Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Don't Mention the War versus Escalating Commitment: Political Party Responses to Military Casualties

Don't Mention the War versus Escalating Commitment: Political Party Responses to Military Casualties Abstract Do political parties change their position when military casualties increase? Several studies demonstrate that once military casualties increase, public support for sometimes even the government itself declines. With this potential backlash, once governing parties are faced with military casualties, do they (1) maintain that intervention was the “right thing” to do and even escalate their commitment by becoming even more pro-military or (2) try to avoid the blame and downplay the issue, i.e., “not mentioning the war”? And do the opposition parties become more negative or more positive about the military? To evaluate this, we measure the position on military issues in parties’ election manifestoes. Our dataset comprises 326 party policy changes in eleven Organisation for Economic Co-operation and Development (OECD) countries and focuses on post-Cold War military interventions. By using pooled time-series cross-sectional analysis, we find that opposition parties and governing parties respond differently. Generally, governing parties become more negative in their manifesto and opposition parties more positive. We also demonstrate important differences between party families and pre/post-9/11. Our analyses show that whether political parties change policy course once confronted with negative outcomes depends on their position in office, and also the direction in which they change policy depends on political ideology. Many Western democracies have—almost permanently—troops committed to some troubled region in the world. With some regularity, soldiers’ lives are lost in these missions abroad, especially in large-scale missions that entail high risks, such as the interventions in Afghanistan and Iraq. Foreign interventions affect incumbent support (Mueller 2005; Norpoth and Sidman 2007; Tir and Singh 2013). In particular, higher (own) casualty levels in military interventions abroad decrease public support for the mission (Eichenberg 2005) or even the incumbent itself (Eichenberg, Stoll, and Lebo 2006; Karol and Miguel 2007; Mueller 1970, 1971). Even though studies have challenged the conventional wisdom that the public is casualty adverse (Larson and Savych 2005; Boettcher and Cobb 2006; Gelpi and Feaver 2006; Berinsky 2007; Kriner and Shen 2012), several studies have shown that elites are. In a survey experiment conducted by Gelpi, Feaver, and Reifler (2009), members of the general public, civilian elites, and military senior officials were asked for three hypothetical intervention scenarios about how many casualties they would tolerate. They found members of the mass public to be more willing than the two groups of elites to accept military casualties (Gelpi, Feaver, and Reifler 2009), suggesting different levels of casualty tolerance between elites and the mass public.1 This finding can be explained by the assumption among civil and military elites that in the post-Cold War era, the public no longer tolerates military casualties (Dunlap 1999; Hyde 2000; Smith 2005). Empirical studies suggest that this conventional wisdom is widespread among elites from different strands: the media, politicians, and their staff (Kull and Destler 1999; Lacquement 2003). Given this, how do political parties respond to this? If a military mission becomes increasingly unpopular due to increasing casualties among the nation's soldiers, governing parties have two choices. First, governing parties defend the intervention and even escalate their commitment by becoming more outspoken on the subject in a positive manner (Staw 1976; Fearon 1994; Boettcher and Cobb 2009; Gennaioli and Voth 2011). We label this as the escalate commitment hypothesis. Second, governing parties shift in the same direction as public opinion (Adams 2012), by becoming less positive or even negative about the military. Public policy analysts claim that if governments carry out unpopular policies, they engage in blame avoidance strategies to evade electoral punishment (Kent Weaver 1986; Pierson 1994; Vis and van Kersbergen 2007; Wenzelburger 2014). According to these lines of reasoning, governing parties try to defuse the issue by focusing attention on other issues and taking a more neutral stance toward the military. We label this as the don't mention the war hypothesis. Opposition parties also face a dilemma: do they ride the wave of public dissatisfaction with an unpopular military intervention and attack the government on the military issue (blame the government hypothesis)? Or would such a move be seen as unpatriotic, and will they instead support the troops in the field (rally-around-the-flag hypothesis)? We approach these questions by analyzing the strategic communication of parties in their election manifestoes. These documents are relevant since parties constitute central actors that link and broker the preferences of voters to policy (Green-Pedersen 2007; Gemenis 2013). Even though the content is not often well known to the public, they constitute the core of party's policy and most of all contain promises that parties can later be held accountable for (Walgrave and Nuytemans 2009). Several studies have shown that an increase in military casualties can influence electoral outcomes (Gartner, Segura, and Barratt 2003; Karol and Miguel 2007; Norpoth and Sidman 2007; Dannenbaum 2011), and often have led to heated debates during electoral campaigns. For instance, numerous presidents felt the need to account for military policies and military casualties during their presidential campaigns, such as President George W. Bush who had to deal with anti-war sentiments during his 2004 campaign (Karol and Miguel 2007; Berinsky 2009), or were criticized for not doing so, such as presidential nominee Hillary Clinton in the 2016 campaign over the events in Benghazi. If parties indeed make a strategic evaluation based on electoral sentiments, we should expect to see preemptive shifts in party policy positions in strategic documents such as party manifestoes (Adams, Haupt, and Stoll 2008). Our paper analyzes 326 changes in issue position of seventy-five parties in eleven advanced democracies (Canada, Denmark, France, Germany, Italy, the Netherlands, Norway, Spain, Sweden, the United Kingdom, and the United States), for the post-Cold War period. Since the Cold War, these democracies have fought in several large-scale wars (Bosnia, Afghanistan, First and Second Gulf War) and engaged in multiple post-conflict and state-building missions (both in earlier mentioned conflict areas and through participation in UN missions). We use content-analyzed election manifesto data from the Manifesto Project (MP; Volkens et al. 2015). This allows measurement of (changes in) party positions on the military. We combine the manifesto data with various datasets measuring the military death toll in a conflict,2 the number of troops committed to an area of conflict (Military Balance data), and the total number of battle deaths in conflict areas around the world (Uppsala Conflict Data Program [UCDP] dataset). We also explore differences in party responses before and after 9/11. We found the more military casualties during the previous electoral term, the more governing parties become negative in their next manifesto while opposition parties become more positive. Strategic Options of Governing Parties Most advanced democracies have troops committed to war zones. The missions for which these troops are deployed are very different: some missions are only observatory, while others supply relief aid to distressed areas. Some missions sought to prevent genocide (e.g., the Kosovo conflict of 1999), while the aim of other missions was to bring about regime change (e.g., Iraq, 2003 and Afghanistan, 2002). Some missions carry almost no life-threatening risks for soldiers; the UN mission United Nations Iran–Iraq Military Observer Group (UNIIMOG, 1988–1991) had only one fatality in 3 years. The Second Battle of Fallujah in Iraq, where US soldiers were confronted with the fiercest urban combat since Vietnam, resulted in fifty-four fatalities between November 7 and 16, 2004. Also, some countries—most obviously the United States—have permanent overseas military presence, whereas others are much more reluctant to commit any troops abroad (such as Germany). Still, it is very rare for these eleven countries not to be involved in a mission abroad between 1990 and 2014. Even small countries like the Netherlands and Sweden commit on average one thousand troops per electoral term to missions all over the world. Larger countries like Canada or Italy commit on average about four thousand troops, and the United States has on average almost four hundred thousand troops per electoral term in the period 1990–2014. With different numbers of troops committed, it is unsurprising that the number of soldiers killed in action also differs markedly. Many countries are confronted with either a few dozen military casualties per electoral term, or no military casualties per electoral term. The United States is again an outlier with almost one thousand military casualties per electoral term. Of course, absolute differences in military casualties between countries matter. Still, the within-country variation in military casualties also influences both public opinion and the tone of the national debate. This, in turn, influences the strategic options that government and opposition parties have when they are forced to respond to increases in military casualties. We first consider the strategic options of governing parties. Because they are responsible for decisions regarding a military mission, the electorate sees them as responsible for military casualties. How can they respond to this? First, we posit the escalate your commitment hypothesis. This hypothesis describes the tendency of decision makers to continue with (and sometimes even invest additional resources into) a failing course of action (Brockner 1992; Staw 1976, 1997) in an attempt to justify the costs made thus far (Levy 1997; McDermott 2004). These so-called sunk costs are being perceived as operating costs and decision makers believe it is unlikely that similar costs would be incurred from new initiatives. Decision makers are even more inclined to escalate their commitment to a failing course of action if they need to justify their decisions to others (Karlsson, Juliusson, and Gärling 2005). One of the most frequently mentioned examples hereof is the US commitment to the Vietnam War (Staw 1976; Rubin and Brockner 1985; Whyte 1986; Ripstein 2004). Despite the warning by the undersecretary of state George Ball, that it was unlikely that the United States would meet its objectives in Vietnam (Gettleman et al. 1995, 283), President Johnson escalated the number of troops from sixty-five thousand to five hundred thirty-seven thousand from 1965 to 1968. Apart from foreign policy actions, politicians can escalate their commitment rhetorically to justify the current course of action despite heavy losses. An example hereof is the American President George W. Bush, who argued that “retreating from Iraq would dishonor the men and women who have given their lives in that country, and mean their sacrifice has been in vain” (Boettcher and Cobb 2009, 678). Groups and individuals (Bazerman, Giuliano, and Appelman 1984; Brockner 1992; Whyte 1993) display this tendency of investing more time and money to a decision made earlier despite increasing costs, and basically “throw good money after bad” (Garland 1990, 728). In the context of military interventions, increasing costs are mainly operationalized by tracking the increase in military casualties over the course of a military mission (Staw 1976; Nincic and Nincic 1995; Boettcher and Cobb 2009). Several case studies of decision-making by the US government regarding the 2003 intervention in Iraq highlight how psychological processes such as groupthink (Mintz and De Rouwen Jr. 2010) and loss aversion (Dyson 2006) led to an escalation of commitment to the intervention in Iraq. With the escalate commitment hypothesis, we aim to test whether governing parties indeed continue or increase their commitment to the military cause, despite obvious losses in terms of casualties. Here, and in the remainder of our article, we will translate increasing commitment as a shift toward a more positive position on the military despite the negative consequences suffered from implementing previous policies. H1: The more military casualties in a conflict, the more pro-military governing parties become in their next election manifesto. Second, we posit an alternative hypothesis, the don't mention the war hypothesis. Analyses of party policy shifts indicate that parties shift in the same direction as public opinion (Adams 2012), especially if public opinion is moving away from a party (Adams et al. 2004). Similarly, Budge and Farlie (1983) report that parties stress those issues that work for them and ignore those issues that work against them. Both findings suggest that in response to increasing military casualties in interventions abroad, parties should become more skeptical about the intervention. An assumption of this literature is that parties know public opinion, even on a specific issue like a military intervention. Several studies, in fact, demonstrate that political elites are biased in their beliefs about the effect of military casualties on public opinion. That is, elites assume the public to be more casualty averse than it actually is (Berinsky and Kinder 2006; Gelpi and Feaver 2006; Tomz 2007; Gartner 2008; Berinsky 2009; Mintz, Yang, and McDermott 2011; Bausch 2015; Druckman, Peterson, and Slothuus 2013; for other views see Berinsky 2007). For example, in a survey experiment conducted by Gelpi et al., members of the general public, civilian elites, and military senior officials were asked for three hypothetical intervention scenarios about how many casualties they would tolerate. They found members of the mass public to be more willing than policy elites to accept military casualties (Gelpi, Feaver, and Reifler 2009), suggesting different levels of casualty tolerance between elites and the mass public. This suggests that even if the public does not respond to an increase in military casualties, politicians still assume that they do and change their position in the direction they believe public opinion is. In this paper, we explicitly test whether this is the case. The political reasons to intervene, and whether intervention was deemed a success or a failure, set aside, do politicians assume military casualties are a reason to change political course when contending in the elections? Additional support for the expectation that governing parties become less outspoken toward the military in response to military casualties comes from the literature that analyzes public policy reforms. This literature analyzes how governments implement painful retrenchment of popular public policy (e.g., welfare) programs. By cutting back on these programs, governments risk electoral punishment and to counter this they employ so-called blame avoidance strategies (Kent Weaver 1986; Pierson 1996; Wenzelburger 2014; Vis 2016). To give an example from the International Relations literature, when new information on the absence of weapons of mass destruction in Iraq came out, the Bush administration tried shifting the blame of their decision to intervene to the intelligence community (Bell 2007; Preston 2011). Another strategy is to try and change the agenda. By downplaying a difficult issue and putting new and more popular issues on the agenda, the public might forget about the unpopular war. Following these lines of reasoning, and contra H1, one can expect governing parties to become more neutral or even negative/positive vis-à-vis the military, as a response to increases in the number of soldiers killed during missions abroad. H2:The more military casualties in a conflict, the more negative governing parties become about the military in the next election manifesto. Strategic Options of Opposition Parties What about the options of opposition parties? Opposition parties are often considered as the losers of the electoral game (Riker 1988). Their lack of access to office, and thus the rents, prestige, and policy access associated with office, should motivate them to change strategy, and indeed these parties—in particular parties systematically out of office—have been identified as issue entrepreneurs (Hobolt and de Vries 2015). If public opinion moves against the government, there is an opportunity for the opposition to blame the government for the mess and reap the electoral rewards (Bueno De Mesquita and Downs 2006; Smith 1998). This expectation falls in line with Arena's observation that wars often “attract widespread support early on, with opposition emerging later” (Arena 2015, 109). The analyses about party policy shifts mentioned earlier do not