Preemption in the Rehnquist Court

April 2, 2004

Unedited transcript prepared from a tape recording

Proceedings:
MR. GASAWAY:  As many of you know, I am Robert Gasaway of Kirkland and Ellis.  And we are here today to hear from Dr. Michael Greve.  Dr. Greve has done an empirical study of preemption in the Rehnquist court and is going to tell us how to and why we win and lose cases.  After Dr. Greve talks, we are going to hear from two commentators, first, Dan Schweitzer of the National Association of Attorneys General.  Dan has one of the more interesting jobs in Washington.  He is Supreme Court Counsel for the National Association of Attorneys General, which means he gets to edit and comment on, in moot court, about 35 to 40 cases per year in the Supreme Court in which states are involved.

And then we have Professor Tom Merrill.  Professor Merrill has not only done empirical work, as a professor now at Columbia Law School and before that in Northwestern, on decisions in the Supreme Court, he was also a deputy solicitor general for several years in the 1980s.

So welcome to Michael, to Dan, and to Tom Merrill.  After Michael talks and we hear from Dan and Tom, we will, obviously, take your comments from the floor.

Michael.

MR. GREVE:  Thank you, Rob, and thanks, Tom and Dan, for agreeing to comment on this. 

What I am going to do today for the first 20, 25 minutes is a massive data dump.  This is why we're doing this at ten o'clock in the morning.  I can do this on adrenalin; you may need caffeine.

There is an empirical study which we have conducted over the past year and a half or so--and I should mention I didn't do this alone.  Ken Hendricks and John Plik [ph] in an earlier stage of this project, Denny Smith, have really done the Lord's work on this stuff.  It takes a lot to scrub these data.

So the first half of the talk I'll just simply describe what the Rehnquist court's preemption decisions look like--the case load, subject matter, the preemption of tort claims, the parties, how much conflict and how much consensus there was on these decisions, and what the outcomes were.  In the second part, I will interpret some of the results with respect to the Supreme Court's role, with respect to the outcomes--and there I will concentrate on participation of states as parties, on the role of the solicitor general, on the role of state amicus briefs--and then I will present some rough regression results.  And finally, we will turn to what everybody wants to know, which is how do the justices vote and are there blocs and swing votes.

So we will start with case load.  There were overall 102 statutory preemption decisions between 1986, which is when Chief Justice Rehnquist assumed his present position, and the 2003 term.  Are there exactly 102 decisions?  I can't tell.  There are about a dozen cases that are really, really hard to classify, where you really have a dispute as to whether they are preemption cases or not.  Dan, in an earlier phone conversation, emphatically insisted that one of the cases in this sample is not a true preemption case.  I gave him another example in return of a case that we excluded, which is arguably a preemption case.  If 24 lawyers went over that set of a dozen cases and tried to decide is that or is that not a true preemption case, they would probably come to varying results.

But this is roughly the right number and they are 95, 98 percent the right cases.  That distinguishes this case sample from all other case samples along similar lines.  And I can tell you that whether you include these borderline cases or not, you are going to get roughly similar results so long as you don't over-interpret the results that I am going to present later.

What we excluded is Section 1983 cases.  Of course, Section 1983 is a statute, but [inaudible]  the world looks very different.  We excluded federal common law preemption, mostly dormant Commerce Clause cases, and we excluded cases involving the role of the judiciary itself.  So what you are dealing with here is the hard core of statutory preemption cases.

The frequency, as you see, preemption cases range from two cases in the 1997 term to a high of 13 in the first year in 1986, and the average is exactly six per term.  They have declined in frequency over time.  That is what the trend line there primitively indicates.  And here I should mention we distinguish throughout between the early Rehnquist Court, which is up to the 1993 term, and then the late Rehnquist Court--which sounds like it died, but what it means is from 1994 on forward, beginning with the appointment of Justice Stephen Breyer.  Since then, of course, the court has sat in its present composition.  That may matter a whole lot, as Professor Merrill has argued in a terrific paper.  It does matter a lot he has chosen the more fortunate terms, perhaps, First Rehnquist Court and Second Rehnquist Court.  I will stick with early and late.

The decline of the case load here doesn't mean a whole lot, because the preemption cases declined in tandem with the Rehnquist Court's civil docket.  Overall, in both periods, preemption cases are about 8 percent of the civil docket.

What are these cases about?  Here they are all by subject matter: labor and employment cases; cases that include ERISA cases, economic regulation, transportation and infrastructure cases, health, safety, environmental cases--we lump those all of those together and call them regulatory cases.  This is the hard core of the regulatory state.  That is about three-quarters of the total docket in both the early and late Rehnquist Court.  And then there is a smattering of cases involving public benefits such as veterans' benefits, tax cases--I was actually surprised there is only one in the late Rehnquist Court--and other cases, that is mostly Indian cases and Federal Arbitration Act cases.

Preemption of tort claims, an intensely controversial subject.  Of course, that question is sort of do-or-die both for the trial bar and for corporate America, and there's a burgeoning literature in this field.  What do we know about tort cases in these areas?  Well, there are 31 cases out of 102 total that deal with the federal preemption of state common law, almost always tort claims.  Twenty-nine of these cases fall into the four regulatory categories there.  And in that universe they are about 40 percent of the case load, with a slight increase for the late Rehnquist Court.

In terms of do these tort cases, as I will call them, look any different from the general preemption universe?  Yes, kind of.  Surprisingly, the tort cases come disproportionately from federal courts, not state courts.  I didn't expect that at all.  Only seven of these tort cases came from state courts.  That is only 17 percent of that universe of cases.  It is twice that for federal courts.

Tort cases yield a disproportionate number of mixed votes and outcomes--that is, cases where the Supreme Court or some individual justices found something to be partially preempted; a disproportionate number of divided votes and opinions--that is to say, these cases are often more controversial or more often controversial than statutory, clean preemption cases dealing with the preemption of state law.

And, I should add, tort cases yield a disproportionate number of pro-preemption rulings.  In the late Rehnquist Court that was about 60 percent, as against 41 percent in cases dealing with the preemption of state statutes.  However, you should not make too much of that spurious finding because, as I will show later, it is most likely the lack of state participation in these cases, rather than the nature of the state law claim, that explains the higher probability of preemption rulings.  States are rarely parties to tort cases.

Parties.  This is such an important insight that we devoted an entire screen to it.  The enforcement of federal preemption through litigation is to a large extent the work of business or other private parties.  Everyone thinks of preemption cases as, well, this is about state-federal relations and so that ought to involve the parties.  No, it doesn't.  And the reason why this is important is it explains a great deal of the outcomes in preemption cases.  I will get back to that.

Who are these parties?  We have ordered the preemption cases by plaintiff and defendant and then by petitioner and respondent.  As you will see in these charts, preemption cases are almost always initiated by a private party, in 90 out of 102 cases.  State governments participated as an original party to a preemption dispute in just over half of all cases--55--and there are only ten cases in which a state initiated a preemption lawsuit.  Curiously enough, nine of these fall into the early Rehnquist Court.

Only six cases involve both the federal and a state government as parties; 47 cases involved exclusively private parties.  You know, I keep these areas there.  The green area, there are--I will later call these wholly private cases, and these are state party cases.  And in some of the analyses we will see later, we have concentrated on those 94 cases to take out the quirky, atypical cases that might bias the results.

Finally, business petitions far out-number state petitions and petitions by non-governmental parties, private and business combined, constitute 70 percent of the case universe that the Supreme Court decides.

How controversial are these decisions?  Fifty-three of the 102 cases were unanimous decisions.  That is much higher of the general degree of unanimity on the Rehnquist Court, which is 41 percent, and that ratio has remained roughly constant.  It has been 50/50 all the way through.

If you add, as we did in these columns, we toyed around with adding to that unanimity universe the cases where Justice Stevens files a lone dissent.  That doesn't mean that the case was controversial.

[Laughter.]

MR. GREVE:  There are a also a large number of opinions there where Justice Thomas insists that the Federal Arbitration Act has no preemptive force.  As far as I am concerned, he is right on that, but eight of his colleagues disagree.

So if you add that to the case universe, then it is about one in four cases that prove to be controversial in the sense that there is a vote differential of three or less.

We have added to this cases with mixed outcomes, that is, partial preemptions yielded nine near-unanimous verdicts on the preemption question--that is to say, there is a question whether state tort claims are preempted.  In one respect, the court splits 5-4, and on another state claim they all say, yes, that is preempted.  We divided these kinds of cases in two and weighted each observation at 50 percent.  That is not a totally clean way of proceeding, but it is the least bad way of proceeding.

If you add those nine unanimous decisions, you end up with this result and it turns out that one in five decisions, roughly, is controversial.  And we will operate with that number lower down.

What are the outcomes?  Preemption litigation, the Supreme Court has proven by and large a 50/50 proposition, and that is true for both the early and late Rehnquist Court.  It appears that preemption outcomes are slightly more probable in contested cases.  That is this there--the ratio seems to be almost 60 percent.  That may be a real result but you are dealing here with very, very few cases, so I don't have a whole lot of confidence in that finding but there may be something to it.  And by a raw measure of case outcomes, the Rehnquist Court does not appear to have become more hostile to federal preemption.  As you will see later, that is actually a bit of a surprise.

