Review of the Literature

Theory

Research discussed in the sections above highlights two important reasons for why we theorize the Court should CVSG to gain more information about executive preferences. First, justices pursue preferences, but they do so in a strategic manner. A means to establish policy attainments is to ensure the implementer of policy, the president, will enforce the Court’s decision. Understanding where the president stands allows the Court to make more calculated decisions about where in the policy space its ruling can successfully land. In publicly charged policy areas, the Court may be seeking a signal of executive support (such as in the Brown cases).Footnote ² Alternatively, the brief could cue the Court to outcomes that may bring potential sanctions. Because the Court wishes to see its policy implemented, and because the Court depends on institutional legitimacy, it wants to know of the possibility of threats to these goals.Footnote ³

Second, and more simply, the justices care about making good policy. Indeed, Hausegger and Baum (Reference Hausegger1999, 182) demonstrate that the Court will ask Congress to review the legal sphere surrounding decisions in order to create good policy “as they see it.” Since the executive branch is responsible for implementation of policy, the Court may want to know about possible problems of implementation. The actors in the executive branch have the (greater) requisite expertise on implementation. Because we argue that the SOTU proxies the executive agenda, and this agenda often includes policy plans, an area receiving less attention may indicate one with less public information about the impacts of changing policy. Thus, in these areas, the Court may wish to CVSG to supplement its information.

We theorize that because the SG does present the president’s position in briefs, the Court can ascertain the executive’s preferences through them. The assertion that SG briefs are reflective of executive interests is supported in extant literature (Salokar Reference Salokar1992; Meinhold and Shull Reference Meinhold and Shull1998; Johnson Reference Johnson2003; Black and Owens Reference Black and Owens2012; Epstein and Walker Reference Epstein and Walker2016). When looking at other appointed officials and bureaucratic agencies, scholars recognize that presidents can influence their agents’ behavior through such means as appointing those with similar ideologies or by disciplining divergence (West and Cooper Reference West and Cooper1990; Wood and Waterman Reference Wood and Waterman1991; Hammond and Knott Reference Hammond and Knott1996; Moore, Maclin, and Kershner Reference Moore, Maclin and Kershner2001; Shipan Reference Shipan2004; Waterman Reference Waterman2009). Likewise, presidents appoint SGs who share their preferences, and presidents can punish SGs who fail to represent the president (Segal Reference Segal1988; Segal and Reedy Reference Segal and Reedy1988; Salokar Reference Salokar1992; Bailey and Chang Reference Bailey and Chang2001). President Reagan is said to have essentially fired Solicitor General Lee, and Nixon forced out Solicitor General Griswold (see Salokar Reference Salokar1992; Bailey and Chang Reference Bailey and Chang2001). By and large, scholars argue that the SG accurately represents the president (Segal Reference Segal1988; Fischer Reference Fischer1990; Clayton Reference Clayton1992; Salokar Reference Salokar1992; Meinhold and Shull Reference Meinhold and Shull1998; Graham Reference Graham2002; Curry, Pacelle, and Marshall Reference Curry, Pacelle and Marshall2008; Black and Owens Reference Black and Owens2012). Black and Owens (Reference Black and Owens2012, 808) note that half of the SG briefs from the 2007 Court term explicitly highlighted executive branch preferences. As such, and in line with others studying CVSGs (Johnson Reference Johnson2003; Black and Owens Reference Black and Owens2012), we argue that the SG represents the president in CVSGs and the Court can determine executive preferences through them.Footnote ⁴

The iconic briefs (the result of a CVSG) in the Brown cases demonstrate just how closely a CVSG might reflect the executive’s preference.Footnote ⁵ President Eisenhower had a direct hand in editing the briefs in the Brown II case.Footnote ⁶ As further evidence of the president’s effort in Brown, Attorney General Brownell called Eisenhower to say the Supreme Court followed the president’s formula almost exactly.Footnote ⁷

We recognize the need for the SG to be cognizant of Court needs and interests while also strategically representing the preferences of the president. While we do not match the content of CVSGs to presidential agenda items, we do examine areas of more and less publicly declared agendas. If the Court CVSGs more often when the president speaks less about a policy area, we believe we capture a desire of the Court to understand more about the president’s interests in the area.

