Bringing the boys back home: Campaign promises and US decision-making in Iraq and Vietnam

This article argues that electoral politics acts as an important constraint on presidential decision-making in war. Going beyond the existing literature’s focus on cases of conflict initiation, it outlines how electoral pressures push and pull presidents away from courses of action which may otherwise be deemed strategically optimal. Importantly, however, these electoral constraints will not just apply on the immediate eve of an election but will vary in strength across the electoral calendar. Together, this conceptual framework helps explain why presidential fulfilment of rhetorical pledges made on the previous campaign trail may be belated and often inconsistent. To probe the plausibility of these arguments, case studies of the closing stages of the wars in Vietnam and Iraq are outlined, drawing on archival and elite interview material. These episodes demonstrate that electoral accountability can be a powerful factor affecting wartime decision-making, but its effect is non-linear, and not easily observed through a narrow focus on particular timeframes.

Defending his December 2018 decision to withdraw troops from Syria, Donald Trump chastised the naysayers who were quick to criticise the move. There was, he maintained, a clear logic to his decision. The president had been elected on a platform in which he had repeatedly signalled his distaste for a continued American military presence in the Middle East. Now, the order to bring some 2000 troops home was simply about following through on his democratic mandate. 'I campaigned on getting out of Syria and other places', tweeted Trump (2018), alluding to reports of a corollary intention to withdraw 7000 troops from Afghanistan. 'Now when I start getting out the Fake News Media, or some failed Generals who were unable to do the job before I arrived, like to complain about me & my tactics, which are working. Just doing what I said I was going to do! ' Trump's claim is notable because of the nakedly political justification of a decision involving the commitment of 'boots on the ground'. To be sure, what Trump actually said about Syria and Afghanistan on the 2016 campaign trail was vaguer than his later statements imply. This may come as no surprise, since Trump's campaign promises can be viewed more as rhetorical vessels designed to reach voters than serious policy proposals (Hall, 2021) and may be the product of his celebrity status more than political nous (Moon, 2019). Yet despite the lack of details, his limited appetite for the sustained presence of combat troops remained a common thread in a more consistent broader narrative which remained heavily critical of recent US interventions in the region. Even when indicating he had a plan to defeat ISIS in Iraq and Syria, for instance, he appeared to prefer airpower to manpower, famously pledging to 'bomb the s-out of them', while calling for other nations to increase their share of the commitment. 'I am going to have very few troops on the ground', he plainly told one interviewer during the campaign (Trump, 2016). If his promises were neither expressed nor subsequently fulfilled with perfect consistency (MacDonald and Parent, 2019), it was not without reason that the president invoked past rhetorical commitments when justifying his controversial decision to expedite the withdrawal from Syria in October 2019 (Lucey, 2019;Trump, 2019). Much the same could be said of the White House's subsequent justification of a deal struck with the Taliban the following February , which contained provisions for the withdrawal of US troops from Afghanistan. Notably, the president reportedly pushed advisers to accelerate the timeline of that drawdown even faster than anticipated in the original agreement, so as to better satisfy his electoral priorities (Gibbons-Neff and Barnes, 2020). The transparent and unapologetically political nature of his rationale for these decisions is striking, and speaks to the point made well in the introduction to this special issue (Lacatus and Meibauer, 2021) that incumbent presidents can expect to be held to account by voters for their ability to act on the rhetorical commitments made on the campaign trail. Like any president, Trump has good electoral reasons for trying to fulfil his promises (Bernstein, 2019;Fishel, 1985).
This article complements contributions to this issue which focus on the construction of Trump's campaign rhetoric by exploring how and when these commitments materially affect the foreign policy of presidents once in office. Placing Trump's record in historical context, it analyses two past cases sharing similar characteristics: Barack Obama's decisions regarding the pace and finality of a troop drawdown schedule in Iraq and Richard Nixon's earlier handling of the denouement of the Vietnam War. These cases are not exhaustive of the range of foreign policy decisions a president is faced with, nor can they be fully representative of the behaviour of all political leaders, each of whom bring particular beliefs and personalities to any given problem. Yet they do stand out as particularly instructive since, like the present incumbent, both Obama and Nixon gained office in part thanks to their opposition to overseas conflict, and both also took far longer than expected to extricate US forces from the battlefield (Boys, 2014;Boys, 2015). By focusing on decisions concerning a conflict in which US forces are already committed, moreover, this article builds on recent attempts to shift existing scholarly attention away from initial decisions to use force and towards in bello decision-making (Payne 2019/2020). Finally, these cases also serve as something of a hard test for the claim that such electoral considerations matter, since we know that leaders who inherit wars and are not seen as 'culpable' for their initiation should be less vulnerable to domestic political punishment for their outcome (Croco, 2011).
