by Dr. Douglas Kriner and Dr. Francis Shen

June 19, 2017

from SSRN Website

 

Abstract

America has been at war continuously for over 15 years, but few Americans seem to notice.

This is because the vast majority of citizens have no direct connection to those soldiers fighting, dying, and returning wounded from combat. Increasingly, a divide is emerging between communities whose young people are dying to defend the country, and those communities whose young people are not.

In this paper we empirically explore whether this divide - the casualty gap - contributed to Donald Trump's surprise victory in November 2016.

The data analysis presented in this working paper finds that indeed, in the 2016 election Trump was speaking to this forgotten part of America. Even controlling in a statistical model for many other alternative explanations, we find that there is a significant and meaningful relationship between a community's rate of military sacrifice and its support for Trump.

Our statistical model suggests that if three states key to Trump's victory - Pennsylvania, Michigan, and Wisconsin - had suffered even a modestly lower casualty rate, all three could have flipped from red to blue and sent Hillary Clinton to the White House.

There are many implications of our findings, but none as important as what this means for Trump's foreign policy. If Trump wants to win again in 2020, his electoral fate may well rest on the administration's approach to the human costs of war.

Trump should remain highly sensitive to American combat casualties, lest he become yet another politician who overlooks the invisible inequality of military sacrifice.

More broadly, the findings suggest that politicians from both parties would do well to more directly recognize and address the needs of those communities whose young women and men are making the ultimate sacrifice for the country.
 

Battlefield Casualties and Ballot Box Defeat: Did the Bush-Obama Wars Cost Clinton the White House?
 

I - Introduction

Imagine a country continuously at war for nearly two decades. Imagine that the wars were supported by both Democratic and Republican presidents.

Continue to imagine that the country fighting these wars relied only on a small group of citizens - a group so small that those who served in theater constituted less than 1 percent of the nation's population, while those who died or were wounded in battle comprised far less than 1/10th of 1 percent of the nation's population. ¹

And finally, imagine that these soldiers, their families, friends, and neighbors felt that their sacrifice and needs had long been ignored by politicians in Washington.

Would voters in these hard hit communities get angry? And would they seize an opportunity to express that anger at both political parties? We think the answer is yes. And the proof is the 2016 victory of Donald J. Trump.

Trump's victory over Hillary Clinton has prompted massive speculation about how the political pundits got it wrong. ² Some suggest it was Hillary's poor strategy and lack of messaging, ³ while others point to Trump's ability to connect emotionally with an angry electorate. ⁴

Still others emphasize macro-level forces like the economy. ⁵

With so much post-election analysis, it is surprising that no one has pointed to the possibility that inequalities in wartime sacrifice might have tipped the election. Put simply: perhaps the small slice of America that is fighting and dying for the nation's security is tired of its political leaders ignoring this disproportionate burden. ⁶

To investigate this possibility, we conducted an analysis of the 2016 Presidential election returns. In previous research, we've shown that communities with higher casualty rates are also communities from more rural, less wealthy, and less educated parts of the country. ⁷

In both 2004 and 2006, voters in these communities became more likely to vote against politicians perceived as orchestrating the conflicts in which their friends and neighbors died. ⁸

The data analysis presented in this working paper finds that in the 2016 election Trump spoke to this part of America. Even controlling in a statistical model for many other alternative explanations, we find that there is a significant and meaningful relationship between a community's rate of military sacrifice and its support for Trump.

Indeed, our results suggest that if three states key to Trump's victory - Pennsylvania, Michigan, and Wisconsin - had suffered even a modestly lower casualty rate, all three could have flipped from red to blue and sent Hillary Clinton to the White House.

There are many implications of our findings. First, the findings should signal to the White House that Trump's 2020 electoral fate may well rest on the administration's approach to the human costs of war.

If Trump wants to maintain his connection to this part of his base, his foreign policy would do well to be highly sensitive to American combat casualties.

Many politicians have exhibited casualty sensitivity of course, but if this segment of the electorate is particularly important to Trump's fortunes in 2020, it may suggest a more powerful democratic brake on foreign wars.