Now I will move to interpretation.  The Supreme Court's role.  What the Supreme Court itself think its role is in federal preemption cases?  And there you can typically distinguish between three roles.  The first is kind of a federal supremacy role.  Guaranteeing federal supremacy, that is what the Supreme Court is about and what it's here for.  Justice White held something like that view.  If that is true, if you follow that role, you probably should see a higher disproportionate number of state court rulings in the bunch, because they have a higher or may have a higher propensity to wander off the reservation.  And you should see a disproportionate number of reviews in cases where the lower courts, federal or state, ruled against preemption.

The second type is something you might want to call an error-correction view in preemption cases, though not necessarily over the full range of cases.  Here is the general intuition.  We should, says Justice Scalia, to the extent possible approach preemption cases without any presumptions that bias the results one way or the other.  But the one presumption that does make sense is that Congress would want the results to be uniform and clear across the full statutory range, and that means that we will pay particular attention to errors below and to splits and that sort of thing.  If you adhere to that kind of ideal style of preemption jurisprudence, you should see a very high reversal rate I think.

And the third view is some view that says in preemption disputes we should be sensitive to federalism questions.  One of the first paragraphs in Justice Steven's Geier dissent starts, "This is a case about federalism".  It is unclear how that intuition should sort of counteract or interact with the federal supremacy claims that are involved, but one thing you could probably say is that there should be increased hostility to preemption over time if the court followed that intuition.

Still on the Supreme Court's role, where do these cases come from, from what courts?  We divide them into federal courts and state courts.  It turns out the Rehnquist Court granted cert to state courts in 40 cases and to lower federal courts in 62 cases.  For comparison, civil cases almost always come from federal rather than state courts.  So for the early period here, where there is an almost even match between federal and state cases, that is very, very atypical from the general run of civil cases.  The ERC, the early Rehnquist Court, granted an almost equal number of certs to state and federal courts.  And then there is a break.  It is a very pronounced break.  The later Rehnquist Court has focused its attention on federal courts.  There it is 32 to 12, and that is more in keeping with the general run of civil cases.

The next statements--you can't see this here, but you will see it shortly--the Rehnquist Court reviewed a much larger number of state court decisions against preemption than state court decisions for preemption.  In cases from federal courts, lower court decisions, for-preemption outnumber those against.  I think that is sort of mild support for a supremacy view.

Reviews of state court decisions against preemption dropped from 22 to six.  That is actually an astounding fact.  That accounts for the entire drop in volume between the early and the late Rehnquist Court.  That might be read as support for a federalism view.

How did the Supreme Court treat these lower court decisions?  Reversed in 60 cases and affirmed in 37.  The general reversal rate over that period and over the Supreme Court's period in general is about 60 percent--59, to be perfectly precise.  So this is consistent with that view.

However, the reversal rate increased substantially for the late Rehnquist Court--near balanced for the ERC and three-to-one reversals and two affirmances for the late Rehnquist Court.  That is a shift to what I have called an error-correction view, and that shift is too big, to my mind, to be wholly accidental.  On the other hand, I am totally unclear about what drives it.  It is also, by the way, again, as the earlier shift I've noted, it coincides with the 1994 appointment.

Here you see the probabilities of affirmance depending on the lower court's disposition.  There is a very high affirmance rate for the early Rehnquist Court, an astoundingly low affirmance rate for the late Rehnquist Court--especially if these cases come from state court--and that is true regardless of which way the lower courts ruled.  What this indicates is either the lower courts may be hopelessly confused or maybe the Supreme Court itself is hopelessly confused, but it is a startlingly high rate.

We now move to the outcomes.  There we start with a simple hypothesis:  In deciding preemption cases, the justices or judges rely on simple signals or fire alarms, such as the parties' institutional identity and their litigation positions, not to decide the cases but to sort of put them in mental boxes and to reduce complexity in the legal environment.  And there are three reasons why I think that is a plausible hypothesis.

First, preemption cases, as we have seen, are a steady diet so there is a premium on having mental boxes to put them in.  You don't have to do every case from scratch.  Second, these cases tend to be very, very, very complex, so you want to have some complexity-reducing device.

And the third and most important factor, to my mind, is that there is a high risk of strategic behavior in these cases.  What I mean by that is the cases are about state-federal relations at some level but the arguments are proffered by trial lawyers and by businesses, neither of whom care one whit, as a matter of litigation, about principled questions of federalism.  They just don't give a damn.  So you have to discount their affirmance on both sides by some margin.

This leads to three hypotheses, or expectations.  First, because a state party's complaint about federal interference is more authentic and credible than a private party's averment to the same effect, rulings against preemption will be substantially more likely in cases in which a state is a party.  Second, in wholly private cases, the presence of state amici will render rulings against preemption more likely.  And third, if the Office of the Solicitor General disclaims preemption, then rulings against preemption will be substantially more likely.

How do these expectations hold up?  Here are the probabilities of preemption by party constellation.  And it turns out that, lo and behold, preemption outcomes are much more likely in wholly private cases, two to one--that is this here--than in state participation cases, and that effect is more pronounced for the late Rehnquist Court than for the early Rehnquist Court.

What you see here, from that 68, is that business does pretty well in these tort cases.  You have to remember that of these 47 cases, private cases, a very large number of these tort cases, and business is doing pretty well and those businesses not doing well when states are parties.

You have the probabilities of preemption assorted by petitioner and respondent.  It turns out that when states act as petitioners, they are very likely to prevail.  So if a state is a petitioner, there is a preemption result in just 26 percent of the cases, meaning they win 74 percent of the time.  Now, you have to remember in interpreting these numbers that, in the general run of cases, about 60 of the time the petitioner will win in any event.  But it turns that, regardless of whether they are petitioners or respondents, the states do better than the average party.

Here, the probabilities of preemption, again by petitioner and respondent, and we have sorted these into early and late Rehnquist Court.  The principal finding here is the substantial advantage enjoyed by the petitioners increased during the late Rehnquist Court, especially for the states.

So to sum up here, the higher rate of preemption outcomes in state-party cases is highly consistent with the signalling theory I mentioned at the outset.

Here comes the solicitor general.  What we looked at is what position did the solicitor general take and was that a Republican or a Democrat.  And there are 61 preemption cases where you are dealing with a Republican OSG and 41 in a Democratic administration.

Look at the two left-most columns first.  If you exempt federal-party cases, that is to say where OSG was a party or a federal agency was a party--we're dealing here with OSG amicus briefs--OSG took a pro-preemption position in just 37.8 percent of all cases, and that is about the same for Democrats and Republicans.  The real difference between Democrats and Republicans here is the abstention rate--this--where Republican SGs abstained.  And it turns out all of those abstentions come in the particular kind of cases, which is--the third and fourth columns are these private cases, where business and tort layers gun it out.  No state is a party there. 

It is very, very pronounced.  Democratic SGs frequently take a position and frequently are in favor of preemption, and Republicans sit these cases out very disproportionately.

Democrats are more likely than even Republicans to defend preemption in wholly private cases and much more likely to side with the state when there is a state party.  That is this.  So there is a much more pronounced pro-state tendency in OSG offices when the officeholder is a Democrat.

And as it turns out, it matters.  And here it comes.  These are the conditional probabilities of a pro-preemption outcome by OSG party.  And again, these are the weighted numbers.

First, the OSG success in preemption cases is about 80 percent.  The overall success rate is 70 percent.  But the record of success is asymmetric.  When OSG argues against preemption, an anti-preemption outcome is pretty much a foregone conclusion.  There, point 12.  OSG argues against preemption 88 percent of the time.  That is what the court says.  In contrast, when OSG argues for preemption, anti-preemption parties, meaning the states, still have a roughly one-in-four chance of prevailing.

Next, an OSG intervention for preemption does not appreciably increase the likelihood of a pro-preemption ruling.  So that is the last line and the first; you compare those.  What this suggests to me is that the Supreme Court views a pro-preemption stance on behalf of the OSG as a kind of default position with little signalling value.

And finally, the partisan affiliation appears to matter.  In all 17 cases where a Republican solicitor argued against preemption, the Supreme Court took the same position.  For Democratic solicitors, in contrast, the odds of having the OSG's position rejected were roughly even, regardless of the position.

I think that is highly consistent with a signalling value, especially for the Republican OSGs.  What the justices seem to be saying to themselves is, look, if the feds themselves don't want preemption, it's highly unlikely that there is preemption.  Moreover, if that position is articulated by a Republican, which is beholden to all of these business interests that want to mold down the states, then it is doubly credible and so therefore.

Moving right along to the state amici.  States participated as amici in 65 percent of the cases--64 to be precise--and that has increased over time.  They are almost always joined by more than one state.  And of the 65 cases where a state participated, 47 cases featured mass briefs with nine or more signatories.  By all measures, state participation has continued to increase.

What else do we know about state amicus briefs?  Well, only, quote-unquote, 51 percent of wholly private preemption cases--that is to say, where there's no state party--feature a state amicus.  In contrast, state amici participated in 72 percent of the cases in which a state is already a party.  And states, according to this chart, have a high propensity to file for petitioners rather than respondents, if you look in particular at that rate there.  That is say, where there is a state petitioner, there is almost always a state amicus.  That is not true when the state is a respondent.