Drawing on archival material and a series of interviews with senior decision-makers, the cases reveal several important insights concerning the relationship between campaign rhetoric, the electoral cycle and decision-making in war. Conceiving of the decisionmaking process as a balancing act between the president's often competing interests as both Commander-in-Chief and holder of the highest elected office, it shows how electoral pressures pushed and pulled each president away from courses of action he deemed strategically optimal. The influence of such constraints was not constant, however, but rather varied across the electoral cycle, with both presidents appearing more sensitive to the opinion of voters as an election approached than in its immediate aftermath. Taken together, the cases make clear that presidents may well be trapped by their rhetorical commitments made on the campaign trail, but that the weight accorded to them may figure less strongly in the beginning of a term before coming back to bite as re-election concerns grow. Conceptually, these findings add nuance to existing studies of electoral accountability, suggesting that the electoral cycle introduces a degree of conditionality which is not accounted for by existing studies, which tend to focus on fixed periods in the immediate lead-up to an election. Empirically, it places Trump's recent decisions squarely within the tradition of past precedent.

Elections, campaign promises and war
In approaching decisions about military strategy in war, the American president has dual responsibilities. On one hand, he is Commander-in-Chief, responsible for pursuing a course of action he deems strategically optimal. On the other hand, as an elected officeholder, he must ensure that any course of action he chooses carries minimal electoral risk to his personal political future. In thinking about how electoral pressures affect in-bello decisionmaking, then, we can employ a useful heuristic which conceptualises the decision-making process as a balancing act between two sets of preferences, as illustrated in Figure 1. Like any conceptual lens, this necessarily draws attention to electoral dynamics at the expense of many other alternative or underlying causes of the outcomes of interest. In practice, moreover, presidents will not assess military strategy in the neat two-step process implied here with electoral pressures isolated from consideration of conditions on the battlefield. Nevertheless, while not purporting to offer an exhaustive explanation of any given course of action, this heuristic does offer a reasonable approximation of the president's often competing interests in the national interest and political survival, and thereby a means through which we can analyse the influence of electoral pressures on decision-making.
On one hand sits the 'strategic preference' of the president. This refers to the course of action on the table which the president assesses to carry optimal characteristics in terms of military utility. What exactly this looks like in practice will of course depend on the nature of the broader war objectives and the president's assessment of the relative costs and benefits of the proposed action, measured in terms of expected advantage on the battlefield, likely diplomatic repercussions, financial cost, and so on. This paper is agnostic on what precisely makes up this preference, and on whether this preference may retrospectively be deemed reasonable. It should be possible, however, to infer what this preference is by a close read of contemporary evidence and triangulation with other archival or interview material. A president may rely on military or civilian advisers for guidance, or instead draw on his own instincts and experience, but it is sufficient for our purposes simply to identify what the president assesses to be the option most likely to yield the best military outcome.
On the other side of this decision-making process is the 'electoral preference', taken here to mean the option on the table which is deemed by the president to carry least risk to his electoral fortunes. This depends on the president's understanding of the likely reaction of the electorate to the military strategy under consideration. While it is plausible that electoral incentives may sometimes encourage risky behaviour, such 'diversionary' use of force is broadly considered the exception to the rule. Beyond the various empirical and methodological problems with the diversionary war literature (Levy, 1989;Meernik, 2004: 157−205), foreign policy tends to be considered a toxic issue which is best avoided by incumbents on the campaign trail, even for those with strong records, as James Boys (2021) illustrates with his discussion of the 1992 campaign. Since electoral pressures are therefore more appropriately considered constraints on decision-making, rather than independent sources of strategic preferences, they are conceptualised here as an intervening variable. While not a silver bullet explanation, electoral pressures may be more appropriately considered as pressures which push and pull the president away from the option which he deems to be strategically optimal.
