Obama's authorization paradox: Syria and Congress's continued relevance in military affairs.

Kriner, Douglas L.

When President Barack Obama stepped into the Rose Garden to address the American people on August 31, 2013, his words caught many by surprise. The announcement of the policy decision itself--that President Obama had decided the United States should take military action to punish the Syrian regime of Bashar al-Assad for its use of chemical weapons against its own people--was expected. For more than a year, the administration had warned Assad that using chemical weapons was a red line that, if crossed, would provoke an American response. What took even longtime politicos by surprise, however, was what the president said next:

But having made my decision as Commander-in-Chief based on what I am convinced is our national security interests, I'm also mindful that I'm the President of the world's oldest constitutional democracy. I've long believed that our power is rooted not just in our military might, but in our example as a government of the people, by the people, and for the people. And that's why I've made a second decision: I will seek authorization for the use of force from the American people's representatives in Congress. (1)

For many, this move posed quite a puzzle. At a meeting of his National Security team on August 24, the president by all accounts appeared to have decided to order a limited military strike against the Assad regime to punish it for the August 21 chemical weapons attacks that killed approximately 1,400 Syrian civilians. As late as August 30, Secretary of State John Kerry publicly labeled Assad "a thug and a murderer," leaving little doubt that American military action was imminent (Memmott 2013) However, over the 24 hours preceding the president's Rose Garden address, Obama changed his mind.

A series of developments on both the international and domestic fronts undoubtedly influenced the president's decision calculus. On Wednesday, August 28, talks at the United Nations Security Council on a formal response to the chemical weapons attack broke down in the face of a Russian and Chinese veto of any resolution for military action. The following day, Prime Minister David Cameron shockingly lost a vote in the House of Commons to authorize the use of force. While France continued to stand with the United States, Cameron announced that he would respect the will of Parliament and not join any military action, despite his personal belief that a military response was required (Faiola 2013). Finally, members of Congress were also increasingly demanding a voice in Syrian policy. As of August 29, 140 members, including 21 Democrats, had signed a letter calling on Obama to seek congressional authorization before ordering a military strike; to do otherwise, the letter claimed, would be unconstitutional (Shabad 2013).

The president did not change his mind because he believed a unilateral strike would have been illegal. Indeed, the administration's lawyers formally concluded that "important national interests" in bringing stability to the Middle East and enforcing international norms against the use of chemical weapons justified a unilateral strike, even absent authorization by Congress or the Security Council (Savage 2013). Moreover, in the Rose Garden, President Obama reiterated that he fervently believed the president has the authority to order a military strike absent congressional authorization. Instead, journalistic accounts have argued that Obama's eleventh hour reversal was a response to concerns, both at home and abroad, about the perceived legitimacy of a unilateral strike. Unlike Libya, for which Obama opted not to seek congressional authorization, a strike against Syria would not have the imprimatur of a UN Security Council mandate or, like Kosovo, of a unified North Atlantic Treaty Organization front.

This emphasis on legitimacy is almost certainly correct--but it is also imprecise. Did Obama change his mind out of fear that Congress would exercise its formal powers to block or terminate an illegitimate military action? For all of the reasons detailed by the unilateral powers literature, this is highly unlikely. (2) Similarly, because the United States wields a veto on the Security Council, the president would never face serious international repercussions for a unilateral strike. Perceptions of the legitimacy of a strike could conceivably affect the president's legacy and the judgment of history; however, it is equally if not more likely that Obama's vacillation in the face of Assad's defiant use of weapons of mass destruction against his own people could harm Obama's legacy more than arcane debates over legalistic minutiae had Obama acted, like virtually all of his predecessors, alone and without congressional sanction. Rather, Obama's reversal seems best explained as an effort to avoid the political costs of going it alone in another military strike in the Middle East.

As such, Obama's Syrian gambit reflects a much larger reality in American politics that political scientists often ignore: the contemporary Congress, despite its many institutional failings, still retains considerable influence over the nation's military affairs. To be sure, Congress no longer declares war. Additionally, for a variety of reasons, the appropriations power is a problematic instrument of congressional influence, one so blunt that it has often proved all but impossible to exercise effectively. (3) Similarly, the War Powers Resolution (WPR) has failed to meet the lofty expectations of its drafters who envisioned it as an instrument for congressional reassertion in war powers without having to use the power of the purse to end military actions of which legislators disapproved. Presidents routinely report their actions to be "consistent with" the WPR, even as they refuse to recognize its constitutionality. Furthermore, presidents consistently refuse to report that American forces are deployed into a zone of "hostilities," which would automatically trigger the WPR's 90-day withdrawal clock, even when American forces are routinely taking fire. And Congress, for its part, has time and again proved unable to start the withdrawal clock on its own initiative (Hinckley 1994).

Despite these repeated legislative failures, several recent studies have argued that Congress may retain, in certain conditions, significant influence over the course of American foreign policy through more informal means. For example, Howell and Pevehouse (2007) show that presidents, when deciding whether or not to respond to international crises arising on the world stage, appear to anticipate Congress's reaction. Their findings indicate that the strength of the president's party in Congress is a strong and significant predictor of the probability of a presidential military response.

Moreover, congressional influence persists after American troops are in the field. Recent scholarship finds that presidents backed by strong co-partisan majorities employ larger scale and longer duration military engagements to pursue their foreign policy goals than do presidents besieged by an opposition-controlled Congress on the other end of Pennsylvania Avenue (Clark 2000; Kriner 2010; Wang 1996). And event history analysis of America's most important post-1945 uses of force suggests that a variety of congressional actions--from introducing and voting on legislative vehicles to curtail a use of force (even when they do not pass), to holding investigative hearings, to publicly speaking out against a war or in favor of it--significantly influence the course of American military interventions (Kriner 2010).