Does it matter?  Here are the outcomes.  State amicus participation appears to have a positive effect on state success.  The numbers all go in the right direction, right, from dot 46 to dot 69 and from dot 30 to 41 percent in the wholly private cases.  If you contrast the most extreme positions, here is state party with a mass state brief, that is 70 percent against preemption, or 70 percent state win; whereas you see in wholly private cases where there is no state at all, the ratio is the reverse.

That, too, is consistent with the signalling theory that I have articulated at the outset--with one exception.  You would expect that the effect of amicus briefs, state amicus briefs, should be higher in these wholly private cases where the court has not already heard from a state.  But it isn't, according to this.  So we will see how this holds up.

Here are the regression results for these three variables.  What effect does it have when the OSG says there is no preemption?  What effect does a state party have and what effect do state amici have?  And it turns out that the effect of OSG and state participation variables is in the expected direction--both make preemption less likely, that the effect is substantial for both and, moreover, statistically significant for all cases.  And these are all cases.

For contested cases only, the OSG variable loses statistical significance.  That is this regression for the contested cases only.  There the significance probably is lost due to small numbers.  There is an R square here of 36, which is respectable, though not terrific.

The state amicus variable here has a small effect, and it does not approach statistical significance.  And that result is very robust and holds for any subset and configuration of cases that we could think of.

Remember what I said earlier about the tort cases, that there seem to be a disproportionate number of pro-preemption rulings in these tort cases?  But it turns out if you plug that in, it turns out the tort variable has a small and statistically insignificant effect and it fails to improve the fit of the model.  Principally what it does, if you add the tort variable, is the state-party variable becomes statistically insignificant.  That is most likely the covariance problem, because, as I mentioned earlier, states don't participate in these tort cases, so the variables here track each other.

I have a chart--if state amicus briefs don't have an effect, why do they participate?  I want to rattle through this quickly, because I don't want to leave the misunderstanding that this is a conclusive finding.  All we can find and show is that--we can't show a statistically significant effect, but for this set of cases that doesn't mean it doesn't matter.  There are other studies, one in particular by Professor Merrill and a colleague that shows that state amicus briefs, especially when they are filed for respondents, do have an effect.  There are reasons to believe that state attorneys general, just as everybody else, files amicus briefs for reasons of consumption values, right?  Just as the ACLU files amicus briefs for fund-raising purposes, so state AGs join with each other to say, well, we are being good soldiers here.

A third factor--and this is not a rigorous explanation, but it is sort of a suggestion--there are low opportunity costs in most of these cases to file an amicus brief, but there is a high potential return.  It's like buying a lottery ticket--it is a very small investment and a huge payoff if you hit the jackpot.  By the same token, that may also be true of amicus briefs in certain cases, and that might explain a good deal of the participation.  And finally, I would suggest that there may be a market test for the general model I have sketched.

If it is true that state amicus briefs do have an effect, then the parties should eventually adjust their behavior.  And what I mean by that is that I would expect in the future the trial bar to lobby for state intervention in wholly private cases much more than they have in the past, especially on the respondent side.  And we would expect, in the future, business to lean much, much more heavily on the OSG, especially Republican OSGs, to shut the heck up when they're tempted to say there is no preemption, because that sinks the case.

And so if you follow those kinds of actual world events, you should get a market test for my hypothesis here.

Finally and at long last, the justices' votes.  What are the probabilities of voting for preemption?  Most justices are about as likely to vote for or against preemption as the court as a whole--surprise, surprise.  That is because most of the cases are unanimous.  In contested cases, of course, the votes diverge more sharply.

And the prominent outlyers here, that is justices with a record of substantial divergence from the court average, are Justice White--long gone on the pro-preemption side--and Justices Breyer, Ginsberg, Thomas, and Stevens on the anti-preemption side.  An interesting finding.  It gets more interesting.

These are the judicial votes in contested cases only, early Rehnquist Court versus late Rehnquist Court.  It turns out that Justice Scalia turns distinctly more pro-preemption than the court as a whole in these cases, especially during the late Rehnquist Court.

Justice Souter has in the late Rehnquist Court turned as hostile to preemption as has Justice Stevens.  And Chief Justice Rehnquist and Justices Kennedy and O'Connor appear to have become a little more preemption friendly.  Right?  They all go up--O'Connor, Rehnquist, Scalia, Kennedy.

But--and this is a chart that should sink a corporate defense counsel's heart--these are the replacements.  And it turns out that all of the justices who replace the outgoing justices are vastly more preemption hostile than their predecessors.  Look at Justice White, right, there is that New Deal war horse who sort of fond preemption everywhere.  He has been replaced by Justice Ginsburg.  That is the most extreme.  And, of course, you have lost Justices Marshall and Brennan, who thought that federalism was kind of an excuse for trampling on black people.  And they have been replaced with sort of people with a more modern anti-preemption attitude.

If you look at this chart, it actually becomes somewhat difficult to explain why preemption outcomes haven't shifted in a more anti-preemption direction.

Finally, are there blocs, are there swing votes?  There are some indications of reversed bloc voting; that is to say, if you take federalism 5, they tend to vote in favor of preemption--with the exception of Justice Thomas, who is sort of the swing vote--and here you have the four liberals, who, vis-a-vis the straightforward federalism cases, seem to have reversed their votes.

But in any given case, it is possible to pick up a vote from an unlikely justice.  These are frequency of voting with the majority in contested cases, late Rehnquist Court only.  And it turns out nobody has a perfect score in these pro-preemption outcomes; that is to say, nobody voted all the time with the majority in these kinds of cases.  And the same is true, by and large, on the anti-preemption side.

So what you can take away from this is it will be impossible for a state to win a preemption case without Justice Ginsburg's or Justice Souter's vote--or Justice Breyer's vote, even--but you can pick up votes elsewhere.  And, conversely for corporate America, it means you can't rely on the conservative bloc to carry you through; you will need a vote someplace else.

The same idea appears here.  This simply shows the likelihood of voting together.  And it turns out the paired observations showed little evidence of bloc voting and no evidence of a swing vote.  And we stopped with the paired votes and didn't look into blocs because if you can't find paired votes, then you certainly can't find blocs.

And that mercifully ends the presentation.  What did we learn from it?  I will spend two minutes on it.

One of the things that I have noticed--I mean, the change in '94 is very, very pronounced, it seems to me.  I have talked about the review pattern and there are other changes that you can notice in '94.  The only thing that didn't change is the outcomes in these cases.  And I really cannot explain that, and I cannot explain in particular why the outcomes haven't changed more than they did.

I will give you two possible explanations.  One is this.  In a crunch, the conservatists will come through and bail out corporate America on brutally policy-oriented grounds.  But they rarely vote as a bloc.  And if there is a free vote, they will vote their true preferences, which is on the other side--or at least one of them will.  The prototypical case, if that analysis is right, is Geier, right?  Justice Breyer flipped to the pro-preemption side, so Justice Thomas basically had a free vote to do as he pleased and that meant go with the states in that case.  Garamendi you can read like that, a number of other cases.  That, of course, won't work if Justice Breyer himself is against preemption.  Then they really have to vote as a bloc.  They rarely do that; they did that only twice.  And that means when Justice Breyer flips, the case flips.

So in other words, to bring that to a point, there is a swing vote.  It turns out to be Justice Breyer, except it's masked because his vote allows other justices, in particular on the conservative side, to vote strategically.  Every data point in this collection is consistent with that explanation.

But there is a rival and more benign interpretation, to my mind, which I cannot exclude and which is also consistent with the data, which is that in 1994 the Supreme Court of the United States at long last got themselves a justice who actually understands these cases.  They are very, very complex, they are very convoluted and very complicated.

You can take my word for it, eight of them don't begin to comprehend what these cases are about.  Steve Breyer does.  And so, of course, everybody looks to him in these kinds of cases and takes cues from him.  And so maybe then the remaining justices sort themselves around that pattern, those alignments, in accordance with their own sense of the scheme before them and in accordance with their intuitions about a whole layer of issues involving statutory interpretation, delegation issues, the role of the courts, all of which are involved in preemption cases.

And I want to emphasize that those intuitions need not have to do with federalism and they need not have to do anything with sort of pro-business or anti-business.  They may just be genuine legal disagreements.  I cannot exclude that possibility at all.

I will end there.  I have gone on for far too long.

MR. GASAWAY:  Thank you, Michael.

Dan?

MR. SCHWEITZER:  Well, I guess it would be entertaining if I now discussed how I disagreed with pretty much everything Michael said and why--except that I don't.  I think it is a terrific compilation of statistics and analysis that Michael has done, and I agree with the vast majority of what he said, both methodologically and his conclusions that he drew from it.

So what I thought might be most helpful is to provide some insights on what I know best, which is why states file amicus briefs, and some comments on whether I think they are effective or not, particularly since Michael did have some thoughts on it.  I have been on the inside for the past eight years, and I thought I would give my take first on my general sense from experience of why states tend to file amicus briefs.

And then, what I did is I looked at the actual cases over the past eight years, I guess it is not all of the late Rehnquist Court, it is the Schweitzer part of the Rehnquist Court, a self-centered perspective on the Supreme Court's history.

And then, if there is time, I thought I would talk a little bit about what the National Association of Attorneys General does on preemption cases, because I know it does come up a little bit in Michael's paper.  And I thought maybe I would clarify about that.

So let me start why a State Attorney General would decide not merely to join the amicus brief but to write an amicus brief.  And I came up with ten reasons, because it is nice to have ten reasons.  I have five what I think are the key reasons and then five are sort of supplemental ones that come into play from time to time.  And, of course, some of these will be interrelated.