In Figure 1 above, Outcome A is the strategically optimal outcome, reflecting the strategic preference in any given case. If electoral considerations have no effect on decisionmaking -akin to the null hypothesis in this analysis -either because they are ignored by the president or align with the strategic preference, we would expect a correlation between the strategic preference and the ultimate outcome. If, however, as is argued here, electoral pressures do matter, we would expect the outcome to look meaningfully different, with the strategic preference pushed off course by such electoral constraints, as represented by Outcome B. The outcome is generally taken to mean a decision to alter the level of military engagement in a war. The wider the gap between the preferences, moreover, the greater adjustment will be required and thus the greater influence electoral constraints may have. In extreme cases, there may be no viable course of action which sufficiently balances the relative strategic benefits and electoral risks. In this case, presidents may be forced to pick between the unpalatable choice of prioritising either strategic optimality or political survival.
In exploring two cases involving the extrication of US forces from an inherited war of choice, this article only focuses on one specific application of this model. It is a particularly pertinent type of case for this special issue's broader focus, however, since it permits a consideration of exactly how rhetorical commitments made in an earlier campaign may affect the decision-making of presidents once they gain office. Studies of public opinion and war clearly imply why promises to 'bring the boys home' from unpopular engagements overseas might be relatively attractive to aspiring candidates. While the public generally knows little and cares less about foreign policy, access to information about the costs of war can activate public attention (Baum and Potter, 2015;Powlick and Katz, 1998). Since casualties are among the most salient and visceral costs of war, and the burden of these costs tends to fall disproportionately on the average voter, public opinion is perceived to be acutely sensitive to rising numbers of body bags (Mueller, 1973). While vibrant academic debates continue over the true extent of public casualty aversion (see, for instance, Gelpi et al., 2009), the claim is intuitive enough to make a pledge to remove troops from harm's way a generally prudent move in electoral terms.
What the existing literature is less helpful for, however, is shedding light on how a successful candidate reconciles the political need to fulfil such a pledge with the strategic requirements of handling a conflict overseas once in office. To be sure, excellent research exists examining the influence of electoral constraints on the initial decision to enter into a conflict (Gaubatz, 1999), and why democratic leaders not facing re-election tend to be more active and indeed belligerent in matters of foreign policy (Conconi et al., 2014;Potter, 2016;Zeigler et al., 2014). Yet the study of decisions to reduce the level of military commitment to an inherited conflict remains a significant gap in the study of elections and war. It is to this task which this paper now turns, through an examination of decisionmaking in the latter stages of the wars in Iraq and Vietnam.

Barack Obama and the drawdown in Iraq
Barack Obama came into office having pledged to end the 'dumb war' in Iraq, offering a highly specific proposal for a 16-month drawdown which would bring American participation in the conflict to a 'responsible end' 1 . What role did electoral constraints play in the subsequent decisions to effect such a withdrawal, and how did their influence wax and wane across the electoral cycle?

Finding the 'sweet spot'
At first glance, Obama appeared to deliver on his promise early in his first term. Just 5 weeks after becoming President, Obama 2 announced the results of his own 30-day strategic review of the war in Iraq. The withdrawal of US troops was to be placed on an expedited timetable, he said, with all combat brigades out by 31 August 2010, to be followed by those remaining by the end of 2011.
If one scratches beneath the surface, however, it transpires that this withdrawal plan constituted a compromise between the need to deliver on his rhetorical commitments of the previous campaign while balancing strategic considerations of feasibility. Indeed, the commanding general, Ray Odierno, had wasted no time in presenting his views to the incoming administration during the transition. 'The overall risk is extremely high', wrote Odierno of the proposed 16-month timetable (Gordon and Trainor, 2012: 567). He had already put on record his concerns that even the existing schedule, which would see US troops leave by the end of 2011, was 'achievable but the risk is high' 3 . Of particular concern was the pace of the troop drawdown envisaged by the Obama transition team, which would proceed on a linear staircase pattern of one brigade a month.
Not facing the voters again for another 4 years, the Obama administration demonstrated considerable flexibility in finding a 'sweet spot' between the campaign plan and what the President was now being told was the militarily optimal course of action. The first key concessions to the military's plan entailed some semantic sleight of hand. Every time Obama spoke about getting troops out, noticed Colin Kahl, Obama's working lead on the strategic review at the Pentagon, he spoke about 'combat' troops. 'That gives you some space', argued Kahl 4 , explaining that 'the only thing that had to change was the "combat" modifier'. Over the next couple of weeks, the idea of introducing an additional milestone into the Iraq drawdown process was debated and adopted. A mid-point -serving as 'a political marker' -between the start of the drawdown and the December 2011 deadline for total withdrawal would be set, whereby the 'combat phase' of the American military's commitment to Iraq would end. After this point, all remaining troops would be given new 'advise and assist' missions. As Odierno's Political Adviser, Emma Sky 5 approvingly notes, 'it's the same soldiers -same guys, different mission'.