Second, the findings are also a lesson for the Democrats and establishment Republicans who are still trying to figure out how to beat Trump.

Our analysis suggests that politicians from both parties would do well to more directly recognize and address the needs of those communities whose young women and men are making the ultimate sacrifice for the country.

Third, the results also raise puzzling questions about the relationship between some of Trump's rhetoric (for instance his highly-publicized argument with a Gold Star family) and his perception amongst communities with higher casualty rates. Further research is required to explore these and other implications.

The paper proceeds as follows.

In Part II, we review the relevant scholarly literature on the political costs of high casualty rates. In Part III, we present our analysis of the relationship between local casualty rates and support for Trump.

In Part IV, we begin to explore the implications of these results for policymaking and campaign strategy.

 

II - Donald Trump and the Politics of War Casualties

Between October 10, 2001 and the 2016 presidential election, almost 7,000 American service members lost their lives in wars in Afghanistan and Iraq.

While the American public initially rallied in support of both conflicts, public support soured as their human costs rose. ⁹
 

Despite the kindling of an Iraqi insurgency and President Bush's embarrassingly premature declaration of "Mission Accomplished," Bush secured reelection in 2004. However, he lost significant electoral ground in states and communities that had paid the heaviest share of the war burden in casualties. ¹⁰

By 2006, the continuing deterioration of the situation in Iraq emboldened Democrats to promise to end the war in the Middle East. That year's midterm elections returned Democrats to power in both chambers of Congress for the first time since before the 1994 Republican Revolution. Underlying this sweeping change was a further erosion in support for the GOP among the constituencies hardest hit by the war.

In both the Senate ¹¹ and the House, ¹² Republican losses were steepest among communities that had suffered disproportionately high casualty rates in Iraq. ¹³

Finally, in the 2008 presidential election one of the starkest points of contrast between Barack Obama and John McCain was their diametrically opposite views on the Iraq War. McCain was a steadfast supporter and argued that the U.S. must assiduously stay the course to ultimate victory. Obama had opposed the war from the start and promised to end the conflict.

Voters ultimately chose Obama in a landslide.

The electoral punishment suffered by Republicans in the 2000s was a story of both casualty and economic inequality. The communities suffering the most from the fighting overseas were communities with lower income and education levels. ¹⁴

These communities, in turn, increasingly turned against political candidates insisting on more combat.

The resulting GOP losses in communities hardest hit by the war echoes findings from previous conflicts. When the United States goes to war, the sacrifice that war exacts in blood is far from uniformly distributed across the country. ¹⁵ And in the Civil War, ¹⁶ Korea, ¹⁷ Vietnam, ¹⁸ and Iraq, ¹⁹ constituencies that have suffered the highest casualty rates have proven most likely to punish the ruling party at the polls.

While previous research tells us much about how incumbent politicians lose votes due to battlefield casualties, it offers few clues as to how a candidate might win back such voters.

In many respects, the bombastic campaign of the billionaire businessman and political neophyte Donald Trump appeared consciously calculated to appeal to communities fed up with fifteen years of costly and inconclusive war.

The core of Trump's nationalist, populist message was to "make America great again." While the details of the message shifted as the campaign developed, Trump regularly praised the military - while also noting that at least some of their efforts seemed to have been for naught.

On the campaign trail, Trump sometimes sounded like a traditional hawk. He repeatedly mocked the Obama administration's passive approach toward the Islamic State and boasted of his intention to "bomb the hell out of ISIS."

Similarly, he derided the Iran nuclear pact as one of the "worst deals" ever and promised a more aggressive posture with increasingly bellicose rhetoric. Channeling his inner Reagan, Trump also called for greater military spending across the board, including on nuclear weapons, even if such moves threatened to trigger a new arms race.

And perhaps above all, Trump regularly pledged in his stump speeches to take care of the military. ²⁰ He noted repeatedly that the military's resources, especially its manpower resources, were "depleted." ²¹

A Trump administration, he promised, would bring fresh manpower and weapons.