The first, and I think the preeminent, reason a State Attorney General office would decide to write an amicus brief is that he or she or he or she's top staff who is making the decision realizes that their office has an institutional interest in the case.  What that usually means is that their office represents an agency that enforces the type of state statute or regulation that is at issue in the preemption case.

Now I was going to say a classic example of this is Rush Prudential v. Moran, an ERISA preemption case, where 37 states had state laws similar to the Illinois independent review provision at issue.  Then it hit me that I am not sure how many of those are actually enforced by the state itself, but I imagine Illinois was actually a respondent in that case, and I imagine some number of states actually are involved in the enforcement there.  But that is an example.

The second reason, and it is very much interrelated, is has the Attorney General office litigated the issue itself.  And example looking down the list of cases is the Director of Revenue of Missouri v. Cobank [ph.], which involved whether federally-created banks for a cooperative are exempt from state taxes.  And states around the country had been litigating it, anxious to get it to the Supreme Court.  It got to the court, and Missouri wrote an amicus brief.  They had litigated before and were happy to do it again.

As you can see, there is an overlap between that and the first issue but it really brings it to a final point when they actually litigated the issue.  It also makes it a little easier to write the brief because they have done some of the research.

And you will notice that these two points which I think are the two most important points would suggest that a state is more likely to write an amicus brief in a Supreme Court preemption case where the law that is arguably preempted is positive law.  It is a state statute or regulation as opposed to common law, where the states typically are not the enforcers or never are the enforcers.

Key reason number three, the case involves an important issue to the Attorney General.  An example of this is the cigarette regulation.  The states filed an amicus brief in Lawlard Tobacco v. Riley [ph.] case, in light of the master settlement agreement in the tobacco litigation around the country.  Obviously, tobacco is very important to the states.

Another example is regulation of HMO.  It is a very important issue to the states, making sure that their citizens are not subject to sharp practices by HMOs.  And so the State Attorneys General have been frequent amici in the ERISA preemption cases, particularly the ones involving HMO regulation.

Reason number four, a client agency, essentially the state AGs client says we care about this, we want you to get involved.  That recommendation was turned down.

Number five, often forgotten but mustn't be underestimated I call fortuity and resources, resources being the key.  Is this an office which has the funds, has the attorneys with the time and ability and expertise to write amicus briefs.  And the fortuity part is do these particular lawyers have the ability to do it or are they just tied down on other more pressing matters.  Amicus briefs are important but to an extent they are something of a luxury.

So those I think are the key five issues.  And if those five line up right, I think generally you will see an amicus brief in any Supreme Court case and particularly a preemption case.

Five other factors that are less important but sometimes come to bear.  First, they will look at this and say, well, we have an institutional interest.  It is important to us but do we have anything to add?  And it is certainly something that they should be looking at if the state isn't.  Is there something to say?  And related to that is what is the skill of the party in the case?  Do they really need help?  And if they do, it makes it more likely that a state will want to jump in and really make sure that the position that matters to them is advocated well in the court.

The next reason is the political philosophy and ideology of the Attorney General.  Taking preemption is an Attorney General who believes strongly in state sovereignty.  This is an Attorney General who believes strongly in supporting plaintiffs and the right of plaintiffs to recover for alleged wrongs.

Number eight is one that Michael mentioned, supporting a sister state.  It is certainly the case where the state is a party and the state will ask other states can you please support us in this important case and sometimes states will want to be collegial as Michael suggested.

Number nine, a state will sometimes want to support a political subdivision.  Sometimes these preemption cases arise from municipalities, the City of Columbus v. Garage [ph.] case.  The case this term, Engine Manufacturers Association v. Southern Air Quality District in California, where California files an amicus brief supporting their local air quality district.

And finally, number ten which is the reason which I guess some people might look to most cynically, but it is the one that I think happens least often is home-cooking.  The party in the case is a local business, is a local citizen who is a plaintiff and calls up the AG and says, please, you have got to help me, I am your constituent.  It is really I think very rare that that is the driving force in a state doing an amicus brief in the Supreme Court, though I will talk a little bit about that later as I go through what has happened in the eight years that I have been with the association.

Let me quickly go through a few more considerations that bear when a state is deciding whether to join an amicus brief, because it does affect whether this is a two-state amicus brief or one of the mass amicus briefs--[tape ends].

--but also considerations are far more mundane.  Was the brief well written?  Was the amicus brief circulated too late to the states to give time to work through the internal state process?  How many other states have joined the brief?  Some states feel more comfortable joining the amicus brief.  That twenty states have joined some are happy to be part of our two or three-state amicus brief.  Does the State Attorney General need the Governor's approval?  In most states they don't and AG is an independently elected officer and does what they do. 

But some states, either as a matter of formal policy or just collegiality, will run it through the Governor's office.  And finally, they will look at the legal issue and say do we agree is this tenable.  It is judicial credibility and we don't want to lose credibility and join a brief that we think is making an argument that just doesn't apply.

And, of course, that is also is a reason of joining a brief.  In writing a brief, going back to that part of it, states don't want to look bad and write a brief in a case they think the state has a losing position, would have a losing position.

Okay, issue number two, on how is this played out in practice?  And I looked at the cases that Michael provided by the preemption by the Rehnquist Court in the past seven terms and then I looked also at this term where there are three preemption cases by my count.  And, as always, there are cases that really are hard to decide.  Is Alaska v. EPA a preemption case?  I don't know.

There were 32 of these cases in the past eight years, during my time in NAAG.  Twenty-two of them had a mass state amicus brief, so about two-thirds.  Of those twenty-two cases, four of them involved common law claims, which is what was being allegedly preempted was a plaintiff's common law claim.  Eighteen of them involved statutory claims, state statutory regulations that were quarterly preempted.

Seven of the twenty-two cases that states were involved in were ERISA preemption cases.  There have just been a lot of them over the past number of years, which is unfortunate for me.

In terms of supporting petitioner versus supporting respondent, during this time ten of the mass amicus briefs supported petitioner, twelve supported respondent.  So a little different then the statistics Michael had.  Maybe it had to do with what the states had to do with the 1994-95 term before I go there or something else.  I don't know.

Of the ten cases where the states abstained from filing mass amicus briefs of preemption cases, they were evenly divided between statutory and common law claims.

Now what conclusions do I draw from this?  One might conclude from these totals that my theory I mentioned earlier is confirmed that the states are, in effect, more likely to participate in preemption cases when the preemption law is positive law-- statute regulations rather than common law.

When I looked at the cases a little more closely, I concluded I couldn't really say that.  What I drew from my review is that looking very closely at these ten cases where the states stood out is that the states pretty much file a mass amicus brief in every preemption case except for what I will obscure ones.  This is a rundown of some of the cases where the states didn't file a brief.

Boggs v. Boggs [ph.], does ERISA preempt a state law allowing a non-participant spouse to transfer by testamentary instrument an interest in an undistributed pension plan benefits.  Foster v. Love [ph.], this is about the unique Louisiana open primary statute under which the federal offices are filled before the federal election date.  ELAL Israel Airlines v. Senegal [ph.], when the Warsaw Convention allows no recovery for an airline injury, does that preclude the passenger from maintaining an action under state tort law?

And the last one after all of that is Entergee v. Louisiana Public Service Commission [ph.].  Does a tariff that gives discretion to the utility to determine the precise cost allocation preempt a state public utility's commission order that adjudges those costs imprudent?

You can understand why the State Attorneys General were not using the resources on these cases, which isn't to say that an argument couldn't be given in these cases for why there is a state interest there, but I think you can understand.

And so it is hard to draw a theory about why the states are sitting out.  The real answer is the states are participating in every darn case where it really appears that this is a possibly important preemption case.

Conclusion two, as I mentioned, Michael noted that states appeared to be filing amicus briefs more often for petitioners than respondents, which appeared counter-intuitive because Professor Merrill's remarkable survey of amicus briefs suggested that amicus briefs, at least state ones, but generally tend to be more helpful when you support a respondent.

And the third that was positive which I think makes absolute sense is that if you file a cert petition--and I hope I am not stepping on what you were going to talk about--when you file a cert petition and it is granted, you have shown at least some ability to litigate well in your Supreme Court.  As a respondent, on the other hand, you already lost what usually respondents prevail in cert granting.  So you have already lost the first step.  So there is a tendency to think that possibly respondent's counsel generally are not as good as petitioner's counsel.  And there are some number of cases where respondent's counsel are not that good.  So it is better to get in and help the poorly represented respondent.

Whether or not that is the case, in any event, the states do appear, at least in recent years, during the Schweitzer tenure and AG, are doing the wise thing and helping respondents more often.

But when Michael did mention the statistics, why the states are helping petitioners more, it did help me ponder why that would be the case.  And I did have a theory, which is that often states will file amicus briefs at the cert stage of Supreme Court cases.  And when they do that, if cert is granted, they will tend to file an amicus brief on the merits supporting that same position.  So there is a built-in mechanism to have some number of cases where the states will not even have to make a decision.  They are going to file the amicus brief supporting the petitioner.

And in the ten cases during my time in NAAG where the states did file a pro-petitioner amicus brief in a preemptive case, five of those they filed the amicus brief previously at the cert stage for what it is worth.

