The Swing Vote and the Attitudinal Model of the Supreme Court: Explaining Ideological Shift in the Court Over Time

Patrick D. Shushereba ABSTRACT

The decisions of the Court are partially grounded in the ideology of the justices themselves, which is the attitudinal model of the Court. I argue that the swing vote justice on the Supreme Court has a substantial effect in determining the ideological direction of the Court over time. If the attitudinal model of the Court is correct, then the swing vote justice on the Court has a disproportionate amount of influence on the direction of the Court over time. I examine the influence of the swing vote justices on the Court since 1981. This project analyzes the voting behavior of two specific justices, Justice Sandra

Day O’Connor, the swing vote on the Rehnquist Court and Justice Anthony Kennedy, the swing vote justice on the Roberts Court. Previous literature has identified the swing vote justice, but only to determine how powerful that justice is relative to the other justices on the Court. My manuscript not only establishes the swing vote justice and the power that they wield on the Court, but also determines how that power is sustained over time. This manuscript examines 5-4 Supreme Court decisions since

O’Connor’s appointment to analyze O’Connor’s role as the swing vote, and determine the extent to which her political ideology as defined under the attitudinal model aligns with the direction of the

Supreme Court during her tenure. I also look at the 5-4 decisions since Justice Kennedy joined the Court to determine his role on the Court during the period that Kennedy and O’Connor served concurrently, as well as Kennedy’s position as the swing vote since O’Connor’s departure. Additionally, I determine which issues were significant issue areas for both Justice O’Connor and Justice Kennedy, and employ a case study analysis to determine what types of case specific considerations the justices use when reaching decisions.

Methodology

In order to quantitatively test my hypotheses I went through all of the Supreme Court decisions from 1981 to the present, and compiled a data set that consisted of all of the 5-4 decisions. I used the website www.oyez.org which is a project of the Chicago-Kent College of Law. The website tracks the voting records of Supreme Court justices in every case back until 1953. I started my research from the

1981 Supreme Court term because that was the year that Justice Sandra Day O’Connor was appointed to the Court. Justice Kennedy was appointed to the Court in 1988, and served with O’Connor until her retirement after the 2005 term. I went through all of the 5-4 decisions and documented the votes of each justice on the Court, whether they were in the majority or the minority for the decision. I concluded my data set with the 2012 term because that is the most recent complete term. There were a couple of different things that I needed to take into consideration when I was starting my data set. One situation that I encountered and had to account for was when the court had a split vote over different issues within a single case. There were instances where both of the split votes were 5-4 decisions. If this was the case, then I counted both votes in my data set because they were unique votes, even if it was on one case. However I did also encounter cases where there was a split vote, sometimes as many as three votes on separate issues within one case. For these instances I only counted the relevant 5-4 votes in my data set. If the vote was any other distribution than 5-4 then it wasn’t counted. I also created variables to account for other circumstances, such as if Justice O’Connor or Justice Kennedy wrote the majority opinion, or if they wrote a concurrence or dissent in the decision. When documenting whether or not Justice O’Connor or Justice Kennedy authored a concurrence I did not make any distinction between a regular concurrence and a special concurrence in my data set. A 5-4 majority opinion that was authored by either Justice O’Connor or Justice Kennedy will provide insight into both the legal reasoning that each justice employs, as well as the ideology of each justice, lending credibility to the attitudinal model. The final major variable in my data set is the issue at hand in the decision. Many of the Court’s 5-4 decisions involve Constitutional issues, which frequently results in the justice’s casting votes in accordance with their political preferences.

In addition to using numerical data to substantiate my hypotheses, another method that I used in order to support my results was an analysis of cases similar to a case study. There are two ways to utilize this method of analysis. The first is to look at individual issue areas such as Eighth Amendment cases. After identifying issue areas where either Justice O’Connor or Justice Kennedy was in the majority a high percentage of time, the next step is to identify cases in those issue areas where Justice Kennedy or Justice O’Connor authored the majority opinion or authored a concurrence in the decision. Looking at the cases in this way will aid in identifying certain nuances of O’Connor and Kennedy’s jurisprudence that causes them to decide cases the way that they do. The second way to take advantage of this method is to look at issue areas, but identify instances where either Justice O’Connor or Justice Kennedy voted with the majority in some cases of this type, but with the minority in other cases. Then it is beneficial to go through those cases and identify what case specific considerations resulted in the Justice making the decision that they ultimately did.