Of direct interest to our study, Johnson (Reference Johnson2003), who argued SOP considerations motivated CVSGs, accounted for presidential agendas using Light’s (Reference Light1999) measure. Johnson (Reference Johnson2003) hypothesized that the Court should be more likely to CVSG in cases on the president’s agenda but found no significant relationship. Light developed an important and popular measure of presidential agenda items. To be included as an agenda item, each item had to be mentioned in the SOTU and be classified as “in accordance” with the president’s program by the Office of Management and Budget (Light Reference Light1999, 9) or, after 1980, be mentioned in the Congressional Quarterly Almanac data (Light Reference Light2004; Eshbaugh-Soha Reference Eshbaugh-Soha2005). Using these or similar methods (see Eshbaugh-Soha [Reference Eshbaugh-Soha2005], who uses Public Papers of the Presidents), researchers offer a list of agenda items per president. Our research uses a different measure of presidential salience that diverges in part from Light’s measure of agenda items.

We prefer to use the content of the SOTU for our analysis to develop our measure of the general presidential agenda. The strength of Light’s measure is that it filters through all the information presented in the SOTU to identify the most important of presidential initiatives, such as a balanced budget amendment or the Clean Air Act (Eshbaugh-Soha Reference Eshbaugh-Soha2005). We do note, however, that Light’s measure and the modern derivatives of it start with the SOTU. For our study, our goal is not to isolate the most prominent presidential initiatives. Instead, we care about how much information the Court likely has about the presidential preferences in the vast policy sphere in which it acts. By using an issue area’s proportional coverage in the SOTU we get a relative measure of information surrounding the topic. Furthermore, this dynamic measure provides variety over the years, not simply presidential terms.

We use the SOTU to proxy the general agenda of the president. The SOTU publicly highlights the general issue and ideological preferences of the president (Light Reference Light1982; Tedin, Rottinghaus, and Rodgers Reference Tedin, Rottinghaus and Rodgers2011). Conventionally and practically, presidents are limited in how much time they can spend speaking during the SOTU. Thus, time on any given issue area diminishes time spent on other issue areas. Because there are disparate amounts of time devoted to different topic areas, there are different amounts of public information available concerning the president’s preferences on areas.Footnote ⁸

Often, many justices attend the SOTU. We argue that the SOTU serves to let political actors, including the courts, know about presidential priorities. We do not argue, however, that a justice must be in attendance to the SOTU to be informed about general executive preferences.Footnote ⁹ But, as important, intelligent, and strategic political actors, justices are aware of the public preferences of other actors.Footnote ¹⁰ Furthermore, some studies indicate that the Court is aware of and sensitive to agenda items present in SOTUs (Wahlbeck Reference Wahlbeck1997; Rogol, Montgomery, and Kingsland Reference Rogol, Montgomery and Kingsland2018). If the president omits or speaks less about an issue area, the public and other political actors are relatively less certain of executive preferences on the issue. If the Court wants to be most informed on the preferences of the president, it should CVSG in cases whose issue area has received less attention. Therefore, we hypothesize that as an issue area’s proportional coverage in the SOTU increases, the likelihood of the Court issuing a CVSG should decrease.

Our theory also explains the lack of CVSGs on issues that are likely of greater importance to the president or that at least have more information surrounding them. Rather than suggesting that the Court should be more likely to request a brief when a topic is an agenda item for the president, we suggest that the Court should be more likely to request a brief when the president has spoken relatively less about a topic area. The Court is motivated to understand where the president sits in these cases; it already knows presidential preferences in the others.

The fact that CVSGs are uncommon, yet occur across terms, and that the SG views CVSGs as a mandate to submit suggests that this is an important strategic tool for the Court. We believe the CVSG is used carefully because the request is not without costs for the executive branch. The SG (or Department of Justice) must devote effort and human capital to writing this brief (Thompson and Wachtell Reference Thompson and Wachtell2009). Consequently, this tool is likely one that requires strategic use by the Court. Just as the SG cannot submit a brief on all cases of interest (Lee Reference Lee1985), neither can the Court request a brief. Thus, our theory demonstrates when cases will be prime for a CVSG, based on the need for information concerning the executive’s interests.

