Shortly before the argument, Justice Roberts, who is recused, sneaks off.  In a nine-justice court, a recusal is indistinguishable from a vote for affirmance (as 4-4 tie votes are affirmed), so the petitioners are already starting a vote down.

 
Carter Phillips, Sidley Austin, arguing for the petitioner Warner-Lambert, says that policing fraud on federal agencies is something entirely within the federal ambit, and that the Michigan statute is preempted by Buckman. 

Scalia asks how far that principle goes: what if there were a tort action alleging conditions of FDA marketing were not met?  Phillips punts the question: the petitioners are simply seeking a narrow ruling that Buckman applies.  But, Scalia, says, Buckman's cause of action was FDA-specific.  Phillips: But juries would be making the same sort of determinations that Buckman held impermissible.

Ginsburg (who perhaps tips her hand when she refers to the petitioner as a "drug dealer") asks isn't the Michigan statute just an "invigorated" version of the regulatory compliance defense?  Stevens: what if there is no statute, just the regulatory compliance defense, rebutted by the plaintiffs with a claim of fraud.  Phillips, yes, the plaintiffs would be preempted from making such an argument to the jury if regulatory authority is implicated. 

Stevens questions the premises of Buckman: what is the evidence of burden to the government?  Phillips: Buckman held that we don't wait until there is serious interference with the agency, we step in in advance.  Stevens asks about the bribery exception to the Michigan statute.  Phillips says that that is not pre-empted.  Later, Daryl Joseffer, arguing for the government, argues that it is pre-empted.

Stevens: If this goes to litigation, what happens?  Don't the plaintiffs have the burden of proving their case?  Phillips: Yes, and the FDA didn't withdraw the drug.  But there is still the problem of state courts pulling on federal prerogatives if that is litigated.

Daryl Joseffer, SG's office, argues for the government.  (And if you want to feel old, watch the oral argument of someone in a frock coat whom you supervised when he was a brand new associate taking his first deposition.)  Souter takes up Ginsburg's question: is there the same issue if regulatory compliance is a defense?  Joseffer: it depends on whether it implicates federal interest and asks the jury to second-guess the regulatory decision.  Preemption applies only to legal theories, not evidence.  Souter: so your objection is to the plaintiffs being required to prove but-for withdrawal?  Joseffer: not just that.  It's the second-guessing of the approval process.  Souter: But why isn't that the same as the compliance defense?  Joseffer: if it's a mere matter of evidence, no problem, but questioning product-specific approval decisions.  Souter: but under Michigan statute, FDA approval is conclusive.  Joseffer: a predicate to liability under the statute is inquiry into the FDA approval process.  That's the same thing forbidden under Riegel and under Buckman.  Stevens: But besides Buckman, there has always been a regulatory compliance defense. What have courts said in those circumstances?  Joseffer: this is a new problem; it's a theory of modern vintage to inquire into specific product approval.  Kennedy: what's your strongest case besides Buckman?  Joseffer: This is so obviously a federal prerogative that there isn't other precedent here.

A good chunk of the arguments were taken up with back-and-forth between Ginsburg and Scalia on the question of severability and how to interpret Michigan law.  Joseffer says the US has no position on the state law question, but notes that the Garcia decision in the Sixth came about because plaintiffs asked for preemption of the entire statute; on rebuttal, Phillips argues that the Court should defer to the Sixth Circuit's findings on Michigan state law (and the Second Circuit's deferral to the Sixth Circuit), but acknowledges that a state supreme court certification on the severability question is possible.  But he disagrees with Ginsburg's premise that there was no bitter with the sweet: if the FDA does make the findings of fraud and withdraws a drug, a cause of action is possible under Michigan law, and that is a reasonable reading of the statute.  Ginsburg argues that the legislature wouldn't have agreed to that bargain.

Allison Zieve, Public Citizen, for respondent plaintiffs: the misrepresentation exception to the Michigan statute does not implicate Buckman.  There is no reason to fear, as Buckman did, that companies would oversubmit evidence to the FDA, because they are required to submit "all available information" about adverse events. 