President Obama's decision to seek congressional authorization for his proposed Syrian air strikes complements this recent research by plainly suggesting that such authorization is politically quite valuable, even if it is legally of little consequence. (4) Specifically, this article proposes two main mechanisms through which an authorization could pay significant political dividends for the president. First, seeking legislative sanction for the use of force may rally popular support for a military mission in the near term among an ambivalent public. Bolstering the institutional legitimacy of the president's actions may persuade some undecided Americans to back the president's decision. Second, and perhaps even more importantly, authorizations may diminish the intensity with which potential adversaries in Congress later criticize the administration's actions should a military action fail to unfold according to plan. President Obama hinted at this in his September 4 press conference justifying his decision to go to Congress. Without an authorization vote, "Congress will sit on the sidelines, snipe. If it works, the sniping will be a little less; if it doesn't, a little more" (Obama 2013). (5) If forced to make a decision on an authorization, members who vote "aye" may feel their hands tied politically and thus may be significantly less likely to criticize or vote against a military intervention in the future, even if the mission's costs are higher than anticipated.

Ex Ante Influence: An Experimental Approach

Past scholarship suggests at least two ways in which seeking a congressional authorization for military action against Syria could pay political dividends, one ex ante and one ex post. I begin with the first. In the lead-up to the attack, first seeking and then (hopefully) securing congressional authorization could bolster public support for military action amidst an environment of public uncertainty.

Despite Congress's historically low approval ratings, the vast majority of Americans still express a desire for the legislature to authorize the use of force before it commences. For example, an NBC News poll conducted on the eve of President Obama's surprise announcement asked "Do you think that President Obama should or should not be required to receive approval from Congress before taking military action in Syria?" Almost 80% of Americans said yes. (6) Similarly in September 2002, only one year removed from the 9/11 terrorist attacks, almost two-thirds of Americans agreed that President George W. Bush "should have to get the approval of Congress before taking military action against Iraq." (7)

Thus, it is at least conceivable that even simply seeking congressional authorization might pay political dividends for the president in terms of rallying a skeptical American public behind another military intervention in the Middle East. However, it is important to remember that presidents do not make such highly public overtures and appeals in a vacuum (e.g., Jacobs and Shapiro 2000). Rather, the efficacy of any sort of presidential public appeal is likely to be contingent on the response of other political actors (e.g., Berinsky 2009; Brody and Shapiro 1989; Zaller 1992). As such, the ultimate effect of Obama's decision to seek congressional authorization on public opinion may well depend on how that gambit was received on the other end of Pennsylvania Avenue.

To explore these dynamics in more detail, I embedded an original survey experiment on an online survey. A convenience sample was recruited via Amazon's Mechanical Turk service. Although the sample is not nationally representative, it is considerably more diverse than undergraduate samples routinely used in many international relations studies of public opinion. Moreover, recent research by Berinsky, Huber, and Lenz (2012) demonstrates that replicating experiments on samples recruited in this way yields very similar results to previously published studies with nationally representative samples. (8) The survey was administered on September 2, 2013--a mere two days after Obama's surprise announcement--and it successfully recruited 867 subjects.

To examine the influence both of Obama's surprise decision to seek congressional authorization for the use of force in Syria and of the reaction to this decision from other political elites on public opinion, the experiment randomly assigned subjects to one of four groups. In the control treatment, subjects read the first part of the president's Rose Garden address, stopping at the point where Obama announced that he had decided to use force against Syria: Ten days ago, the world watched in horror as men, women and children were massacred in Syria in the worst chemical weapons attack of the 21st century. Yesterday the United States presented a powerful case that the Syrian government was responsible for this attack on its own people. Now, after careful deliberation, I have decided that the United States should take military action against Syrian regime targets. This would not be an open-ended intervention. We would not put boots on the ground. Instead, our action would be designed to be limited in duration and scope. But I'm confident we can hold the Assad regime accountable for their use of chemical weapons, deter this kind of behavior, and degrade their capacity to carry it out.

Subjects assigned to the authorization treatment read the same excerpt from the speech, but then also received an additional paragraph announcing President Obama's decision to first seek congressional authorization for military action. But having made my decision as Commander-in-Chief based on what I am convinced is our national security interests, I'm also mindful that I'm the President of the world's oldest constitutional democracy. I've long believed that our power is rooted not just in our military might, but in our example as a government of the people, by the people, and for the people. And that's why I've made a second decision: I will seek authorization for the use of force from the American people's representatives in Congress.

Subjects assigned to the next two treatments also read the three paragraphs above. However, those in the congressional support treatment also read the following expression of support for the decision from Senate Minority Leader Mitch McConnell:

Senator McConnell: Today the President advised me that he will seek an authorization for the use of force from the Congress prior to initiating any combat operations against Syria in response to the use of chemical weapons. The President's role as commander-in-chief is always strengthened when he enjoys the expressed support of the Congress.

By contrast, those in the congressional criticism treatment were given the following view attributed to Senator John McCain:

Senator McCain: President Obama is abdicating his responsibility as commander-in-chief and undermining the authority of future presidents. The President does not need Congress to authorize a strike on Syria. The President doesn't need 535 Members of Congress to enforce his own red line. (9)

Nothing in either treatment explicitly stated that a congressional leader supported or opposed the use of force in Syria. Rather, both treatments focused on congressional support for or opposition to the president's decision to seek congressional authorization before acting militarily.

Comparing levels of support for both President Obama and his handling of the crisis in Syria across the control group and the authorization treatment allows us to estimate the influence of the decision to seek authorization itself, isolated from potential confounding factors, on public opinion. The final two treatments allow us to examine how any such effect of the decision to seek the authorization is conditional on the reaction of other political elites to the president's gambit.

Table 1 estimates the effect of each of the experimental treatments on four dimensions of public opinion: support for unspecified military action against Syria, support for a limited air strike and cruise missile attack against Syria, approval of President Obama's job performance in the Syria crisis, and President Obama's feeling thermometer rating. (10)

Because the first two questions asked respondents simply whether they supported or opposed military action, the models in columns 1 and 2 are logits. The approval of Obama's handling of Syria question was measured using a five-point likert scale ranging from strongly approve to strongly disapprove; as a result, model 3 is an ordered logit. Finally, the Obama feeling thermometer asked respondents to rate their feelings toward the president on a 0 to 100 scale, and therefore model 4 presents the results of an OLS regression. (11)

In each model, the explanatory variables of interest are indicators for each of the experimental treatments: the authorization treatment, McConnell support treatment, and McCain criticism treatment. The control group is the omitted baseline. The models also control for each respondent's partisanship, gender, age, and educational attainment. (12)

As shown in column 1, none of the experimental treatments had any effect on subjects' willingness to support the use of force in Syria, broadly defined. In the control group, only 30% of the sample supported "taking military action against Syria in response to the use of chemical weapons," and none of the experimental treatments significantly raised or lowered this level of support from the baseline. However, when asked to consider a more precise scenario--a limited military response with cruise missiles and air strikes--the baseline level of support in the control group rose substantially, and at least one of the experimental treatments had a significant effect on support for the use of force.

