Is the Supreme Court more liberal? Or are the cases more conservative?

A few days before the Obamacare and same-sex marriage cases dropped, the New York Times published a fascinating data analysis showing that the Supreme Court was on track to have one of the most liberal terms in years.

But is the Court getting more liberal or are the cases getting more conservative?

The Times’ analysis relies on the Supreme Court Database, which codes whether decisions are liberal or conservative (an explanation of their approach to doing this is here.) But while this kind of analysis can tell you whether the Court is finding more often for the pro-affirmative action side, or for the pro-choice side, or for unions, it can’t tell you why — and that makes it hard to say whether the Court really has become more liberal or more conservative.

So imagine a finding that the court is ruling for affirmative action more often. One reason might be that the Court has become more liberal on affirmative action cases. Another reason might be that the Court has become more conservative, and so is hearing challenges to affirmative action laws that it wouldn’t even have considered in past terms, even though it is ultimately siding with the status quo.

In that case, it might look like the Court is getting more liberal, as it is ruling for the pro-affirmative action side, but it would actually be getting more conservative.

In a post at the Upshot, Brendan Nyhan outlines some research that suggests more conservative Supreme Courts take more conservative cases, and so can end up leading to more liberal-looking rulings:

In a 2009 article, the political scientists Kevin T. McGuire, Georg Vanberg, Charles E. Smith Jr. and Gregory A. Caldeira proposed a theory that provides an alternate explanation to liberal drift. They predicted that conservatives would press their luck to take advantage when they had a majority on the court, appealing more cases they lost in lower courts. (Conversely, liberals would be less likely to appeal cases because they were more likely to prefer lower-court decisions and to fear creating damaging precedents.)

Mr. McGuire and his co-authors then showed empirically that this process increased the number of conservative reversals of lower-court rulings but also increased the number of cases in which a more liberal ruling was affirmed because litigants guessed wrong about how far the court was willing to go.

In other words, when conservatives have a majority on the Supreme Court, they send more conservative cases to the Supreme Court — and so, naturally, they end up losing more cases before the Supreme Court, which makes the Court look more liberal.

The Obamacare ruling is a good example. One way to read the outcome of that case is that the Court sided with liberals, and that’s evidence of a more liberal term. But another way to read that case is that it only made it to the Supreme Court because the Court has become so conservative — any other Court wouldn’t have bothered, and so the proper interpretation is that King v. Burwell is evidence of the Court’s conservatism. Nyhan quotes Eric Citron, a former Supreme Court clerk, making exactly this point:

This [decision] is a good example of what’s problematic with the proposition that this is a “liberal” term. On the one hand, this is a huge victory for the left wing of the Court on a contentious issue; on the other hand, this issue does not get granted with a less conservative Court. It all depends on your baseline.

But there’s another thing all these measurements miss: the importance of various rulings. The same-sex marriage ruling is a liberal ruling of enormous, even historic, magnitude. The scores will count it as one case, equal to any other case, but it isn’t — and its presence alone will ensure that liberals long remember this Court.

Most Supreme Court decisions, after all, don’t get this reaction from the other branches of government:

Correction: I misunderstood the methodology behind the Supreme Court Database when I first read it. I read it as relying on Segal-Cover and Martin-Quinn scores to come to its conclusions as to which rulings are liberal and which are conservative. In fact, it’s the reverse: Segal-Cover and Martin-Quinn scores used the Supreme Court Database’s coding to ground their conclusions. Thanks to the NYT’s Jeremy Bowers for walking me through this. The text has been updated throughout.