However, other Trump campaign themes were decidedly iconoclastic. While few Republicans openly lauded the Iraq War in 2016, Trump vehemently denounced it and the Republican president who waged it. ²²

In a nationally televised debate before the South Carolina primary, Trump minced few words:

"I want to tell you. They lied. They said there were weapons of mass destruction, there were none. And they knew there were none." ²³

Again and again on the campaign trail, Trump labeled Iraq a disaster and pledged to keep the United States out of stupid wars.

As an example of this approach, when asked how to grapple with the quagmire in Syria, Trump sang the virtues of allowing Russia to play the lead role, as it would keep the United States out of another costly and unnecessary foreign war. ²⁴

This theme was consistent with a similar sentiment from his kick-off speech, where he both criticized the war in Iraq and recognized the sacrifice of American troops:

"We spent $2 trillion in Iraq, $2 trillion. We lost thousands of lives, thousands in Iraq. We have wounded soldiers, who I love, I love - they're great - all over the place, thousands and thousands of wounded soldiers." ²⁵

In sum, Trump promised a foreign policy that would be both simultaneously more muscular and more restrained.

Trump promised to rebuild and refocus the military:

"Our active duty armed forces have shrunk from 2 million in 1991 to about 1.3 million today. … Our military is depleted, and we're asking our generals and military leaders to worry about global warming."

And he also promised to be much more reticent in its use:

"Our friends and enemies must know that if I draw a line in the sand, I will enforce it. However, unlike other candidates for the presidency, war and aggression will not be my first instinct.

 

You cannot have a foreign policy without diplomacy. A superpower understands that caution and restraint are signs of strength." ²⁶

III - Assessing Trump's Electoral Performance in High Casualty Constituencies

In one sense, all Americans have been affected by fifteen years of nearly continuous war in Iraq and Afghanistan.

Americans of all stripes have watched each conflict's developments unfold through extensive media coverage, movies, and personal stories from veterans returning from combat.

Indeed, so great are its posited effects on American society that some analysts have proclaimed the emergence of an "Iraq Syndrome," echoing the public skepticism about the efficacy of the use of force and the growing popular reluctance to employ it that emerged after Vietnam. ²⁷
 

However, on another, very tangible dimension, some Americans have experienced the costs of war much more acutely than others.

Most directly, of course, the costs of war have been concentrated on those men and women who fought and died in foreign theaters and on their families.

But Americans' exposure to these costs has also varied significantly according to the experience of their local communities. In the Iraq and Afghanistan wars, for example, seven states have suffered casualty rates of thirty or more deaths per million residents. By contrast, four states have suffered casualty rates of fifteen or fewer deaths per million.

As a result, Americans living in these states have had different exposure to the war's human costs through the experiences of their friends and neighbors and local media coverage. ²⁸

At lower levels of aggregation, the disparities are often even more extreme.

For example,

as of the 2016 election, just over 50% of U.S. counties had experienced a casualty rate in Iraq and Afghanistan of 1 or fewer deaths per 100,000 residents.

 

However, more than a quarter of counties had experienced a casualty rate more than 3.5 times greater, and 10% of counties had suffered casualty rates of more than 7 deaths per 100,000 residents.

 

Voters in such communities increasingly abandoned Republican candidates in a series of elections in the 2000s. ²⁹

To examine whether the Trump campaign was able to reverse the GOP's earlier losses among those constituencies hardest hit by the nation's recent wars, we conduct analyses at both the state and county level.

Following previous research on the electoral impact of local casualties, ³⁰ we operationalize the dependent variable as the change in the two party-vote share received by the Republican candidate from 2012 to 2016.

This allows us to examine where Trump out-performed Mitt Romney four years prior. Moreover, using the change in vote share from one election to the next provides an important measure of statistical control as many factors that affect the GOP vote share in a constituency should have remained roughly unchanged over this short four-year period.

