Signing statements, gridlock, and presidential strategy.

Author:Kennedy, Joshua B.

During President George W. Bush's second term, controversy emerged over the president's perceived handling of hundreds of duly enacted congressional statutes. Phillip Cooper (2005) was among the first to call attention to Bush's propensity for using signing statements as a way to construe the intent of a bill he had just signed, and others quickly followed suit. Charlie Savage of the Boston Globe reported in early 2006 that "President Bush has quietly claimed the authority to disobey more than 750 laws enacted since he took office," and a year later, the American Bar Association (2006, 1) released a report decrying "the misuse of presidential signing statements," urging a different course of action that would limit undue presidential influence.

A surge in research on presidential signing statements followed in the wake of the controversy (Berry 2009; Conley 2011; Kelley and Marshall 2008, 2009, 2010; Korzi 2011; Ostrander and Sievert 2013; Sollenberger and Rozell 2011). The belief that Bush's usage of signing statements was "unprecedented" (Savage 2006) was significantly overstated, but nevertheless, focus among many presidential scholars, especially scholars of unilateral presidential power, began to shift to the causes and consequences of signing statements. Signing statements are not new, but the perception at least is that their use is consequential, prompting many to ponder when and why presidents will use these statements.

Extant research has advanced two theories about the use of presidential signing statements. The first is a political perspective, which holds that signing statements are at least in part a reaction to ideological differences with Congress (Berry 2009; Kelley and Marshall 2008, 2010; Whitford 2012). The second view is a separation of powers perspective, which takes the signing statement as a presidential reaction to the infringement by Congress on the powers of the president (Evans 2011; Ostrander and Sievert 2012). This view holds that political considerations can be secondary to more fundamental presidential concerns about the encroachment of Congress on the president's prerogatives. As Ostrander and Sievert (2012, 10) write, "presidents and Congress are interested not only in policy outcomes but also in their relative balance of power vis-a-vis one another." These two views are not mutually exclusive. Berry (2009) and Kelley and Marshall (2008) contend that institutional prerogatives matter to the president when it comes to the use of signing statements. Thus, the signing statement is, as Ostrander and Sievert suggest, a story of multiple considerations, meeting together to shape presidential strategy.

However, scholars have not adequately considered ideological differences between the president and Congress as well as gridlock within Congress when developing models to test the usage of signing statements. Principally, measures of separation between the two branches have largely been dichotomized to represent the differences between unified and divided government (Berry 2009; Kelley and Marshall 2008; Ostrander and Sievert 2012). When ideology is considered (Kelley and Marshall 2010), it is done independent of gridlock within Congress. This article argues that varying levels of legislative gridlock interact with the nature of Congress's relationship to the president (i.e., whether government is unified or divided) to change presidential motivations and the use of signing statements.

Gridlock and Presidential Strategy

Gridlock and divided government, when visited in conjunction with the study of legislative productivity and government action, received quite a beating in the realm of public opinion and among scholars for years before David Mayhew's (1991) groundbreaking study. James Sundquist (1988, 629) captured the conventional wisdom of the time when he described divided government as both "unhealthy" and "debilitating." However, with Mayhew (1991) came the argument that divided government hampers legislative productivity (at least in the face of important legislation) either not at all or not as much as is expected. With the exception of some challengers (Edwards, Barrett, and Peake 1997), Mayhew's general conclusion that divided government is not as crippling as people think is largely supported, though occasionally with caveats (Fiorina 2003; see also Howell et al. [2000] for a methodological objection to Mayhew).

This is not to say, however, that the sources of gridlock and its additional consequences are unimportant or that the debate is at all settled. Mayhew (1991) and others (Binder 1999, 2003) have cautioned political scientists to consider what gridlock means in some depth; that is to say, gridlock is not just about which party controls which branch. Binder (1999, 530) concisely sums up this sentiment when she writes, "The distribution of policy views within and across the two major political parties has predictable and important effects on the legislative performance of Congress and the president."

