Votes That Matter
Our commenter David Nieporent remains very confused about the argument I’ve been making for years about Sam Alito, although it’s very straightforward:
The point is that this is, well, an incredibly stupid metric. Why is failing to cast a “liberal swing vote” the sign of lack of moderation? It’s a cherry-picked non-stat created by throwing out most of the data.
Now, if he’s the sole ‘conservative’ vote, or one of two — that is, if he’s on the losing side of a bunch of 8-1 or 7-2 decisions — then that might show lack of moderation. But 5-4 votes, almost by their nature, don’t show anything at all.
If the question is, “Will Alito ever vote against the Chamber of Commerce?”, then the answer, as defined by this study, is yes. So why isn’t that “moderation”?
I don’t see what’s complicated about the argument. There are two sets of data in question. The first set of data shows that Alito is more likely to cast pro-business votes than any other justice, which in itself is inconsistent with assertions that Alito is a relative moderate (assertions that have never had a shred of evidence in their favor to begin with.)
The second set of data is additionally important and isn’t merely arbitrary, because it shows the likelihood that a justice will cast meaningful vote in a heterodox direction. I suspect it is true, looking at cases overall, that Thomas is the justice most likely to write a solo dissent or a more-extreme concurrence in a politically salient case, but these votes matter very little — if you’re bringing a lawsuit you don’t really care if you win 9-0 or 8-1. (It’s not that these votes don’t show a “lack of moderation,” just that they show a less consequential lack of moderation.) On the other hand, if you’re a plaintiff in a tort suit or a criminal defendant, there’s a chance that Scalia and Thomas might actually cast a decisive vote in your favor, while Alito’s vote is in play only if you’ve already won. In other words, this second set of data doesn’t show that Alito never votes against the Chamber of Commerce’s position, and I never claimed that it did. What it does show is that Alito has yet to cast a vote of any consequence against a position advanced by the Chamber of Commerce. Which goes to reinforce the always-obvious point that if you’re a class of person whose interests are disfavored by the contemporary Republican Party and are bringing a lawsuit, you’d actually rather have Scalia than Alito.
As it happens, yesterday’s decisions provide another case in point. Alito did cast a superfluous vote in a habeas corpus case making the result 7-2. On the other hand, there was another case in which Alito could have joined the Court’s most liberal member and three of its other conservative embers to hold that if the state wants to require that restitution be paid to victims the state has to follow its own explicit statutory guidelines. But of course he didn’t, because as long as his vote matters Alito will always side with the state in a criminal case. It’s a consistent pattern.
Nieporent is “confused” because his paycheck depends on his not understanding the obvious, at least in this case.
I really hate the Overton window sometimes. Sam Alito is exactly who all of us DFHs knew him to be.
I think the commenter has a point. By their nature 5-4 votes are difficult to decide, and so you’ve shown that Alito is a difficult vote for the anti-Chamber of Commerce position to get in that situation. You follow Bill James — it’s a little like his argument about good or bad teams in lopsided and close contests. (That is, dominant teams do less well in close contests, almost by definition the weaker team has a better chance if you restrict the discussion to close contests.) Hindsight logic tells us that the players who get hits in those games have produced more “meaningful” hits, but in reality the meaningful players are the ones that consistently put the team ahead 5-0 and give the other team no chance to win. It’s not a perfect parallel, but I get the sense that this idea of “meaning” is similar.
That said, I buy the general argument completely — Alito is not a moderate — and I also buy the first part of the argument, that he is simply the most pro-business justice. I think trying to parse whether a vote is more meaningful in close votes is ambiguous at best. It might be better to see him not as the 5th vote but as the 1st vote in such situations. The premise that he is the most pro-business justice is merely somewhat corroborated by the close votes, they’re not an additional piece of evidence as such.
As you say, 5-4 cases are in some sense hard to decide, which is what makes them a critical variable in this analysis. Hard to decide cases should produce at least somewhat inconsistent votes by justices, particularly moderate ones. Where there is clear consistency, as in Alito’s case, then the person is clearly not a moderate and is, in fact, a politically motivated activist. This was Scott’s original point, FWIW.
I don’t think the analogy holds, because the evidence suggests that the distribution of run-producing events is essentially random, while Supreme Court voting is both strategic and differs across cases. So looking at the data in all cases ignores cases in which a justice may be willing to go along with an already-solid majority but wouldn’t cast a swing vote, and also ignores the very real cases in which Scalia and Thomas but not Alito might act as the median vote in a given case.