To measure variation in communities' exposure to wartime casualties, we accessed data from the Defense Casualty Analysis System of the Department of Defense on 6,856 American soldiers killed pursuant to operations in Iraq and Afghanistan. ³¹

Of these service members, 6,732 listed home of record information from one of the 50 states, and 10 hailed from the District of Columbia. From this data, we constructed casualty counts for each state and divided them by state population to construct a casualty rate per million residents. ³²

For the vast majority of these soldiers, the DoD also provided a home county of record. ³³ To capture the greater nuance in the uneven geographic allocation of casualties across the country, we constructed casualty counts for each county and then divided them by each county's population to create a casualty rate per 10,000 residents. ³⁴

Because the relationship is easiest to visualize at the state level, we first constructed a scatter plot showing each state's casualty rate on the x-axis and the change in GOP vote share from 2012 to 2016 on the y-axis (Figure 1).

Trump out-performed Romney in forty of fifty states.

However, the clear positive relationship shown in the scatter plot illustrates Trump's ability to make electoral inroads among high casualty states. ³⁵

Figure 1

Trump's Electoral Success in High Casualty States

How to read Figure 1:

 

Figure 1 illustrates that there is a direct relationship between a state's combat casualty rate and the state's support for Donald Trump.

 

As discussed in the main text, states that experienced greater military sacrifice in the war in Iraq were more likely to vote for Trump.

 

Additional statistical analysis confirms that this relationship is robust, even when controlling for alternative explanations.

 

Support for Trump, on the y-axis, is measured as Trump's improvement (or decline) in state vote share as compared to Mitt Romney in 2012. For example, 5% on the y-axis means that Trump won 5% more of the state's votes in 2016 as compared to Romney in 2012.

 

The state casualty rate, on the x-axis, is measured as the per- capita (per 1 million) rate of soldiers from each state who died in combat between 2001 and the 2016 election.

 

In an additional analysis, we found that the same relationship holds when we measure the total number of soldiers killed and wounded in battle.

Could Trump's gains among high casualty states have tipped the balance?

The data suggests it is possible. After all, Trump's victory in the Electoral College depended on razor-thin margins in a handful of key states. Central to Trump's victory was his ability to flip three reliably blue states:

Pennsylvania, Michigan, and Wisconsin.

Trump carried each of these states by less than 1%.

In terms of their share of wartime sacrifice, all three of these states experienced casualty rates in Iraq and Afghanistan that placed them in the middle of the distribution, nation- wide. Michigan's casualty rate was the national median, while Pennsylvania's casualty rate was just above the median and Wisconsin's just below it.

What if each of these states had suffered a lower casualty rate - for example, that of neighboring New York?

Figure 2 presents the estimates obtained from a simple regression model. ³⁶ In each state, our analysis predicts that Trump would have lost between 1.4% and 1.6% of the vote if the state had suffered a lower casualty rate.

As illustrated in Figure 2, such margins would have easily flipped all three states into the Democratic column.

Trump's ability to connect with voters in communities exhausted by more than fifteen years of war may have been critically important to his narrow electoral victory.
 

Figure 2

How Lower Casualty Rates in

Wisconsin, Pennsylvania, and Michigan

might have cost Clinton the election

How to read Figure 2:

 

Figure 2, which is based on the predictive statistical model discussed in the text, graphically examines what would have happened in the 2016 Presidential election if Wisconsin, Pennsylvania, and Michigan had experienced a lower casualty rate.

 

The darker red and darker blue bars on the left plot the actual vote percentage for Trump and for Clinton.

 

The lighter red and lighter blue bars on the right plot the predicted vote percentage, if each of these states had a lower casual rate.

 

Our models suggest that - if there had been a lower casualty rate in each state - Trump would have lost all three.

However, most states are large, heterogeneous places.

The wartime experiences and direct exposure to war costs of residents of upstate and western New York, for example, may look very different from those living in the New York City suburbs.

To account for these intra- state differences and to paint a more nuanced picture, we conducted a follow-up analysis of the relationship between Iraq and Afghanistan war casualties and Trump's electoral success at the county level. The first column in Table 1 presents the results of a bivariate ordinary least squares regression of the change in GOP vote share from 2012 to 2016 on a county's casualty rate.