After aggregating all of the data and running the appropriate tests, I expected the results to confirm a number of different hypotheses. I believe that the results will show that Justice O’Connor was in the majority more than any of the other justices on the Court.

Although voting with the majority more than any other justice is not indicative of a justice being the swing vote on the Court. Another important component to look at is the frequency that the justice in question voted with each “block” of justices, either liberal or conservative. The data set that I have compiled will allow me to show exactly how often Justice O’Connor, and later Justice Kennedy voted with each side on cases where the votes were obviously partisan. I also anticipate the results showing me a similar situation with Justice Anthony Kennedy.

However, for a number of different reasons I expect Kennedy’s position as the swing vote to be less pronounced than O’Connor during the period where the two served on the Court simultaneously. First,

Kennedy was appointed to the Supreme Court in 1988, and served concurrently with Justice O’Connor until her retirement in 2005. Justice Kennedy was not widely considered to be the swing vote until after

Justice O’Connor left the Court. I expect to see both O’Connor and Kennedy in the majority for a large number of cases, but that Justice Kennedy will step into the sole swing vote position after O’Connor retires. Additionally, I expect Justice Kennedy’s position to carry less significance partly because Kennedy is still currently serving on the Court. My data set encapsulates O’Connor’s full tenure on the Court, but for Kennedy it is incomplete.

Prominent literature suggests that there may be some flaws in my research design which need to be addressed. For example, it is difficult in general to explain or predict judicial behavior based solely on data from voting records. To address that concern, it needs to be clarified that the point of this paper is not to propose a new model to explain judicial behavior. It is to look at the influence of the swing vote within the context of the attitudinal model. Another way that I have worked to alleviate this concern is by not relying on quantitative data from voting records alone. The addition of the analysis of majority opinions will help demonstrate that, given the accuracy of the attitudinal model, the swing vote justice has a substantial amount of influence over the ideological direction of the Court, and a direction that you can see changing over time through the decisions.

Another problem that could be raised is that there is a bias in the cases the Supreme Court hears even before the oral arguments. This claim derives from the fact that the justices on the Court have discretion over what cases will be heard. The argument therefore, is that the justices on the Court will exercise their ideological preferences when deciding what cases to hear, creating a bias that could affect the way that the justice votes after oral arguments are heard. I argue that this concern is not as significant as it may initially appear. The reason this is not a genuine concern is that when the justices are deciding what cases to hear, the threshold is much lower than when they are reaching a decision on the merits after oral argument. When the justices are deliberating on whether or not to hear a case, they use the “rule of four”. Essentially, four of the justices have to agree that the case has merit and presents a problem that the Court needs to address. There is a very large difference between voting to hear oral arguments on a case because there appears to be a constitutional issue and casting a vote on a case based on the merits after oral arguments where each justice has had a chance to ask questions of the attorney’s arguing for each side.

Results

Using SPSS to run statistical tests, I was able to reach a lot of different conclusions. First,

I found that Justice O’Connor was in the majority for 66% of the 5-4 decisions while she was on the Court, and Justice Kennedy has been in the majority for 74.6% of the cases since he began on the Court. Such a high percentage by these justices strongly suggests that they can be identified as the swing vote, as they were in the majority of more 5-4 decisions than any other justices.

Additionally, there were time-dependent patterns that emerged. According to my data, Justice

O’Connor became more influential as she spent more time on the Court. Kennedy’s influence was also trending upward, but not nearly at the rate of O’Connor.