Basically, I agree. The better analogy is that those supposedly “moderate” votes in lopsided cases are like big hits in games where one’s team is already way behind — in a context where the batter gets to choose when he gets big hits. Thinking about it, maybe the logic is exactly reversed, just as you say.
I think where it gets confusing to people is that the close votes inherently mean that the conservative side has *some* plausible merit. So Alito votes his conservatism when conservatives seem to have a good case — what could possibly be wrong with that? Again, not arguing here, just explaining how some of the assumptions aren’t 100% obvious.
Right — it is true that Alto being the very worst justice from a liberal perspective is context-dependent. A Court where Thomas is the median vote would probably be worse than a Court where Alito is the median vote: there would be a few more liberal outcomes, but also some conservative outcomes more radical than anything you would get from the current Court. But in the actual context of the Roberts court, a Scalia or Thomas might produce liberal results Alito won’t, while their more radical opinions won’t get 5 votes anyway.
Extremism in defense of the rich is no vice. Moderation in the pursuit of business interests is no virtue.
David Nieporent lacks a fundamental grasp of how statistics works. Unless Alito’s pro-business decisions were a strict superset of every other judge’s, SOME metric is needed and inevitable. The metric of close votes is intuitive and has at least some value. To complain about this metric requires, at a minimum, proposing a competing metric. What metric do you think we should use instead, David Nieporent? (What do you think THAT metric will show about Alito?!) Criticizing a metric won’t make the data change. Show us how we have misunderstood the data.
In particular, I wasn’t buying the accusation of cherry-picking. Scott Lemieux may not have the best metric ever, fine. Present yours, we’ll argue over them. But claiming he’s doing bad things with the data? That requires a bit more proof! Since cherry-picking is not a classical Greek fallacy, it has very nebulous criteria… is David Nieporent proposing that he could select the data in such a way that Alito appears to be the most liberal judge? Anti-business? In favor of women’s rights? If there is only one way to “cherry-pick” the data, it is not cherry-picking. So once again, I find that David Nieporent is failing to do the minimum amount of effort required.
I thought I made my position clear: one shouldn’t cherry-pick the data at all; rather, one should use all the data.
If one wants to argue that Alito is not moderate because he votes more pro-business than other justices, fine. (Though I wince at the identities pro-CoC = pro-business = conservative = not moderate.) Then we can debate how often one needs to vote, er, anti-business to be moderate. But to “prove” Alito’s lack of moderation by dating he never does something, exploiting the shock value of the word “never,” and then getting to that factoid by tossing out most of his votes is not legitimate.
And I simply reject the notion, as explained in my comments above, that what Scott is measuring is “moderation” at all.
If someone says to you: “a statistical analysis shows that players in professional basketball are getting taller over time, so Americans must be getting taller”…
One way of attacking this hypothesis is to attack the link between height of pros and the average American. But a second way to attack the hypothesis is to dispute that the statistics about the pro players were properly done.
=======
When you use the term “cherry-picking” you are making the second kind of criticism. It carries certain burdens. Given that you are wandering back and forth across this distinction, I am skeptical that you are prepared to acknowledge, let alone meet, these burdens. I am not even sure what conclusion about Alito you think SHOULD be reached from this data. Do you think he is misranked relative to Thomas or something?
The data says what Scott says it does. He is not cherry-picking; the trend he is weighting does clearly exist in the larger data. Alito is the most extreme when considering all the data. Scott’s metric is applied faithfully to all the justices. Claims to the otherwise require… you know… numbers.
=======
By all means feel free to have the first kind of debate with Scott, about whether pro-business is immoderate or whatever. But leave the statistics part out of it. Or step up.
Scott, are you making the accusation that Alito is engaging in strategic voting? That, given the exact same case, Alito might vote for the “liberal” position if the lineup is already 6-2 in favor of that position, but that if it’s 4-4 he will refuse to do so?
Because if that were an accurate description of Alito’s jurisprudence it would certainly support your hypothesis; the problem is that looking at “swing votes” doesn’t provide evidence for that, because (to state what should be obvious but apparently isn’t) cases where the rest of the lineup is 4-4 are not the same as cases where the rest of the lineup is 6-2.
Well, that in itself is a puzzling argument; being “pro-business” isn’t moderate? Is being “moderate” kind of like being David Broder, to you? You have to position yourself in the middle, half the time favoring business and half the time opposing it?