As in the state-level analysis, the relationship is positive and statistically significant. Trump was even more successful in surpassing Romney's 2012 performance in communities that had suffered disproportionately high casualty rates.

Prior research has shown that Iraq and Afghanistan war casualties are not randomly distributed across the country. Rather, they correlate significantly with other demographics that might also identify communities particularly receptive to Trump's candidacy. ³⁷

To insure that county casualty rates are not just serving as a proxy for another characteristic identifying counties predisposed to support Trump to a greater degree than Romney, we estimated a second regression model including a number of control variables.

Perhaps most importantly, because prior research has shown that recent war casualties have hailed disproportionately from communities with lower levels of income and educational attainment, we control for each county's median family income and percentage of adult residents with a college degree.

Exit polls from 2016 showed that Trump performed well among voters without a college degree; as a result, this is a particularly important control. ³⁸

In addition to income and education, we also included three variables indicating each county's racial composition: the percentage of residents that were white, black, or Latino. Trump struggled to connect with African American voters, and his hard-line immigration policies alienated him from many Latinos.

As a result, we expect Trump to struggle making electoral inroads in counties with large non-white populations.

Finally, we control for the percentage of each county's population that lives in rural areas, as well as the percentage of each county's population that are military veterans. The results are presented in column 2 of Table 1.

Even after including all of these demographic control variables, the relationship between a county's casualty rate and Trump's electoral performance remains positive and statistically significant.

Trump significantly outperformed Romney in counties that shouldered a disproportionate share of the war burden in Iraq and Afghanistan. ³⁹

 

IV - Looking Ahead - An Electoral Check on Military Adventurism?

When President Obama won in 2008, pundits regularly discussed frustration with the Iraq War as a factor motivating voters.

Yet when Obama won re-election in 2012 the wartime narrative was not as prominent. And in the post-election analysis of the 2016 cycle, discussion of war fatigue has been all but absent.

This oversight may plausibly be due to the fact that most American elites in the chattering class have not, at least in recent years, been directly affected by on-going conflicts. Children of elites are not as likely to serve and die in the Middle East, and elite communities are thus less likely to make this a point of conversation.

The costs of war remain largely hidden, and an invisible inequality of military sacrifice has taken hold. ⁴⁰

Our analysis in this paper suggests that Trump recognized and capitalized on this class-based divergence. His message resonated with voters in communities who felt abandoned by traditional politicians in both parties.

If our interpretation of the data is correct, what does this mean for the future of policymaking in the Trump administration? Trump's surprise victory has raised pressing questions about how the political neophyte will exercise his newfound political power.

During the campaign, scores of national security experts, including many prominent Republicans, publicly denounced Trump, warning that he possessed neither the knowledge base nor the temperament to lead the world's most powerful military. ⁴¹

In his first months in office, Trump's continued cavalier rhetoric concerning nuclear weapons and a renewed arms race, coupled with controversial national security staffing decisions - such as removing the chairman of the Joint Chiefs and the Director of National Intelligence from the National Security Council's principals committee and to elevate former Breitbart CEO and political adviser Steve Bannon to the same body - did little to assuage such concerns. ⁴²

As of this writing in June 2017, Trump has significantly increased bombing of ISIS targets in Iraq and Syria. ⁴³

This includes dropping the Massive Ordnance Air Blast (MOAB) bomb, known as the "mother of all bombs." While these actions were criticized by some, they also drew bi-partisan support because some of the bombs were in reaction to gas attacks carried out by Syrian President Bashar al-Assad.

Congress and the courts are unlikely to offer a significant check on President Trump's unilateral authority to direct the nation's military policy. While Congress possesses the constitutional powers needed to provide such a check, perhaps foremost the power of the purse, it often lacks the political will to use them. This will almost certainly be the case for the foreseeable future with Republicans in charge of both chambers of Congress.

Courts can, and have, struck down some executive actions that exceed constitutional limits on executive power, even in the military realm.

However, these cases are limited in number and scope. As a result, public opinion and, ultimately, the ballot box may be the strongest check on presidential recklessness.

