Changing institutional and electoral dynamics in Congress in the last several decades have placed greater limitations on presidents' ability to influence roll call outcomes. Presidents' floor success rates have been a casualty of the interaction of split-party control of the presidency and Congress, or "divided government," with heightened intra-party cohesion on Capitol Hill (Fleisher and Bond 2000). Recent presidents who have confronted assertive opposition majorities in Congress have adapted to these conditions by turning to a powerful tool in the bargaining process: the veto. Although critics may contend that "frequent use of the veto is difficult to reconcile with the Neustadtian imperative to govern by persuasion" (McKay 1994, 449), presidents' veto success was exceptional in the closing decade of the twentieth century. President George Bush sustained 28 of the 29 regular vetoes he cast from 1989 to 1992. Similarly, only one of Bill Clinton's 36 regular vetoes was successfully challenged in Congress from 1995 to 2000.
Scholars have consequently taken a renewed interest not only in presidents' ability to halt legislation through the veto power but also in their strategic use of implied vetoes to mold policy outcomes. Chief executives' success in wresting concessions from Congress through veto threats is a central test of legislative influence (Wayne 1978; Watson 1993; Deen and Arnold 2002). Emerging scholarship on this aspect of presidential behavior has been based largely on veto threats that are issued publicly (Cameron 2000). As Bond, Fleisher, and Krutz (1996, 134) remind us, however, public veto threats may not constitute a representative sample of the actual signals presidents employ. Presidents may have incentives to issue veto threats privately or behind the scenes in order to escape public and media scrutiny and avoid confrontation with Congress.
With the benefit of archival data uncovered at the George Bush Presidential Library in College Station, Texas, this article compares the effects of "private" and "public" veto threats on executive-legislative bargaining and policy outcomes in the 102d Congress (1991-1992). Archival research revealed 82 veto threats that were initially issued by the Bush administration outside the public eye. Thirty-nine of these threats were never publicly disclosed in media sources. The other 43 threats were later made public. This subset of veto threats is distinguishable from the 162 veto threats that were initially reported in the national media and for which no archival documentation was uncovered.
This article develops a typology of veto threats and compares the impact of Bush's threats on legislative outcomes according to the policy significance of bills. The results of the analysis do not so much dispute formal models of veto threats that emphasize the importance of the president's public rhetoric as they show such models to be incomplete. Private veto threats "killed" a host of minor legislation to which Bush objected. Veto threats that started off privately and served as early signals to the Democratic majority, only to be made public via "inside the beltway" reporting later, led to a high probability of executive-legislative compromise on ordinary bills such as appropriations measures. Consistent with Cameron's (2000) analysis, however, public threats on the most salient legislation, such as Mayhew's (1991) "landmark" bills, were most likely to result in vetoes and inter-branch confrontation. Indeed, public threats on important legislation often led to "blame game" politics between the White House and Capitol Hill in the 102d Congress.
The analysis progresses in four stages. The first section contrasts an "insider" approach to veto threats with the assumptions of formal theories of veto threats that accentuate the necessity of "going public" (Kernell 1997). The second section provides an overview of the archival data and methods employed. …