Let’s grant this to be true. What does it have to do with the price of tea in China? This is measuring something, but not “moderation.” And, similarly, the fact that Scalia or Thomas might cast a decisive vote in your favor doesn’t make them “moderate.” You’re making a category error here. You’re confusing (to use your own term) heterodoxy (*) with moderation.
It simply makes no sense to say, “Well, he voted for the criminal defendant in case A, but that’s not a sign of moderation because some other conservatives did too; he voted against the criminal defendant in Case B, and since all the other conservatives did, that shows lack of moderation.”
To put it concretely, let’s assume we have a criminal defendant where all 4 liberals vote in the criminal defendant’s favor. If this defendant also picks up (say) Kennedy, Scalia, and Alito’s votes, then Alito is voting more liberally than Thomas or Roberts. But being more liberal than them, in your analytic approach, doesn’t count at all for anything; you throw out those cases entirely in your assessment. On the other hand, if he votes exactly the same as all four of them, you mark it as a demerit.
(*) For the sake of argument, I’m accepting your premise that the conservative position is to be anti-tort plaintiff and anti-criminal defendant.
Scott, are you making the accusation that Alito is engaging in strategic voting?
“Accusation” is the wrong word, both because I don’t think there’s anything normatively wrong with strategic voting, and because until the papers are released (or someone leaks conference votes to Bob Woodward) we simply don’t know. It wouldn’t surprise me if Alito switched votes after conference in, say, Reading — but obviously we don’t know. All I’m saying is that the possibility should be considered.
Well, that in itself is a puzzling argument; being “pro-business” isn’t moderate?
In this context, where it means “supporting the business interest where the law is ambiguous the vast majority of the time,” yes.
Is being “moderate” kind of like being David Broder, to you?
In this context, essentially yes; I don’t think this is terribly controversial. I note that this doesn’t mean that being moderate is
better; obviously, if you’re a conservative voting like Alito is desirable!
you throw out those cases entirely in your assessment.
Nobody’s throwing out anything. There are multiple ways of looking at the data, that’s all. If I were bringing a lawsuit, it would be worth knowing that the Thomas and Scalia are more likely to be swing votes. That’s not the only way of measuring a judge’s ideological valence, but it’s worth knowing. Whether you call this “heterodoxy” or “moderation” — who cares? I think the argument is clear.
What’s odd about your defensiveness, of course, is that by both measures Alito shows up as the most conservative member of the Court, rendering your “cherry-picking” charge transparently bogus.
Ironically, in his attempt to prove your metric to be incredibly stupid, he has uncovered a further use for it… It seems that your metric might be an interesting tool for detecting strategic behavior. Can we get the Freakonomics guys on this?!
Thank goodness you are a humble political science professor willing to consider judges as political agents. How long has this metric been undiscovered by traditional straw-person caricatures of law professors who are concerned only with the “correctness” of decisions!!
Indeed, one of the “related posts” below — “A Rare Victory For Civil Rights Enforcement” — illustrates exactly the silliness of your approach. Alito voted “liberally” in each of these cases, but because some other conservatives did also, you ignore the data points entirely.
Being precisely as liberal as Kennedy, in other words, gets Alito nothing in your book; he has to be more liberal than Kennedy, voting “liberal” when Kennedy votes conservative in order to get credit.
Again, nobody’s “ignoring” the data points “entirely.” The data measuring all cases is there to see; I’ve linked to it; I’ve discussed it. I still think it’s worth pointing out that Thomas and Scalia are sometimes swing votes, and Alito never is.
[...] Votes That Matter : Lawyers, Guns & Money [...]
Here’s another way to make the argument that (as I take it) Scott is making:
Alito’s vote is only in play in the easier Court cases — that is to say, the cases in which precedent, statutory text, etc. give the most guidance. When the legal materials give guidance that leans heavily in favor of an anti-CoC result, then Alito may indeed vote for the anti-CoC result. (Some of the ERISA cases the court sees are like this. Alito’s actually pretty good on ERISA.)
But where, as in most cases, the legal materials give little guidance — where, as the legal realists say, the case is radically indeterminate — Alito’s vote is simply not in play. He will go for the pro-CoC result because he can — that is to say, he can do so without losing legal face.
This seems to me to be the clearest way of putting it.
To put it differently and more tersely: the strongest, not the weakest, evidence in favor of Alito’s pro-business disposition is precisely his votes in the 5-4 cases.