Also, I was able to single out specific issue areas where each justice seemed to vote with the majority a high percentage of the time. For example O’Connor had a strong disposition towards Eighth Amendment issues. Of the 28 Eighth Amendment issues that came before her, she was in the majority for 21 of them, and wrote the majority opinion for 7 of those. There is still a vast amount of data I have left to analyze and draw conclusions about. Table 1: Percentage in Majority by Year 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 Warren 37.5% 76.5% 65.4% 62.5% 65.7% Burger Byron 59.4% 76.5% 80.8% 62.5% 61.1% 68.3% 86.7% 78.8% 83.3% 58.3% 57.1% 50.0% White Lewis 56.3% 70.6% 69.2% 75.0% 77.8% 75.6% Powell William 40.6% 55.9% 61.5% 56.3% 63.9% 63.4% 73.3% 78.8% 69.4% 70.8% 28.6% 72.2% Rehnquist William 59.4% 41.2% 42.3% 43.8% 33.3% 39.0% 33.3% 24.2% 33.3% Brennan Thurgood 65.6% 38.2% 46.2% 37.5% 38.9% 39.0% 33.3% 21.2% 33.3% 41.7% Marshall Harry 65.6% 35.3% 42.3% 50.0% 36.1% 48.8% 33.3% 33.3% 33.3% 45.8% 64.3% 38.9% Blackmun John Paul 78.1% 41.2% 46.2% 62.5% 44.4% 41.5% 40.0% 27.3% 44.4% 45.8% 71.4% 38.9% Stevens Sandra Day 40.6% 64.7% 57.7% 50.0% 77.8% 56.1% 66.7% 75.8% 66.7% 66.7% 57.1% 50.0% O’Connor Antonin 100% 68.3% 60.0% 75.8% 66.7% 54.2% 35.7% 77.8% Scalia Anthony 73.3% 84.8% 69.4% 54.2% 64.3% 77.8% Kennedy David 62.5% 92.9% 22.2% Souter Clarence 28.6% 72.2% Thomas Cases 32 34 26 16 36 41 15 33 36 24 14 20 per term

1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004

Byron White

William Rehnquist 69.2% 62.5% 53.8% 64.7% 56.3% 53.3% 73.7% 64.0% 70.0% 50.0% 47.1% 58.8%

Harry Blackmun 38.5%

John Paul Stevens 38.5% 50.0% 46.2% 47.1% 43.8% 46.7% 26.3% 36.0% 35.0% 50.0% 52.9% 41.2% Sandra Day O’Connor 53.8% 68.8% 69.2% 70.6% 62.5% 66.7% 84.2% 76.0% 80.0% 85.7% 76.5% 64.7%

Antonin Scalia 69.2% 56.3% 53.8% 58.8% 50.0% 60.0% 78.9% 64.0% 70.0% 50.0% 52.9% 52.9%

Anthony Kennedy 92.3% 81.3% 76.9% 82.4% 81.3% 60.0% 73.7% 76.0% 70.0% 50.0% 58.8% 70.6%

David Souter 46.2% 37.5% 46.2% 47.1% 37.5% 53.3% 31.6% 44.0% 35.0% 64.3% 52.9% 52.9%

Clarence Thomas 53.8% 50.0% 53.8% 58.8% 68.8% 60.0% 84.2% 64.0% 70.0% 50.0% 64.7% 58.8% Ruth Bader Ginsburg 38.5% 50.0% 53.8% 29.4% 43.8% 46.7% 26.3% 40.0% 30.0% 50.0% 52.9% 47.1%

Stephen Breyer 43.8% 46.2% 41.2% 56.3% 40.0% 21.1% 36.0% 40.0% 50.0% 41.2% 52.9%

Cases per term 13 17 15 17 16 16 21 25 20 14 18 17

2005 2006 2007 2008 2009 2010 2011 2012 John Paul 40.0% 24.0% 63.6% 43.5% 46.2% Stevens Sandra Day 100% O’Connor Antonin 80.0% 68.0% 54.5% 69.6% 76.9% 64.3% 53.3% 55.0% Scalia Anthony 60.0% 100% 63.6% 82.6% 61.5% 85.7% 80.0% 85.0% Kennedy David 40.0% 28.0% 54.5% 40.9% Souter Clarence 70.0% 68.0% 63.6% 69.6% 84.6% 71.4% 60.0% 70.0% Thomas Ruth Bader 40.0% 24.0% 54.5% 43.5% 23.1% 42.9% 40.0% 40.0% Ginsburg Stephen 30.0% 36.0% 45.5% 34.8% 38.5% 35.7% 53.3% 50.0% Breyer John G. 70.0% 76.0% 54.5% 56.5% 61.5% 57.1% 60.0% 65.0% Roberts Samuel 62.5% 76.0% 45.5% 60.9% 76.9% 57.1% 53.3% 60.0% Alito Sonia 0.0% 30.8% 42.9% 53.3% 35.0% Sotomayor Elena 42.9% 46.7% 40.0% Kagan Cases per 10 25 11 23 13 13 14 15 term