Simply reading President Obama's decision to seek congressional authorization for a Syrian strike appears to have had no effect on support for a limited military action; the relevant coefficient, while positive, is small and statistically insignificant. Instead, the influence of Obama's decision to seek congressional authorization on public opinion appears to be dependent on the reaction of other political elites. For example, support for air strikes is highest among those who learned of Senator McConnell's praise for Obama's decision to go to Congress for authorization. First differences derived from simulations holding all other variables constant at their means or medians suggest that the McConnell treatment increased support for a military strike from the control baseline by about 9%. However, when Obama's decision is paired with criticism of it from Senator McCain, Obama receives no increase in support for a limited strike from his decision; indeed, the coefficient for this treatment switches sign and is negative.

A similar pattern emerges when examining respondents' assessments of Obama's handling of the Syrian crisis. By itself, there is little evidence that the decision to seek congressional authorization led Americans to evaluate the president's handling of the situation in the Middle East in a more favorable light. The coefficient for the authorization treatment is positive, but it is not statistically significant. When the president's decision is coupled with a statement from Senator McConnell praising the decision to go to Congress, support for Obama's handling of the crisis rises considerably. First differences suggest that this experimental treatment raises the probability of a respondent approving or strongly approving of Obama's handling of Syria by almost 12%, all else being equal. However, when news of Obama's decision is coupled with criticism of it from Senator McCain, this benefit disappears.

Finally, there is strong evidence that the experimental treatments also affected subjects' feelings toward President Obama more generally. Here, the authorization treatment by itself did have a significant, positive effect on attitudes toward Obama, increasing his feeling thermometer rating by almost four points, on average, from the control group baseline, all else being equal. The McConnell support treatment yielded an even larger positive increase in President Obama's feeling thermometer ratings--more than five points above the control group baseline. (13) However, among subjects who read both of Obama's decision to go to Congress and of Senator McCain's critique of this move, the president's feeling thermometer rating was no higher than in the control group. (14)

Thus, this simple experiment speaks to the potential ex ante benefits presidents may reap in terms of bolstering presidential support for themselves and their foreign policies merely by seeking congressional authorization for the use of force. However, the results also suggest that this benefit is highly conditional on the reaction of other political elites. When members of Congress praise the president for his decision, public support increases substantially; however, congressional criticism of the decision to follow the Constitution and first seek congressional authorization can seriously erode or even eliminate any short-term increase in public support. (15)

Ex Post Influence: Muting Criticism on the Floor and in the Public Sphere

Thus, presidents may seek authorization in the hope of securing immediate benefits in terms of increased public support. Such support may be particularly likely to emerge if the authorization succeeds. However, the most important advantages of a congressional authorization--even of a failed authorization vote--may materialize months afterward as a military venture unfolds. A wealth of scholarship has emphasized Congress's failure in all but the rarest of cases to use the power of the purse, the WPR, or any other measure at its disposal to legislatively compel a president to abandon his preferred military policy course (Fisher 1995; Rudalevige 2005; Schlesinger 1973). However, this does not mean that Congress has remained silent when confronted with military policies with which it disagrees. Rather, on a wide range of military actions from major wars in Korea, Vietnam, and Iraq, to smaller interventions in Lebanon, Somalia, Haiti, and Bosnia, members of Congress have introduced legislation to constrain the commander in chief, investigated presidential policies, and denounced the administration's handling of foreign policy before the television cameras. Recent research suggests that through such actions, even though they do not have the force of law, Congress has exerted significant influence on the duration of major military actions in the post--World War II era (Kriner 2010).

Congressional criticism in each of these venues is politically costly. It can precipitate both real and anticipated shifts in public opinion (Baum and Groeling 2010; Berinsky 2009; Howell and Pevehouse 2007), and it can force presidents to expend political capital in foreign policy that will then be unavailable for other initiatives (Neustadt 1990). Moreover, congressional criticism sends signals of American disunity to our adversaries, which can also affect the military costs that must be paid to achieve the administration's objectives (e.g., Auerswald 2000; Schultz 2001). As a result, presidents facing legislative challenges or even just intense public criticism of their military policies on Capitol Hill may conclude that staying the course militarily is no longer worth the heightened costs and will therefore adjust their policy course accordingly.

Influence on Voting Behavior

If congressional criticism during the course of a military venture is costly, then presidents have strong incentives to do everything possible to minimize such criticism. Seeking congressional authorization for the use of force may be one particularly powerful way of doing so. Even failed authorizations may pay political dividends if those members who vote for the authorization find it more difficult to criticize the president's conduct of a war as the conflict unfolds. Both the 2004 general election and 2008 Democratic primaries provide suggestive evidence of such a dynamic. John Kerry famously voted for the $87 billion supplemental Iraq War appropriation before he voted against it, and he was constantly dogged by allegations of flip-flopping on the war. Similarly, Hillary Clinton and John Edwards had trouble gaining traction with the antiwar left in 2008, both hampered by their 2002 Senate votes to authorize the use of force against Iraq.

But are these examples emblematic of a larger pattern? As an initial inquiry, I first model legislators' vote choices on two initiatives to curtail the use of force in Iraq. The first roll call was on an amendment to H.R. 1815 offered by Lynn Woolsey (D-CA) in 2005 calling on the Bush administration to develop a plan for withdrawing American forces from Iraq. The second, potentially more serious initiative launched under the newly minted and democratically controlled 110th Congress was H.R. 2237, a bill sponsored by Jim McGovern (D-MA) that would compel the president to withdraw American troops from Iraq expeditiously. The key independent variable in the logistic regressions is whether or not each member had previously voted to authorize the war in Iraq in 2002. (16) To examine whether any effect for a previous vote to authorize the Iraq War is limited only to Democratic members, we also include an interaction between the previous authorization vote measure and a dummy variable indicating whether or not each member is a Republican. The models also include a full range of control variables drawn from Kriner and Shen (2014). Specifically, the models control for each member's partisanship, whether he or she was a party leader, each member's number of foreign policy committee memberships, gender, race, and the number of Iraq War casualties sustained within 50 miles of his or her constituency. Table 2 presents the results. (17)