Kennedy: but wouldn't discovery be intrusive upon the FDA?  Zieve: No, because discovery would be the same as a typical product liability case.

This sets Justice Breyer off on a long speech: Why isn't that worse?  You're alleging that the drugs have side-effects that hurt people.  But there is a risk on the other side, the loss of the drug that can help people.  On balance, is the drug going to save lives or cost lives?  Who should make that decision?  The FDA, or a jury of twelve people that sees only the hurt people?  That is a decision for Congress, and Buckman holds that Congress said the agency decides.  (Scalia smiles at this.)

Zieve tries to disagree: product liability simply compensates injured patients, and thus complements FDA regulation.  We're not seeking a regulatory remedy, just compensation.  Kennedy: but your premise is that the drug should not have been sold.  Zieve, in response, picks up on Ginsburg's "invigorated defense" argument: the Michigan statute simply holds that approval is reliable evidence of duty.  If a company lied to or bribed the FDA, then approval is no longer reliable.  The statute doesn't police the FDA enforcement, or take a drug off the market, just decides whether plaintiff can get damages like in the other states.  Kennedy: but the jury will have to decide the drug should have been off the market.  Zieve: Yes, in Michigan, we'll have to show that, though we think we can.  Stevens: could you still lose if the exception applies?  Zieve: Yes.  We'd have to go through the same process as Rezulin litigation throughout the rest of the country.  There is already a lot of discovery done of Warner-Lambert and the FDA.

Alito: Is the government wrong that having a jury second-guess approval interferes with the regulatory mission?  Zieve: In this case there is no interference, because the drug is off the market.  But even as a general matter, there's no interference, because the effect of making a showing in court has no regulatory effect.  It's the same as any other state tort litigation, which doesn't police regulatory compliance either.  Alito: But wouldn't discovery into the internal decisionmaking of the FDA be problematic?  Zieve: Plaintiffs seek that now, but Michigan courts can be trusted to control it.  Alito: What evidence would you need to prove exception?  Ginsburg: Does it obstruct the FDA to take their depositions?  Zieve: No more so than in any other state.  Rezulin discovery was negotiated with the FDA, including deposition of the medical officer.  Vioxx discovery was contested, and the MDL ordered discovery.  Other cases, the FDA successfully blocks discovery.  There's no evidence it is a burden on the agency or judges are overwhelming the FDA.

Scalia asks if there can be a distinction between drugs on the market and off the market; after all, if it is on the market, the FDA has incentive to retrospectively review the data in a way that it doesn't if it is off the market.  Zieve: But all the FDA could do then is to sanction the manufacturer.  Scalia: Wouldn't that solve your problem?  You then get to show the exception applies.  Zieve: But we can't force the FDA to make a finding, and the FDA finding doesn't get us compensation. 

Breyer: But the drug will be withdrawn.  What if a jury is wrong? Who's more likely to be right?  He is confused: you mean that it is possible for the FDA to approve use, that the label is okay, the doctor follows the label, and plaintiffs get compensation from the manufacturer for a disclosed side effect?  Zieve and Ginsburg answer yes, and Zieve explains that the Michigan statute narrows, rather than expands tort law.  Breyer: So you're saying that if the tort system isn't preempted, this isn't preempted.

On rebuttal, Phillips largely takes up the question of the appropriateness of severability.  He blunders a bit when he argues that the FDA can order restitution to consumers after a drug is withdrawn and Stevens spanks him noting that that is hardly full remedial compensation for a wrongful death claim.  But he closes noting that the FDA has full incentive to go after fraud, even post-withdrawal, and that Public Citizen petitioned the FDA to have Rezulin withdrawn on this sort of evidence, and the FDA rejected the petition. 

Defense attorneys may be concerned that Phillips's argument never gave a clear answer to the question of where to place the line between every-day tort law and the Michigan statute and frustrated that he hewed to a narrow line of application of Buckman rather than a larger theory of when preemption was appropriate.  Zieve takes the position--which she may regret later in Wyeth--that Michigan is just a special example of state tort law.  A number of justices seemed genuinely surprised that it is common in product liability litigation to depose FDA officials about their administrative decisions.