Justice O’Connor’s departure from the Supreme Court in January 2006 left Justice Kennedy as the sole “swing” voter. Political scientists often to refer to a swing voter as a “pivotal voter” or the “median voter” because in theory, the pivotal voter’s vote provides a sufficient condition for rendering a collective outcome favored by the median voter. Kennedy leans toward the conservative side of the ideological spectrum (see below) but has joined the liberal justices in some high profile cases.
Kennedy’s swing vote status on the Court has been highlighted by commentators in the context of the replacement for Justice Stevens. It is important to remember that the appointment of Elena Kagan would not transform the ideological balance of the Court (assuming Kagan would not turn out to be more conservative than Kennedy once on the Court). Replacing Stevens with Kagan leaves Justice Kennedy as the swing voter on most issues. Thus, in the run-up to the Kagan announcement, many liberals emphasized the importance of replacing Justice Stevens with someone who would be able to influence Kennedy. As noted by Darren Hutchinson, many argue that Kagan provides the perfect fit (see here as well), given her track record for reaching out to conservatives and seeking to form consensus and coalitions across the ideological divide. Political scientist and Supreme Court scholar Paul Wahlbeck, who has done extensive research on collegial interaction and influence within the Court, forecasts that Kagan’s capacity for wielding influence will be limited. Moreover, in a recent post, I suggested that Kennedy had “become more solidly aligned with the four more staunch conservatives.” Jonathan Adler took issue with my generalization, pointing out cases — such as Mass v. EPA, Boumediene, Wyeth, and Kennedy v. LA. — that contradicted my claim.
I thought I would bring some data to bear on these topics, particularly examining Justice Kennedy’s voting agreement history with the liberal and conservative wings of the Court. The data offer a generalized, empirical foundation to Kennedy’s swing vote status, offer more fine-grained conclusions than the assertion I made about Kennedy (that Jonathan Adler corrected), and offer something of a glimpse of what we can expect in the future from Justice Kennedy. All of the data reported and examined here come from the Supreme Court Database.
Kennedy’s Voting Behavior
First, to get a general sense of how conservative Kennedy’s voting behavior has been over time, Figure 1 below plots the percentage of the time in each term Kennedy has cast a conservative vote in a case.1 The graph plots his behavior in all formally-decided cases (including both unanimous and non-unanimous decisions) as well as closely decided cases in which there was a 6-3 or 5-4 outcome. I begin the analysis with the 1994 term instead of the 1987 term2 (Kennedy’s first term) because it marks the beginning of the long, 11-year natural court (no personnel change) from the 1994-2004 terms, and it includes many of the justices with whom Justice Kennedy currently serves. Recall that Chief Justice Roberts replaced Chief Justice Rehnquist at the beginning of the 2005 term. Justice Alito replaced Justice O’Connor toward the beginning of the 2005 term (January 2006). [Note: The Court's terms begin in October of the term year and are effectively over by the end of the following June. For instance, the 1994 term lasted from October 1994 until June 1995. This analysis goes up to the end of the last term, which ended in June 2009.]
Figure 1: Justice Kennedy’s Voting Behavior, 1994-2008 Terms
As can be seen, Kennedy has maintained a rather conservative — but not overwhelmingly conservative — voting record, though there are some anomalies (e.g., the 1995 and 2007 terms). In 6-3 and 5-4 decisions, Justice Kennedy has been more prone to be on the conservative side. While his conservatism hits the floor in the 2007 term, the 2008 term found Kennedy once again a mostly reliable conservative vote.
Voting Agreement Between Kennedy and Fellow Justices
To gain further traction on Kennedy’s behavior, I have calculated “voting agreement scores” between Kennedy and each of his fellow justices from the 1994-2008 terms. These scores represent the percentage of the time Justice Kennedy votes the same way as the other justices. I have calculated these scores for non-unanimous cases (where there was at least some disagreement) and for closely divided (5-4 and 6-3) decisions. First off, Table 1 below summarizes the data for the 1994-2004 period, as well as the 2005-2008 period. Justices color-coded as red represent the conservative wing of the Court, while blue represents the liberal wing.
What is immediately clear from Table 1 is that Kennedy has voted more consistently with the conservative wing of the Court. This is not too surprising. Interestingly, though, in the closely divided decisions, Kennedy has sided with the conservatives roughly twice as often as he has sided with the liberals. From the 1994-2004 terms, Rehnquist was Kennedy’s most consistent voting partner, for both non-unanimous decisions and closely divided ones. But O’Connor — the other swing justice on the Court during the 1994-2004 period — had the lowest voting agreement score with Kennedy among the conservatives in closely divided decisions. In closely divided decisions, Scalia and Thomas fell in between Rehnquist and O’Connor. Since Roberts and Alito joined the Court, Alito has replaced Rehnquist as Kennedy’s most frequent voting partner; Roberts is a close second.