Our work advances the earlier work of Johnson and Black and Owens in a few ways. Black and Owens reject the SOP argument in favor of the Court, using CVSGs to determine their usefulness in a case. Black and Owens’s (Reference Black and Owens2012) measure of “Separation-of-Powers Concern” is an important one, but we feel it is too narrow to justify writing off all Court SOP considerations. Black and Owens (Reference Black and Owens2012, 809) hypothesize that “a justice will be more likely to CVSG as the outcome that she prefers is increasingly likely to trigger a congressional override attempt that requires a presidential veto to survive.” Additionally, we are not predicting individual justice’s votes but rather the institutional calculation of the Court. Thus, our study accounts for the institutional decision of the Court to request information reflective of varying amounts of public information regarding executive preferences. Our study controls for similar considerations that Johnson (Reference Johnson2003) finds important. But, our SOTU measure of presidential agenda is more reflective of relative degrees of information about presidential preferences and broadly captures the publicly expressed agenda of which most political actors should be aware.

We do not view the perspectives, SOP (as studied by Johnson) and policy or legal motivations (the view supported in Black and Owens [Reference Black and Owens2012]), as mutually exclusive; when considering implementation concerns, one should consider the other branches and vice versa. In this case, perhaps some of the policy impact questions are reduced when there is more political (i.e., SOTU) conversation about a policy topic. Thus, the need for the Court to CVSG for implementation concerns could also be reduced. This study is certainly not the final say in the exploration of the executive-judicial relationship and the Court’s concern for other branches. However, the results of our analysis lend support to the SOP argument in the sense that we see the seeking of information about the president’s preference when it is lacking. The most intuitive explanation for this behavior would be one that suggests the Court, at times, cares about the preferences of the other branches.

Data and Method

We argue that the Court desires information about the executive’s preferences, and if it is lacking information, it should be more likely to induce the SG to submit a brief. Using the Spaeth et al. Supreme Court Database, we examine Supreme Court cases accepted for cert and decided during the 1956–2009 terms. We exclude cases when the government was a party to it. We also exclude cases in which the SG chose to submit a brief without an invitation, and therefore the Court would not have the opportunity to CVSG (see also Johnson Reference Johnson2003).Footnote ¹¹

The dependent variable is a dichotomous measure of whether or not the Court issued a CVSG on the merits. These data were gathered by Solowiej and Collins (Reference Solowiej and Collins2009), updated through 2005 by Corley, Steigerwalt, and Ward (Reference Corley2013), and updated for this project through 2009. The variable is coded as 1 if the case had a CVSG and 0 otherwise. Because the dependent variable is dichotomous, we use a logit. We cluster on the year argued to account for nonindependence across groups.

Independent Variables

To measure the president’s publicly stated agenda, we use SOTU addresses. The Comparative Agendas Project, by Frank Baumgartner and Bryan Jones, offers the State of the Union Addresses Dataset, which codes statements made during the SOTU according to policy area.Footnote ¹² The Comparative Agendas Project’s Supreme Court Dataset recodes the Spaeth et al. Supreme Court Database to align with the Baumgartner and Jones topic areas.Footnote ¹³ Thus, we have a bridge between data sets on issue area. Baumgartner and Jones identify 20 major policy areas, including macroeconomics; civil rights, minority issues, and civil liberties (one category); defense; and social welfare.Footnote ¹⁴

Our primary independent variable of interest is the proportion of the SOTU the president devotes to a specific topic area per year. For every issue area, we count the number of statements made about that topic area and divide by the total number of statements in that SOTU. Therefore, every topic area for every year has a value that indicates the proportion it made up of that year’s SOTU. And, every case has the associated proportion for its issue area’s prominence for that year. Each SOTU is delivered in the beginning of the year. Court terms run from October to June/July. As such, we match the Proportion of SOTU variable to the year the case was argued at the Supreme Court. For example, if a case was argued in November 1957, it gets its Proportion of SOTU measure from the 1957 SOTU address.