The data is strongly consistent with the hypothesis that a prior vote to authorize the use of force diminishes a member's willingness to vote against that war at a later date, even when the very premise upon which the war was launched proved to be false. In both models, the coefficient for the previous authorization variable is negative and statistically significant. In the 2005 roll call, the coefficient for the previous authorization * Republican interaction is small and statistically insignificant. This shows that members of both parties who voted to authorize the war in 2002 were significantly less likely to vote to curtail it in 2005 than were those who did not. In the 2007 vote on the McGovern bill, all 130 Republicans who had voted to authorize the Iraq War in 2002 voted against the bill; Ron Paul (R-TX) and John Duncan (R-TN), both of whom voted against the 2002 authorization, were the only two Republicans to vote for the McGovern bill. As a result, the interaction is not included in the second model, but the data again suggests that prior authorization votes depressed a member's willingness to later break with the administration on a floor vote for members of both parties.

First differences derived from simulations illustrate the substantive size of the estimated effects. For example, the median Democrat who did not vote to authorize the Iraq War in 2002 has a predicted probability of voting for the Woolsey amendment calling for withdrawal from Iraq of almost 0.75. For Democrats who voted for the authorization, this predicted probability is less than 0.30. Two years later, the predicted probability of voting for the McGovern bill for the median Democrat who had not previously voted to authorize the war was 0.89 versus only 0.38 for those Democrats who had voted to authorize the war in 2002.

Table 3 examines four roll calls involving military engagements in Lebanon, Bosnia, and Kosovo to explore whether this strong negative relationship between whether a member previously voted to authorize a conflict and his or her later willingness to vote to curtail it on the floor also held in prior conflicts. Following rather tumultuous debates, both the House and Senate authorized the American presence in Lebanon in the fall of 1983 before the Beirut barracks bombing. In the weeks and months following the bombing, multiple legislative vehicles designed to either encourage or outright compel President Ronald Reagan to withdraw the marines began moving through both chambers. One of these initiatives, an amendment to H.R. 4185 sponsored by Clarence Long (D-MD) that would cut off funding for the marine mission as of March 1, 1984, received a floor vote in the House before Reagan reversed course and withdrew the marines. The Senate authorized the use of force in Bosnia through Senate Joint Resolution 45 in December of 1995 (the resolution was never brought to the floor for a vote in the House). In 1999, the Senate held two roll calls to curtail operations in Kosovo: the Smith amendment to S. 2057 that would cut off funds for operations in Bosnia after March 31, 1999, without additional congressional authorization and the Hutchinson amendment to S. 2132 that would require reductions in troop concentrations over the next year. Finally, the Senate authorized American military action in Kosovo via Senate Concurrent Resolution 21 (which failed in the House on a 213-213 vote). Later that year, the Senate considered the Smith amendment to S. 1059 that would cut off funding for military operations in Kosovo as of October 1, 1999, without explicit additional authorization from Congress.

The factors driving individual members' calculations on each of these votes are considered in Kriner (2010). Table 3 replicates the models presented there with a new variable: whether each member had previously voted to authorize the conflict. In addition to this explanatory variable of interest and its interaction with member partisanship, the models also control for each member's partisanship, veteran status, foreign policy committee memberships, and constituency characteristics, including the presidential vote share in each member's constituency, the member's vote share in the last election, the number of military personnel in each member's constituency, and whether the constituency is located in the South. (18)

For votes in each conflict, the models show strong and statistically significant negative relationships between previously having voted to authorize a conflict and later voting to curtail it--and this is true even when the authorization never became law, as in Bosnia and Kosovo. Members of both parties who voted to authorize a conflict were significantly less likely than their peers to vote for its early termination after the fact. Substantively, the relationship is also quite large. For example, the smallest coefficients were observed in the two votes to curtail the peacekeeping mission in Bosnia. Simulations reveal that the predicted probability of a Republican member voting to constrain Bill Clinton decreases from 0.69 among those who had not voted to authorize the use of force to 0.32 among those who had voted to authorize military action in 1995.

Of course, it is possible that this relationship is not causal. Members who voted to authorize a use of force may believe in the policy goals being pursued more than those who did not. This may render such members more likely to support staying the course, all else being equal. Alternatively, members who voted to authorize a war may simply hold more hawkish views about the efficacy of military action in American foreign policy. This, too, could make them less likely to vote to curtail an intervention downstream. As a result, these views and assessments, not a politically constraining prior authorization vote, could be driving the observed differences across authorizers and non-authorizers.

However, the exceedingly large splits in voting behavior between those who voted to authorize a war and those who did not, even within the opposition party, are striking. All opposition party members stand to gain by attacking the president's handling of military affairs when the costs of its policies are higher than anticipated. In so doing, they can seek to damage the presidential party's brand name (Cox and McCubbins 1993), weaken the president's standing among the public (Brody 1991), and damage his party's prospects in the next election (Campbell and Sumners 1990; Jacobson 2004; Waterman, Oppenheimer, and Stimson 1991). By declining to challenge the president, these members are foregoing potentially significant political advantages. Perhaps the most plausible explanation for this decision is that such members perceive that they might also pay significant political costs for criticizing a military course of action that they themselves voted to authorize.

Influencing Rhetoric

Finally, voting for legislation that would curtail an ongoing use of force is only one way in which members of Congress can raise the costs of militarily staying the course for the president. Prior research suggests that even by simply criticizing the president's policies in the public sphere, members of Congress can rally public opinion against the White House and impose significant political costs on the president for pursuing his preferred policy course (e.g., Baum and Groeling 2009; Howell and Pevehouse 2007; Kriner 2010; Mayhew 2000). (19) Are members who previously voted to authorize a military action also more reluctant to criticize it and the administration's conduct of military policy in the future if the military venture has failed to meet expectations?