How, then, can we define gridlock? Krehbiel (1) (1998, 20-39) puts forth a theoretical construct of this concept that is dependent upon various "pivotal" actors within the legislature. The veto pivot is the member of Congress whose views are ideologically arranged in such a way as to give him or her the final say in overturning a presidential veto. That is to say, if the president vetoes a bill, a veto override cannot succeed unless the veto pivot is in favor of the override. The median voter occupies the ideological midpoint of the legislature. If the median voter is not on board with a bill, it cannot muster a simple majority and subsequently leave Congress for presentation to the president. The filibuster pivot is the member who falls into such ideological placement as to be the deciding vote to invoke cloture. If the filibuster pivot is on the side of the majority, the bill will pass. If the filibuster pivot is on the side of the minority, the bill may enter limbo, lacking the requisite votes to invoke cloture and approve passage.

How gridlock affects presidential strategy is not as well understood as it is in other fields of political science. One notable exception is Howell's Power without Persuasion (2003). In this book, the author devises a particular measure based upon the theory of Krehbiel's (1996, 1998) gridlock interval in examining the usage of executive orders, finding that more gridlock gives the president more latitude to "go it alone" with unilateral directives. I subject signing statements, a different unilateral presidential power, to similar conditions. In so doing, I attempt to elucidate how gridlock affects presidential strategy with respect to signing statements.

The Use of Signing Statements

The history of the signing statement has been covered in depth (Kelley 2007a, 2007b; Kelley and Marshall 2010). Though these statements have been commonly used as a tool of direct presidential action since the Reagan administration (Kelley 2007b; Kelley and Marshall 2010), increased scholarly attention to signing statements is due in large part to perceptions about how George W. Bush used them to advance his agenda and preserve his institutional prerogatives (Cooper 2005).


Signing statements often reflect concerns about seemingly unconstitutional statutory provisions (Berry 2009; Ostrander and Sievert 2012; Sollenberger and Rozell 2011) and presidential attempts "to assert that [the provisions] are unconstitutional encroachments on the president's role as chief executive" (Berry 2009, 267). However, presidents can and do use signing statements, as Kelley and Marshall (2008, 255) put it, "to bend policies toward [their] preferred interpretation." In these cases signing statements do not solely reflect the president's concerns about the institutional capacity of his office, but also his concerns for policy outcomes.

Figure la shows rates of signing statement usage in two categories: the total number of signing statements issued and the number of substantive signing statements (i.e., signing statements that contain at least one provision dealing with the implementation of a statute) issued. Figure lb shows rates of the number of substantive actions (i.e., how many provisions the president construes or how many directives are issued) taken within all signing statements. These figures are based on a reading of signing statements as catalogued by The American Presidency Project (various years). (2)

Variation on the total number of signing statements seems largely idiosyncratic, but substantive signing statements begin to spike in the Reagan administration (as discussed, for example, by Kelley [2007a]) and have endured as a presidential tool since that time. During both Bush administrations, substantive signing statements comprised more than half of all signing statements. The reason George W. Bush has generated such controversy is because he frequently (though not universally) listed each individual provision to which his administration objected or planned to construe counter to the wishes of Congress. This is evidenced by the dramatic spike in substantive actions taken between 2001 and 2009- However, his predecessors were not always as forthcoming. Consider, for example, this statement from George H. W. Bush (emphasis not in original): "A number of provisions in the Act condition the President's authority.... Accordingly, I will treat them as having no legal force or effect in this or any other legislation in which they appear" (Bush 1991). Compare that statement to the following from George W. Bush:

Provisions of the Act that purport to require congressional committee or individual leaders' approval prior to the execution of the law shall be construed as calling solely for notification. ... These provisions include language under the headings "United States Fish and Wildlife Service, Administrative Provisions," "National Park Service, Construction," "Departmental...

To continue reading