On the question of timelines, the President readily jumped at Bob Gates' suggestion that he accept a compromise of 19 months. 'I'm okay with that', Obama told Gates, adding, 'It's also good politically' (Gates, 2014: 324). By the time the month-long strategy review was completed, the 19 would become 18, a point which was skated over very quickly in Obama's 6 Camp Lejeune speech. The staircase pattern of the drawdown was quickly dropped, too, with Odierno granted authority to stagger the pace to mitigate strategic risks associated with the upcoming Iraqi elections. Finally, Obama would surprise the military by acknowledging Odierno's ideal-case request for the size of the follow-on force that would remain after the so-called 'combat phase' ended: in his speech he would talk of a range of 35,000 and 50,000 troops.
In this opening phase of Obama's presidency, then, the President's electoral promises offered only mild constraints which were quickly and relatively painlessly massaged to satisfy concerns about strategic optimality. 'What this set of compromises allowed us to do was essentially take Obama's campaign pledge to be all out over the course of 16 months, to reducing from 14 brigades to 6 with a different mission over 18 months', recalls Kahl 7 , adding that 'it could be argued that it wasn't a violation of anything he said during the campaign, because he said there would be a residual force, they would have these missions, and he'd listen to his commanders. And so that seemed to kind of square the political circle'.

Closing down
That all US troops would in fact leave Iraq at the end of 2011 was not, however, foreordained. From late 2010 until the autumn of 2011 there remained a fierce debate inside the administration about the size of a follow-on force that was widely expected to remain in theatre. In resolving this debate, the influence of electoral considerations rose exponentially as the 2012 campaign season loomed.
In the administration, it was widely accepted that the strategically optimal course of action comprised the retention of a significant component of troops in Iraq. Both Secretaries of Defence during this period were in favour of such a residual presence (Gates, 2014: 553;Panetta, 2014: 356−357, 393). The commanding general, Lloyd Austin, was adamant on the point, as his initial request for 20,000-24,000 troops made clear. While sensitive to the higher numbers, the senior civilian leadership in the Pentagon was also eminently committed to a follow-on force of some capacity 8 . 'I always thought it was going to be a question of how much or how little would remain, as opposed to leaving', recalls John Brennan 9 , adding, 'I was concerned that things were going to go south if we got out'. Even if the use of drones could help reduce the number of 'boots on the ground', something which Brennan noted 'was in keeping with what President Obama campaigned for, in terms of reducing our engagement in these foreign wars', it was not a panacea which negated the need for any follow-on force. 'It's not just a drone in a box', explains Brennan, adding that 'there is a tremendous, tremendous upstream capability you need', meaning 'a US military presence on the ground was critically important in order to be able to have the infrastructure, the hardware, and the capabilities that are necessary'.
Importantly, the White House seemed to agree that some sort of residual ground commitment was strategically optimal. 'The President really was open to this smaller footprint in terms of a follow-on force', insists Tony Blinken 10 . As National Security Adviser to the Vice President, who had been delegated responsibility for Iraq policy by Obama, he is certain on what the administration saw as the strategic necessity: 'making sure that we continued to have some forces on the ground that could work with the Iraqis, to continue to train them, to prosecute the counterterrorism mission'. Perhaps even stronger proof is the fact that Obama himself chaired a series of NSC meetings through the spring, the net result of which was an approval on 19 May of a provisional plan to leave 10,000 troops in Iraq.
The problem, however, was that this strategic preference was significantly misaligned with the electoral preference of the President. Having run on a specific pledge to withdraw from Iraq, he knew that voters would surely hold him to account if he appeared to abandon that pledge. This explains, for instance, why the administration rejected the commander's initial proposals for a 20,000-24,000 strong force. In the meeting in which General Austin proposed these figures to Pentagon officials apprised of the White House's preference, this became starkly apparent. As soon as the headline figure was mentioned, according to Colin Kahl 11 , Austin was warned: 'A president who campaigned on not leaving a Korea style presence in perpetuity in Iraq is not going to leave a Korea style presence in Iraq. That's not going to happen. This is dead on arrival and it's actually going to hurt you in the White House'.