Is there home cooking?  Have the states been filing amicus briefs supporting home plaintiffs or businesses?  Only three times during these past eight terms did I find that to be the case in a preemption matter.  And two of them were cases where it was a legislative enactment statute of the state that was arguably preempted, the example being this term's Aetna v. Devilla [ph.] case where Texas filed an amicus brief and actually argued amicus during oral argument to defend state law which made it easier for HMO customers to sue the HMOs for decisions not to provide medical service.

So I am not sure I want to see too much home cooking there even if the plaintiff happened to be a Texas resident.  And there is only one other preemption case that I find where the lead state was the home state of the plaintiff.

Just a couple of other quick points during my time.  Are there too many state amicus briefs?  You know, some have speculated that the value of state amicus briefs is diluted by the states filing so many of them.  And in terms of micro signalling, I think that is certainly a reasonable concern.  State amicus briefs would certainly have a more powerful signal to the Supreme Court if the states filed them few and far between where they say that this is the time where we think it is such an intrusion on state prerogatives on state preemption that we want to bring it to your attention.

On the other hand, as Michael showed, the real powerful signalling is when a government files an amicus brief that is against its presumed interest, like when the Solicitor General says there is not preemption here.  It has been very rare for states to file amicus briefs saying there is preemption, and I think it will be very rare in the future.  So in terms of signalling, I am not sure it really will matter that much in the end.

I think that then leads us to the broad question:  Are state amicus briefs as a general matter helpful to the court?  This is something of the $64,000 question.  As a general matter with respect to amicus briefs, it is always hard to tell.  And this is what Professor Merrill's monumental study looked at.  You know, it is hard to tell.

You know, I looked at it on a case-by-case basis.  Someone says you have got to file an amicus brief in this case.  And I say, all right, what are we going to write about?  And in every case, it is really a great challenge to say what are you going to write about that is not going to be in the brief of the parties?

And so, I have generally found that there are two kinds of amicus briefs by states that are very helpful, putting to the side Solicitor General briefs that are always important and helpful.  One is, as I said, when you look at the party and their lawyer just isn't very good, and, alas, even in the U.S. Supreme Court it happens a fair amount.  And when I help states on amicus briefs that I thought were real nice solid briefs, and they were the only good nice solid brief on that side of the case.  So that is a situation where it will be helpful.

The other one is I would say the Brandeis brief, even if it is not a pure Brandeis brief, at least a brief that gives the court some real world example of the consequences of an adverse ruling.  The example that really became prominent last year was in the affirmative action case where there was the brief from the military officers who said this is going to affect the military if you hold that affirmative action is unconstitutional across the board.

I have heard the justices says they acknowledged they live in something of an ivory tower.  They want to hear what is going to happen in the real world and that kind of amicus brief is always going to be useful.

If there are a couple of minutes and I am not being elbowed by Professor Merrill, I will quickly say what the role of NAAG is, because I think there might sometimes be a little bit of confusion about what the National Association, this collection of State Attorneys General does, generally but I will talk about it in terms of amicus briefs.

The Association itself consists of the 56 Attorneys General of the states and territories.  But I and some of my colleagues are the staff of NAAG.  And what we do is help the Association, help the Attorneys General.

So one thing NAAG does not ever do is to make the decision of whether or not to file a Supreme Court amicus brief.  I will facilitate, I will discuss, I will advise, but I will never make the decision about filing the brief.  Each individual State Attorney General will make that decision, whether to write and whether to join on his or her own.  So what do I do?

A large part on amicus brief is logistics.  If the state is going to be writing an amicus brief, they will use me to help spread the word.  They will send me a memo.  I will then e-mail it out to all the amicus contacts in each state, letting them know that a brief is being written.  When they finish the draft, they will e-mail the draft to me, and I will e-mail that out to all the states so that they can have the draft to decide whether to join.

Yes, I went through three years of law school to do that.  It is not the most taxing part of my job, but it is an important service to the states and it doesn't take up too much of the day.

The funner part of what I do is editing state amicus briefs.  Some states give me their briefs to edit.  It is based on personal relationships, on the needs of their office other states don't.  I offer my editing services to all states.

If there are states on both sides of a case, I will still provide the circulation role.  Every state understands when I circulate an amicus memo or circulate a draft brief, that is not an advocacy of the brief by NAAG; it is just the role we play.  On the other hand, I won't edit an amicus brief if there are states on both sides of it, because that seems to be going over the neutrality that I do institute in what I do and that NAAG generally does.

And finally, we will sometimes organize conference calls of interest to State Attorneys to discuss whether an amicus brief should be written and, if so, what it should talk about.  But again, I will not be the motivator behind it, but I will try to facilitate the conference call.

Now that you know more about state amicus briefs than you ever possibly thought you would or probably ever wanted to, I will turn things over to Professor Merrill.

MR. GASAWAY:  Thank you, Dan.

Why don't we go straight to Mr. Merrill.

MR. MERRILL:  Thank you.

It is very much a pleasure to be here and to be able to respond to Mike's paper.  I was telling Mike before we started that he will never be thanked enough for the labor that he put into this paper.  I personally know how much effort it is to do these quantitative judicial studies and to do them with the care that Mike has done, or try to do them with the care that Mike has done.  He will be praised, but not nearly as much as he should be for the effort that he has done, because this adds tremendously to our knowledge about a very important set of cases in the Supreme Court.

My comments are going to be sort of miscellaneous.  I will organize them into three categories.  First, I am going to talk a little bit about the model of judging that underlies Michael's study and how that might be sharpened a little bit and how it might be used to provide maybe some additional explanations for some of the puzzles that he uncovers.

Secondly, I will have a few comments about some of the empirical findings and the interpretation of those findings.  And lastly, I will very tentatively throw out a thought about sort of normative theory of preemption or the legal document we apply in resolving preemption cases and maybe how we might take some baby steps toward doing a little better than we do in the way we argue brief preemption controversies.

Okay, first on to the model of judging.  Now this may seem a little bit academic and I hope it is not overly-academic.  But political scientists have developed a variety of theories of what motivates judges to behave the way they do.  And these theories are referenced a little bit in Michael's paper and I think they are utilized but not as explicitly as they might be or as carefully as they might be.

Michael's basic thesis here is that we can explain these preemption cases at least in part to a signalling theory.  Sometimes political scientists call this Q theory.  But the question that that raises is:  What is being signalled and why is the signal being read the way it is?  And to answer those sorts of questions, we need to have a theory about what makes judges do what they do.

Now as Michael explains in the paper, a political scientist had developed something called the attitudinal theory, which drives most lawyers nuts in the way it is usually expressed, which is basically the judges simply try to maximize their personal policy preferences.  They talk law but the reason they vote the way they do is because they have certain ideological things that they are trying to advance in their decision-making, and the law is just sort of a cover for that.  So that is one theory.

Another theory is what most lawyers would espouse, certainly at a confirmation hearing, which is what the political scientists call the legal theory, which is the idea that judges are just conscientiously trying to get the law right.  They are trying to decide these cases correctly, given the complicated norms and conventions of the American legal system.

And the third theory which is the most recent one to be developed by political scientists, and really sorts of builds on the earlier two is the sort of strategic actor theory.  And this approach starts with the insight that judges don't act in splendid isolation.  If all we had was the attitudinal theory or the legal theory, you would think that judges would act reflectively in cases.  They would read the briefs, listen to the arguments, and then they would just vote based on either their political preferences or their sincere convictions about what the law requires.  But the strategic model points out that in a collegial court and certainly in a Supreme Court it takes multiple judges to produce a legally binding precedent.  So at the Supreme Court level, that is typically five.  You have to get five votes to make a binding precedent.  So this entails a need for certain cooperation among judges and justices, and they have to temper their reflective reactions to cases, whether they be attitudinal reactions or legal reactions in order to put together a coalition in support of a single position.

Other political scientists pointed out even further that in order to make these decisions stick, in order to really have your legal or attitudinal preferences realized in the world, you need the cooperation of other branches of government.  The Executive has to enforce the judgments.  Congress has to be dissuaded from overruling them.  The states have to comply.

And so, the strategic model would only suggest not only that there be an internal focus among judges toward cooperative alliances but also that there would be some attention paid to the way other branches of government react.

I think Michael's paper either explicitly or implicitly relies on each of these three theories.  But it kind of shifts from one theory to the other as he goes through the explanations.  Maybe that is okay, but I think he should be a little more explicit about exactly which theory he is using and why this kind of shifting is permissible.

Now when he talks about the Solicitor General and in particular about the signal that is sent when the Solicitor General disclaims preemption in an amicus brief, the explanation is based on the legal model of judging.  The explanation is that the justices think that the SG is largely captive of federal agencies that are trying to expand their power.  And so, lo and behold, if the Solicitor General comes along and disclaims preemption, wow, the law must be really clear on that point that preemption is not appropriate and, therefore, the justices take that as a signal that the correct legal view is the anti-preemption view.

So after that part of the story, Michael is sort of using the legal model to explain how signalling work.  When he gets to the state as a party, it is a little less clear.  But I think here he is sort of invoking an attitudinal model, because what he says here, if I understand correctly, is if a state is appearing in a Supreme Court in a case arguing against preemption, that the justices will respond by thinking that the state, you know, is really seriously concerned about its prerogatives in federalism values and because the justices, at least a large number of them, are sympathetic to those federalism values, they tend to defer to the state somewhat.  So that is an attitudinal explanation I think that many justices are sort of sympathetic to states' rights and, therefore, take the signal of the state as a party to be a sign that that value is implicated by the state.