The type of individual statements coded varies between broad or sweeping statements and statements promoting a specific policy. Take, for example, the issue area of civil rights and liberties. Sometimes presidents offer broad statements promoting equality and decrying discrimination. In the 1999 SOTU, President Clinton said, “Discrimination or violence because of race or religion, ancestry or gender, disability or sexual orientation, is wrong, and it ought to be illegal” (see Baumgartner and Jones’s State of the Union Addresses Dataset). Alternatively, statements may be more specific and allow ideological leanings to be on display. In his 1988 SOTU, President Reagan said, “The Congress opens its proceedings each day, as does the Supreme Court, with an acknowledgment of the Supreme Being—yet we are denied the right to set aside in our schools a moment each day for those who wish to pray” (see Baumgartner and Jones’s State of the Union Addresses Dataset). While we do not denote the difference in these types of statements in this study, we are confident that the proportion of statements devoted to a topic area are indicative of the president’s public agenda. Likewise, more time devoted to a topic area indicates there is more available information regarding presidential preferences in that area.

We control for a variety of other traits that extant literature demonstrates may influence the Court’s decision to CVSG. First, we control for the saliency of the case. Salient cases are often viewed as politically important. It is likely that the executive branch has preferences on these cases. A Court that wishes to remain cognizant of the other branch’s preferences, for fears of sanctions, may be more motivated to request a brief in salient cases. We use Epstein and Segal’s (Reference Epstein and Segal2000) measure of case salience.Footnote ¹⁵ This variable captures whether the case was discussed in a New York Times front-page article the day after the decision was handed down. Cases covered are coded 1 and all others 0.

We control for Ideological Distance between the president and the Court. We suggest the Court should be more likely to CVSG when it is further from the executive. If a Court is further from the president, indicating likely different preferences, yet still seeks to understand the president’s positions (again, to avoid or understand the risk of sanction), it should be more likely to CVSG. We use the Common Space Scores to place the Court and president on one ideological scale. Then, we calculate the absolute value of distance between the president and the median justice on the Court (raw data available from Epstein et al. [Reference Epstein, Martin, Segal and Westerland2007] and Lewis et al. [Reference Lewis, Rosenthal, Boche, Rudkin and Sonnet2017]).

We also control for whether the case contained a constitutional question. We hypothesize that the Court should be less likely to CVSG in constitutional cases. A previous study notes that the Court is less likely to CVSG “on a constitutional question, as the Court possesses more than sufficient expertise to answer constitutional questions” (Thompson and Wachtell Reference Thompson and Wachtell2009, 281). We also suspect that the Court wants to protect its power of constitutional review.

Like Johnson (Reference Johnson2003), we control for various SOP considerations that may influence the Court’s decision to CVSG. In periods of greater presidential capital, the Court may be more deferential to the president or more likely to CVSG. We include four variables that Johnson (Reference Johnson2003) argues are important: Election Year, Presidential Margin of Victory, Divided Government, and Presidential Approval. Election years are coded as 1 and all others 0. (Like Johnson [Reference Johnson2003], we use the year argued to determine election year.) The Presidential Margin of Victory variable is the difference between the president and the other major party’s candidate. Divided government is coded as 1 when the president’s party does not fully control Congress and 0 otherwise. Finally, presidential approval is measured by the American Presidency Project’s average approval rating for a given year.Footnote ¹⁶

We examined a total of 3,649 US Supreme Court cases between 1956 and 2009. The primary question in this exploration is when and why the Court decides to request an amicus brief in such cases. Overall, the rate of CVSG is low; the court requested a brief 7.34% of the time.Footnote ¹⁷ We propose that the Court is motivated to CVSG when it desires to gather information about the president’s preferences when it is not clear where the president stands—that is, in this context, to acquire more information in the issue areas that the president covered comparatively less in the SOTU.

There is variation in our independent variable Proportion of SOTU. The minimum proportion for a topic area is 0. In fact, 18 of the 20 topic areas (defense and macroeconomics being the exceptions) experienced at least one year when they were not covered in the SOTU. Proportion of SOTU’s maximum value is .447. In 1971, President Nixon spent 45% of his SOTU covering the topic area of government operations. He discussed a proposed reorganization of the cabinet departments and revenue sharing between the federal and state governments. Additionally, there is variation among the topic areas themselves. For example, the minimum proportion for civil rights and liberties is 0, and the maximum is 11. There is also variation among presidents on the topic areas. In 1964, President Johnson spent 11% of his speech covering civil rights and liberties. However, in 1967, he spent a dramatically smaller 0.4% on civil rights and liberties. Additional summary statistics by topic area for Proportion of SOTU can be found in table A2.Footnote ¹⁸