differentiate between opposition and governing parties. They provide the same hypothesis here: with public opinion moving against the intervention, parties will shift in the same direction as public opinion. Hypothesis 3, labeled as the blame the government hypothesis, formalizes these expectations. H3:The more military casualties in a conflict, the more negative opposition parties become about the military in the next manifesto. The previous hypothesis is premised on the assumption that public opinion eventually moves against a military intervention if more and more soldiers are killed. However, shortly after the start of a military intervention, public support generally rises in favor of the intervention. Mueller (1970) analyzed the effect of foreign policy events on (US) presidential approval rates over time, such as the effects of the Korean and Vietnam wars. He found that support for military interventions in the United States peaks in the beginning of a war but steadily declines over time. This is the so-called rally-around-the-flag effect, or rally effect in short (Mueller 1971, 2005). Rally effect generally recedes during an electoral term. But this does not mean that it is politically advantageous to criticize the mission. For instance, when John Kerry criticized President Bush’s handling of the Iraq War during the electoral campaign of 2004, Democratic Senator Zell Miller, stated that “while young Americans are dying in the sands of Iraq and the mountains of Afghanistan, our nation is being torn apart and made weaker because of the Democrats' manic obsession to bring down our Commander-in-Chief” (McCroskey 2006, 324). Also outside campaign time, it is often seen as unpatriotic to criticize an intervention. Content analyses of newspaper editorials during the Iraq War (Chesebro 1997; Dardis 2006) identified a post-9/11 rally effect (Aday, Cluverius, and Livingston 2005; Dimitrova and Strömbäck 2005; Dimitrova and Connolly-Ahern 2007) in Western media. This effect has even been identified as one of the most important determinants of presidential (and wartime) vote in 2004 (Norpoth and Sidman 2007). The rally effect is short lived in general (30 days, Stoll 1987 or 1–2 months, Lian and Oneal 1993); the post-9/11 period marked a relatively long-term increased public support for governments as an exception (over 13 months, Schubert, Stewart, and Curran 2002; Hetherington and Nelson 2003). In other words, post-9/11 criticism toward the mission may actually cause a backlash, and therefore, opposition parties motivated to recapture office must show their patriotic credentials and preemptively support the mission especially if soldiers are getting killed during missions abroad. Here, opposition parties are aware of the rally effect and act accordingly since they have nothing to gain from opposing the government over its policies. In addition, the open political competition within democracies has an effect on the behavior of both governing and opposition parties during international crises. In order for a threat to be credible, democracies can show resolve when threats show strong domestic support and thus an active opposition is absent (Schultz 2013). Moreover, increasing legislative unity influences the likelihood of a government escalating in a conflict (Foster and Gibler 2016). We label this as the rally-around-the-flag hypothesis. H4:The more military casualties in a conflict, the more pro-military opposition parties become in the next manifesto. Ideology, 9/11, and Commitment Party policy shifts take place in a specific context. We consider three variables here that are likely—from a theoretical perspective—to influence the relationship between military casualties and party policy shifts: 9/11, the initial commitment of a party to a mission, and the party's ideology. First, there are important differences in the nature of the interventions before and after 9/11. In the 1990s, most interventions (e.g., those in former Yugoslavia) were humanitarian missions with relatively little risk for soldiers. But after 9/11 most interventions were large-scale large-risk ground operations to first bring about regime change, and eventually quell counterinsurgencies. The former were primarily run by the American army; European armies were more deeply involved in the latter. This brings about three effects, in both Europe and North America. (1) The post-9/11 interventions have a much higher death toll than those in the 1990s. (2) Public opinion toward humanitarian missions may differ from missions aimed at regime change. (3) The political opportunities differ between humanitarian missions and regime change missions. For instance, Dutch opposition parties requested more military intervention in Bosnia in the early 1990s and were critical of the government standing idly by. Instead, opposition over the interventions in Afghanistan and Iraq focused on questions surrounding the legal basis for the interventions, civilian casualties, and the length (prolonging of mandates) of the interventions. In addition, the change in American foreign policy (Schubert et al. 2002; Litwak and Litwak 2007) and grand strategy (Biddle 2005) after 9/11, bypassing the UN Security Council, the creation of ad hoc alliances in both Afghanistan and Iraq (Williams 2008), and the subsequent international debates on transatlantic partnership (Gordon 2002; Noetzel and Schreer 2009; Sperling and Webber 2009; Rupp 2016) may have further motivated political controversy around military interventions beyond the United States. In sum, we expect the impact of our independent variables to be conditioned by 9/11. For that reason, we run a robustness check in which we separate our analyses in a split sample: the pre-9/11 and the post-9/11 period. Second, party responses to events may be driven by party ideology. Right-wing parties, reversely, are traditionally seen as pro-military. In particular, Green parties and socialist parties have anti-military and even pacifistic policy positions. Left-wing parties are less outspoken on the issue. If your initial belief is that interventions do not work, then military casualties reinforce your initial belief. If your initial belief is that interventions do work, you are likely to interpret military casualties as an argument for putting more effort into the intervention. We consider these arguments and to account for differences not only between but also within left and right parties, we include party family as a variable. The reason for using party family rather than the left–right axis is that we do not expect a linear relationship between positions on the military along this axis, and make a distinction between far right, right wing, left wing, and socialist/ecologist parties (Wagner et al. 2017). Third, what about a party's initial commitment to a mission? Governing parties may inherit conflicts, while they have been opposed to the mission in the past (Croco 2011; Croco and Weeks 2016). In practice this rarely happens. Most military interventions are in fact widely supported by parliament. Radical parties are more likely to oppose missions, but these rarely govern. The Online Appendix 3 provides an overview of voting behavior for large-scale expeditionary missions. Currently, we do not theorize additional strategic interactions between governing parties and opposition parties. Taking a position against or in favor of a military mission may be more attractive depending on the positions that rival parties take (Adams and Somer-Topcu 2009; Williams 2015). Following second and third points in this section, we expect party responses to be primarily explained by party ideology and party's initial commitments to the mission, rather than possible external pressures that arise from the behavior of other parties. Design and Methods Our sample comprises political parties elected to parliament (or in the United States, to Congress) from eleven Organisation for Economic Co-operation and Development (OECD) countries (Canada, Denmark, France, Germany, Italy, the Netherlands, Norway, Spain, Sweden, the United Kingdom, and the United States) from 1990 to 2014.3 The sample thus comprises mature Western democracies where we would expect the same incentives in order to secure political survival (Bueno de Mesquita et al. 2002; Valentino, Huth, and Croco 2010), while at the same time differentiating on the national institutional level (e.g., presidential and non-presidential systems) and international institutional level (e.g., mixed membership EU and NATO). We focus on the period since the Cold War and the collapse of the Soviet Union (i.e., 1990), because since then, democracies have fought in several large-scale wars (Bosnia, Afghanistan, First and Second Gulf War) and engaged in post-conflict and state-building missions (such as long-term UN missions), so there are sufficient interventions to study. Description of the Dependent Variable We will analyze changes in party positions on the military using data from the MP. Looking at election manifestoes allows us to track changes in political strategic investment by government and opposition parties over a long period of time (1990–2014). This dataset contains quantitative content analyses of election programs and indicates the percentage of positive or negative references in election manifestoes to a wide range of policy issues, including the military. In addition, it allows us to measure changes over the time period we are interested in—the period between the end of the Cold War and 9/11, and the post-9/11 period. An alternative, such as the Chapel Hill expert surveys (Bakker et al. 2015), ranges from 1999 to 2014 maximum and measures in this time period political stances of parties only on EU foreign and security policy. One item measures parties’ support for peace and security missions also outside the EU context, but since this question was asked in two waves it only provides a single cross-section of change in position. Most military missions take place outside this context. Comparing political responses in and outside the US context over a longer period of time (since the Cold War) is of theoretical relevance. Almost all inferences on how politicians respond to military casualties are based on high-profile cases (such as the Vietnam or the Iraq War) or are based on the US sample alone.4 The MP dataset allows us to look at changes in policy preferences from election to election, over a longer period of time for all the countries in our sample making it more fit for our purposes (Volkens et al. 2015). Hence, the MP dataset provides data on how parties respond in their own communication. We will analyze the change in party position on the military (Y1) as expressed in its election manifesto.5 We measure this position by subtracting the number of negative references from the number of positive references. Positive references are sentences in the manifestoes that express the need to maintain or increase military expenditure; to secure adequate manpower in the military; to modernize armed forces; or for rearmament and self-defense. Negative references are those referring to decreasing military expenditure, disarmament, and reducing (or abolishing) conscription. Because we are interested in change, our dependent variable expresses the difference in percentage of references to the military between the current election manifesto (t) and the previous election manifesto (t−1). A positive number indicates that a party has become more positive about the military, while a negative number indicates that a party has become more negative. The sample mean is 0.057 (see table 1) indicating that parties on average become more positive about the military in our sample. Table 1. Operationalization of main (in)dependent variables Dependent variables . Indicator . Data source . Mean . St. Dev. . Min. . Max. . ∆Position Positiont—position(t−1) Manifesto data (position = per104–per105) 0.014 2.62 −29.25 8 Independent variables Indicator Data source Mean St. Dev. Min. Max. Military casualties Cumulative number of soldiers killed between elections (log transformed) See the Online Appendix 1 2.45 1.33 0 8.12 ∆Seatshare ∆Seatshare = seatsharet—seatshare(t−1) ParlGov 0.000 0.127 −0.566 0.590 War deaths War deaths are the sum of battle deaths + genocide (log transformed) UCDP dataset/United Nations 2.88 5.48 −3.24 13.8 Observation per category In office 1 = Governing party ParlGov 1 = 27.74% 0 = Opposition party 0 = 72.26% Party family 0 = Left wing MP dataset 0 = 21.63% 1 = Right wing 1 = 54.88% 2 = Socialist/ecologist 2 = 23.49% Post-9/11 Post-9/11 party manifestoes 1 = 21.78% 1 = post; 0 = ante 0 = 78.22% Dependent variables . Indicator . Data source . Mean . St. Dev. . Min. . Max. . ∆Position Positiont—position(t−1) Manifesto data (position = per104–per105) 0.014 2.62 −29.25 8 Independent variables Indicator Data source Mean St. Dev. Min. Max. Military casualties Cumulative number of soldiers killed between elections (log transformed) See the Online Appendix 1 2.45 1.33 0 8.12 ∆Seatshare ∆Seatshare = seatsharet—seatshare(t−1) ParlGov 0.000 0.127 −0.566 0.590 War deaths War deaths are the sum of battle deaths + genocide (log transformed) UCDP dataset/United Nations 2.88 5.48 −3.24 13.8 Observation per category In office 1 = Governing party ParlGov 1 = 27.74% 0 = Opposition party 0 = 72.26% Party family 0 = Left wing MP dataset 0 = 21.63% 1 = Right wing 1 = 54.88% 2 = Socialist/ecologist 2 = 23.49% Post-9/11 Post-9/11 party manifestoes 1 = 21.78% 1 = post; 0 = ante 0 = 78.22% Open in new tab Table 1. Operationalization of main (in)dependent variables Dependent variables . Indicator . Data source . Mean . St. Dev. . Min. . Max. . ∆Position Positiont—position(t−1) Manifesto data (position = per104–per105) 0.014 2.62 −29.25 8 Independent variables Indicator Data source Mean St. Dev. Min. Max. Military casualties Cumulative number of soldiers killed between elections (log transformed) See the Online Appendix 1 2.45 1.33 0 8.12 ∆Seatshare ∆Seatshare = seatsharet—seatshare(t−1) ParlGov 0.000 0.127 −0.566 0.590 War deaths War deaths are the sum of battle deaths + genocide (log transformed) UCDP dataset/United Nations 2.88 5.48 −3.24 13.8 Observation per category In office 1 = Governing party ParlGov 1 = 27.74% 0 = Opposition party 0 = 72.26% Party family 0 = Left wing MP dataset 0 = 21.63% 1 = Right wing 1 = 54.88% 2 = Socialist/ecologist 2 = 23.49% Post-9/11 Post-9/11 party manifestoes 1 = 21.78% 1 = post; 0 = ante 0 = 78.22% Dependent variables . Indicator . Data source . Mean . St. Dev. . Min. . Max. . ∆Position Positiont—position(t−1) Manifesto data (position = per104–per105) 0.014 2.62 −29.25 8 Independent variables Indicator Data source Mean St. Dev. Min. Max. Military casualties Cumulative number of soldiers killed between elections (log transformed) See the Online Appendix 1 2.45 1.33 0 8.12 ∆Seatshare ∆Seatshare = seatsharet—seatshare(t−1) ParlGov 0.000 0.127 −0.566 0.590 War deaths War deaths are the sum of battle deaths + genocide (log transformed) UCDP dataset/United Nations 2.88 5.48 −3.24 13.8 Observation per category In office 1 = Governing party ParlGov 1 = 27.74% 0 = Opposition party 0 = 72.26% Party family 0 = Left wing MP dataset 0 = 21.63% 1 = Right wing 1 = 54.88% 2 = Socialist/ecologist 2 = 23.49% Post-9/11 Post-9/11 party manifestoes 1 = 21.78% 1 = post; 0 = ante 0 = 78.22% Open in new tab Description of the Independent Variables The cost of war will be operationalized by the number of soldiers of a country killed during missions abroad between elections (see figure 1). This variable adds the number of these military casualties by month up until the month before the elections and is based on the number of military casualties on a monthly basis. These data have been collected through open data sources, such as army archives, rolls of honor, and veteran websites.6 Military casualties are not normally distributed due to the large differences between countries such as the United States and Norway in terms of the number of soldiers deployed. As figure 1 visualizes, we have used a log transformation of the variable. For the same reasons, we have transformed the total number of war-related deaths (the sum of battle-related deaths and genocide victims) into a log variant. Figure 1. Open in new tabDownload slide Distribution of military casualties variable Figure 1. Open in new tabDownload slide Distribution of military casualties variable Our second independent variable, difference in seatshare, measures the change in seat share between the previous election (t−1) and the election before that (t−2). Several studies demonstrate that losing elections spurs party change (Budge 1994; Somer-Topcu 2009). Therefore, we add it as a control variable to see whether changes in the size of a government coalition have an effect. War deaths, our third independent variable, accounts for the number of civilians and soldiers killed of all striving parties during regular warfare. According to UCDP, battle-related deaths occur in what can be described as “normal” warfare involving the armed forces of the warring parties. This includes traditional battlefield