To gain insight into this question I exploit a massive database of more than 7,500 speeches given on the floor of the U.S. House concerning the Iraq War between March 19, 2003, and December 31, 2010 (Kriner and Howell 2013; Kriner and Shen 2014). Table 4 presents results from a negative binomial event count model of the number of speeches criticizing the Iraq War given by each representative in each Congress from the beginning of the war in 2003 through the end of combat operations in Iraq under President Obama in 2010. (20) The independent variable of interest is whether or not each member had previously voted to authorize the Iraq War in 2002 and the interaction of this variable with partisanship. The model also includes all of the additional variables from Kriner and Shen (2014) as controls: partisan affiliation, constituency casualties, whether or not a member is a party leader; each member's foreign policy committee memberships, seniority in the House, and demographic characteristics including veteran status, gender, and race.

Strongly consistent with the hypothesis, the event count models show that members of both parties who previously voted to authorize the Iraq War were much less willingly to vociferously criticize the war on the floor than were members who did not vote to authorize the invasion. The coefficient for the main effect is negative and statistically significant, while the coefficient for the partisan interaction is substantively small and statistically insignificant, indicating no significant differences in the relationship between Democratic and Republican members. Substantively, the effect is also considerable. For example, first differences derived from simulations show that the median Democrat in the 108th Congress who did not vote to authorize the war in Iraq is predicted to give over six floor speeches critical of the Iraq War and the administration's conduct of it. By contrast, the predicted number of antiwar speeches for the median Democrat who voted to authorize the conflict in 2002 is less than two.

Thus, the analysis of Iraq War rhetoric is also strongly consistent with, if not conclusive proof for, the argument that authorizations may pay lasting dividends for presidents. The data suggests that authorizations may give presidents valuable political cover years after the vote occurs by suppressing criticism among those who voted for the authorization, even members of the opposition party who should have strong incentives to expose failings in the administration's policies.

Conclusion

While much foreign policy scholarship has downplayed Congress's importance in shaping the nation's military affairs, presidential actions, such as President Obama's decision to seek congressional authorization before using force against Syria, speak to the continued influence that Congress exerts in foreign affairs through informal means. Rather than legislatively barring the president from deploying troops abroad or cutting off funds for their continued use once deployed, members of Congress most often seek to constrain the commander in chief by ratcheting up the political costs of pursuing his preferred military policy course.

Clearly, one way to minimize congressional opposition and its political ramifications is to pursue military policies that expeditiously achieve their objectives at minimal costs. However, given the unpredictability inherent in any military mission, presidents are often forced to rely on other strategies, such as deciding to seek legislative sanction for their policies in an effort to mute congressional criticism downstream. Consistent with the logic behind such a stratagem, this analysis finds that members of Congress who previously voted to authorize the use of force--even those from the opposition party--are less willing to vote to curtail the military action or criticize it publicly than their co-partisan peers who did not vote to authorize the war. This relationship holds for both presidential co-partisans and members of the partisan opposition. All members of the opposition party stand to gain politically by attacking the president when his military policies fail to go according to plan. However, members who previously voted to authorize a conflict are also thereby linked to it, making it more politically perilous for them to criticize the president and his conduct of military affairs even when a window of opportunity to do so opens.

When viewed through this lens, the decision by President Obama (and similar decisions by his predecessors) to seek congressional authorization for military action (even while maintaining that the president as commander in chief has the authority to order our armed forces abroad unilaterally) makes sense as a conscious effort to tie legislators' hands politically. Indeed, Senator Obama's full-throated criticism of the Iraq War helped catapult him onto the national stage, and as a candidate for the Democratic nomination, Obama used his early opposition to the war while in the Illinois state senate to distinguish himself from his main rivals, both of whom had voted for the congressional authorization in 2002. More than most, President Obama understood the politics of congressional war criticism and the political benefits of muting it if possible. By making members who vote for an authorization more hesitant to vote or speak out against the war in the future, the president can endeavor to blunt potentially costly opposition from the other end of Pennsylvania Avenue.

To be sure, the decision to seek congressional authorization for a Syrian strike was a calculated risk by the president. Interviews with administration officials in the immediate wake of Obama's reversal suggest that many if not most in the administration believed (perhaps mistakenly) that Congress would ultimately authorize the strike-- beliefs undoubtedly buoyed by the strong public support for military action by leading Republicans including John McCain and Lindsey Graham. (21) If the administration had ultimately moved forward with the vote and lost, the vote could have damaged the president and weakened his hand, particularly if he nonetheless decided that a military response was still required. However, this is not necessarily the case. For example, during the war in Kosovo, President Clinton won an authorization vote in the Senate but lost in the House. In hindsight, it appears the support from a number of prominent Senate Republicans outweighed the costs of the failed House vote. Moreover, political commu nications scholarship has emphasized that the media gives disproportionate emphasis and attention to support for presidential policies from the opposition party (Groeling 2010). Such "costly signals" also have disproportionate impact on public opinion (Baum and Groeling 2009). Obama may well have reasoned that forcing more Republicans, particu larly in the Senate, off the fence may have brought many to McCain's and Graham's side. Such a development, aided by the media, may well have brought additional pressure to bear on the House, rallied public opinion behind the use of force, and tied promi nent Republicans publicly to the administration's Middle Eastern policies. The political rewards may well have outweighed the anticipated risks.

References

Ackerman, Bruce, and Oona Hathaway. 2011. "Limited War and the Constitution: Iraq and the Crisis of Presidential Legality." Michigan Law Review 109 (4): 447-517.

Auerswald, David. 2000. Disarmed Democracies: Domestic Institutions and the Use of Force. Ann Arbor: University of Michigan Press.

Baum, Matthew, and Timothy Groeling. 2009- War Stories: The Causes and Consequences of Public Views of War. Princeton, NJ: Princeton University Press.

Baum, Matthew, and Timothy Groeling. 2010. "Reality Asserts Itself: Public Opinion on Iraq and the Elasticity of Reality." International Organization: 64 (3): 443-79.

Berinsky, Adam. 2009. In Time of War: Understanding American Public Opinion from World War II to Iraq. Chicago: University of Chicago Press.

Berinsky, Adam, Greg Huber, and Gabriel Lenz. 2012. "Evaluating Online Labor Markets for Experimental Research: Amazon.com's Mechanical Turk." Political Analysis 20 (3): 351-68.

Brody, Richard. 199T Assessing the President: The Media, Elite Opinion, and Public Support. Stanford, CA: Stanford University Press.

Brody, Richard, and Catherine Shapiro. 1989- "Policy Failure and Public Support: The Iran-Contra Affair and Public Assessments of President Reagan." Political Behavior 11 (4): 353-69.