Even the 10,000-figure to which Obama signed up, which at that stage appeared big enough to satisfy the minimal strategic requirements of continued counterterrorism operations yet small enough to be politically acceptable, soon came under pressure as the 2012 campaign approached. By August, Obama had retreated to a plan for 5000-strong followon force, and in October decided to pull the plug on the whole endeavour, going to zero. While the administration's official explanation for this outcome pointed to difficulties in getting a covering agreement through the Iraqi parliament, even those close to Obama admit that this 'ended up being an excuse, or a public explanation, for what I believe was a policy choice' 12 since 'it's hard to count the number of places where we've accepted a degree of ambiguity or just a complete absence of an agreement to do what we thought was necessary for our mission'.
In fact, what drove the President here was clear to several officials. The Chairman of the Joint Chiefs, for instance, recalls how 'it was very clear; you go back to his campaign promise, and it was going to be zero, it's just this question of how we were going to get there' 13 . 'Look at what he said in the campaign', agrees Doug Lute 14 , who served as 'war czar' on the NSC, noting that 'presidents tend to do what they said'. Perhaps most telling were Obama's public comments during the third presidential debate of the 2012 campaign, when he flatly and falsely refuted Mitt Romney's suggestion that he ever tried to keep troops on in Iraq. 'That is not true', said Obama, 'what I would not have done is left 10,000 troops in Iraq that would tie us down'. 15 In reality, this was precisely the policy Obama pursued between May and August of 2011. Walking back from this position in 2012 would seem to support the view that as the 2012 election drew closer, the idea of remaining in Iraq became a political liability for a candidate running on a platform which reified his extrication of American forces from Iraq. John Brennan 16 summarises the pattern well: 'Throughout this process he was weighing a number of factors and considerations, not just in terms of security requirements, but also some of those political dimensions, as it got closer to the elections'.
This episode captures well the balancing process between a president's strategic and electoral preferences and indicates the non-linear nature of sensitivity to audience costs associated with appearing to renege on a campaign pledge. Whereas in 2009 Obama was willing to defer to military judgements regarding the viability of his electoral preference to withdraw in 16 months, in 2011 his patience for similar compromise was limited. While Obama appears to have agreed with his advisers' perceptions of the strategic necessity of a follow-on force of at least 10,000 troops, this became politically untenable as the shadow of his campaign pledge grew larger. By October, there appears to have been no viable course of action which satisfied both his strategic preference and electoral preference, resulting in a fateful decision by Obama to prioritise his own electoral prospects over the future stability of Iraq.

Richard Nixon and the pursuit of 'peace with honour'
Like Obama, Richard Nixon inherited a long and costly war which he had pledged to end in his presidential campaign, promising to secure 'peace with honour' in Vietnam. In reality, it would be 4 years before he would sign the Paris Peace Accords. What can a focus on the electoral cycle reveal about the key decisions that shaped this protracted path to peace?

"We can't restrain him when he's angry'
That Nixon's strategic preference, at least until late 1970, was to escalate the war in order to achieve victory on the battlefield is made plainly apparent by the president's own words to Kissinger less than 2 weeks after assuming office. 'It seems vitally important to me at this time', wrote Nixon 17 'that we increase as much as we possibly can the military pressure on the enemy in South Vietnam'. 'I call it the Madman Theory, Bob', he later confided to his Chief of Staff, explaining: I want the North Vietnamese to believe I've reached the point where I might do anything to stop the war. We'll just slip the word to them that, 'for God's sake, you know Nixon is obsessed about communism. We can't restrain him when he's angry -and he has his hand on the nuclear button' and Ho Chi Minh himself will be in Paris in two days begging for peace. (Haldeman, 1978: 122) Electorally, of course, this strategic preference flew directly against the spirit if not letter of his pledge to end the war in Vietnam. While the vagueness of his rhetorical commitment gave him some latitude -unlike Obama, he had made no specific promises about exactly how or when he would extricate US forces from the war -the audience costs associated with overt escalation did seem to limit his room for manoeuvre in crafting policy in this early phase of the war. This explains, for instance, why the bombing of Cambodian sanctuaries in March 1969 was a secret operation, with American participation in such indiscriminate attacks denied by the administration acutely aware of the antiwar sentiment at home.