All presidents consider the likely judgment of voters, both for their own reelection and for the prospects of a co-partisan successor who can defend their legacies. However, the significant inroads that Trump made among constituencies exhausted by fifteen years of war - coupled with his razor thin electoral margin (which approached negative three million votes in the national popular tally) - should make Trump even more cautious in pursuing ground wars.

Trump, of course, has already proven in his first 100 days that conventional wisdom (and conventional political theory) may not apply to his administration. However, Trump has plainly demonstrated keen electoral instincts and may well think twice before taking actions that risk alienating an important part of his base.

Our results also have important implications for Democrats.

Currently the Democratic Party is engaging in a period of fitful soul searching in a quest to understand its inability to connect with many working class and rural voters who abandoned the party of Roosevelt for Trump.

Much of this introspection has focused on the party's position on trade policy, economic inequality, and emphasis on identity politics.

However, Democrats may also want to reexamine their foreign policy posture if they hope to erase Trump's electoral gains among constituencies exhausted and alienated by fifteen years of war.
 

Table 1:

County Casualty Rates and Change in GOP Vote Share, 2012-2016

Note:

Table 1 presents the results of ordinary least squares regression in which dependent variable is the change in GOP share of the two-party vote from 2012 to 2016. Standard errors in parentheses. All significance tests are two-tailed.

* p < 0.05
** p < 0.01

References

1. For data on the number of Americans who died or were wounded in the Iraq and Afghanistan wars, see: https://www.dmdc.osd.mil/dcas/pages/casualties.xhtml. For an estimate of the number of Americans who have served in theater, see: http://watson.brown.edu/costsofwar/costs/human/veterans
    

2. For an analysis of the election forecast models, see James E. Campbell, et al, A Recap of the 2016 Election Forecasts, 50 POLITICAL SCIENCE & POLITICS 331 (2017).
    

3. AMIE PARNES & JONATHAN ALLEN, SHATTERED: INSIDE HILLARY CLINTON'S DOOMED CAMPAIGN (2017)
    

4. On Clinton's messaging failures, see Molly Ball, Why Hillary Clinton Lost (The Atlantic, Nov 15, 2016); on Trump's connection with an angry electorate, see: Jeff Guo, A New Theory for Why Trump Voters Are So Angry - That Actually Makes Sense (Washington Post, Nov 8, 2016).
    

5. Brad Schiller, Op-Ed: Why did Trump win? The Economy, Stupid (Los Angeles Times, Nov 9, 2016).

    

6. Even prior to the election, we were on record as suggesting this might be the case. As one of us said in a radio interview in September, "… it will be very interesting to see after the election … the extent to which this group [overlooked, primarily white, working class veterans] and others like them found a voice in the Trump campaign… Trump is speaking, in part, to a group who hasn't found their voice heard by other politicians." http://www.accessminnesotaonline.com/2016/09/28/invisible-inequality-in-the-military/
    

7. DOUGLAS L. KRINER & FRANCIS X. SHEN, THE CASUALTY GAP (2010).
    

8. See, infra, Part II.

    

9. Richard Eichenberg, Richard Stoll & Matthew Lebo, War President: The Approval Ratings of George W. Bush, 50
   J. CONFLICT RESOL. 783 (2006); Christopher Gelpi, Peter D. Feaver & Jason Reifler, Success Matters: Casualty Sensitivity and the War in Iraq, 30 INT'L SECURITY 7 (2005/2006); Erik Voeten & Paul Brewer, Public Opinion, the War in Iraq, and Presidential Accountability, 50 J. CONFLICT RESOL. 809 (2006); Matthew A. Baum & Tim Groeling, Reality Asserts Itself: Public Opinion on Iraq and The Elasticity Of Reality, 64 INTERNATIONAL ORGANIZATION 443 (2010).

    

10. David Karol & Edward Miguel, The Electoral Cost of War: Iraq Casualties and the 2004 U.S. Presidential Election, 69 J. Pol. 633 (2007).
     