Buhrmester, Michael, Tracy Kwang, and Samuel D. Gosling. 2011. "Amazon's Mechanical Turk: A New Source of Inexpensive, Yet High-Quality Data?" Perspectives on Psychological Science 6 (1): 3-5.

Campbell, James, and Joe Sumners. 1990. "Presidential Coattails in Senate Elections." American Political Science Review 84 (2): 513-24.

Clark, David. 2000. "Agreeing to Disagree: Domestic Institutional Congruence and U.S. Dispute Behavior." Political Research Quarterly 53 (2): 375-401.

Cox, Gary, and Matthew McCubbins. 1993. Legislative Leviathan: Party Government in the House. Berkeley: University of California Press.

Druckman, James, and Cindy Kam. 2011. "Students as Experimental Participants: A Defense of the Narrow Data Base.'" In Handbook of Experimental Political Science, eds. James N. Druckman, Donald P. Green, James H. Kuklinski, and Arthur Lupia. New York: Cambridge University Press, 41-57.

Edwards, George. 2006. On Deaf Ears: The Limits of the Bully Pulpit. New Haven, CT: Yale University Press.

Faiola, Anthony. 2013. "British Prime Minister David Cameron Loses Parliamentary Vote on Syrian Military Strike." Washington Post, August 29. http://articles.washingtonpost.com/2013-08-29/ world/41576515_1_tougher-response-syria-president-bashar (accessed January 29, 2014).

Fisher, Louis. 1995. Presidential War Power. Lawrence: University Press of Kansas.

Groeling, Timothy. 2010. When Politicians Attack! Party Cohesion in the Media. New York: Cambridge University Press.

Hinckley, Barbara. 1994. Less than Meets the Eye: Foreign Policy Making and the Myth of the Assertive Congress. Chicago: University of Chicago Press.

Howell, William. 2003. Power without Persuasion: The Politics of Direct Presidential Action. Princeton, NJ: Princeton University Press.

Howell, William, and Jon Pevehouse. 2007. While Dangers Gather: Congressional Checks on Presidential War Powers. Princeton, NJ: Princeton University Press.

Jacobs, Lawrence, and Robert Shapiro. 2000. Politicians Don't Pander: Political Manipulation and the Loss of Democratic Responsiveness. Chicago: University of Chicago Press.

Jacobson, Gary. 2004. The Politics of Congressional Elections. New York: Pearson Longman.

Kittur, Aniket, Ed. H. Chi, and Bongwon Suh. 2008. "Crowdsourcing User Studies with Mechanical Turk." In CHI '08 Proceedings of the SIGCHI Conference on Human Factors in Computing Systems. New York: ACM Press, 453-56.

Kriner, Douglas. 2010. After the Rubicon: Congress, Presidents, and the Politics of Waging War. Chicago: University of Chicago Press.

Kriner, Douglas, and Francis Shen. 2014. "Responding to War on Capitol Hill: Battlefield Casualties, Congressional Response, and Public Support for the War in Iraq." American Journal of Political Science 58 (1): 157-74.

Kriner, Douglas, and William Howell. 2013. "Congressional Leadership of War Opinion? Backlash Effects and the Polarization of Public Support for War." In Congress Reconsidered, 10th ed., eds. Lawrence Dodd and Bruce Oppenheimer. Washington, DC: CQ Press, 377-400.

Mayhew, David. 2000. America's Congress: Actions in the Public Sphere, James Madison through Newt Gingrich. New Haven, CT: Yale University Press.

Memmott, Mark. 2013. "Kerry Says Assad, A Thug and a Murderer,' Was Behind Attack." NPR, August 29. http://www.npr.org/blogs/thetwo-way/2013/08/30/217211589/coming-up-kerry -statement-about-the-crisis-in-syria (accessed January 29, 2014).

Munoz, Carlo. 2013. "Syria Strikes Cannot Be Merely Cosmetic,' Say McCain, Graham." The Hill, August 30. http://thehill.eom/blogs/defcon-hill/operations/319755-syria-strikes-cannot-be -merely-cosmetic-say-mccain-grahamm (accessed January 29, 2014).

Neustadt, Richard. 1990 [I960]. Presidential Power and the Modern Presidents. New York: Free Press.

Obama, Barack. 2013. "The President's News Conference with Prime Minister John Fredrik Reinfeldt of Sweden in Stockholm, Sweden." September 4. http://www.presidency.ucsb.edu/ws/?pid= 104040 (accessed January 29, 2014).

Rudalevige, Andrew. 2005. The New Imperial Presidency: Renewing Presidential Power after Watergate. Ann Arbor: University of Michigan Press.

Savage, Charlie. 2013. "Obama Tests Limits of Power in Syrian Conflict." New York Times, September 8. http://www.nytimes.eom/2013/09/09/world/middleeast/obama-tests-limits-of-power-in-syrian -conflict.html (accessed January 29, 2014).

Schlesinger Jr., Arthur. 1973. The Imperial Presidency. Boston: Houghton Mifflin.

Schultz, Kenneth. 2001. Democracy and Coercive Diplomacy. New York: Cambridge University Press.

Shabad, Rebecca. 2013. "140 House Members Say Obama Needs Approval From Congress on Syria." The Hill, August 29. http://thehill.com/blogs/blog-briefing-room/news/319127-55-house -members-say-obama-needs-approval-from-congress-in-syria-strikes (accessed January 29, 2014).

Todd, Chuck. 2013. "The White House Walk-and-Talk that Changed Obama's Mind on Syria." NBCNews.com, August 31. http://firstread.nbcnews.com/_news/2013/08/31/20273128-the-white -house-walk-and-talk-that-changed-obamas-mind-on-syria (accessed January 29, 2014).

Wang, Kevin. 1996. "Presidential Responses to Foreign Policy Crises: Rational Choice and Domestic Politics." Journal of Conflict Resolution 40 (1): 68-97.

Waterman, Richard, Bruce Oppenheimer, and James Stimson. 1991- "Sequence and Equilibrium in Congressional Elections: An Integrated Approach ." Journal of Politics 53 (2): 372-93.

Zaller, John. 1992. The Nature and Origins of Mass Opinion. New York: Cambridge University Press.