It also appears to explain why Nixon chose to embark on a course of 'Vietnamization', whereby combat duties would progressively transfer from United States to South Vietnamese forces. This was not strategically consistent with Nixon's preferred plan to win the war by military means. Far from it. The national security apparatus unanimously concluded that South Vietnamese armed forces 'cannot now, or in the foreseeable future, stand up to both the VC and sizable North Vietnamese forces' 18 . Instead, the troop withdrawals were his PR plan, the 'salted peanuts' which partially satiated the public's appetite for an end to the war (Kissinger, 1979(Kissinger, : 1480(Kissinger, −1482. 'We buy time with troop withdrawals', Nixon thus reassured General Abrams in an NSC meeting 19 . The ability of electoral considerations to restrain Nixon's strategic preference was notably limited, however. Secretly carpet bombing a neutral country combined with token withdrawals cannot be considered a good example of the effects of electoral accountability. As with Obama, the stage of the electoral cycle appears to have been a crucial reason why Nixon felt able to make such minor concessions to his rhetorical commitments during this period of weak political sensitivity. As Nixon himself explained, 'The enemy had caught him in the beginning of his term with three years more to run' (FRUS: no. 137).

'Your re-election is really important'
The weight Nixon accorded to electoral considerations significantly increased as the next election season approached, much as it did for Obama in Iraq. Through 1971, it became apparent that there was no longer any viable course of action available which satisfied Nixon's desire for peace through military escalation at minimal electoral risk. Instead, he chose to prioritise his political future.
This becomes most evident from a 10 March 1971 conversation between Nixon and Henry Kissinger in the aftermath of the failure of a US-backed invasion of Laos. Framed as a test of the 'Vietnamization' policy which neither Nixon nor Kissinger really placed much hope in to begin with, the outcome of the operation was a chaotic retreat by AVRN forces. 'Henry, I have become completely fatalistic about the goddamn thing', Nixon admitted, concluding that 'we've been heroes long enough' 20 . Abandoning his previous preference for victory on the battlefield, the new route forward was clear: 'we're going to get the hell out and hope and pray that nothing happens before 1972'. Nixon and Kissinger agreed to look to settle with a deal entailing a ceasefire and the return of American prisoners of war (POWs) in return for an American unilateral withdrawal timed to keep South Vietnam on life support until November 1972. Why that date? 'Above all', explained Kissinger, 'your re-election is really important'. If there was a precise moment when the balance between strategic and electoral preferences shifted in favour of the latter, this conversation was very likely it.
This, in effect, was the so-called 'decent interval' strategy, whereby the United States would agree to withdraw unilaterally in return for a promise from Hanoi not to overrun South Vietnam until a sufficient time after the US 1972 election. This would enable Nixon to declare he had won 'peace with honour', fulfilling his campaign pledge and guaranteeing his election. A week after Nixon had warmed to the idea, he repeated it back to Kissinger in another private conversation. In a recording of their discussion 21 , Nixon relished the day when he could 'announce the whole damn thing. And that's that. And then the war's dead as an issue. [Snaps his fingers.] Like that. Out!' Kissinger agreed, adding only that the President needed to factor in the decent interval, too, because 'if we get out after all the suffering we've gone through . . . we can't have it knocked over-brutally, to put it brutally -before the election'. Nixon agreed, 'That's why this strategy works pretty well, doesn't it?' Two months after these discussions, the 'decent interval' was made a realistic prospect by the terms offered in Kissinger's May 31 meeting in the ongoing peace talks in Paris. Dropping his demand for mutual withdrawal, Kissinger effectively signalled that the United States would unilaterally withdraw. In return, a ceasefire-in-place, which allowed Vietcong forces to remain in South Vietnam, would simply mortgage Saigon's future for a short-term truce. 'That's important, because we don't want South Vietnam to fall', Nixon clarified in a briefing with Kissinger 2 days prior 22 . At least not before November, anyway, since 'our major goal is to get our ground forces the hell out of there long before the elections'. Kissinger subsequently clarified the purpose of the deal as follows 23 : Kissinger: So we get through '72. I'm being perfectly cynical about this, Mr. President. Nixon: Christ, yes. Kissinger: If we can, in October '72, go around the country saying, 'We ended the war and the Democrats wanted to turn it over to the Communists'-Nixon: That's right. Kissinger: -then we're in great shape.
This is not merely straw-in-the-wind evidence; similar remarks can be found on at least three other occasions in this period 24 . Throughout these conversations, there is precious little mention of the need to guarantee the survival of President Thieu in Saigon's government, which Nixon (1978: 348) later claimed was a red line. Nor was there any mention by Kissinger of his previously-held belief that a ceasefire-in-place was tantamount to surrender (Kissinger, 1969).