When he gets to the amicus briefs, at least when he gets to the states, the NAAG enterprise that Dan Schweitzer has described, filing amicus briefs in these cases, he is not explicit.  But I think implicitly what he is invoking is the kind of external strategic model which would suggest that the more states sign on to an amicus brief or the more state amicus briefs there are in a particular case, the more likely it is that the justices will rule against preemption.  Now this finding is not statistically significant, but nevertheless it is in the paper.

And the explanation, at least implicitly, seems to be that the justices are impressed by the larger number of states that are filing and it must be that the states really care about this.  So it is more costly for us to rule against them because we are going to make a lot of states mad.  And to the extent we want states to be cooperative and to think that our judgments are worth following, there would be some costs if we rule against the states where 48 of them think that is preemption is not warranted.

So I think the various signalling explanations are all linked up to different theories of judging.  And Michael should perhaps work a little bit on tightening that aspect of the paper up.

Now the strategic model of judging is of interest to me.  Let me suggest very briefly that I think the strategic model in particular supplies the best explanation for the largest paradox that this study presents, which, of course, is the paradox of how is it that what I call the second Rehnquist Court and Michael calls the late Rehnquist Court and Dan calls the Schweitzer Rehnquist Court, how is it that court of late has produced a very stable coalition of five justices who routinely vote on the state's right side, Tenth Amendment, Eleventh Amendment, Thomas clause, Section 5 of the Fourteenth Amendment type cases, but does not produce similar behavior in preemption cases?

If you are at all theoretically inclined, you find this to be baffling because if federalism matters, it would seem that preemption cases, if anything, are more important than the Eleventh Amendment, because preemption determines the relative scope of regulatory authority between the states and the federal government.  If you believe in regulatory competition, if you believe in the states as laboratories of experiment, preemption probably matters more than the Eleventh Amendment does.  So why do we have this great anomaly?

And my answer, and I will try to give this as briefly as possible because I don't want to get too diverted into the official federalism cases, is that I think the justices are being strategic in those cases, but they are not being strategic in the preemption cases.  And the story is a little bit complicated.  But if you go through each of the justices in the majority, you will find evidence that some of them I think are true believers in states' rights.  Chief Justice Rehnquist I think falls in this category, and a number of have demonstrated over the years that they have either proclivities or twinges in that direction like Justice O'Connor and Justice Kennedy.  Justice Thomas seems to be ideologically committed to states' rights, but Justice Scalia is a true anomaly in view, because throughout his career as a federal civil servant, as an academic, as a D.C. Circuit Judge, he never showed any inclination towards states' rights.  He has always been a kind of nationalist.  He is a Hamiltonian conservative, not the Jeffersonian conservative.  And his behavior in preemption cases which Michael has documented also suggests a lack of support for states' rights.

I have argued in my lecture that Michael made brief reference to is that Justice Scalia has been behaving strategically in the last ten years by consistently voting with the other four conservatives in these official states' rights cases.  And that has created a solid majority that we have seen.  And if you like, in the question period I can give you some more evidence in support of that.

But the point here would be that the difference between the official federalism cases and the preemption cases are ones where I think the justices are not behaving particularly strategically.  They are behaving attitudinally or legally.  They are being reflective.  They are just sort of responding to the cases I think they individually respond.  And so it is a situation where we see what the court would look like without strategic behavior.  It is a mess.

In English parliamentary terms, it is like the coalition has declared that preemption cases are a free vote without party discipline.  And so everyone is free to vote their conscience.

And so what I think we see in the preemption cases is a very complicated picture where all these different values and attitudes, attitudes about tort suits and the tort bar, attitudes about statutory interpretation and the correct interpretation, federalism values are sort of put into the blender and we get the sort of mismatch that we get.  So that is what I want to say about theories of judging and how they might apply here.

Secondly, empirics, I have three quick points to make about the empirics.  One has to do with the relative importance of state as a party and tort suit in explaining these outcomes.  Now, Michael does this one regression which he put up there in one of the slides, where he puts in both tort suit and the state as a party into the equation, and shows that the state as a party loses significance when that happens.  And he says that he thinks this is because there are very few cases where states are parties when tort suits are implicated, and what is really going on here is that the state as a party is a signal.

Now I think that is not so clear, and that you could argue it the opposite way.  You could say that what really matters is whether or not it is a tort suit, and that the justices are more kind of interested and committed to either furthering or retarding the tort bar than they are in supporting or not supporting states in preemption cases, and that, therefore, the reason we see the state as a party as being significant is that that is a proxy for it not being a tort suit.

So I think there are dual explanations there that would point to different variables as being explanatory.  And I am not sure that the statistical paraphernalia that Mike has used are sophisticated to weed out which one is actually doing the laboring here.

Not to send Mike out to do even more research, but one interesting thing that you might think about doing or that somebody might think about doing would be to try to do a study of the justices' attitudes towards tort cases in the non-preemption context.  Now the court decides a number of cases that are tort like.  FEAL cases are the clearest example but also Federal Torts Claims Act cases, Superfund cases and so forth.  So perhaps somebody could look at all the tort-like cases that justices decide that don't involve preemption and see if you could come up with some kind of model of how the justices behave in those cases and then use that to sort of try to differentiate out between whether it is a tort suit or state as a party that is driving the outcomes in the cases where states are parties. 

The second point has to do with the very interesting finding that Mike put up on his screen that the affirmance rate is going down in the second Rehnquist Court in preemption cases; in other words, without regard to the lower court outcome, the court is reversing a much higher percentage of these preemption cases than it was during the first Rehnquist Court.

And I would like to suggest, this goes back to my lecture in St. Louis again, I would like to suggest that a possible explanation for this is that the second Rehnquist Court is a judicial machine, the efficiency of which we have not seen for a long, long time.  And there are many reasons why it is efficient.  One is that the Chief Justice is pretty efficient.  But the main reason why it is so efficient I think is that the nine Justices have been sitting together now for almost nine years without any change in personnel.  And the reason that makes that court so efficient is that every case the court decides, there is a little experiment where each justice is looking at the either other justices and sort of learning a little bit more about their values and attitudes, whether they be sort of policy preferences or legal theories or whatever.

And so after a number of years go by and you sat with the same people over and over again, you begin to know a huge amount of information about how these people react in different circumstances.  And so I think a lot of what has happening with the Rehnquist Court was the declining case load, for example.  One of the findings that I made in my St. Louis piece was that the number of plurality decisions has gone way down.  I think the explanation is that court when it looked at a candidate is just better able today to predict--each justice is better to predict how the others are going to react to this case.  And so they can basically behave more efficiently in the sense that they have a much higher percentage of reversals.

They take cases when they are going to reverse.  If they are going to affirm, to some extent it is a wasted effort.  You might resolve a conflict, perhaps, but to some cases taking a case to affirm is not accomplishes as much as taking a case to reverse.  So I think the overall efficiency of the court has kind of explained the somewhat puzzling finding that the affirmance in preemption cases is going.

Lastly, on the empirics, this gets in a little bit to the puzzle about amicus briefs that Dan was discussing.  I agree with what Michael says and what Dan says about this.  And let me just throw in a couple of other ideas about what appears to be the puzzling phenomenon that all of these people are busy filing these amicus briefs and there is no evidence that they have any significant effect on the outcome at the Supreme Court.

The first explanation which I think justifies Dan's enterprise is it may be like an arms' race.  It may be that if amicus briefs were only filed on one side but not on the other side, that that would, in fact, have a significant effect on the outcome in favor of the side that has the amicus support.  But people have figured this out and so the petitioner anticipates that the respondent is likely to file an amicus brief.  So the petitioner solicits amicus briefs.  And so you get amicus briefs on both sides of the case.  And whatever additional information or signalling it may provide is kind of neutralized or cancelled out through that process.

And our study in the Penn Law Review did show that there is a great symmetry in these cases.  The highest difference between respondent briefs and petitioner briefs in a Supreme Court case is zero.  I mean, the highest number is where they are exactly equal on both sides, and then the next highest is the disparity of one and it falls off quite rapidly.  So the number of amicus briefs tends to be very similar on both sides, and I think that is some evidence in support of the kinds of arms' race type theory.

But what that means is that it is totally rational for Dan and his Attorneys General to sort of file briefs in these cases, because, if they didn't, the preemption forces on the other side would file amicus briefs and they might increase their win rate.  So even if you cannot see in the statistics in effect an outcome, it may still be rational to file.

The second point would be just a suggestion that to some extent this phenomenon may be lawyer driven rather than client driven.  And again, I think Dan made a few comments somewhat consistent with this, whether you are interested and you got free time kind of point.  And I would suggest this, that Supreme Court litigation is glamorous work.  Most lawyer litigators would love to be involved with this if they could, but we have a Supreme Court that is deciding half as many cases than it did during the Burger court years.  And so there are fewer of these cases to go around.  And there are more lawyers who would like to do this glamorous work.

And so one way to do it is to sort of gin up amicus brief clients and file amicus brief as a surrogate for being a real Supreme Court litigator.  And there may be various rational reasons for doing this.  This is the psychic income of being able to tell your friends that you are doing Supreme Court work.  There is advantage in recruiting associates to you office if you can point to all the Supreme Court work you do, and there may be advantages to clients to point out your Supreme Court filing.