11. Douglas L. Kriner & Francis X. Shen, Iraq Casualties and the 2006 Senate Elections, 32 LEGIS. STUD. Q. 507, 516-23 (2007); Scott Sigmund Gartner & Gary M. Segura, All Politics Are Still Local: The Iraq War and the 2006 Midterm Elections, 41 POL. SCI. & POL. 95 (2008).
     

12. Christian Grose & Bruce Oppenheimer, The Iraq War, Partisanship, and Candidate Attributes: Explaining Variation in Partisan Swing in the 2006 U.S. House Elections, 32 LEGIS. STUD. Q. 531 (2007)

     

13. This pattern is not unique to the Iraq War. Previous research has shown how voters in high casualty constituencies have punished incumbents associated with the war in conflicts ranging from the Civil War (Carson, Jenkins, Rohde, and Souva 2001), to Korea (Kriner and Shen 2010), to Vietnam (Gartner, Segura, and Barratt 2004; Kriner and Shen 2010).
     

14. Scholarship on military recruiting has long emphasized the importance of economic incentives, in addition to patriotism. Even today, as the Army struggles to retain experienced soldiers, it is significantly increasing re-enlistment bonuses, while the Air Force is considering resorting to "stop-loss" orders to compel pilots to remain in the force even after their terms of service conclude. Lolita Baldor, "Needing Troops, Army Offers up to $90k Bonuses to Reenlist," (Associated Press, June 6, 2017), http://www.startribune.com/needing-troops-army-offers-up-to-90k-bonuses-to-re-enlist/426780681/. John Donnelly, "Stop-Loss an Option for Air Force to Keep Departing Pilots," (Roll Call¸ April 10, 2017), http://www.rollcall.com/news/policy/stop-loss-option-air-force-keep-departing-pilots.
     

15. Kriner & Shen (2010); DENNIS LAICH, SKIN IN THE GAME: POOR KIDS AND PATRIOTS (2013); KATHY ROTH- DOUQUET & FRANK SCHAEFFER, AWOL: THE UNEXCUSED ABSENCE OF AMERICA'S UPPER CLASSES FROM MILITARY SERVICE - AND HOW IT HURTS OUR COUNTRY (2007).
     

16. Jamie Carson et al., The Impact of National Tides and District-Level Effects on Electoral Outcomes: The U.S. Congressional Elections of 1862-63, 42 Am. J. Pol. Sci. 887 (2001)
     

17. Kriner & Shen (2010).
     

18. Scott Sigmund Gartner, Gary M. Segura & Bethany A. Barratt, War Casualties, Policy Positions, and the Fate of Legislators, 53 POL. RES. Q. 467 (2004); Kriner & Shen (2010).
     

19. Douglas L. Kriner & Francis X. Shen, Invisible Inequality: The Two Americas of Military Sacrifice, 46 U. MEM. L. REV. 545 (2016).

     

20. For instance, in his speech Trump said: "We have an Army that hasn't been in this position since World War II, in terms of levels and in terms of readiness and in terms of everything else. We are not capable like we have to be. This will be one of my most important elements. When I talk cost cutting, I do for so many different departments where the money is pouring and they don't even know what to do with it. But when it comes to the military we have to enhance our military. It's depleted. That's the word I tend to use. It's a depleted - we have a very depleted military. We have great people, we have a depleted military. I told you about the jet fighters. Well it's like that with so many other things. So we are going to take care of our military. We're going to take care of our military - the people in our military, the finest people we have." Remarks at a panel hosted by the Retired American Warriors PAC in Herndon, Va., Oct 3, 2016. Online: http://time.com/4517279/trump-veterans-ptsd-transcript/
     

21. See also Trump's speech on April 27, 2016 at an event hosted by the National Interest: http://nationalinterest.org/feature/trump-foreign-policy-15960
     

22. Whether or not Trump had actually been against the war originally was a matter of dispute. See, e.g., Tim Murphy, What Did Donald Trump Say on the Iraq War and When Did He Say It? (Mother Jones, Sept 26, 2016).