The 'decent interval' strategy now in-play, Nixon carefully calibrated a complementary troop withdrawal schedule. Being able to declare 'peace with honour' necessitated a domestic signal of total disengagement by November 1972. Since total withdrawal also spelled doom for South Vietnam, it could not be timed too early for fear that Saigon would collapse too close to the election. The net result was a series of staggered drawdowns. Three announcements between April 1971 and January 1972 saw troop numbers incrementally drop to 69,000 by May 1972. Though he publicly stated that AVRN improvements made these drawdowns viable, privately Nixon knew the opposite was true, having concluded that the South Vietnamese forces were not up to scratch as early as March 1971, after the botched invasion of Laos. American withdrawal was a question of 'when', not 'if'; 'without question we are gonna get out -cut off this fucker [Thieu]', he reiterated in June 25 . When it came to the issue of timing, of course, 'everything has to be played in terms of how we [i.e. Nixon] survive' 26 .
As Chief of Staff of the Army under Nixon, and with the experience of having commanded all US forces in Vietnam until 1968, General William Westmoreland 27 was wellplaced to assess whether this withdrawal schedule was conditions-based. His view is clear: 'They withdrew on a straight-line basis, regardless of the justification. I mean it was just a withdrawal, a planned withdrawal regardless of other factors. It was all tied into the forthcoming elections . . . it was all politics'. Though Westmoreland's own prescriptions for how to proceed in Vietnam might seem equally unwise in retrospect, Nixon's electorally timed withdrawal schedule effectively gave a steadily deflating lifebuoy to a drowning man. Saigon's long-term future was mortgaged to the short-term goal of Nixon's re-election.
While the precise timings of Nixon's strategy would be the matter of further debate throughout 1972 (Armacost, 2015: 74−79;Hughes, 2015), the evidence presented here demonstrates that by early 1971, Nixon's decision-making on Vietnam had already become driven almost entirely by his electoral preference. This strongly indicates that electoral politics do not only matter in an election year. Furthermore, it offers a powerful example of a case when the balancing process between strategic and electoral preferences yields no satisfactory course of action, forcing the president to choose between incompatible priorities. In Nixon's case, he chose to sacrifice the future of South Vietnam at the altar of electoral expediency.

Conclusion
This paper has argued that electoral considerations act as an important constraint on presidential decision-making in war. Furthermore, it noted that we may usefully conceive of electoral pressures as part of a balancing act between the strategic preference and the electoral preference of the president, whereby electoral considerations push and pull the president away from what is deemed the strategically optimal course of action. After applying this heuristic to past cases of wartime decision-making in Iraq and Vietnam, it seems clear that campaign pledges to 'bring the boys home' exert a strong influence over incumbent presidents when assessing military strategy. This influence may not be constant, however, but rather varies across the electoral cycle. Taken together, these cases help explain why presidential fulfilment of such rhetorical pledges made on the previous campaign trail may be belated and often inconsistent. Viewed in this light, Trump's delayed yet transparent efforts to fulfil his promise to extricate US forces from the 'endless wars' in the Middle East fit squarely within the mould established by his predecessors.

Reviewing the cases
In the case of Obama and Iraq, the decision-making process behind the Camp Lejeune speech suggests that electoral considerations may play a mild role early in a president's term. An electorally preferable course of action would have entailed the fulfilment of Obama's 16-month linear drawdown plan, as articulated in the campaign. In the event, however, the White House seems to have accepted the advice of its military advisers and adapted the plan to offset the perceived strategic risks of the campaign's proposal.
In the decision to abandon plans for a follow-on force after 2011, we find support instead for the idea that electoral constraints begin to bite as the electoral cycle wears on. Though never a supporter of the war in Iraq, Obama did signal a clear willingness to retain troops after 2011, in line with the overwhelming majority view of his advisers. The electoral costs of doing so, however, would have been significant. With the 2012 campaign around the corner, the electoral preference won out, pulling the President away from efforts to keep any troops in Iraq at all. For Nixon, the pattern is similar. Early in his term, Nixon appeared intent on seeking victory in Southeast Asia through escalation almost regardless of the potential domestic political costs of doing so. The somewhat vague pledge to secure 'peace with honour' may have encouraged him to keep much of this belligerence covert or notionally hidden from voters, accompanied by token troop withdrawals, but it was Nixon's strategic preference of victory through military means that won out.
In the second half of his first term, however, Nixon's attention shifted towards his own re-election prospects. Calculating that continued escalation was inconsistent with his political interests, he resolved instead to prioritise his personal fortunes. Devising a strategy of diplomatic manoeuvres and troop withdrawals carefully calibrated to yield maximum political benefit at home, he aimed to finally deliver on the promise of peace, whatever the strategic cost.