fighting, guerrilla activities, bombardments of military units, cities, and villages etc. The targets are usually the military itself and its installations, or state institutions and state representatives, but there is often substantial collateral damage in the form of civilians killed in crossfire, indiscriminate bombings etc. All deaths—military as well as civilian—incurred in such situations are counted as battle-related deaths (Eriksson 2015). If we would measure only the (own) military casualties suffered during conflicts, it is not clear whether political parties respond to these or (political debate on) the casualties brought on by a conflict in general. For instance, despite relatively low (own) military casualties, the ongoing war in Syria is leading to heated political debates (MacAskill 2013). The UCDP dataset does not include one-sided violence directed toward citizens as battle-related deaths, so we have included data on the Rwanda Genocide (1994), Srebrenica massacre (1995), and Darfur (2004–2007) and have chosen the average of the low and high estimates indicated by international organizations and Non-governmental organizations. The reason to include these numbers is that genocides have caused heated political debates on UN mandates, and the episodes in Rwanda and Srebrenica even inspired the so-called “responsibility to protect” doctrine aimed at the protection of the civilian populations during peacekeeping missions (Tardy 2011; Hultman 2013). This can have consequences on the level of party politics as well. For instance, Dutch opposition parties blamed the government for “not doing enough” during the Yugoslav civil war (see the Online Appendices 1 and 3). Our fourth independent variable, in office, comes from the ParlGov dataset (Döring and Manow 2016). In office codes whether a political party was in office before the election (1) or not (0). A second dummy variable differentiates between elections before 9/11/2001 and after 9/11/2001, which allows us to re-run the analysis for a pre- and post-9/11 sample. The fifth independent variable, party family, is a restructured version of the original party family variable from the MP dataset. Following from our discussion in the theory section, we distinguish between the far right (ethnic, regionalist), the center right (nationalist, agrarian, conservative, liberal, and Christian democratic parties), left wing (social democratic parties), and socialist/ecological parties. Since none of the far-right parties in our sample ever took office, these are dropped from the analysis. As mentioned in the theory section, and illustrated by the differences in party position per party family in figure 2, we do not expect a linear effect for left-right ideology. In addition, left–right positions contain much more information than is necessary to test our hypothesis. Party family allows us to account for outliers on both the right side (ethnic and regionalist parties) and the left side (socialist and green parties). Even though we would have preferred to employ a separate category for both far right and far left, there are not enough observations in our data to do so. For instance, there are only four observations in which nationalist parties are in office. Such a small n would be problematic when treating this as a separate category in an interaction model. To ensure the far-right parties do not influence our result, we ran a separate analysis to check whether the effects for right-wing parties change. They do not. In addition, there are no observations for ethno-territorial parties that were in office in our sample. So, these parties are dropped from the analysis. Figure 2. Open in new tabDownload slide Box plot of position on the military by party family Note: The box plot presents the spread of position on the military per party family. Operationalization of position on the military is discussed in the methods section. Figure 2. Open in new tabDownload slide Box plot of position on the military by party family Note: The box plot presents the spread of position on the military per party family. Operationalization of position on the military is discussed in the methods section. We do not use public opinion about a mission as an independent variable. From a practical perspective, these data are not available for the time frame and country selection we are interested in. From a theoretical perspective (as noted earlier) what is more relevant than actual public opinion shifts is the belief that public opinion is shifting. We noted there that is likely to occur if military casualties increase. Estimation Method We conduct cross-sectional time-series analyses since this model will allow us to account for the dynamics between our dependent variable, the difference in party positions, and our independent variables. Such models have issues with autocorrelation (Beck and Katz 1995). Even though lagged variables incorporate feedback over time in our model, it will introduce the issue of heteroskedastic error terms at the same time (Stimson 1985). In addition, the biases of lagged variables associated with trends within our independent variables and error term can annul the effect of the theoretical model (Plümper, Troeger, and Manow 2005). We want to avoid the lagged dependent variable taking up part of the explanatory variance within the theoretical. Also since every country in our sample does not hold its election simultaneously, our panel is unbalanced. We correct for this in two ways: (1) by estimating autocorrelation within panels as well as heteroskedasticity across panels by applying Feasible Generalized Least Squares (FGLS) regression corrected for autocorrelated errors and (2) by constructing confidence intervals around relative changes and specifying the type of standard error reported robust to misspecification. Therefore, FGLS will be our estimation technique since it can estimate autocorrelation (AR[1]) within panels, as well as heteroskedasticity across panels. Another reason to opt for FGLS is a theoretical one. Fixed-effect models do not allow for time invariant variables (such as party family), and such variables become especially problematic from a theoretical point of view when using interactions. We are interested in the question whether party family has an effect on our main effect, and not whether party family varies over time. Our hypotheses assume a difference in the response of government and opposition parties to military casualties. We model this by means of an interaction effect between the government variable and the military casualties variable. We also discussed that party responses may differ per party family. To model this, we add a three-way interaction effect between the government variable, the military casualties variable, and party family. To this end, we follow the recommendations put forward by Brambor, Clark, and Golder (2006). Do Parties Become More Positive or More Negative about the Military? Do parties change position on the military, that is, do they become more positive or negative when there are more military casualties? Table 2 presents the results of our two regression analyses. Since the main hypotheses are tested with interaction effects, it is more informative to interpret the marginal effects of these interactions. Figure 3 displays the marginal effects of military casualties for government and opposition parties in general (to the left in figure 3, based on model 1 in table 2) and per party family (to the right in figure 3, based on model 2 in table 2). First, let us interpret the findings from model 1. We report a significant negative effect of military casualties for governing parties (b = −0.23), and a significant positive effect of military casualties for opposition parties (b = 0.2). To illustrate this, we turn to Denmark in 2011. It suffered thirty-six military casualties between the 2007 and 2011 elections (log = 3.6). In this situation, an opposition party is predicted to become more positive about the military (3.6 × 0.2 = 0.72) and a governing party to become more negative (3.6 × −0.23 = −0.83). The model perfectly predicts the shift of the governing party, the Conservatives (−0.83), while the other governing party, the Liberal Party, did not shift. The main opposition party—the Social Democratic Party—became more positive (0.85), as predicted by the model. A similar result can be found in the United States. In 2008, confronted with increasing military casualties from Iraq and Afghanistan, the Republicans (governing party) reduce positive mentions of the military (−0.28), and Democrats (opposition party) increase positive mentions (3.75). For example, the Democratic Party platform contained promises such as “we will rebuild our armed forces to meet the full spectrum needs of the new century” and “we support plans to increase the size of the Army by 65,000 troops and the Marines by 27,000 troops.” It is remarkable that the Democrats become so much more pro-military, while at the same time advocating a retreat from Iraq. Consider a few sentences from the “Ending the War in Iraq” section of the Democrats’ 2008 party platform. “Our men and women in uniform have performed admirably while sacrificing immeasurably. Our civilian leaders have failed them.” “We will give our military a new mission: ending this war and giving Iraq back to its people.” What this suggests is that a radical break with policy (retreat from Iraq) can only be justified if a commitment to the military is shown. Figure 3. Open in new tabDownload slide Average marginal effect of military casualties on changes in party position Note: *p < .05. Dots are average marginal effects from the interaction between government × military casualties (model 1, see table 2) and government × military casualties × party family (model 2, see table 2). Bars represent 95 percent confidence intervals. Points to the left of the vertical bar stem from model 1, while those to the right stem from model 2. Figure 3. Open in new tabDownload slide Average marginal effect of military casualties on changes in party position Note: *p < .05. Dots are average marginal effects from the interaction between government × military casualties (model 1, see table 2) and government × military casualties × party family (model 2, see table 2). Bars represent 95 percent confidence intervals. Points to the left of the vertical bar stem from model 1, while those to the right stem from model 2. Table 2. FGLS regression analyses of the number of military casualties and change in position on the military . Model 1: Basic model two-way interaction . Model 2: Three-way interaction (party family) . ∆Position . Β . SE . B . SE . ∆Military casualties 0.196* 0.044 0.250* 0.089 Governing party 0.974* 0.218 1.23* 0.486 Gov party × military casualties −0.423* 0.070 −0.400* 0.169 Party family (Left wing as ref) Right-wing parties −0.369* 0.121 0.041 0.333 Ecological/socialist 0.033 0.159 1.16* 0.441 Party family × military casualties Right-wing parties −0.044 0.105 Ecological/socialist −0.442* 0.146 Gov party × party family Right-wing parties −0.439 0.560 Ecological/socialist 2.31 3.50 Gov party × party family × military casualties Right-wing parties −0.068 0.192 Ecological/socialist −0.502 1.20 Post-9/11 0.312* 0.106 0.354* 0.102 War deaths −0.063* 0.007 −0.065* 0.007 ∆Seatshare 1.75* 0.517 −1.38 * 0.537 Constant −0.223 0.180 −0.513 0.302 N (obs) 326 326 N (groups) 65 65 Wald 336.02* 371.19* . Model 1: Basic model two-way interaction . Model 2: Three-way interaction (party family) . ∆Position . Β . SE . B . SE . ∆Military casualties 0.196* 0.044 0.250* 0.089 Governing party 0.974* 0.218 1.23* 0.486 Gov party × military casualties −0.423* 0.070 −0.400* 0.169 Party family (Left wing as ref) Right-wing parties −0.369* 0.121 0.041 0.333 Ecological/socialist 0.033 0.159 1.16* 0.441 Party family × military casualties Right-wing parties −0.044 0.105 Ecological/socialist −0.442* 0.146 Gov party × party family Right-wing parties −0.439 0.560 Ecological/socialist 2.31 3.50 Gov party × party family × military casualties Right-wing parties −0.068 0.192 Ecological/socialist −0.502 1.20 Post-9/11 0.312* 0.106 0.354* 0.102 War deaths −0.063* 0.007 −0.065* 0.007 ∆Seatshare 1.75* 0.517 −1.38 * 0.537 Constant −0.223 0.180 −0.513 0.302 N (obs) 326 326 N (groups) 65 65 Wald 336.02* 371.19* Note: *p < 0.05. Time-series cross-sectional regression analyses of party position changes, using an AR(1) specification. We use observations from sixty-five parties (N [groups]), with a total of 326 party position changes. Open in new tab Table 2. FGLS regression analyses of the number of military casualties and change in position on the military . Model 1: Basic model two-way interaction . Model 2: Three-way interaction (party family) . ∆Position . Β . SE . B . SE . ∆Military casualties 0.196* 0.044 0.250* 0.089 Governing party 0.974* 0.218 1.23* 0.486 Gov party × military casualties −0.423* 0.070 −0.400* 0.169 Party family (Left wing as ref) Right-wing parties −0.369* 0.121 0.041 0.333 Ecological/socialist 0.033 0.159 1.16* 0.441 Party family × military casualties Right-wing parties −0.044 0.105 Ecological/socialist −0.442* 0.146 Gov party × party family Right-wing parties −0.439 0.560 Ecological/socialist 2.31 3.50 Gov party × party family × military casualties Right-wing parties −0.068 0.192 Ecological/socialist −0.502 1.20 Post-9/11 0.312* 0.106 0.354* 0.102 War deaths −0.063* 0.007 −0.065* 0.007 ∆Seatshare 1.75* 0.517 −1.38 * 0.537 Constant −0.223 0.180 −0.513 0.302 N (obs) 326 326 N (groups) 65 65 Wald 336.02* 371.19* . Model 1: Basic model two-way interaction . Model 2: Three-way interaction (party family) . ∆Position . Β . SE . B . SE . ∆Military casualties 0.196* 0.044 0.250* 0.089 Governing party 0.974* 0.218 1.23* 0.486 Gov party × military casualties −0.423* 0.070 −0.400* 0.169 Party family (Left wing as ref) Right-wing parties −0.369* 0.121 0.041 0.333 Ecological/socialist 0.033 0.159 1.16* 0.441 Party family × military casualties Right-wing parties −0.044 0.105 Ecological/socialist −0.442* 0.146 Gov party × party family Right-wing parties −0.439 0.560 Ecological/socialist 2.31 3.50 Gov party × party family × military casualties Right-wing parties −0.068 0.192 Ecological/socialist −0.502 1.20 Post-9/11 0.312* 0.106 0.354* 0.102 War deaths −0.063* 0.007 −0.065* 0.007 ∆Seatshare 1.75* 0.517 −1.38 * 0.537 Constant −0.223 0.180 −0.513 0.302 N (obs) 326 326 N (groups) 65 65 Wald 336.02* 371.19* Note: *p < 0.05. Time-series cross-sectional regression analyses of party position changes, using an AR(1) specification. We use observations from sixty-five parties (N [groups]), with a total of 326 party position changes. Open in new tab As for the other effects reported in model 1, the “post-9/11” variable is positive, suggesting that after 9/11 parties on average have become more positive about the military. The war deaths variable is negative; thus, the more battle-related deaths in conflicts over the world, the more negative parties become about the military. Even though empirical studies consistently find the public to deem one's “own” casualties more important than civilian casualties abroad in times of conflict (Gaines et al. 2007; Kahl 2007; Hultman 2013), a higher number of war-related deaths leads nonetheless to a more negative stance on the military in party manifestoes. For party family (with left-wing parties as a reference), the effect is negative and significant for right-wing parties suggesting that on average right-wing parties become less pro-military in our sample. This perhaps is a consequence of the fact that our sample starts shortly after the end of the Cold War. Parties that have increased their seat share in the previous election become more positive about the military in the next election. Now we move to model 2 and the three-way interaction effects with party family to the right of the vertical bar in figure 3. For left-wing parties, we find that opposition parties, like the general trend, become more positive about the military if there are more military casualties. The effect for left-wing governing parties is insignificant. Right-wing parties fully comply with the findings in the general model. Lastly, we find no effect for ecologist/socialist party. Very few of them are governing parties. But also the general effect of ecologist/socialist party is insignificant in model 1. In conclusion, these parties are unresponsive to military casualties. This might fit into the larger phenomenon that such parties (e.g., niche parties) are less responsive than mainstream parties (Ezrow et al. 2011). Robustness Checks We have conducted several robustness checks: (1) excluding the United States, (2) whether parties have changed their strategy before and after 9/11, (3) absolute numbers rather than the log variants of our independent variable of military casualties, (4) differences in composition of the coalition, (5) using a left–right position variable instead of party family, and (6) control for popularity. The Online Appendix 2 contains the full regression tables. We exclude the United States because it is a statistical outlier in our