DOUGLAS L. KRINER

Boston University

(1.) http://www.whitehouse.gov/the-press-office/2013/08/31/statement-president-syria (accessed January 29, 2014).

(2.) Inter alia, Howell (2003).

(3.) For a thorough overview, see Ackerman and Hathaway (2011)

(4.) Of course, one could easily argue that whether American military actions are properly authorized as articulated in the Constitution is of considerable normative importance. But as a matter of law, it is very unlikely that presidential action absent congressional authorization will result in successful legal action overturning the president, either by Congress or by the courts.

(5.) At the time of his decision, President Obama also seemed confident that he could secure the necessary votes in Congress (Todd 2013).

(6.) NBC News Poll conducted by the polling organizations of Peter Hart (D) and Bill McInturff (R). August 28-29, 2013. N = 700 adults nationwide. Another question asked "In your view, who should have final authority for deciding whether the United States should conduct military air strikes against Syria: Congress or President Obama?" 61% said Congress versus only 30% saying that the decision should rest with President Obama. Pew Research Center/USA Today, September 4-8, 2013

(7.) CBS News /New York Times Poll, September, 2002. http://www.ropercenter.uconn.edu/data_access/ ipoll/ipoll.html (accessed September 13, 2013).

(8.) On the use of Mechanical Turk samples, see also, inter alia, Kittur, Chi, and Suh (2008), and Buhrmester, Kwang, and Gosling (2011). For a thorough review of subject recruitment procedures and an assessment of the generalizability of experimental results obtained from a sample recruited via Mechanical Turk, see Berinsky, Huber, and Lenz (2012). In brief, subjects are recruited from an online marketplace via an advertisement requesting workers to take a survey in exchange for modest financial compensation. For this survey, I paid $.50 for completing the approximately five-minute survey. Participants had to be 18 or older, reside in the United States, and have a greater than 95% acceptance rate for their work on other jobs undertaken through Mechanical Turk. Subjects who met these criteria and accepted the task (or HIT in Turk parlance) were then directed to an online survey hosted on Qualtrics. As with other convenience samples, the demographics of this sample are not nationally representative. Specifically, the sample was somewhat younger and more liberal than the population as a whole. However, the sample is considerably more demographically and geographically diverse than student samples still often used in political science research (for a defense of student samples, see Druckman and Kam 2011) and samples recruited from Mechanical Turk generally compare favorably in terms of demographics with another high-quality Internet study, the unweighted American National Election 2008-2009 Panel Study (Berinsky, Huber, and Lenz 2012, 356-61). Most importantly, Berinsky, Huber, and Lenz have found that replicating a series of classic experiments on samples recruited through Mechanical Turk yields results very similar to those found in published research (2012, 361-65). An additional concern regards the quality of the answers given by an online convenience sample. As a measure of quality control, the survey concluded with an attention filter. Embedded in a paragraph of text was an instruction for respondents to ignore the question itself and to check the other box and enter the numeric sequence 1,2,3 instead. Eighty-two percent of subjects answered the attention filter correctly. Results are virtually identical if those who did not answer the attention filter correctly are excluded from the analysis.

(9.) This quotation was actually taken from Peter King (R-NY); Senator McCain would express similar sentiments the following day. Subjects were informed of this in a debriefing at the conclusion of the survey. The statement was attributed to Senator McCain so that both congressional cues would come from prominent members of the U.S. Senate.

(10.) The precise questions regarding Syria were as follows: Do you think the United States should take military action against the Syrian government in response to the use of chemical weapons or not? Now, more specifically, if U.S. military action in Syria were limited to air strikes using cruise missiles launched from U.S. naval ships that were meant to destroy military units and infrastructure that have been used to carry out chemical attacks, would you support or oppose this U.S. military action in Syria? (This first pair of questions is taken directly from NBC/Wall Street Journal polling, which also observed much more support for the former than the latter). Do you generally approve or disapprove of the job Barack Obama is doing in handling the situation in Syria?

(11.) As a measure of quality control, unobserved timers were used to measure how long each respondent viewed the screen with the experimental treatments. The median respondent read this screen for approximately 53 seconds, and the average time was just over a minute. Less than 5% of the sample read this screen for less than 10 seconds. Given that it was virtually impossible to read even the control group treatment in this amount of time, these respondents are excluded from the analysis in Table 1. Results are similar regardless of whether these respondents are included or if alternative thresholds are used. Finally, 23 subjects refused to answer the feeling thermometer questions and are also therefore excluded from the analysis.

(12.) Simple differences in means tests across experimental treatments yield similar results.

(13.) Wald tests suggest the increase above and beyond that observed in the authorization treatment is not statistically significant.

(14.) As a further robustness check that the effects are real, I conducted a placebo test on support for another Democrat who was included in the list of feeling thermometers: Hillary Clinton. If the experimental treatments are truly causing the differences observed across groups, then they should only affect attitudes toward President Obama. Consistent with this, reestimating the model in column 4 for Hillary Clinton's feeling thermometer rating yields three statistically insignificant coefficients for the three experimental treatment variables.

(15.) The mixed results from the experiment reflect patterns in the observational polling data. Polling immediately after the president's speech indicates that a clear majority of Americans approved of the decision to ask Congress. And despite the assertions of Senator Lindsey Graham and other critics that the decision made Obama appear weak and indecisive, a majority of Americans told pollsters that seeking congressional authorization made Obama appear strong, versus only a third who said that it made him look weak. (When asked "regardless of your opinion on missile strikes, do you approve or disapprove of Barack Obama's decision to ask Congress to authorize U.S. military action against the Syrian government?" 57% approved (ABC News/Washington Post Poll. September 4-8, 2013). Similarly, when asked whether "President Obama's request for congressional approval for a U.S. military response to Syria's use of chemical weapons makes the president look strong or makes the president look weak?" 52% replied strong versus only 33% weak (McClatchy-Marist Poll. September 7-8, 2013). Nevertheless, President Obama's Gallup Daily approval rating showed virtually no change in the weeks before and after his decision to seek congressional authorization. This is consistent with a wealth of presidential scholarship on the limited power of the bully pulpit, more generally (e.g., Edwards 2006).

(16.) Of course, many members of the House in 2005 or 2007 did not serve in 2002. Limiting the analysis only to those members who were also in Congress at the time of the 2002 vote yields very similar results.