These cases are of course not identical. Perhaps most notably, the nature of the strategic and electoral preferences were in opposing directions: for Obama, a quicker and more final troop drawdown favoured his electoral prospects at the expense of perceived strategic optimality; for Nixon, a slower pace of drawdown was critical to guarantee his election even though it was understood to spell strategic defeat for Vietnam in the long-term. This suggests that the direction in which such constraints will push the president -towards continued military engagement or away from itwill depend on the incumbent's assessment of the electoral risks at stake and the stage in the electoral cycle.

Implications for studies of elections and war
In demonstrating the significance of electoral pressures in the decision-making process, this article responds to and echoes recent calls to take elections seriously in analyses of American foreign policy (Johnstone and Priest, 2017;Schwartz, 2009). More specifically, it offers support for what we might term the 'accountability' school, whereby elections are seen as a mechanism of democratic constraint, encouraging a cautious approach towards the use of force (Bueno de Mesquita et al., 2003;Conconi et al., 2014). To be sure, existing research suggests that presidents may be able to reduce the pressure from voters on their policy deliberations by seeking to 'lead' public opinion (Jacobs and Shapiro, 2000), and in some cases may be able to convince voters of the necessity of war (Katz, 2017). Yet while 'selling' a war may continue to attract the lion's share of attention in the broader literature on public opinion and US foreign policy (Casey, 2008;Holland, 2012;Osgood and Frank, 2010), the more pressing concern, at least for incumbents who inherited an unpopular conflict like Obama and Nixon, appears to be the need to be able to 'sell' a peace at the ballot box, even if this entails making decisions on strategy which are assessed to be sub-optimal in terms of military utility.
By focusing on the redemption of campaign pledges, this study also builds on existing work suggesting that it is not simply upcoming elections that matter, but previous ones, too (Potter, 2013). Further, it indicates that the pressures associated with these promises do not exert a linear influence, but vary across the electoral cycle. This implies that electoral accountability has a conditional influence on decision-making, a point which is missed by existing studies which focus primarily on election years, or parts thereof. In what might be considered a 'honeymoon' phase of a first term, the rhetorical commitments made on the campaign trail may exert a relatively weak influence on the decisionmaking behaviour of a newly elected president. Since audience costs are lower at this stage in the cycle (Chiozza, 2017;Fearon, 1994), it makes sense that a president is relatively free to pursue a course of action which satisfies his or her strategic preference. As time goes on, however, the political need to redeem those earlier pledges rises, particularly as presidents tend to make 'promises kept' a key theme of future campaign strategies, much as Trump did even in the 2018 midterm race, as Cora Lacatus (2021) shows. This may be especially true if a president's pledges contained highly specific policy proposals, rather than generalised or ambiguous (see Meibauer, 2021). Though this of course holds for an election year itself, the cases here demonstrate that the anticipation of an upcoming election may have a similarly strong impact on decision-making; both Obama's withdrawal decision and Nixon's adoption of a 'decent interval' strategy came in the penultimate year of each president's first term.
This implication points to the need for more nuance and a wider temporal lens in the study of elections and war, drawing on earlier work which examined the relationship between the electoral cycle and foreign policy more broadly (Armacost, 2015;Nincic, 1992;Quandt, 1986). If we restrict our analysis to an election year, or even the three months preceding an election, as much existing research does, we are liable to mistake absence of evidence for evidence of absence thanks to a too-narrow timeframe. As elected officials and Commanders-in-Chief, American presidents balance the competing preferences dictates by their dual roles, but they do so with one eye on the political calendar.
Building on this paper's findings, three areas for future research stand out. First, while this article has focused on issues involving 'boots on the ground', it might well be fruitful to explore a number of other foreign policy 'tools' which the president has available to him. As Gustav Meibauer's (2021) article in this volume suggests, alternative policies like 'no fly zones' carry their own electoral dynamics. Second, since this paper's arguments are to some extent driven by the president's desire for re-election, it stands to reason that this may look markedly different for second-term presidents, with concerns about legacy plausibly altering the incentive structure and a 'lame duck' status reducing a president's sensitivity to audience costs. Third, noting that both cases examined in this article focus on wars in which the United States struggled to achieve the initial objectives, additional studies of wars fought with a greater degree of success may yield further insight on how changing strategic conditions may influence the strength of electoral constraints.

Funding
The author is grateful for financial support for this research from Merton College and the Department of Politics and International Relations at the University of Oxford, as well as the Economic and Social Research Council.