sample. In the analysis excluding the United States, the results are in the same direction, but the effect of military casualties for left-wing, opposition parties disappears. The significant effect in our initial analyses must have been driven by party position changes of the Democratic Party (US). The effects for right-wing parties are stable. Opposition, right-wing parties become more positive, while government, left-wing parties become more negative. This means that only our results for right-wing parties hold on both sides of the pond. We ran a second robustness analysis that compares the period before 9/11 with the one after. There are two important differences: pre-9/11 there were far fewer military casualties, and the missions post-9/11 are about regime change rather than humanitarian missions. Our results from the post-9/11 analysis are similar to our analysis of the general sample, regardless of whether we include or exclude the American parties. Including American parties in our pre-9/11 analysis shows similar results as well. When excluding American parties from our pre-9/11 sample, the effects are no longer significant, which means that the effects before 9/11 hold only for non-US democracies. With the third robustness analysis, which is for absolute numbers rather than the log variants, the effects are the same but become less significant. Only the effects for left-wing and right-wing opposition parties hold, while those for governing parties disappear. Interestingly, we also found different effects between parties that have won or lost the previous election. In our main model, we found winners tend to become more positive about the military in their next party manifesto. The same effect of increased security is seen in coalition governments in general: the larger the majority enjoyed by parties in government, the more positive they become about the military. This, in combination with the mixed effects found on a party family level, suggests that future research could benefit from looking deeper into the mechanisms driving these results, and establishes in which cases position in office (and how strong that position is) trumps ideological differences. This allows researchers to disentangle position in office from ideology on a more detailed level and identify the preconditions for blame avoidance strategies. In our fifth robustness check, we omitted the party family variable and replaced it with a variable that indicates the left–right position of the party. For this, we used the rile variable from the MP database (Volkens et al. 2015). This is a composite variable that includes party stances on a range of different issues, which can be aligned along a left–right dimension. The results are highly similar to the results using the party family measure: we find positive effects for right-wing opposition parties, negative effects for right-wing governing parties, positive effects for left-wing opposition parties, but no effect for left-wing governing parties. And finally extreme values on the left end of the left–right scale are associated with a negative effect but only for opposition parties. Arguably, this is the area in which the socialist and green parties are placed. In conclusion, we have controlled for popularity and found parties that are losing in the polls to become (on average) more pro-military. This could indicate that parties that are the victim of being blamed by voters for bad (national) policies shift their attention to foreign policy in an attempt to avoid blame. Since we do not have enough observations for each party family, it is difficult to entangle other effects that might be at play in the background. This might be, however, interesting for future research. Conclusion and Discussion To summarize our findings (also see table 3), we reject our commitment hypothesis. Governing parties do not increase their commitment to the military when military casualties increase. We accept our don't mention the war hypothesis. Governing parties reduce their commitment to the military when military casualties increase. However, this effect is strongly driven by the American parties and the post-9/11 period. For the opposition parties, we reject H3 (they do not reduce their commitment); instead, we accept H4 (opposition parties increase their commitment to the military). Again, this effect is strongest when including the American parties and in the post-9/11 period. As for ideological differences, the general findings are the clearest for right-wing parties. For left-wing parties we have more null findings. Table 3. Summary of effects of the number of military casualties . Including United States . Excluding United States . . Opposition . In office . Opposition . In office . All parties + − 0 0 Left wing + 0 0 0 Right wing + − + − Ecologist/socialist 0 0 0 0 . Including United States . Excluding United States . . Opposition . In office . Opposition . In office . All parties + − 0 0 Left wing + 0 0 0 Right wing + − + − Ecologist/socialist 0 0 0 0 Open in new tab Table 3. Summary of effects of the number of military casualties . Including United States . Excluding United States . . Opposition . In office . Opposition . In office . All parties + − 0 0 Left wing + 0 0 0 Right wing + − + − Ecologist/socialist 0 0 0 0 . Including United States . Excluding United States . . Opposition . In office . Opposition . In office . All parties + − 0 0 Left wing + 0 0 0 Right wing + − + − Ecologist/socialist 0 0 0 0 Open in new tab How do political parties respond to increasing military casualties in military interventions? Several studies demonstrate that with increasing military casualties, public support for the military presence in that conflict declines. In line with Budge and Farlie (1983), opposition parties respond to this by becoming more positive on the military, while governing parties nuance their position, and pay less attention to it. For opposition parties, increasing military casualties are a good opportunity to talk about the military and the failure of the government to use the military effectively. But they are not critical about the military itself. For governing parties, it is better not to talk about the military at all. In contrast to studies of decision-making, we do not find that governments escalate their commitment in their communication. They do not mention the military in a more positive way in an attempt to maintain that the intervention was “the right thing” to do. This suggests that governing parties try to avoid blame for the casualties primarily by mentioning the war. Opposition parties changing their tone after 9/11 could be explained by either the fact that the nature of military operations changed after 9/11, or that these missions were far more deadly (with higher soldiers-killed-to-troop ratios). Before 9/11, a large part of the military interventions undertaken by Western democracies were third party, humanitarian interventions and consisted of a mandate to intervene between two (or more) already existing striving parties. After 9/11, Western democracies themselves became one of the striving parties by initiating conventional military conflict, by invading Afghanistan and Iraq leading to far higher death tolls. Here, political elites seem to agree (despite their ideological orientation) on their strategy: avoid blame once in government, and escalate commitment once in opposition. Future research could benefit from further unpacking these findings, by looking into the composition of government coalition. This can be done by looking at ideology, or (as our results indicate) by position in office. It would be interesting to test whether secure governments (with large coalition majorities and public support) use different tactics when faced with military casualties than insecure governments. One would expect (based on findings from the diversionary use of force literature and prospect theory; see, e.g., Farnham 1992; Levy 1996; Mercer 2005; Brulé and Williams 2009) that insecure governments are more prone to escalatory policies, and thus might become more positive about the military. There is evidence that public opinion is not always responsive to military casualties (Boettcher and Cobb 2006; Gelpi and Feaver 2006; Berinsky 2009). This has consequences for the strategies of political parties. Still, we find that political parties do respond to military casualties even if public opinion does not. Even though we hypothesized active damage control from governing parties when confronted with military casualties, we found no evidence to support this. The most interesting finding is the consistent finding of governing parties becoming more negative, and indications of parties from different party families adopting similar strategies as opposition or governing party. However, even though party manifestoes can reveal underlying policy preferences over time, they only provide 4-year-interval snapshots and thus capture only part of the dynamics between government and opposition, namely, close before the elections. In addition, by design, our study does not include a content analysis on the tone, therefore limiting the outcomes only to positive or negative references to the military in general. Another point worth making in this context is the fact that the decision to intervene is largely a co-decision between government and opposition parties in multiparty systems and it has two consequences. First, it does not allow for our models to adequately capture strategic interaction on the domestic level between political parties. That is, the strategies of government and opposition parties are likely dependent not only upon events such as military casualties but also on the behavior of opponents. Second, it is hard to specify to what extent the conflict is inherited (or not). This is especially the case for the post-9/11 large-scale interventions in Afghanistan (average duration of deployment 60 months with a maximum of 225 months) and Iraq (average duration of deployment 46 months with a maximum of 180 months). Since governments fear public backlash over military casualties, one would expect a higher incentive for shifting the blame to the previous government who decided to intervene in the first place. For instance, after being confronted with two increasingly unpopular and costly wars (Afghanistan and Iraq) President Obama promised during his campaign (in both 2008 and 2012) to end the war in Afghanistan, and took credit in 2012 for retreating American troops from Iraq (which already began under President Bush since September 2007). Unfortunately, our data do not allow to differentiate between “inherited” and “chosen” conflicts (see the Online Appendix 3). Still, despite the high amount of military casualties that are only to be found in the United States, we were able to systematically test and compare party strategies and conclude that political strategy for right-wing parties does not differ much on both sides of the pond. Footnotes 1 Even though it is not possible to control for this within the European context, it is worth noting that US studies that controlled for the number of soldiers killed per state found casualty tolerance to be lower among voters in those particular states that endured heavy military losses in wartime (Hayes and Myers 2009; Kriner and Shen 2014), and could vary over time as well (Althaus, Bramlett, and Gimpel 2011). 2 Data gathered through open sources (databases of national departments of Defense, rolls of honor, veteran websites etc.); for more information see the Online Appendix. 3 The reason for limiting the sample to these eleven countries is two-fold. First, and for theoretical reasons, full-fledged Western democracies outside the transatlantic geographical area were excluded. Second, for reasons of data availability and time constraints, not all European countries were included in the sample. In our sample, there are countries with presidential systems (e.g., the United States), semi-presidential systems (e.g., France), proportional representation systems (e.g., Denmark and the Netherlands), and majoritarian systems (e.g., the United Kingdom). Despite these differences in specifics of government arrangement and party systems, they share broad similarities in the functioning of the democratic process. 4 Empirical work on public opinion and military casualties has been done on, e.g., Canada, France, the United Kingdom, and Israel (see, for instance, Cornish 2003; Clements 2013; Gribble et al. 2015). However, these empirical tests have mainly been done on high-profile events (Afghanistan, Iraq, Libya) and on—from a European perspective—military outliers. 5 Party position on peace is not taken into consideration since this variable measures to what extent parties are in favor of peace, but do not give any information on the preferred means to achieve this (e.g., diplomacy or the military). 6 Even though different countries measure the number of soldiers killed differently (e.g., whether soldiers who die of a heart attack are considered “killed during a mission” or not), the number of soldiers killed in the dataset is based on the numbers reported by the national governments as such. References Adams J. 2012 . “ Causes and Electoral Consequences of Party Policy Shifts in Multiparty Elections: Theoretical Results and Empirical Evidence .” Annual Review of Political Science 15 : 401 – 19 . Google Scholar Crossref Search ADS WorldCat Adams J. , Clark M., Ezrow L., Glasgow G.. 2004 . “ Understanding Change and Stability in Party Ideologies: Do Parties Respond to Public Opinion or to Past Election Results ?” British Journal of Political Science 34 ( 4 ): 589 – 610 . Google Scholar Crossref Search ADS WorldCat Adams J. , Haupt A. B., Stoll H.. 2008 . “ What Moves Parties? The Role of Public Opinion and Global Economic Conditions in Western Europe .” Comparative Political Studies 42 ( 5 ): 611 – 39 . Google Scholar Crossref Search ADS WorldCat Adams J. , Somer-Topcu Z.. 2009 . “ Policy Adjustment by Parties in Response to Rival Parties’ Policy Shifts: Spatial Theory and the Dynamics of Party Competition in Twenty-Five Post-War Democracies .” British Journal of Political Science 39 ( 4 ): 825 . Google Scholar Crossref Search ADS WorldCat Aday S. , Cluverius J., Livingston S.. 2005 . “ As Goes the Statue, So Goes the War: The Emergence of the Victory Frame in Television Coverage of the Iraq War .” Journal of Broadcasting & Electronic Media 49 ( 3 ): 314 – 31 . Google Scholar Crossref Search ADS WorldCat Althaus S. L. , Bramlett B. H., Gimpel J. G.. 2011 . “ When War Hits Home: The Geography of Military Losses and Support for War in Time and Space .” Journal of Conflict Resolution 56 ( 3 ): 382 – 412 . Google Scholar Crossref Search ADS WorldCat Arena P. 2015 . “ Crisis Bargaining, Domestic Opposition, and Tragic Wars .” Journal of Theoretical Politics 27 ( 1 ): 108 – 31 . Google Scholar Crossref Search ADS WorldCat Bakker R. , De Vries C., Edwards E., Hooghe L., Jolly S., Marks G., Vachudova M. A.. 2015 . “ Measuring Party Positions in Europe: The Chapel Hill Expert Survey Trend File, 1999–2010 .” Party Politics 21 ( 1 ): 143 – 52 . Google Scholar Crossref Search ADS Google Scholar Google Preview WorldCat COPAC Bausch A. W. 2015 . “ Democracy, War Effort, and the Systemic Democratic Peace .” Journal of Peace Research 52 ( 4 ): 435 – 47 . Google Scholar Crossref Search ADS Google Scholar Google Preview WorldCat COPAC Bazerman M. H. , Giuliano T., Appelman A.. 1984 . “ Escalation of Commitment in Individual and Group Decision Making .” Organizational Behavior and Human Performance 33 ( 2 ): 141 – 52 . Google Scholar Crossref Search ADS WorldCat Beck N. , Katz J. N.. 1995 . “ What to Do (and Not to Do) with Time-Series Cross-Section Data .” American Political Science Review 89 ( 3 ): 634 – 47 . Google Scholar Crossref Search ADS WorldCat Bell S. 2007 . “ The UK's Risk Management Approach to National Security .” The RUSI Journal 152 ( 3 ): 18 – 23 . Google Scholar Crossref Search ADS WorldCat Berinsky A. J. 2007 . “ Assuming the Costs of War: Events, Elites, and American Public Support for Military Conflict .” Journal of Politics 69 ( 4 ): 975 – 97 . Google Scholar Crossref Search ADS WorldCat Berinsky A. J. . 2009 . In Times of War: Understanding American Public Opinion from World War II to Iraq . Chicago & London: University of Chicago Press . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Berinsky A. J. , Kinder D. R.. 2006 . “ Making Sense of Issues through Media Frames: Understanding the Kosovo Crisis .” Journal of Politics 68 ( 3 ): 640 – 56 . Google Scholar Crossref Search ADS WorldCat Biddle S. 2005 . American Grand Strategy after 9/11: An Assessment. Challenges ( vol. 1 ). US Army War College: Strategic Studies Institute . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Boettcher W. A. , Cobb M. D.. 2006 . “ Echoes of Vietnam? Casualty Framing and Public Perceptions of Success and Failure in Iraq .” The Journal of Conflict Resolution 50 ( 6 ): 831 – 54 . Google Scholar Crossref Search ADS WorldCat Boettcher W. A. , Cobb M. D.. 2009 . ““ Don't Let Them Die in Vain”: Casualty Frames and Public Tolerance for Escalating Commitment in Iraq .” Journal of Conflict Resolution 53 ( 5 ): 677 – 97 . Google Scholar Crossref Search ADS WorldCat Brambor T. , Clark W. R., Golder M.. 2006 . “ Understanding Interaction Models: Improving Empirical Analyses .” Political Analysis 14 ( 1 ): 63 – 82 . Google Scholar Crossref Search ADS WorldCat Brockner J. 1992 . “ The Escalation of Commitment to a Failing Course of Action: Toward Theoretical Progress .” Academy of Management Review 17 ( 1 ): 39 – 61 . Google Scholar Crossref Search ADS WorldCat Brulé D. J. , Williams L. K.. 2009 . “ Democracy and Diversion: Government Arrangements, the Economy, and Dispute Initiation .” Journal of Peace Research 46 ( 6 ): 777 – 98 . Google Scholar Crossref Search ADS WorldCat Budge I. 1994 . “ A New Spatial Theory of Party Competition : Uncertainty, Ideology and Policy Equilibria Viewed Comparatively and Temporally .” British Journal of Political Science 24 ( 4 ): 443 – 67 . Google Scholar Crossref Search ADS WorldCat Budge I. , Farlie D. J.. 1983 . Explaining and Predicting Elections: Issue-effects and Party Strategies in 23 Democracies . London : Allen & Unwin . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Bueno De Mesquita B. , Downs G. W.. 2006 . “ Intervention and Democracy .” International Organization 60 ( 3 ): 627 – 49 . OpenURL Placeholder Text WorldCat Bueno de Mesquita B. , Morrow J. D., Siverson R. M., Smith A.. 2002 . “ Political Institutions, Policy Choice and the Survival of Leaders .” British Journal of Political Science 32 ( 4 ): 559 – 90 . Google Scholar Crossref Search ADS WorldCat Chesebro J. W. 1997 . Analyzing Media: Communication Technologies as Symbolic and Cognitive Systems. College Composition and Communication (vol. 48 ). New York & London: Routledge . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Clements B. 2013. “ Public Opinion and Military Intervention: Afghanistan, Iraq and Libya .” The Political Quarterly 84 ( 1 ): 119 – 31 . Crossref Search ADS WorldCat Cornish P. 2003 . “ Myth and Reality: US and UK Approaches to Casualty Aversion and Force Protection .” Defence Studies 3 ( 2 ): 121 – 8 . Google Scholar Crossref Search ADS WorldCat Croco S. E. 2011 . “ The Decider's Dilemma: Leader Culpability, War Outcomes, and Domestic Punishment .” American Political Science Review . http://doi.org/10.1017/s0003055411000219 . OpenURL Placeholder Text WorldCat Croco S. E. Weeks J. L. P. 2016 . “ War Outcomes and Leader Tenure .” World Politics . http://doi.org/10.1017/s0043887116000071 . OpenURL Placeholder Text WorldCat Dannenbaum T. 2011 . “ Bombs, Ballots, and Coercion: The Madrid Bombings, Electoral Politics, and Terrorist Strategy .” Security Studies 20 ( 3 ): 303 – 49 . Google Scholar Crossref Search ADS WorldCat Dardis F. E. 2006 . “ Marginalization Devices in U.S. Press Coverage of Iraq War Protest: A Content Analysis .” Mass Communication & Society 9 ( 2 ): 117 – 35 . Google Scholar Crossref Search ADS WorldCat Dimitrova D. V. Connolly-Ahern C.. 2007 . “ A Tale of Two Wars: Framing Analysis of Online News Sites in Coalition Countries and the Arab World During the Iraq War .” Howard Journal of Communications 18 ( 2 ): 153 – 68 . Google Scholar Crossref Search ADS WorldCat Dimitrova D. V. Strömbäck J.. 2005 . “ Mission Accomplished? Framing of the Iraq War in the Elite Newspapers in Sweden and the United States .” International Communication Gazette 67 ( 5 ): 399 – 417 . Google Scholar Crossref Search ADS WorldCat Döring H. Manow P.. 2016 . “ Parliaments and Governments Database (ParlGov): Information on Parties, Elections and Cabinets in Modern Democracies .” Development version. Druckman J. N. Peterson E. Slothuus R.. 2013 . “ How Elite Partisan Polarization Affects Public Opinion Formation .” American Political Science Review 107 ( 1 ): 57 – 79 . Google Scholar Crossref Search ADS WorldCat Dunlap C. J. Jr. 1999 . “ Kosovo, Casualty Aversion, and the American Military Ethos: A Perspective .” Journal of Legal Studies 10 : 95 . OpenURL Placeholder Text WorldCat Dyson S. B. 2006 . “ Personality and Foreign Policy: Tony Blair's Iraq Decisions .” Foreign Policy Analysis 2 ( April 2005 ): 289 – 306 . Google Scholar Crossref Search ADS WorldCat Eichenberg R. C. 2005 . “ Victory Has Many Friends: U.S. Public Opinion and the Use of Military Forde, 1981–2005 .” International Security 30 ( 1 ): 140 – 77 . Google Scholar Crossref Search ADS WorldCat Eichenberg R. C. Stoll R. J. Lebo M.. 2006 . “ War President: The Approval Ratings of George W. Bush .” Journal of Conflict Resolution 50 ( 6 ): 783 – 808 . Google Scholar Crossref Search ADS WorldCat Eriksson M. 2015 . “ UCDP/PRIO Armed Conflict Dataset Codebook .” Upsala Conflict Data Programs (UCDP) ( 4 ): 22 . OpenURL Placeholder Text WorldCat Ezrow L. De Vries C. Steenbergen M. Edwards E.. 2011 . “ Mean Voter Representation and Partisan Constituency Representation: Do Parties Respond to the Mean Voter Position or to Their Supporters ?” Party Politics 17 ( 3 ): 275 – 301 . Google Scholar Crossref Search ADS WorldCat Farnham B. 1992 . “ Roosevelt and the Munich Crisis: Insights from Theory Prospect .” Poitical Psychology 13 ( 2 ): 205 – 35 . Google Scholar Crossref Search ADS WorldCat Fearon J. D. 1994 . “ Domestic Political Audiences and the Escalation of International Disputes .” American Political Science Review 88 ( 3 ): 577 – 92 . Google Scholar Crossref Search ADS WorldCat Foster D. M. Gibler D. M.. 2016 . Is Anyone Listening? Crisis Signaling by the British House of Commons, 1918–2004 . Working Paper, University of Alabama . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Gaines B. J. Kuklinski J. H. Quirk P. J. Peyton B. Verkuilen J.. 2007 . “ Same Facts, Different Interpretations: Partisan Motivation and Opinion on Iraq .” Journal of Politics 69 ( 4 ): 957 – 74 . Google Scholar Crossref Search ADS WorldCat Garland H. 1990 . “ Throwing Good Money After Bad: The Effect of Sunk Costs on the Decision to Escalate Commitment to an Ongoing Project .” Journal of Applied Psychology 75 ( 6 ): 728 – 31 . Google Scholar Crossref Search ADS WorldCat Gartner S. S. 2008 . “ The Multiple Effects of Casualties on Public Support for War: An Experimental Approach .” American Political Science Review 102 ( 1 ): 95 – 106 . Google Scholar Crossref Search ADS WorldCat Gartner S. S. Segura G. M. Barratt B. A.. 2003 . “ War Casualties, Policy Positions, and the Fate of Legislators .” Political Research Quarterly 53 ( 3 ): 467 – 77 . Google Scholar Crossref Search ADS WorldCat Gelpi C. Feaver P. D.. 2006 . “ Success Matters Casualty Sensitivity and the War in Iraq .” International Security 30 ( 3 ): 7 – 46 . Google Scholar Crossref Search ADS WorldCat Gelpi C. Feaver P. D. Reifler J.. 2009 . Paying the Human Costs of War: American Public Opinion and Casualties in Military Conflicts . Princeton: Princeton University Press . Google Scholar Crossref Search ADS Google Scholar Google Preview WorldCat COPAC Gemenis K. 2013 . “ What to Do (and Not to Do) with the Comparative Manifestos Project Data .” Political Studies 61 : 23 – 43 . Google Scholar Crossref Search ADS WorldCat Gennaioli N. Voth H.-J.. 2011 . State Capacity and Military Conflict . Working Paper . Google Scholar Crossref Search ADS Google Scholar Google Preview WorldCat COPAC Gettleman M. E. Young M. Franklin J. Franklin H. B.. 1995 . Vietnam and America: A Documented History . New York : Grove Press. Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Gordon P. H. 2002 . “ NATO and the War on Terrorism a Changing Alliance .” The Brookings Review 20 ( 3 ): 36 – 8 . OpenURL Placeholder Text WorldCat Green-Pedersen C. 2007 . “ The Growing Importance of Issue Competition: The Changing Nature of Party Competition in Western Europe .” Political Studies 55 ( 3 ): 607 – 28 . Google Scholar Crossref Search ADS WorldCat Gribble R. , Wessley S., Klein S., Alexander D. A., Dandeker C., Fear N. T.. 2015 . “ British Public Opinion After a Decade of War: Attitudes to Iraq and Afghanistan .” Politics 35 ( 2 ): 128 – 50 . Google Scholar Crossref Search ADS WorldCat Hayes A. F. Myers T. A.. 2009 . “ Testing the “Proximate Casualties Hypothesis”: Local Troop Loss, Attention to News, and Support for Military Intervention .” Mass Communication and Society 12 ( 4 ): 379 – 402 . Google Scholar Crossref Search ADS WorldCat Hetherington M. J. Nelson M.. 2003 . “ Anatomy of a Rally Effect: George W. Bush and the War on Terrorism .” PS: Political Science & Politics . http://doi.org/10.1017/S1049096503001665 . OpenURL Placeholder Text WorldCat Hobolt S. B. de Vries C. E.. 2015 . “ Issue Entrepreneurship and Multiparty Competition .” Comparative Political Studies 48 ( 9 ): 1159 – 85 . Google Scholar Crossref Search ADS WorldCat Hultman L. 2013 . “ UN Peace Operations and Protection of Civilians: Cheap Talk or Norm Implementation ?” Journal of Peace Research 50 ( 1 ): 59 – 73 . Google Scholar Crossref Search ADS WorldCat Hyde C. K. 2000 . “ Casualty Aversion .” Air & Space Power Journal 14 ( 2 ): 17 . OpenURL Placeholder Text WorldCat Kahl C. H. 2007 . “ In the Crossfire Or the Crosshairs? Norms, Civilian Casualties, and U.S. Conduct in Iraq .” International Security 32 ( 1 ): 7 – 46 . Google Scholar Crossref Search ADS WorldCat Karlsson N. Juliusson Á. Gärling T.. 2005 . “ A Conceptualisation of Task Dimensions Affecting Escalation of Commitment .” European Journal of Cognitive Psychology 17 ( 6 ): 835 – 58 . Google Scholar Crossref Search ADS WorldCat Karol D. Miguel E.. 2007 . “ The Electoral Cost of War: Iraq Casualties and the 2004 U.S. Presidential Election .” Journal of Politics 69 ( 3 ): 633 – 48 . Google Scholar Crossref Search ADS WorldCat Kent Weaver R. 1986 . “ The Politics of Blame Avoidance .” Journal of Public Policy 6 ( 4 ): 371 – 98 . Google Scholar Crossref Search ADS WorldCat Kriner D. L. Shen F. X.. 2012 . “ How Citizens Respond to Combat Casualties .” Public Opinion Quarterly 76 ( 4 ): 761 – 70 . Google Scholar Crossref Search ADS WorldCat Kriner D. L. Shen F.. 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 . Google Scholar Crossref Search ADS WorldCat Kull S. Destler I. M.. 1999 . Misreading the Public: The Myth of a New Isolationism . Washington, DC: Brookings Institution Press . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Lacquement R. A. 2003 . “ The Casualty Aversion Myth .” Naval War College Review LVII ( 1 ): 40 – 57 . OpenURL Placeholder Text WorldCat Larson E. V. Savych B.. 2005 . American Public Support for US Military Operations from Mogadishu to Baghdad: Technical Appendixes. Rand Corporation. Levy J. S. 1996 . “ The Diversionary Theory of War .” In Handbook of War Studies , edited by Midlarsky Manus I., 259 – 88 . Boston: Unwin Hyman . OpenURL Placeholder Text WorldCat Levy J. S. . 1997 . “ Prospect Theory, Rational Choice, and International Relations .” International Studies Quarterly 41 ( 1 ): 87 – 112 . Google Scholar Crossref Search ADS WorldCat Lian B. Oneal J. R.. 1993 . “ Presidents, the Use of Military Force, and Public Opinion .” Journal of Conflict Resolution 37 ( 2 ): 277 – 300 . Google Scholar Crossref Search ADS WorldCat Litwak R. S. Litwak R.. 2008 . Regime Change: U.S. Strategy through the Prism of 9/11. Baltimore: JHU Press. Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC MacAskill E. 2013 . “ Obama Faces Tough Decisions on Syria As Pressure for Intervention Mounts .” The Guardian , p. 2 . OpenURL Placeholder Text WorldCat McCroskey J. C. 2006 . An Introduction to Rhetorical Communication: A Western Rhetorical Perspective . New York : Routledge . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC McDermott R. 2004 . “ Prospect Theory in Political Science: Gains and Losses from the First Decade .” Political Psychology 25 ( 2 ): 289 – 312 . Google Scholar Crossref Search ADS WorldCat Mercer J. 2005 . “ Prospect Theory and Political Science .” Annual Review of Political Science 8 : 1 – 21 . Google Scholar Crossref Search ADS WorldCat Mintz A. De Rouwen jr K. . 2010 . Understanding Foreign Policy Decision Making . New York: Cambridge University Press . Google Scholar Crossref Search ADS Google Scholar Google Preview WorldCat COPAC Mintz A. Yang Y. McDermott R.. 2011 . “ Experimental Approaches to International Relations .” International Studies Quarterly 55 ( 2 ): 493 – 501 . Google Scholar Crossref Search ADS WorldCat Mueller J. E. 1970 . “ Presidential Popularity from Truman to Johnson .” American Political Science Review 64 ( 1 ): 18 – 34 . Google Scholar Crossref Search ADS WorldCat Mueller J. E. . 1971 . “ Trends in Popular Support for the Wars in Korea and Vietnam .” The American Political Science Review 65 ( 2 ): 358 – 75 . Google Scholar Crossref Search ADS WorldCat Mueller J. E. . 2005 . “ The Iraq Syndrome .” Foreign Affairs 84 ( 6 ): 44 – 54 . Google Scholar Crossref Search ADS WorldCat Nincic D. J. Nincic M.. 1995 . “ Commitment to Military Intervention: The Democratic Government As Economic Investor .” Journal of Peace Research 32 ( 4 ): 413 – 26 . Google Scholar Crossref Search ADS WorldCat Noetzel T. Schreer B.. 2009 . “ Does a multi-tier NATO Matter? the Atlantic Alliance and the Process of Strategic Change .” International Affairs 85 ( 2 ): 211 – 26 . Google Scholar Crossref Search ADS WorldCat Norpoth H. Sidman A. H.. 2007 . “ Mission Accomplished: The Wartime Election of 2004 .” Political Behavior 29 ( 2 ): 175 – 95 . Google Scholar Crossref Search ADS WorldCat Pierson P. 1994 . Dismantling the Welfare State: Reagan, Thatcher, and the Politics of Retrenchment in Britain and the United States . Cambridge : Cambridge University Press . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Pierson P. . 1996 . “ The New Politics of the Welfare State .” World Politics 48 ( 2 ): 143 – 79 . Google Scholar Crossref Search ADS WorldCat Plümper T. Troeger V. E. Manow P.. 2005 . “ Panel Data Analysis in Comparative Politics: Linking Method to Theory .” European Journal of Political Research 44 ( 2 ): 327 – 54 . Google Scholar Crossref Search ADS WorldCat Preston T. 2011 . Pandora's Trap: Presidential Decision Making and Blame Avoidance in Vietnam and Iraq . Plymouth : Rowman and Littlefield Publishers. Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Riker W. H. 1988 . “ Liberalism Against Populism: A Confrontation Between the Theory of Democracy and the Theory of Social Choice .” In Liberalism Against Populism: A Confrontation between the Theory of Democracy and the Theory of Social Choice , 1 – 19 . Prospect Heights, Illinois : Waveland Press . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Ripstein A. 2004 . Critical Notice Too Much Invested To Quit. Economics and Philosophy (vol. 20 ). Oxford: Pergamon Press . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Rubin J. Z. Brockner J.. 1985 . “ Social Influence and Entrapment .” In Entrapment in Escalating Conflicts , 57 – 100 . New York, NY: Springer. Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Rupp R. E. 2016 . NATO after 9/11: An alliance in continuing decline. NATO After 9/11: An Alliance in Continuing Decline . New York: Springer . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Schubert J. N. Stewart P. A. Curran M. A.. 2002 . “ A Defining Presidential Moment: 9/11 and the Rally Effect .” Political Psychology 23 ( 3 ): 559 – 83 . Google Scholar Crossref Search ADS WorldCat Schultz K. 2013 . “ Domestic Politics and International Relations .” In Handbook of International Relations , 478 – 502 . London: SAGE Publications Ltd. Google Scholar Crossref Search ADS Google Scholar Google Preview WorldCat COPAC Smith A. 1998 . “ International Crises and Domestic Politics .” The American Political Science Review 92 ( 3 ): 623 – 38 . Google Scholar Crossref Search ADS WorldCat Smith H. 2005 . “ What Costs Will Democracies Bear? A Review of Popular Theories of Casualty Aversion .” Armed Forces & Society 31 ( 4 ): 487 – 512 . Google Scholar Crossref Search ADS WorldCat Somer-Topcu Z. 2009 . “ Timely Decisions: The Effects of Past National Elections on Party Policy Change .” The Journal of Politics 71 ( 1 ): 238 – 48 . Google Scholar Crossref Search ADS WorldCat Sperling J. Webber M.. 2009 . “ NATO: From Kosovo to Kabul .” International Affairs 85 ( 3 ): 491 – 511 . Google Scholar Crossref Search ADS WorldCat Staw B. M. 1976 . “ Knee-Deep in the Big Muddy : A Study of Escalating Commitment to a Chosen Course of Action .” Organizational Behavior and Human Performance 16 ( 1 ): 27 – 44 . Google Scholar Crossref Search ADS WorldCat Staw B. M. . 