(17.) Kriner and Shen (2014) also examine two additional roll calls from 2003 and 2004 (the first is the 2003 $87 billion supplemental appropriations bill, and the second is a bill declaring that the United States was made safer by the removal of Saddam Hussein). Because these bills were not really efforts to curtail the military operation in Iraq, they are not considered here.

(18.) Where two votes are pooled together, the models also include a dummy variable identifying each vote. Results are very similar if separate models are estimated for each roll call.

(19.) Such public criticism can also send costly signals of American disunity to the enemy, as Vice President Cheney was fond of reminding congressional critics during the Iraq War.

(20.) Results ate very similar if focusing only on the 108th Congress, the temporally closest to the authorization.

(21.) See Todd (2013) and Munoz (2013).

Douglas L. Kriner is an associate professor of political science at Boston University. He is the author of After the Rubicon: Congress, Presidents, and the Politics of Waging War and (with Francis Shen) of The Casualty Gap: The Causes and Consequences of American Wartime Inequalities. TABLE 1 Effect of Experimental Cues on Attitudes toward Syrian Strike and Obama Use Limited Obama Obama Force Strike Handling Thermometer Authorization 0.09 0.09 0.16 3.97 * treatment (0.22) (0.20) (0.18) (2.36) McConnell support 0.22 0.37 * 0.45 ** 5.29 ** treatment (0.22) (0.20) (0.18) (2.23) McCain criticism 0.01 -0.14 0.06 0.97 treatment (0.22) (0.20) (0.18) (2.23) Democrat 0.82 *** 0.99 *** 0.89 *** 24.30 *** (including (0.23) (0.21) (0.16) (2.31) leaners) Republican 0.70 ** 0.51 ** -0.29 -12.69 *** (including (0.27) (0.25) (0.22) (2.83) leaners) Female -0.64 *** -0.32 ** -0.32 ** 0.35 (0.17) (0.15) (0.14) (1.74) Age 0.02 ** 0.01 0.01 0.06 (0.01) (0.01) (0.01) (0.09) Education 0.05 0.04 -0.01 1.15 (0.09) (0.09) (0.08) (1.02) Constant -2.02 *** -1.33 *** 31.66 *** (0.44) (0.41) (4.81) 810 810 810 810 R-squared 0.32 Note: The first and second columns report results from logistic regressions. The third column reports results from an ordered logit model. The fourth column reports results from an OLS regression. Robust standard errors in parentheses. *** p < 0.01, ** p < 0.05, * p < 0.1. TABLE 2 Effect of Prior Authorization on Willingness to Vote to Curtail Iraq War 2003 H.R. 2007 H.R. 1813 2237 Voted co authorize in 2002 -1.94 *** -2.68 *** (0.43) (0.51) Voted to authorize * Republican -0.11 (1.03) Ln casualties within 50 miles 0.51 *** 0.87 *** (0.14) (0.19) Leader -1.33 -1.15 (0.99) (1.20) Foreign policy committee memberships -0.66 ** -0.43 (0.32) (0.36) Seniority in chamber 0.09 0.26 *** (0.06) (0.05) Veteran -0.08 0.06 (0.54) (0.51) Black 0.99 ** 1.52 ** (0.45) (0.74) Latino -0.44 -0.43 (0.59) (0.81) Female -0.13 0.75 (0.38) (0.50) Republican -3 40 -5.51 *** (0.68) (0.65) Constant -0.73 -2.32 *** (0.52) (0.65) Observations 428 426 Note: All 130 Republicans who voted to authorize the Iraq War in 2002 who cast a vote in 2007 voted against the McGovern bill; therefore, the Voted to authorize * Republican interaction is not included in this model. Robust standard errors in parentheses. *** p < 0.01, ** p < 0.05, * p < 0.1. TABLE 3 Effect of Prior Authorization on Curtailing Votes in Other Conflicts Lebanon Bosnia Kosovo Voted to authorize -3.23 *** -2.04 ** -3.83 *** (0.35) (1.04) (1.16) Voted to authorize * Republican 0.25 0.39 (0.74) (1.16) Republican -2.03 *** 1.98 *** 0.86 (0.49) (0.70) (1.50) Presidential vote share -2.95 * -6.00 * -5.7 6 (1.79) (3.19) (5.58) Member's vote share -2.04 * -0.09 -13.31 ** (1.24) (3.10) (5.98) Veteran -0.42 -0.91 ** -0.16 (0.30) (0.46) (0.72) Foreign Relations Committee 0.07 0.53 -1.52 (0.47) (0.56) (0.98) Armed Services Committee -0.56 0.24 1.10 (0.51) (0.57) (0.96) Military personnel in -14.77 1.26 -7.25 constituency (24.37) (4.50) (8.05) South -0.57 -0.39 0.61 (0.39) (0.56) (0.88) Constant 4.88 *** 2.08 10.00 * (1.48) (2.63) (5.20) Observations 426 195 98 Note: All 15 Republicans who voted to authorize the use of force in Kosovo voted against the Smith amendment; as a result, the Voted to authorize * Republican is not included in the third model. Robust standard errors in parentheses. The Bosnia model pools data from two votes; the model includes an unreported dummy variable identifying the vote on the Smith amendment. All significance tests are two-tailed. *** p < 0.01, ** p < 0.05, * p < 0.1. TABLE 4 Authorization Votes Reduce Subsequent Floor Criticism of Iraq War (1) Voted to authorize in 2002 -1.29 *** (0.15) Voted to authorize * Republican 0.22 (0.36) Republican -2.84 *** (0.35) Ln casualties w/in 50 miles of O.19 *** district (0.05) Ln casualties w/in 50 miles of 0.08 district * Republican (0.11) Leader -0.06 (0.17) Foreign policy committee 0.39 *** memberships (0.10) Seniority in chamber 0.07 *** (0.01) Military veteran 0 64 (0.18) Female 0.28 * (0.14) Latino -1.58 *** (0.17) African American -0.42 *** (0.15) 109th Congress -0.08 (0.16) 110th Congress 0.30 ** (0.14) 111th Congress -1.36 *** (0.28) Constant 0.86 *** (0.19) Observations 1,769 Note: Robust standard errors in parentheses. All significance tests are two-tailed. *** p < 0.01, ** p < 0.05, * p < 0.1.