So I don't have any hard data on this, but I suspect that there may be some subtle lawyer solicitation of these briefs and some maybe price discounting that goes on by lawyers to file these briefs.  So to some extent the amicus filings may be lawyer driven.

All right.  Point number three--and I will try to be quick--is normative preemption theory.  It seems to me beyond controversy that the current legal document on preemption cases is largely worthless.  Everyone is familiar with the litany.  You have expressed preemption versus implied preemption, and under implied preemption you have conflict and field preemption, and you have various canons such as the traditional police powers of the states are not to be set aside absent clear and convincing evidence from Congress that intend to preempt and blah, blah, blah.  And there is no evidence that any of this is anything but boilerplate and has much influence on outcomes that the courts reach at all.

So is there any way that we might perhaps do better?  And I am not being utopian here.  I am not suggesting that somehow we can invent a legal document that would actually control the outcomes of these cases, because, as Michael has demonstrated, they are very complicated and there are a lot of things going on here.  But maybe we can at least do a little bit better in the sense that having a doctrine that would point lawyers and judges towards factors that would have some constraining force in the way these cases are talked about and resolved.

One suggestion I would throw out is that perhaps we could look at the pattern of express preemption provisions that Congress has actually enacted and express savings clauses.  And looking at those Congressional enactments, we can try to derive some generalizations about which Congress tends to think preemption is appropriate and tend preemption is not appropriate, and we could use those generalizations to create some canons or presumptions that would be applied by courts in ambiguous cases in trying to resolve preemption controversies.

So let me be a little more specific and I will give you the only example I have.  And it comes from the environmental area.  And if you look at environmental statutes--this just happens to be something I know something about--if you look at environmental statutes and their preemption and savings clauses, I think you can discern that there is something like a one-state rule that applies.  Congress thinks to think that in a circumstance where a phenomenon, if the states had power to regulated, would likely be regulated by only one state.  But that is a circumstance where preemption is not appropriate and Congress will pass some kind of savings clause.

On the other hand, if it is a circumstance where the phenomenon would likely be regulated by multiple states with possibly inconsistent outcomes, then Congress tends to pass an express preemption clause.  So under the FIFRA statute, the Federal Insecticide, Fungicide and Rodenticide Act, Section 24 has both an express preemption clause and express savings clause.  And what is expressly are state laws and regulations governing the use and application of pesticides.

And so, Congress may have imagined that applying a pesticide means spraying it on a field.  The field is likely to be located some place in one state and, therefore, only one state or one local government is likely to have authority to regulate the use and application of pesticides.  So that was specifically saved from federal preemption.

Then there is another clause that expressly preempts state and local laws dealing with the labeling and packaging of pesticides.  Now labeling and packaging, you can imagine, is something that happens at the factory before the pesticides have been shipped across the country and if the states could regulate that, you would have a bulkanization with different states having different requirements that would complicate the production process.  And Congress may decide that the benefits of state law are outweighed by the costs in that circumstance.

And you can see similar things in the Clean Air Act.  The Clean Air Act saves state regulation of stationary sources.  Again, Congress might have imagined stationary sources to be fixed and regulated by one state.  Mobil sources with some complications about California that I won't get into are expressly preempted.  No state regulation of mobil sources.

So if you took this pattern and you drew the generalization of the kind of canon or presumption in circumstances where one state can regulate, we would tend not to preempt, but if more than one state could regulate we will, you might be able to use that to decide contested cases.  And the one example I would give you is International Paper v. Lett [ph.], one of the cases in Mike's database, where the Supreme Court was confronted with the question of whether under the Clean Water Act a victim of pollution in another state could use the victim's home state common law to sue the source across the state line.

And the Supreme Court said no.  They said that that suit is preempted and that the victim can sue but has to use the common law of the source state, not of the victim's state.  And the court gave some reasonable policy arguments in support of this, but they didn't really cite anything in the statute that supports the decision.

But my suggested canon about one state law as being applicable it seems to me would also yield the same conclusion in that case, because if a victim's common law could apply, then also the common law of the source state could apply.  And if there are victims two states down the road, their law could apply.  So you could have multiple states applying if you allowed the victim to apply the victim's state common law.  And so that would suggest that under this generalization from what we see in enacted preemption statutes that preemption under that circumstances is appropriate.

So I have one suggestion on how we could get started on crafting a substantive preemption law that would focus us on real issues that differentiate when we should and not preempt in perhaps a better way than the official document does right now.

Thank you.

MR. GASAWAY:  Thank you, Tom.

Mike, are you burning to respond, or should we take some questions?

MR. GREVE:  I agree with almost everything, especially the suggestion that this paper is under-theorized.

MR. GASAWAY:  Are there any questions from the floor?

QUESTION:  They are all fascinating thoughts.  I am not sure which question to ask, but I will jump in with one, mostly to Michael.

I found especially fascinating the SG participation rate, particularly the sort of abstention rates between Democrats and Republicans.  But there are a few other questions I have about the SG participation.  It also is sort of exogenous as to the merit of the case being presented.  I mean, the SG staff supposedly actually reviews these cases and makes some decisions.

Thomas mentioned that the possible signalling effect is particularly strong in cases where the SG says there is no federal preemption regardless what the theory is.  Might it not also be the case in those that the SG joins in because the facts are really strong, the case is really an easy one?  And it is an easy one regardless of whether the result was right or wrong.

Now I forgot what I used to know about regression and whether you have taken this into account.  But it didn't appear to me on the quick run-through that you factored in.  Another factor that would cut against the signalling factor in the SG saying no preemption.

There may be a federal interest as opposed to the states that never seem to see an interest in federal preemption.  I don't see why you assume it is the case that that there is another explanation that cuts against the signalling.  The federal government may in some cases want states to be in the business.  It may be in the federal interest that they not have a preemption.

MR. GREVE:  I suppose it is possible.  I mean, one thing that cannot be explained on the basis of the legal factor is the striking partisan differences in participation rates.  Unless they were looking at very, very different universes of cases, that is really a striking pattern there.  And that must have to do with factors other than the legal niceties of these cases.

With respect to everything else, I agree with these theories and the signalling theory is both under-theorized--Tom is right about that, as I said--and reductionist.  That is the nature of the beast here.  It cannot exclude the legal factors here at all.  I mean, one of the messages of this, quote/unquote, "theory" is that I cannot tell the signalling value from the informational value that is being transported here.  And that is true with respect to amicus briefs, state participation and overseas participation.

So I readily admit.  This was a first step to bring some order to this universe and start with some reasonable hypothesis.  I made a point at the end of my talk to emphasize I cannot exclude the legal model, as it is called, at all.

MR.          :  Let me just comment briefly on the SG because one of the unanswered questions in the studies of the SG, all the studies shows that the SG has a huge success rate, most of the litigants and as an amicus.

But somebody really needs to do some more disaggregation, because the unanswered question is:  To what extent is the SG success rate based on the fact that this is just sort of a tenth justice theory, that staff attorneys in the Solicitors General office have become captured by the Supreme Court and are sort of acting like a law clerk to the Supreme Court, having internalized the court's own preferences and legal theories and biases.  And so they simply push the federal government to adopt whatever position the court is going to find the most reasonable position to be and, lo and behold, they win.

I mean, I was in the SG's office and I think this is not entirely implausible, because there are some cases that the SGs are preoccupied and the top deputies are preoccupied with, and there are always these preemption cases that come through, and the assistant to the Solicitor General has a large amount of discretion in deciding on what the position of the government is going to be in those cases.  They can say that the law is quite clear on this side and, therefore, even though our client agencies want to do the opposite or all of these lawyers from expensive law firms are in here lobbying for preemption, unfortunately the law just doesn't support this, and the higher-ups may just go along with that.

And so the SG may be highly successful simply because a large chunk of the SG filings are simply aping the court's own values and preferences and it is all kind of circular.  And we really I don't think have had any definitive studies to disclaim that possibility.

MR.          :  Mike, let me make one more point on the SG.  What you might do following up on Dan's suggestion is look at the cases where they filed with the certs.  I mean, that abstention rate for Republican SGs is amazing and somewhat disconcerting, especially when you think that there are probably SG briefs from a SVSG at the cert stage.  And in a lot of cases, they grant it.  And just like Dan was saying if you file a cert brief, you are more likely to file a demerit brief.

You know, in many of those cases where they did file, they were roped and collared into filing at the cert stage, and then they are probably going to file at the demerit stage.  So if you wipe out all the cases where Republicans were forced to file and you sort of had the discretionary filing cases and looked at the abstention rate, you are going to be looking at 50 percent abstentions or something like that, I would guess, which is truly amazing when you consider the practical effect of these cases.  It is not as sexy as the Pledge of Allegiance or school prayer or whatever, but Geier is a hugely important case for the economy.

QUESTION:  I was just wondering, on abstention,  is if you had a Democratic SG during the first years of the court taking fewer cases, and I think the SG as a general matter falls in a much higher percentage of cases because, like the lawyers themselves, they're looking for work to do.  I just wonder if that plays into the figures, that Democrats abstain less because there were fewer cases during the Clinton years.

MR. GREVE:  I would have to look up the numbers.  My recollection is that--[end tape].
--it's roughly even.