1997 . “ The Escalation of Commitment: An Update and Appraisal .” In Organizational Decision Making , edited by Shapira Z., 191 – 215 . Cambridge: Cambridge University Press . Google Scholar Crossref Search ADS Google Scholar Google Preview WorldCat COPAC Stimson J. 1985 . “ Regression Models in Space and Time: A Statistical Essay .”American Journal of Political Science 29 ( 4 ): 914 – 47 . OpenURL Placeholder Text WorldCat Stoll R. J. 1987 . “ The Sound of Guns: Is There a Congressional Rally Effect After US Military Action ?”American Politics Research 15 ( 2 ): 223 – 37 . OpenURL Placeholder Text WorldCat Tardy T. 2011 . “ A Critique of Robust Peacekeeping in Contemporary Peace Operations .”International Peacekeeping 18 ( April 2015 ): 152 – 67 . OpenURL Placeholder Text WorldCat Tir J. Singh S. P.. 2013 . “ Is It the Economy Or Foreign Policy, Stupid? The Impact of Foreign Crises on Leader Support .”Comparative Politics 46 ( 1 ): 83 – 101 . OpenURL Placeholder Text WorldCat Tomz M. 2007 . “ Domestic Audience Costs in International Relations: An Experimental Approach .” International Organization 61 ( 4 ): 821 – 40 . Google Scholar Crossref Search ADS WorldCat Valentino B. A. , Huth P. K., Croco S. E.. 2010 . “ Bear Any Burden? How Democracies Minimize the Costs of War .” Journal of Politics . http://doi.org/10.1017/S0022381609990831 . OpenURL Placeholder Text WorldCat Vis B. 2016 . “ Taking Stock of the Comparative Literature on the Role of Blame Avoidance Strategies in Social Policy Reform .” Journal of Comparative Policy Analysis: Research and Practice 18 ( 2 ): 122 – 37 . Google Scholar Crossref Search ADS WorldCat Vis B. , van Kersbergen K. 2007 . “ Why and How Do Political Actors Pursue Risky Reforms? ” Journal of Theoretical Politics 19 ( 2 ): 153 – 72 . Google Scholar Crossref Search ADS WorldCat Volkens A. , Lehmann P., Matthiess T., Merz N., Regel S., Werner A.. 2015 . “ The Manifesto Data Collection. Manifesto Project .” Manifesto Project (MRG/CMP/MARPOR) . Wagner W. , Herranz-Surrallés A., Kaarbo J., Ostermann F.. 2017 . “ Politicization, Party Politics and Military Missions: Deployment Votes in France, Germany, Spain, and the United Kingdom .” Wissenschaftszentrum Berlin Für Sozialforschung , Discussion (January) . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Walgrave S. , Nuytemans M. 2009 . “ Friction and Party Manifesto Change in 25 Countries, 1945–98 .” American Journal of Political Science 53 ( 1 ): 190 – 206 . Google Scholar Crossref Search ADS WorldCat Wenzelburger G. 2014 . “ Blame Avoidance, Electoral Punishment and the Perceptions of Risk .” Journal of European Social Policy 24 ( 1 ): 80 – 91 . Google Scholar Crossref Search ADS WorldCat Whyte G. 1986 . “ Escalating Commitment to a Course of Action: A Reinterpretation .” The Academy of Management Review 11 ( 2 ): 311 – 21 . Google Scholar Crossref Search ADS WorldCat Whyte G. 1993 . “ Escalating Commitment in Individual and Group Decision Making: A Prospect Theory Approach .” Organizational Behavior and Human Decision Processes 54 : 430 – 55 . Google Scholar Crossref Search ADS WorldCat Williams E. 2008 . “ Out of Area and Very Much in Business? NATO, the U.S., and the Post-9/11 International Security Environment .” Comparative Strategy 27 ( 1 ): 65 – 78 . Google Scholar Crossref Search ADS WorldCat Williams L. K. 2015 . “ It's All Relative: Spatial Positioning of Parties and Ideological Shifts .” European Journal of Political Research 54 ( 1 ): 141 – 59 . Google Scholar Crossref Search ADS WorldCat © The Author(s) (2020). Published by Oxford University Press on behalf of the International Studies Association. This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/open_access/funder_policies/chorus/standard_publication_model) http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Foreign Policy Analysis Oxford University Press

Don't Mention the War versus Escalating Commitment: Political Party Responses to Military Casualties

Loading next page...
 
/lp/oxford-university-press/don-t-mention-the-war-versus-escalating-commitment-political-party-KA9KSw0bjR

References (120)

Publisher
Oxford University Press
Copyright
© The Author(s) (2020). Published by Oxford University Press on behalf of the International Studies Association.
ISSN
1743-8586
eISSN
1743-8594
DOI
10.1093/fpa/oraa003
Publisher site
See Article on Publisher Site

Abstract

Abstract Do political parties change their position when military casualties increase? Several studies demonstrate that once military casualties increase, public support for sometimes even the government itself declines. With this potential backlash, once governing parties are faced with military casualties, do they (1) maintain that intervention was the “right thing” to do and even escalate their commitment by becoming even more pro-military or (2) try to avoid the blame and downplay the issue, i.e., “not mentioning the war”? And do the opposition parties become more negative or more positive about the military? To evaluate this, we measure the position on military issues in parties’ election manifestoes. Our dataset comprises 326 party policy changes in eleven Organisation for Economic Co-operation and Development (OECD) countries and focuses on post-Cold War military interventions. By using pooled time-series cross-sectional analysis, we find that opposition parties and governing parties respond differently. Generally, governing parties become more negative in their manifesto and opposition parties more positive. We also demonstrate important differences between party families and pre/post-9/11. Our analyses show that whether political parties change policy course once confronted with negative outcomes depends on their position in office, and also the direction in which they change policy depends on political ideology. Many Western democracies have—almost permanently—troops committed to some troubled region in the world. With some regularity, soldiers’ lives are lost in these missions abroad, especially in large-scale missions that entail high risks, such as the interventions in Afghanistan and Iraq. Foreign interventions affect incumbent support (Mueller 2005; Norpoth and Sidman 2007; Tir and Singh 2013). In particular, higher (own) casualty levels in military interventions abroad decrease public support for the mission (Eichenberg 2005) or even the incumbent itself (Eichenberg, Stoll, and Lebo 2006; Karol and Miguel 2007; Mueller 1970, 1971). Even though studies have challenged the conventional wisdom that the public is casualty adverse (Larson and Savych 2005; Boettcher and Cobb 2006; Gelpi and Feaver 2006; Berinsky 2007; Kriner and Shen 2012), several studies have shown that elites are. In a survey experiment conducted by Gelpi, Feaver, and Reifler (2009), members of the general public, civilian elites, and military senior officials were asked for three hypothetical intervention scenarios about how many casualties they would tolerate. They found members of the mass public to be more willing than the two groups of elites to accept military casualties (Gelpi, Feaver, and Reifler 2009), suggesting different levels of casualty tolerance between elites and the mass public.1 This finding can be explained by the assumption among civil and military elites that in the post-Cold War era, the public no longer tolerates military casualties (Dunlap 1999; Hyde 2000; Smith 2005). Empirical studies suggest that this conventional wisdom is widespread among elites from different strands: the media, politicians, and their staff (Kull and Destler 1999; Lacquement 2003). Given this, how do political parties respond to this? If a military mission becomes increasingly unpopular due to increasing casualties among the nation's soldiers, governing parties have two choices. First, governing parties defend the intervention and even escalate their commitment by becoming more outspoken on the subject in a positive manner (Staw 1976; Fearon 1994; Boettcher and Cobb 2009; Gennaioli and Voth 2011). We label this as the escalate commitment hypothesis. Second, governing parties shift in the same direction as public opinion (Adams 2012), by becoming less positive or even negative about the military. Public policy analysts claim that if governments carry out unpopular policies, they engage in blame avoidance strategies to evade electoral punishment (Kent Weaver 1986; Pierson 1994; Vis and van Kersbergen 2007; Wenzelburger 2014). According to these lines of reasoning, governing parties try to defuse the issue by focusing attention on other issues and taking a more neutral stance toward the military. We label this as the don't mention the war hypothesis. Opposition parties also face a dilemma: do they ride the wave of public dissatisfaction with an unpopular military intervention and attack the government on the military issue (blame the government hypothesis)? Or would such a move be seen as unpatriotic, and will they instead support the troops in the field (rally-around-the-flag hypothesis)? We approach these questions by analyzing the strategic communication of parties in their election manifestoes. These documents are relevant since parties constitute central actors that link and broker the preferences of voters to policy (Green-Pedersen 2007; Gemenis 2013). Even though the content is not often well known to the public, they constitute the core of party's policy and most of all contain promises that parties can later be held accountable for (Walgrave and Nuytemans 2009). Several studies have shown that an increase in military casualties can influence electoral outcomes (Gartner, Segura, and Barratt 2003; Karol and Miguel 2007; Norpoth and Sidman 2007; Dannenbaum 2011), and often have led to heated debates during electoral campaigns. For instance, numerous presidents felt the need to account for military policies and military casualties during their presidential campaigns, such as President George W. Bush who had to deal with anti-war sentiments during his 2004 campaign (Karol and Miguel 2007; Berinsky 2009), or were criticized for not doing so, such as presidential nominee Hillary Clinton in the 2016 campaign over the events in Benghazi. If parties indeed make a strategic evaluation based on electoral sentiments, we should expect to see preemptive shifts in party policy positions in strategic documents such as party manifestoes (Adams, Haupt, and Stoll 2008). Our paper analyzes 326 changes in issue position of seventy-five parties in eleven advanced democracies (Canada, Denmark, France, Germany, Italy, the Netherlands, Norway, Spain, Sweden, the United Kingdom, and the United States), for the post-Cold War period. Since the Cold War, these democracies have fought in several large-scale wars (Bosnia, Afghanistan, First and Second Gulf War) and engaged in multiple post-conflict and state-building missions (both in earlier mentioned conflict areas and through participation in UN missions). We use content-analyzed election manifesto data from the Manifesto Project (MP; Volkens et al. 2015). This allows measurement of (changes in) party positions on the military. We combine the manifesto data with various datasets measuring the military death toll in a conflict,2 the number of troops committed to an area of conflict (Military Balance data), and the total number of battle deaths in conflict areas around the world (Uppsala Conflict Data Program [UCDP] dataset). We also explore differences in party responses before and after 9/11. We found the more military casualties during the previous electoral term, the more governing parties become negative in their next manifesto while opposition parties become more positive. Strategic Options of Governing Parties Most advanced democracies have troops committed to war zones. The missions for which these troops are deployed are very different: some missions are only observatory, while others supply relief aid to distressed areas. Some missions sought to prevent genocide (e.g., the Kosovo conflict of 1999), while the aim of other missions was to bring about regime change (e.g., Iraq, 2003 and Afghanistan, 2002). Some missions carry almost no life-threatening risks for soldiers; the UN mission United Nations Iran–Iraq Military Observer Group (UNIIMOG, 1988–1991) had only one fatality in 3 years. The Second Battle of Fallujah in Iraq, where US soldiers were confronted with the fiercest urban combat since Vietnam, resulted in fifty-four fatalities between November 7 and 16, 2004. Also, some countries—most obviously the United States—have permanent overseas military presence, whereas others are much more reluctant to commit any troops abroad (such as Germany). Still, it is very rare for these eleven countries not to be involved in a mission abroad between 1990 and 2014. Even small countries like the Netherlands and Sweden commit on average one thousand troops per electoral term to missions all over the world. Larger countries like Canada or Italy commit on average about four thousand troops, and the United States has on average almost four hundred thousand troops per electoral term in the period 1990–2014. With different numbers of troops committed, it is unsurprising that the number of soldiers killed in action also differs markedly. Many countries are confronted with either a few dozen military casualties per electoral term, or no military casualties per electoral term. The United States is again an outlier with almost one thousand military casualties per electoral term. Of course, absolute differences in military casualties between countries matter. Still, the within-country variation in military casualties also influences both public opinion and the tone of the national debate. This, in turn, influences the strategic options that government and opposition parties have when they are forced to respond to increases in military casualties. We first consider the strategic options of governing parties. Because they are responsible for decisions regarding a military mission, the electorate sees them as responsible for military casualties. How can they respond to this? First, we posit the escalate your commitment hypothesis. This hypothesis describes the tendency of decision makers to continue with (and sometimes even invest additional resources into) a failing course of action (Brockner 1992; Staw 1976, 1997) in an attempt to justify the costs made thus far (Levy 1997; McDermott 2004). These so-called sunk costs are being perceived as operating costs and decision makers believe it is unlikely that similar costs would be incurred from new initiatives. Decision makers are even more inclined to escalate their commitment to a failing course of action if they need to justify their decisions to others (Karlsson, Juliusson, and Gärling 2005). One of the most frequently mentioned examples hereof is the US commitment to the Vietnam War (Staw 1976; Rubin and Brockner 1985; Whyte 1986; Ripstein 2004). Despite the warning by the undersecretary of state George Ball, that it was unlikely that the United States would meet its objectives in Vietnam (Gettleman et al. 1995, 283), President Johnson escalated the number of troops from sixty-five thousand to five hundred thirty-seven thousand from 1965 to 1968. Apart from foreign policy actions, politicians can escalate their commitment rhetorically to justify the current course of action despite heavy losses. An example hereof is the American President George W. Bush, who argued that “retreating from Iraq would dishonor the men and women who have given their lives in that country, and mean their sacrifice has been in vain” (Boettcher and Cobb 2009, 678). Groups and individuals (Bazerman, Giuliano, and Appelman 1984; Brockner 1992; Whyte 1993) display this tendency of investing more time and money to a decision made earlier despite increasing costs, and basically “throw good money after bad” (Garland 1990, 728). In the context of military interventions, increasing costs are mainly operationalized by tracking the increase in military casualties over the course of a military mission (Staw 1976; Nincic and Nincic 1995; Boettcher and Cobb 2009). Several case studies of decision-making by the US government regarding the 2003 intervention in Iraq highlight how psychological processes such as groupthink (Mintz and De Rouwen Jr. 2010) and loss aversion (Dyson 2006) led to an escalation of commitment to the intervention in Iraq. With the escalate commitment hypothesis, we aim to test whether governing parties indeed continue or increase their commitment to the military cause, despite obvious losses in terms of casualties. Here, and in the remainder of our article, we will translate increasing commitment as a shift toward a more positive position on the military despite the negative consequences suffered from implementing previous policies. H1: The more military casualties in a conflict, the more pro-military governing parties become in their next election manifesto. Second, we posit an alternative hypothesis, the don't mention the war hypothesis. Analyses of party policy shifts indicate that parties shift in the same direction as public opinion (Adams 2012), especially if public opinion is moving away from a party (Adams et al. 2004). Similarly, Budge and Farlie (1983) report that parties stress those issues that work for them and ignore those issues that work against them. Both findings suggest that in response to increasing military casualties in interventions abroad, parties should become more skeptical about the intervention. An assumption of this literature is that parties know public opinion, even on a specific issue like a military intervention. Several studies, in fact, demonstrate that political elites are biased in their beliefs about the effect of military casualties on public opinion. That is, elites assume the public to be more casualty averse than it actually is (Berinsky and Kinder 2006; Gelpi and Feaver 2006; Tomz 2007; Gartner 2008; Berinsky 2009; Mintz, Yang, and McDermott 2011; Bausch 2015; Druckman, Peterson, and Slothuus 2013; for other views see Berinsky 2007). For example, in a survey experiment conducted by Gelpi et al., members of the general public, civilian elites, and military senior officials were asked for three hypothetical intervention scenarios about how many casualties they would tolerate. They found members of the mass public to be more willing than policy elites to accept military casualties (Gelpi, Feaver, and Reifler 2009), suggesting different levels of casualty tolerance between elites and the mass public. This suggests that even if the public does not respond to an increase in military casualties, politicians still assume that they do and change their position in the direction they believe public opinion is. In this paper, we explicitly test whether this is the case. The political reasons to intervene, and whether intervention was deemed a success or a failure, set aside, do politicians assume military casualties are a reason to change political course when contending in the elections? Additional support for the expectation that governing parties become less outspoken toward the military in response to military casualties comes from the literature that analyzes public policy reforms. This literature analyzes how governments implement painful retrenchment of popular public policy (e.g., welfare) programs. By cutting back on these programs, governments risk electoral punishment and to counter this they employ so-called blame avoidance strategies (Kent Weaver 1986; Pierson 1996; Wenzelburger 2014; Vis 2016). To give an example from the International Relations literature, when new information on the absence of weapons of mass destruction in Iraq came out, the Bush administration tried shifting the blame of their decision to intervene to the intelligence community (Bell 2007; Preston 2011). Another strategy is to try and change the agenda. By downplaying a difficult issue and putting new and more popular issues on the agenda, the public might forget about the unpopular war. Following these lines of reasoning, and contra H1, one can expect governing parties to become more neutral or even negative/positive vis-à-vis the military, as a response to increases in the number of soldiers killed during missions abroad. H2:The more military casualties in a conflict, the more negative governing parties become about the military in the next election manifesto. Strategic Options of Opposition Parties What about the options of opposition parties? Opposition parties are often considered as the losers of the electoral game (Riker 1988). Their lack of access to office, and thus the rents, prestige, and policy access associated with office, should motivate them to change strategy, and indeed these parties—in particular parties systematically out of office—have been identified as issue entrepreneurs (Hobolt and de Vries 2015). If public opinion moves against the government, there is an opportunity for the opposition to blame the government for the mess and reap the electoral rewards (Bueno De Mesquita and Downs 2006; Smith 1998). This expectation falls in line with Arena's observation that wars often “attract widespread support early on, with opposition emerging later” (Arena 2015, 109). The analyses about party policy shifts mentioned earlier do not differentiate between opposition and governing parties. They provide the same hypothesis here: with public opinion moving against the intervention, parties will shift in the same direction as public opinion. Hypothesis 3, labeled as the blame the government hypothesis, formalizes these expectations. H3:The more military casualties in a conflict, the more negative opposition parties become about the military in the next manifesto. The previous hypothesis is premised on the assumption that public opinion eventually moves against a military intervention if more and more soldiers are killed. However, shortly after the start of a military intervention, public support generally rises in favor of the intervention. Mueller (1970) analyzed the effect of foreign policy events on (US) presidential approval rates over time, such as the effects of the Korean and Vietnam wars. He found that support for military interventions in the United States peaks in the beginning of a war but steadily declines over time. This is the so-called rally-around-the-flag effect, or rally effect in short (Mueller 1971, 2005). Rally effect generally recedes during an electoral term. But this does not mean that it is politically advantageous to criticize the mission. For instance, when John Kerry criticized President Bush’s handling of the Iraq War during the electoral campaign of 2004, Democratic Senator Zell Miller, stated that “while young Americans are dying in the sands of Iraq and the mountains of Afghanistan, our nation is being torn apart and made weaker because of the Democrats' manic obsession to bring down our Commander-in-Chief” (McCroskey 2006, 324). Also outside campaign time, it is often seen as unpatriotic to criticize an intervention. Content analyses of newspaper editorials during the Iraq War (Chesebro 1997; Dardis 2006) identified a post-9/11 rally effect (Aday, Cluverius, and Livingston 2005; Dimitrova and Strömbäck 2005; Dimitrova and Connolly-Ahern 2007) in Western media. This effect has even been identified as one of the most important determinants of presidential (and wartime) vote in 2004 (Norpoth and Sidman 2007). The rally effect is short lived in general (30 days, Stoll 1987 or 1–2 months, Lian and Oneal 1993); the post-9/11 period marked a relatively long-term increased public support for governments as an exception (over 13 months, Schubert, Stewart, and Curran 2002; Hetherington and Nelson 2003). In other words, post-9/11 criticism toward the mission may actually cause a backlash, and therefore, opposition parties motivated to recapture office must show their patriotic credentials and preemptively support the mission especially if soldiers are getting killed during missions abroad. Here, opposition parties are aware of the rally effect and act accordingly since they have nothing to gain from opposing the government over its policies. In addition, the open political competition within democracies has an effect on the behavior of both governing and opposition parties during international crises. In order for a threat to be credible, democracies can show resolve when threats show strong domestic support and thus an active opposition is absent (Schultz 2013). Moreover, increasing legislative unity influences the likelihood of a government escalating in a conflict (Foster and Gibler 2016). We label this as the rally-around-the-flag hypothesis. H4:The more military casualties in a conflict, the more pro-military opposition parties become in the next manifesto. Ideology, 9/11, and Commitment Party policy shifts take place in a specific context. We consider three variables here that are likely—from a theoretical perspective—to influence the relationship between military casualties and party policy shifts: 9/11, the initial commitment of a party to a mission, and the party's ideology. First, there are important differences in the nature of the interventions before and after 9/11. In the 1990s, most interventions (e.g., those in former Yugoslavia) were humanitarian missions with relatively little risk for soldiers. But after 9/11 most interventions were large-scale large-risk ground operations to first bring about regime change, and eventually quell counterinsurgencies. The former were primarily run by the American army; European armies were more deeply involved in the latter. This brings about three effects, in both Europe and North America. (1) The post-9/11 interventions have a much higher death toll than those in the 1990s. (2) Public opinion toward humanitarian missions may differ from missions aimed at regime change. (3) The political opportunities differ between humanitarian missions and regime change missions. For instance, Dutch opposition parties requested more military intervention in Bosnia in the early 1990s and were critical of the government standing idly by. Instead, opposition over the interventions in Afghanistan and Iraq focused on questions surrounding the legal basis for the interventions, civilian casualties, and the length (prolonging of mandates) of the interventions. In addition, the change in American foreign policy (Schubert et al. 2002; Litwak and Litwak 2007) and grand strategy (Biddle 2005) after 9/11, bypassing the UN Security Council, the creation of ad hoc alliances in both Afghanistan and Iraq (Williams 2008), and the subsequent international debates on transatlantic partnership (Gordon 2002; Noetzel and Schreer 2009; Sperling and Webber 2009; Rupp 2016) may have further motivated political controversy around military interventions beyond the United States. In sum, we expect the impact of our independent variables to be conditioned by 9/11. For that reason, we run a robustness check in which we separate our analyses in a split sample: the pre-9/11 and the post-9/11 period. Second, party responses to events may be driven by party ideology. Right-wing parties, reversely, are traditionally seen as pro-military. In particular, Green parties and socialist parties have anti-military and even pacifistic policy positions. Left-wing parties are less outspoken on the issue. If your initial belief is that interventions do not work, then military casualties reinforce your initial belief. If your initial belief is that interventions do work, you are likely to interpret military casualties as an argument for putting more effort into the intervention. We consider these arguments and to account for differences not only between but also within left and right parties, we include party family as a variable. The reason for using party family rather than the left–right axis is that we do not expect a linear relationship between positions on the military along this axis, and make a distinction between far right, right wing, left wing, and socialist/ecologist parties (Wagner et al. 2017). Third, what about a party's initial commitment to a mission? Governing parties may inherit conflicts, while they have been opposed to the mission in the past (Croco 2011; Croco and Weeks 2016). In practice this rarely happens. Most military interventions are in fact widely supported by parliament. Radical parties are more likely to oppose missions, but these rarely govern. The Online Appendix 3 provides an overview of voting behavior for large-scale expeditionary missions. Currently, we do not theorize additional strategic interactions between governing parties and opposition parties. Taking a position against or in favor of a military mission may be more attractive depending on the positions that rival parties take (Adams and Somer-Topcu 2009; Williams 2015). Following second and third points in this section, we expect party responses to be primarily explained by party ideology and party's initial commitments to the mission, rather than possible external pressures that arise from the behavior of other parties. Design and Methods Our sample comprises political parties elected to parliament (or in the United States, to Congress) from eleven Organisation for Economic Co-operation and Development (OECD) countries (Canada, Denmark, France, Germany, Italy, the Netherlands, Norway, Spain, Sweden, the United Kingdom, and the United States) from 1990 to 2014.3 The sample thus comprises mature Western democracies where we would expect the same incentives in order to secure political survival (Bueno de Mesquita et al. 2002; Valentino, Huth, and Croco 2010), while at the same time differentiating on the national institutional level (e.g., presidential and non-presidential systems) and international institutional level (e.g., mixed membership EU and NATO). We focus on the period since the Cold War and the collapse of the Soviet Union (i.e., 1990), because since then, democracies have fought in several large-scale wars (Bosnia, Afghanistan, First and Second Gulf War) and engaged in post-conflict and state-building missions (such as long-term UN missions), so there are sufficient interventions to study. Description of the Dependent Variable We will analyze changes in party positions on the military using data from the MP. Looking at election manifestoes allows us to track changes in political strategic investment by government and opposition parties over a long period of time (1990–2014). This dataset contains quantitative content analyses of election programs and indicates the percentage of positive or negative references in election manifestoes to a wide range of policy issues, including the military. In addition, it allows us to measure changes over the time period we are interested in—the period between the end of the Cold War and 9/11, and the post-9/11 period. An alternative, such as the Chapel Hill expert surveys (Bakker et al. 2015), ranges from 1999 to 2014 maximum and measures in this time period political stances of parties only on EU foreign and security policy. One item measures parties’ support for peace and security missions also outside the EU context, but since this question was asked in two waves it only provides a single cross-section of change in position. Most military missions take place outside this context. Comparing political responses in and outside the US context over a longer period of time (since the Cold War) is of theoretical relevance. Almost all inferences on how politicians respond to military casualties are based on high-profile cases (such as the Vietnam or the Iraq War) or are based on the US sample alone.4 The MP dataset allows us to look at changes in policy preferences from election to election, over a longer period of time for all the countries in our sample making it more fit for our purposes (Volkens et al. 2015). Hence, the MP dataset provides data on how parties respond in their own communication. We will analyze the change in party position on the military (Y1) as expressed in its election manifesto.5 We measure this position by subtracting the number of negative references from the number of positive references. Positive references are sentences in the manifestoes that express the need to maintain or increase military expenditure; to secure adequate manpower in the military; to modernize armed forces; or for rearmament and self-defense. Negative references are those referring to decreasing military expenditure, disarmament, and reducing (or abolishing) conscription. Because we are interested in change, our dependent variable expresses the difference in percentage of references to the military between the current election manifesto (t) and the previous election manifesto (t−1). A positive number indicates that a party has become more positive about the military, while a negative number indicates that a party has become more negative. The sample mean is 0.057 (see table 1) indicating that parties on average become more positive about the military in our sample. Table 1. Operationalization of main (in)dependent variables Dependent variables . Indicator . Data source . Mean . St. Dev. . Min. . Max. . ∆Position Positiont—position(t−1) Manifesto data (position = per104–per105) 0.014 2.62 −29.25 8 Independent variables Indicator Data source Mean St. Dev. Min. Max. Military casualties Cumulative number of soldiers killed between elections (log transformed) See the Online Appendix 1 2.45 1.33 0 8.12 ∆Seatshare ∆Seatshare = seatsharet—seatshare(t−1) ParlGov 0.000 0.127 −0.566 0.590 War deaths War deaths are the sum of battle deaths + genocide (log transformed) UCDP dataset/United Nations 2.88 5.48 −3.24 13.8 Observation per category In office 1 = Governing party ParlGov 1 = 27.74% 0 = Opposition party 0 = 72.26% Party family 0 = Left wing MP dataset 0 = 21.63% 1 = Right wing 1 = 54.88% 2 = Socialist/ecologist 2 = 23.49% Post-9/11 Post-9/11 party manifestoes 1 = 21.78% 1 = post; 0 = ante 0 = 78.22% Dependent variables . Indicator . Data source . Mean . St. Dev. . Min. . Max. . ∆Position Positiont—position(t−1) Manifesto data (position = per104–per105) 0.014 2.62 −29.25 8 Independent variables Indicator Data source Mean St. Dev. Min. Max. Military casualties Cumulative number of soldiers killed between elections (log transformed) See the Online Appendix 1 2.45 1.33 0 8.12 ∆Seatshare ∆Seatshare = seatsharet—seatshare(t−1) ParlGov 0.000 0.127 −0.566 0.590 War deaths War deaths are the sum of battle deaths + genocide (log transformed) UCDP dataset/United Nations 2.88 5.48 −3.24 13.8 Observation per category In office 1 = Governing party ParlGov 1 = 27.74% 0 = Opposition party 0 = 72.26% Party family 0 = Left wing MP dataset 0 = 21.63% 1 = Right wing 1 = 54.88% 2 = Socialist/ecologist 2 = 23.49% Post-9/11 Post-9/11 party manifestoes 1 = 21.78% 1 = post; 0 = ante 0 = 78.22% Open in new tab Table 1. Operationalization of main (in)dependent variables Dependent variables . Indicator . Data source . Mean . St. Dev. . Min. . Max. . ∆Position