     

23. "The CBS News Republican Debate Transcript: Annotated." February 13, 2016, Washington Post, https://www.washingtonpost.com/news/the-fix/wp/2016/02/13/the-cbs-republican-debate-transcript-annotated/

     

24. https://www.washingtonpost.com/news/post-politics/wp/2015/09/25/donald-trump-let-russia-fight-the-islamic-state-in-syria/
     

25. Trump's campaign announcement speech on June 16, 2015: http://time.com/3923128/donald-trump-announcement-speech/
     

26. Remarks on Foreign Policy at the National Press Club on April 27, 2016: http://www.presidency.ucsb.edu/ws/?pid=117813

     

27. John Mueller, The Iraq Syndrome, 84 FOREIGN AFFAIRS 44 (2005).
     

28. Scott L. Althaus, et al, When War Hits Home: The Geography Of Military Losses And Support For War In Time And Space, 56 JOURNAL OF CONFLICT RESOLUTION 382 (2012).
     

29. David Karol & Edward Miguel, The Electoral Cost of War: Iraq Casualties and the 2004 U.S. Presidential Election, 69 J. POL. 633 (2007); Douglas L. Kriner & Francis X. Shen, Iraq Casualties and the 2006 Senate Elections, 32 LEGIS. STUD. Q. 507 (2007); Grose & Oppenheimer, supra note 12. Gartner and Segura, supra note 11.

     

30. Karol & Miguel, supra note 10; Kriner & Shen (2007), supra note 29; Kriner & Shen (2010), supra note 7.

     

31. Specifically, we use the casualty lists provided by the DoD for Operation Enduring Freedom; Operation Freedom's Sentinel; Operation Iraqi Freedom; and Operation New Dawn. https://www.dmdc.osd.mil/dcas/pages/casualties.xhtml
     

32. State population data taken from the 2016 U.S. Census, Population Division.
     

33. Military rules stipulate that the home of record is each soldier's home at the time of enlistment. By contrast, a soldier's "legal residence" can be changed to the location in which they are stationed if they intend to remain there. https://www.army.mil/article/160640. The DoD records provided (or we were able to identify if missing) home county data for 6,475 service members. For most of the remaining 257 service members, the DoD reported their home county as "multiple," indicating that their home city of record spanned multiple counties.
     

34. County-level population estimates were obtained from the Census Bureau's 2015 American Community Survey.

     

35. Utah represents a clear outlier in the scatter plot. Because the dependent variable is the change in the two-party vote share, this is not due to Evan McMullen's success as a third party candidate in the state. Rather, it reflects Romney's exceptional strength in heavily Mormon Utah in 2012, and Trump's failure to connect with the same constituency in 2016. However, excluding Utah from the analysis yields virtually identical results; for example, the bivariate correlation coefficient decreases only slightly from r = .35 with Utah to r = .31 excluding it.
     

36. Estimates obtained from a bivariate regression illustrated by the best-fit line in the scatter plot presented in Figure 1. Column 2 of Table 1 presents results from a multivariate regression using county-level data.

     

37. Kriner & Shen (2010), supra note 7; Kriner & Shen (2016), supra note 19.

     

38. https://www.washingtonpost.com/graphics/politics/2016-election/exit-polls/

     

39. Although not the focus of our present investigation, it is worth noting that the coefficients for many of the control variables also accorded with expectations. Trump significantly over-performed Romney in counties with greater percentages of residents who did not hold a college degree. He under-performed Romney in counties with higher African American and Latino populations, but over-performed in counties with larger white populations. Finally, Trump ran ahead of Romney in rural communities as well as in communities with large shares of military veterans.

     

40. Kriner & Shen (2016), supra note 19.
     

41. https://www.nytimes.com/interactive/2016/08/08/us/politics/national-security-letter-trump.html
     

42. http://europe.newsweek.com/shuffle-national-security-council-players-lambasted-549892. Note: This decision has since been reversed under new National Security Adviser, H.R. MacMaster.
     

43. http://www.afcent.af.mil/Portals/82/Airpower Summary - March 2017.pdf?ver=2017-04-13-023039-397