In closely divided decisions, Kennedy sided with each of the four liberals around one-third of time from 1994-2004. Kennedy’s agreement rate with Breyer increased to nearly 50% during 2005-2008 period, but it does not increase much at all for Souter, Ginsburg, and Stevens in the 2005-2008 period. Thus, the table reinforces the notion that Kennedy has tended to vote with the conservative justices more often than with the liberals. In closely divided decisions, he has been with the conservatives substantially more often.
Table 1: Voting Agreement with Kennedy in Non-unanimous and Closely Divided (5-4 and 6-3) Decisions, 1994-2008 Terms
I have also created voting agreement scores for each term in order to visualize any trends. Click HERE to view the full results, which include tables with the voting agreement scores for each term; page 1 includes scores for non-unanimous decisions, and page 2 includes scores for closely decided cases (5-4 and 6-3 decisions). The graphs below depict a more generalized view of the tables. They show in general the extent to which Kennedy has voted with the liberal and conservative wings of the Court from the 1994-2008 terms. In each graph, the red plot represents the voting agreement score averaged among the conservative justices for each term. The blue plot represents voting agreement with Kennedy averaged among the liberal justices. The black, vertical bar in each graph marks the entry of Chief Justice Roberts (replacing Rehnquist at the beginning of the 2005 term) and Justice Alito (replacing O’Connor toward the beginning of the 2005 term) onto the Court. Figure 2 includes non-unanimous decisions, while Figure 3 includes closely divided (6-3 and 5-4) decisions.
Figure 2: Voting Agreement With Kennedy in Non-unanimous Decisions, 1994-2008 Terms
Figure 3 : Voting Agreement With Kennedy in Closely Divided (5-4 and 6-3 ) Decisions, 1994-2008 Terms
Figures 2 and 3 indicate that Kennedy’s voting alignment with the conservative wing peaked in the late 1990s and early 2000s. In closely divided cases, he was quite solidly with the conservatives during this time — about 2 to 1 on average. But on these closely divided decisions, his voting agreement with the liberals began to rise after the 2003 term, before Rehnquist’s death and O’Connor’s retirement. In the 2007 term, Kennedy split about evenly between the liberal and conservative wings in close decisions. The 2008 term found Kennedy once again quite solidly aligned with the conservatives in closely decided cases.
Looking at the voting agreement scores (the raw data) for the closely divided decisions (on p. 2), Rehnquist had the highest voting agreement score with Kennedy in 8 out of 11 terms from 1994-2004. During this same time period, in closely divided decisions, O’Connor had the lowest agreement score with Kennedy among the conservatives in 8 out of the 11 terms. This reflects the fact that in many of the 5-4 decisions in which the liberals prevailed, O’Connor was more likely to join the liberals than Kennedy. The 2007 term again seems to be an anomaly. In closely divided decisions, Breyer was Kennedy’s most frequent voting partner during this term. More generally, Kennedy split his votes about evenly between the liberal and conservative wings. The 2008 term saw a “return to normalcy” for Kennedy, as he went back to his usual pattern of voting with the conservative wing substantially more than often than with the liberal wing; though again, his agreement score with Breyer is nearly 50%, which partly reflects Breyer’s increasing propensity (relative to his liberal colleagues) to join the conservatives.
Where does this analysis leave us in terms of the issue of a future Justice Kagan influencing Justice Kennedy once on the bench? First off, the data are fairly clear that Kennedy, over time, has tended to join the conservatives much more frequently than the liberals. Yes, there are some anomalies to this strong trend, like the 2007 term. But even there, Kennedy split his votes roughly evenly between the two wings.
It is hard to believe that Kagan could bring Kennedy over to the liberal wing any more frequently than he has joined them in the past. What we can most likely expect is that Kennedy will continue to join the conservatives more often than he joins the liberals, including the “close” decisions. Of course, I have ignored the specifics of cases here that each term brings and have instead focused on broader trends. These trends make clear, I think, a fairly steady pattern of behavior on the part of Kennedy that I can’t imagine will undergo a wholesale reversal as a result of Kagan’s arrival on the Court. It is important to note, too, while there are some strong trends, it is still possible that in a number of high profile cases in the future, Kennedy will likely join the liberals. He has done just that, of course, in the past in a number of important cases.
1 To code liberal and conservative decisions, I follow the widely-used convention employed by the Supreme Court Database and numerous political scientists. For instance, in civil liberties and rights cases, a liberal vote favors an individual challenging a government restriction of a civil liberty or right, while a conservative vote favors the government restriction. For a further explanation of other issues, see the Supreme Court Database codebook, variable 36 (p. 45).
2 Justice Kennedy took his seat in February 1988, roughly midway through the 1987 term.