QUESTION:  Michael, what do you mean by "abstention," exactly?  You mean they just stood out the case, they don't file?

MR. GREVE:  Yeah.  There's no brief.

MR. GASAWAY:  Judge Williams?

QUESTION:  I just wanted to put together things that Tom and Michael said, starting with Tom's observation that the doctrine is worthless.  I'm so glad to hear that, because--

MR. MERRILL:  Only in preemption cases.

QUESTION:  I mean, I've never been able to make head or tail of it.  --with Michael's observation that, or his sort of discernment of a Breyer bloc emerging from Breyer's being uniquely capable of grasping them.  So that a question arises, is there a developing Breyer Doctrine that maybe sheds light on it and could explain--in the future could be adopted and explain outcomes?

MR. GREVE:  This is risky, because I'd really have to go back over the opinions.  But my general sense is that the better an appreciation a justice or judge has of the particular regulatory regime at hand, the less likely it is that he's going to crank out some grandiose preemption doctrine that works for a telecom case today and an [inaudible] case tomorrow.  I mean, these are, really, very unique regulatory regimes.  The political dynamics vary greatly.  And at the end of the day, you may ask yourself--I mean, one of the reasons why preemption law, quote-unquote, is so messy, my hunch is that there may be no such thing as a generic preemption case.  I mean, it's all tangled up with--it's always about individual, very complex and complicated statutory regimes.

So if there is an emerging Breyer Doctrine, maybe it's the doctrine that we should be very circumspect before we leap to grandiose theories rather than taking a look at the individual regulatory regimes.  That doesn't mean ad hoc-ery; that just means if you want to decide a case like that, you had better damn well understand it first.

MR.          :  One of the things I--I'm not an expert on Justice Breyer, but one of the things that I would point out about Justice Breyer is that he's skeptical--I don't think he likes juries, and I don't think, for that reason, he's so crazy about tort law.  I mean, he's written, you know, like "Breaking the Vicious Circle," his book on risk assessment and so forth.  He's very keen on having these complicated cases about, you know, social risks and regulation of risks being decided by experts--you know, people with Ph.D.s and fancy titles in administrative agencies.

And that's a rather peculiar perspective, which I don't think anybody else on the court shares.  But what it does in some of these preemption tort cases, like Geier, is that it makes him an oddball ally with the sort of, you know, anti-tort-lawyer, pro-business justices.  So it kind of makes him a kind of free radical in the tort cases that can be used to form these coalitions, with Clarence Thomas being much more kind of a firmly committed states rights ideological justice.

Another case that tends to confirm this is, there was a case last term, I think it was called Norfolk & Western Railway, which not a preemption case, it's an FELA case.  And if I remember right, the majority opinion was written by Ruth Bader Ginsburg in favor of an expansive tort doctrine.  It had to do with whether or not you can recover for fear of cancer if you has asbestosis.  And she writes for the majority that says yes.  And it's joined by, of course, Stevens and Souter, but also by Scalia and Thomas.  And then the dissenting opinion by Kennedy is joined by Rehnquist and O'Connor and Stephen Breyer.

So that, again, suggests that Breyer is--and this is one reason why I was suggesting this theory of the study of tort cases, that Breyer has a kind of elitist skepticism about tort suits and juries and so forth that causes him to make common cause with some of the kind corporate bar-oriented conservatives.  And it may not be anything more profound than that, that he doesn't have any sort of great theories of preemption, but that just may be what's going on.

MR. GREVE:  Because he commented on the empirics earlier, maybe John Plik remembers what we actually did there.  We ran the number on state participation and tort in various permutations and then concluded, no, it's really the state-participation variable that drives this.

Having said that, I wouldn't want to rely on the math.  I think your suggestion about looking at FELA cases, Superfund cases, tort-styled cases over the same period here, it's a terrific suggestion, and that case illustrates it.

QUESTION:  So Professor Merrill's suggestion that some of the SGs' decision about whether or not to file might depend on agency preferences.  But you could also say that under a legal model of judging especially, especially for some justices with a broad view of deference to agency decisions, that that could explain--I mean, it might be worth investigating whether that explains some of the outcomes or some of the votes.  Was there any thought given to tracking agency--sort of agency interpretation of the preemptive scope of the law as against votes and outcomes?

MR. GREVE:  Thought was given.  No work was done.  It's a perfectly sensible suggestion.  It just--you know, you try to handle a data set like that, that we didn't pursue that.

MR.          :  Do you have any idea how many cases, though?  I mean, Geier brings to mind, as a case where there was a position that the NHTSA had taken on air bag product suits.  But my impression would be that that's kind of rare, for a federal agency to actually have spoken to the issue that comes up in a kind of a, you know, tort context.  Because frequently, you know, the agency regulations are focused on agency implementation enforcement, which doesn't completely track the causes of action that, you know, come up in the tort context.  But maybe it would be a good suggestion, if there are enough cases, to look at those.

QUESTION:  I guess this is tying in comments from all three of you, so I'd just jump in whomever.  One of the things that was mentioned earlier was the sort of uniqueness of someone like Scalia.  And Scalia in this particular area is kind of rare, because he's publicly talked about this issue.  He has come out publicly and really criticized members of Congress and state legislatures for not reading their constitution.  At some level, these preemption cases represent a failure of the law-making process.  They represent failure by legislators to anticipate properly what the court might do, anticipate what the balance is here between the federal and state power relationship in these areas.

And I guess one thing that--I mean, this is certainly beyond the scope of this project, but in the end, we're taking all these preemption cases as sort of equal units of one, where they're coming out of different states, where they're coming out of different, sort of, dynamics going on in those states in terms of to what extent each of those cases represents a failure in those states.  I mean, at some level, some of these states are much more constitutionally deferential to the court and the Constitution than others.

Going to the points raised by Dan from the NAAG, at some level in some of these states, another factor that might influence whether or not the AG steps in is was he there initially telling the state legislature through an advisory opinion, this is good?  Right?  And then if it comes up and it gets challenged, there's enormous pressure on that AG to come in and defend that statute because he sort of green-lighted it.

So I guess I'd be interested in hearing your thoughts on, you know, kind of all that comes before this, before these cases even come to the court in terms of them being by-products of a very sort of messy legislative process involving trying to anticipate what the court and the Constitution will say on these matters.

MR.          :  Well, I guess what that question highlights for me is now much, as has been mentioned, preemption law is the law of statutory construction in a particular context--which is why it's so hard to find a unified theory of this.  And it then raises questions about the controversy between the new textualists and the contextualists, I suppose, about the approach to statutory construction.  Your question sort of brings to mind, I guess, a criticism of the new textualist, which is this notion that the legislature should be so clever as to write a statute that can answer every single question that's going to arise free of any ambiguity.  And I think it's a legitimate concern that it's really impossible to do so and that sometimes, perhaps, it's placing too strict an onus on legislators, possibly AG offices when they're giving their advice, to write the perfect statute that gives a non-ambiguous answer to every question and so we wouldn't have any of these preemption cases because we'd have an answer.

In the end, I think even the new textualist will find some tough cases; even a contextualist will find, you know, lots of tough cases.  And I'm not sure we're ever going to be able to solve the problem by improved legislative draftsmanship--though obviously the point you raise is an excellent one, that the legislators have to be aware of the constitutional federalism jurisprudence as well as the court's jurisprudence on statutory construction when they go about writing their statutes.

MR. GREVE:  I want to comment briefly, and I hope--well, I don't care if it turns out to be controversial; I'll say it nonetheless.

In some of these cases, it's quite obviously true that there's a whole lot of punting involved by the Congress.  As Frank Easterbrook would say, there was no bargain at all; they just papered it over.  I mean, one day, you know, the automobile industry runs to John Dingel and says cover everything and he says yes.  The next day Ralph Nader runs up to him and says cover nothing and they say yes.  And so you have two clauses that are mutually exclusive and cover the same range.  And that is pretty much Geier.  And so you get those kinds of applications.  And then it's a legitimate instinct, I think, for a federal judge to say, Look, now you feds want me to bail you out for imposing on the states what you failed to do the first time around, and do it explicitly.

On the other hand--and this is not always easy to distinguish--one has to remember and be aware that preemption doesn't simply operate on states.  It operates on states that have every incentive and every reason and every tendency to weasel around--right?--what you just did.  That is to say, the minute it goes on the books, there are 50 states that try to figure out how do we find a way around this?  The notion that you can preempt that once and for all and it will never come up, right, and you will never have to mobilize a federal judge to enforce the original meaning and good sense of the statute, I think, is just a fantasy.  So that instinct sort of cuts in the opposite direction.

And I'll say one more thing about that, and this is along the lines of Tom Merrill, about sort of one-state situations, right, which are the sort of traditional police power situations, and then the multistate, where what you're really trying to do is you're trying to get a handle by federal statute on interstate externalities.

The first observation about that is that I think it is lamentable that modern preemption doctrine does not even remotely allude to that kind of a consideration, and I think that is actually something to worry about.  And over that comment, I've forgotten what I wanted to say in the first place.

MR.          :  Let me just say quickly that, you know, I have a lot of problems with textualism myself.  It's either utopian or else it's kind of mischievous.  But one thing, like in the Chevron context, for the court to be textualist, what that means is the court will frequently conclude that statutes are ambiguous and therefore the agency interpretation should prevail.  Because then at least you have a default interpreter who's ab