v. Stroh Companies, Inc., 952 F.2d 942, 946 (6th Cir. 1992) is one such decision and other courts have also recognized the principle. Pennington v. City of Huntsville, 261 F.3d 1262, 1270 (11th Cir.2001) ("[w]here a decisionmaker conducts his own evaluation and makes an independent decision, his decision is free of the taint of a biased subordinate employee."); Willis v. Marion County Auditor's Office, 118 F.3d 542, 547 (7th Cir.1997) ("[W]hen the causal relationship between the subordinate's illicit motive and the employer's ultimate decision is broken, and the ultimate decision is clearly made on an independent and a legally permissive basis, the bias of the subordinate is not relevant."). In Cariglia v. Hertz Equip. Rental Corp., 363 F.3d 77, 87 n.4 (1st Cir. 2004) while the court held a jury trial was required but the court also observed that the employer would have won had it given the accused employee a "meaningful chance to address the allegations against him").

Madden amounts to a crack-down on what the court held was a less than adequate investigation. Remember the crucial footnote? So, with an "unclear" investigation which was "limited," an employer hangs itself out to dry even if the decision is made by an undisputedly non-biased decisionmaker. It is hard to say what the employer did in the investigation but the court clearly faulted the employer for not finding out whether other employees had been allowed to engage in similar conduct without being disciplined.

The decision in Martin v. Toledo Cardiology Consultants, Inc. shows what happens when the court conclude the employer's investigation amounted to a fait accompli. The employee had worked in the same doctor's office for years but then a new doctor assumed control of the practice. The new doctor didn't exactly follow the best employment practices; he reorganized the office, designating "favorites" (all of whom were under 40) as "team leaders" and telling employees his treatment of the "favorites" would not be subject to question. The doctor complained to the plaintiff about some of her conduct (it was clear the court found these complaints petty to the point of harassment) and later confronted her with evidence that she had used of racial slur to describe a patient. The slur was allegedly heard by another doctor and one of the "favorites." During the meeting, plaintiff signed a document admitting to the racial slur. As a result, she received a salary reduction, was put on probation and ultimately fired over a dispute about the work she wanted to perform. At her deposition, however, she flatly denied making the racial slur.

Like Madden this case largely turned on the quality of the investigation. (Unlike Madden, of course, this case didn't concern an "independent" investigation within the meaning of the "cat's paw" cases because here the decisionmaker was supposedly biased against the plaintiff.) The decisionmaker made a number of mistakes but what seemed to most trouble the court was that he didn't seem to listen to the plaintiff when he asked if she had uttered the racial slur. To be sure, there was evidence she had said it and she did sign a document saying she had said something (the majority and dissenting opinions conflicted on this point) but the court described the plaintiff being presented with "documents that she could sign, quit, or be fired." And the plaintiff testified that she was single and helping to support her mother, so she needed the job and therefore signed the papers. So, despite evidence from two witnesses that plaintiff had referred to a patient in a racially derogatory manner, and despite the fact that she had signed a document admitting to using a racial slur, the court nevertheless found plaintiff's deposition testimony created a factual dispute as to whether or not the employer made a reasonably informed decision in disciplining her for the racial slur. (The investigation into the reasons given for firing plaintiff was not any better but I don't need to bury my point.)

In theory the quality of the investigation shouldn't matter but practically it does because a good investigation can remove any dispute over the facts. I easily found three fairly recent Sixth Circuit decisions where the employer won because it conducted a full investigation and fairly resolved any factual disputes which arose during it. By doing a full and fair investigation, the employers were able to show that the decision was honestly held based upon the facts it knew at the time. Abdulnour v. Campbell Soup Supply Co., LLC, 502 F.3d 496, 502 (6th Cir. 2007); Michael v. Caterpillar Fin. Servs. Corp., 496 F.3d 584, 588 (6th Cir. 2007); Wright v. Murray Guard, Inc., 455 F.3d 702, 707-709 (6th Cir. 2006).

So while none of these decisions say so, the quality of the investigations played a very significant role in the outcome. In Martin evidence that the outcome was predetermined helped kill the employer's defense. In Madden, while the plaintiff was interviewed, that alone was not enough. And Madden shows that an employer must always determine whether or not there have been prior similar incidents by other employees outside of the protected class. Had the employer in Madden shown that it had asked the employee whether or not he knew of any other times other employees had thrown firecrackers, no matter what the answer, the employer certainly would have been much better off in making the decision and/or in defending the lawsuit. And worst case scenario, if an employer ever has to defend a decision to a jury, a thorough and fair investigation unquestionably helps persuade jurors that no discrimination motivated the decision.