Positiont—position(t−1) Manifesto data (position = per104–per105) 0.014 2.62 −29.25 8 Independent variables Indicator Data source Mean St. Dev. Min. Max. Military casualties Cumulative number of soldiers killed between elections (log transformed) See the Online Appendix 1 2.45 1.33 0 8.12 ∆Seatshare ∆Seatshare = seatsharet—seatshare(t−1) ParlGov 0.000 0.127 −0.566 0.590 War deaths War deaths are the sum of battle deaths + genocide (log transformed) UCDP dataset/United Nations 2.88 5.48 −3.24 13.8 Observation per category In office 1 = Governing party ParlGov 1 = 27.74% 0 = Opposition party 0 = 72.26% Party family 0 = Left wing MP dataset 0 = 21.63% 1 = Right wing 1 = 54.88% 2 = Socialist/ecologist 2 = 23.49% Post-9/11 Post-9/11 party manifestoes 1 = 21.78% 1 = post; 0 = ante 0 = 78.22% Dependent variables . Indicator . Data source . Mean . St. Dev. . Min. . Max. . ∆Position Positiont—position(t−1) Manifesto data (position = per104–per105) 0.014 2.62 −29.25 8 Independent variables Indicator Data source Mean St. Dev. Min. Max. Military casualties Cumulative number of soldiers killed between elections (log transformed) See the Online Appendix 1 2.45 1.33 0 8.12 ∆Seatshare ∆Seatshare = seatsharet—seatshare(t−1) ParlGov 0.000 0.127 −0.566 0.590 War deaths War deaths are the sum of battle deaths + genocide (log transformed) UCDP dataset/United Nations 2.88 5.48 −3.24 13.8 Observation per category In office 1 = Governing party ParlGov 1 = 27.74% 0 = Opposition party 0 = 72.26% Party family 0 = Left wing MP dataset 0 = 21.63% 1 = Right wing 1 = 54.88% 2 = Socialist/ecologist 2 = 23.49% Post-9/11 Post-9/11 party manifestoes 1 = 21.78% 1 = post; 0 = ante 0 = 78.22% Open in new tab Description of the Independent Variables The cost of war will be operationalized by the number of soldiers of a country killed during missions abroad between elections (see figure 1). This variable adds the number of these military casualties by month up until the month before the elections and is based on the number of military casualties on a monthly basis. These data have been collected through open data sources, such as army archives, rolls of honor, and veteran websites.6 Military casualties are not normally distributed due to the large differences between countries such as the United States and Norway in terms of the number of soldiers deployed. As figure 1 visualizes, we have used a log transformation of the variable. For the same reasons, we have transformed the total number of war-related deaths (the sum of battle-related deaths and genocide victims) into a log variant. Figure 1. Open in new tabDownload slide Distribution of military casualties variable Figure 1. Open in new tabDownload slide Distribution of military casualties variable Our second independent variable, difference in seatshare, measures the change in seat share between the previous election (t−1) and the election before that (t−2). Several studies demonstrate that losing elections spurs party change (Budge 1994; Somer-Topcu 2009). Therefore, we add it as a control variable to see whether changes in the size of a government coalition have an effect. War deaths, our third independent variable, accounts for the number of civilians and soldiers killed of all striving parties during regular warfare. According to UCDP, battle-related deaths occur in what can be described as “normal” warfare involving the armed forces of the warring parties. This includes traditional battlefield fighting, guerrilla activities, bombardments of military units, cities, and villages etc. The targets are usually the military itself and its installations, or state institutions and state representatives, but there is often substantial collateral damage in the form of civilians killed in crossfire, indiscriminate bombings etc. All deaths—military as well as civilian—incurred in such situations are counted as battle-related deaths (Eriksson 2015). If we would measure only the (own) military casualties suffered during conflicts, it is not clear whether political parties respond to these or (political debate on) the casualties brought on by a conflict in general. For instance, despite relatively low (own) military casualties, the ongoing war in Syria is leading to heated political debates (MacAskill 2013). The UCDP dataset does not include one-sided violence directed toward citizens as battle-related deaths, so we have included data on the Rwanda Genocide (1994), Srebrenica massacre (1995), and Darfur (2004–2007) and have chosen the average of the low and high estimates indicated by international organizations and Non-governmental organizations. The reason to include these numbers is that genocides have caused heated political debates on UN mandates, and the episodes in Rwanda and Srebrenica even inspired the so-called “responsibility to protect” doctrine aimed at the protection of the civilian populations during peacekeeping missions (Tardy 2011; Hultman 2013). This can have consequences on the level of party politics as well. For instance, Dutch opposition parties blamed the government for “not doing enough” during the Yugoslav civil war (see the Online Appendices 1 and 3). Our fourth independent variable, in office, comes from the ParlGov dataset (Döring and Manow 2016). In office codes whether a political party was in office before the election (1) or not (0). A second dummy variable differentiates between elections before 9/11/2001 and after 9/11/2001, which allows us to re-run the analysis for a pre- and post-9/11 sample. The fifth independent variable, party family, is a restructured version of the original party family variable from the MP dataset. Following from our discussion in the theory section, we distinguish between the far right (ethnic, regionalist), the center right (nationalist, agrarian, conservative, liberal, and Christian democratic parties), left wing (social democratic parties), and socialist/ecological parties. Since none of the far-right parties in our sample ever took office, these are dropped from the analysis. As mentioned in the theory section, and illustrated by the differences in party position per party family in figure 2, we do not expect a linear effect for left-right ideology. In addition, left–right positions contain much more information than is necessary to test our hypothesis. Party family allows us to account for outliers on both the right side (ethnic and regionalist parties) and the left side (socialist and green parties). Even though we would have preferred to employ a separate category for both far right and far left, there are not enough observations in our data to do so. For instance, there are only four observations in which nationalist parties are in office. Such a small n would be problematic when treating this as a separate category in an interaction model. To ensure the far-right parties do not influence our result, we ran a separate analysis to check whether the effects for right-wing parties change. They do not. In addition, there are no observations for ethno-territorial parties that were in office in our sample. So, these parties are dropped from the analysis. Figure 2. Open in new tabDownload slide Box plot of position on the military by party family Note: The box plot presents the spread of position on the military per party family. Operationalization of position on the military is discussed in the methods section. Figure 2. Open in new tabDownload slide Box plot of position on the military by party family Note: The box plot presents the spread of position on the military per party family. Operationalization of position on the military is discussed in the methods section. We do not use public opinion about a mission as an independent variable. From a practical perspective, these data are not available for the time frame and country selection we are interested in. From a theoretical perspective (as noted earlier) what is more relevant than actual public opinion shifts is the belief that public opinion is shifting. We noted there that is likely to occur if military casualties increase. Estimation Method We conduct cross-sectional time-series analyses since this model will allow us to account for the dynamics between our dependent variable, the difference in party positions, and our independent variables. Such models have issues with autocorrelation (Beck and Katz 1995). Even though lagged variables incorporate feedback over time in our model, it will introduce the issue of heteroskedastic error terms at the same time (Stimson 1985). In addition, the biases of lagged variables associated with trends within our independent variables and error term can annul the effect of the theoretical model (Plümper, Troeger, and Manow 2005). We want to avoid the lagged dependent variable taking up part of the explanatory variance within the theoretical. Also since every country in our sample does not hold its election simultaneously, our panel is unbalanced. We correct for this in two ways: (1) by estimating autocorrelation within panels as well as heteroskedasticity across panels by applying Feasible Generalized Least Squares (FGLS) regression corrected for autocorrelated errors and (2) by constructing confidence intervals around relative changes and specifying the type of standard error reported robust to misspecification. Therefore, FGLS will be our estimation technique since it can estimate autocorrelation (AR[1]) within panels, as well as heteroskedasticity across panels. Another reason to opt for FGLS is a theoretical one. Fixed-effect models do not allow for time invariant variables (such as party family), and such variables become especially problematic from a theoretical point of view when using interactions. We are interested in the question whether party family has an effect on our main effect, and not whether party family varies over time. Our hypotheses assume a difference in the response of government and opposition parties to military casualties. We model this by means of an interaction effect between the government variable and the military casualties variable. We also discussed that party responses may differ per party family. To model this, we add a three-way interaction effect between the government variable, the military casualties variable, and party family. To this end, we follow the recommendations put forward by Brambor, Clark, and Golder (2006). Do Parties Become More Positive or More Negative about the Military? Do parties change position on the military, that is, do they become more positive or negative when there are more military casualties? Table 2 presents the results of our two regression analyses. Since the main hypotheses are tested with interaction effects, it is more informative to interpret the marginal effects of these interactions. Figure 3 displays the marginal effects of military casualties for government and opposition parties in general (to the left in figure 3, based on model 1 in table 2) and per party family (to the right in figure 3, based on model 2 in table 2). First, let us interpret the findings from model 1. We report a significant negative effect of military casualties for governing parties (b = −0.23), and a significant positive effect of military casualties for opposition parties (b = 0.2). To illustrate this, we turn to Denmark in 2011. It suffered thirty-six military casualties between the 2007 and 2011 elections (log = 3.6). In this situation, an opposition party is predicted to become more positive about the military (3.6 × 0.2 = 0.72) and a governing party to become more negative (3.6 × −0.23 = −0.83). The model perfectly predicts the shift of the governing party, the Conservatives (−0.83), while the other governing party, the Liberal Party, did not shift. The main opposition party—the Social Democratic Party—became more positive (0.85), as predicted by the model. A similar result can be found in the United States. In 2008, confronted with increasing military casualties from Iraq and Afghanistan, the Republicans (governing party) reduce positive mentions of the military (−0.28), and Democrats (opposition party) increase positive mentions (3.75). For example, the Democratic Party platform contained promises such as “we will rebuild our armed forces to meet the full spectrum needs of the new century” and “we support plans to increase the size of the Army by 65,000 troops and the Marines by 27,000 troops.” It is remarkable that the Democrats become so much more pro-military, while at the same time advocating a retreat from Iraq. Consider a few sentences from the “Ending the War in Iraq” section of the Democrats’ 2008 party platform. “Our men and women in uniform have performed admirably while sacrificing immeasurably. Our civilian leaders have failed them.” “We will give our military a new mission: ending this war and giving Iraq back to its people.” What this suggests is that a radical break with policy (retreat from Iraq) can only be justified if a commitment to the military is shown. Figure 3. Open in new tabDownload slide Average marginal effect of military casualties on changes in party position Note: *p < .05. Dots are average marginal effects from the interaction between government × military casualties (model 1, see table 2) and government × military casualties × party family (model 2, see table 2). Bars represent 95 percent confidence intervals. Points to the left of the vertical bar stem from model 1, while those to the right stem from model 2. Figure 3. Open in new tabDownload slide Average marginal effect of military casualties on changes in party position Note: *p < .05. Dots are average marginal effects from the interaction between government × military casualties (model 1, see table 2) and government × military casualties × party family (model 2, see table 2). Bars represent 95 percent confidence intervals. Points to the left of the vertical bar stem from model 1, while those to the right stem from model 2. Table 2. FGLS regression analyses of the number of military casualties and change in position on the military . Model 1: Basic model two-way interaction . Model 2: Three-way interaction (party family) . ∆Position . Β . SE . B . SE . ∆Military casualties 0.196* 0.044 0.250* 0.089 Governing party 0.974* 0.218 1.23* 0.486 Gov party × military casualties −0.423* 0.070 −0.400* 0.169 Party family (Left wing as ref) Right-wing parties −0.369* 0.121 0.041 0.333 Ecological/socialist 0.033 0.159 1.16* 0.441 Party family × military casualties Right-wing parties −0.044 0.105 Ecological/socialist −0.442* 0.146 Gov party × party family Right-wing parties −0.439 0.560 Ecological/socialist 2.31 3.50 Gov party × party family × military casualties Right-wing parties −0.068 0.192 Ecological/socialist −0.502 1.20 Post-9/11 0.312* 0.106 0.354* 0.102 War deaths −0.063* 0.007 −0.065* 0.007 ∆Seatshare 1.75* 0.517 −1.38 * 0.537 Constant −0.223 0.180 −0.513 0.302 N (obs) 326 326 N (groups) 65 65 Wald 336.02* 371.19* . Model 1: Basic model two-way interaction . Model 2: Three-way interaction (party family) . ∆Position . Β . SE . B . SE . ∆Military casualties 0.196* 0.044 0.250* 0.089 Governing party 0.974* 0.218 1.23* 0.486 Gov party × military casualties −0.423* 0.070 −0.400* 0.169 Party family (Left wing as ref) Right-wing parties −0.369* 0.121 0.041 0.333 Ecological/socialist 0.033 0.159 1.16* 0.441 Party family × military casualties Right-wing parties −0.044 0.105 Ecological/socialist −0.442* 0.146 Gov party × party family Right-wing parties −0.439 0.560 Ecological/socialist 2.31 3.50 Gov party × party family × military casualties Right-wing parties −0.068 0.192 Ecological/socialist −0.502 1.20 Post-9/11 0.312* 0.106 0.354* 0.102 War deaths −0.063* 0.007 −0.065* 0.007 ∆Seatshare 1.75* 0.517 −1.38 * 0.537 Constant −0.223 0.180 −0.513 0.302 N (obs) 326 326 N (groups) 65 65 Wald 336.02* 371.19* Note: *p < 0.05. Time-series cross-sectional regression analyses of party position changes, using an AR(1) specification. We use observations from sixty-five parties (N [groups]), with a total of 326 party position changes. Open in new tab Table 2. FGLS regression analyses of the number of military casualties and change in position on the military . Model 1: Basic model two-way interaction . Model 2: Three-way interaction (party family) . ∆Position . Β . SE . B . SE . ∆Military casualties 0.196* 0.044 0.250* 0.089 Governing party 0.974* 0.218 1.23* 0.486 Gov party × military casualties −0.423* 0.070 −0.400* 0.169 Party family (Left wing as ref) Right-wing parties −0.369* 0.121 0.041 0.333 Ecological/socialist 0.033 0.159 1.16* 0.441 Party family × military casualties Right-wing parties −0.044 0.105 Ecological/socialist −0.442* 0.146 Gov party × party family Right-wing parties −0.439 0.560 Ecological/socialist 2.31 3.50 Gov party × party family × military casualties Right-wing parties −0.068 0.192 Ecological/socialist −0.502 1.20 Post-9/11 0.312* 0.106 0.354* 0.102 War deaths −0.063* 0.007 −0.065* 0.007 ∆Seatshare 1.75* 0.517 −1.38 * 0.537 Constant −0.223 0.180 −0.513 0.302 N (obs) 326 326 N (groups) 65 65 Wald 336.02* 371.19* . Model 1: Basic model two-way interaction . Model 2: Three-way interaction (party family) . ∆Position . Β . SE . B . SE . ∆Military casualties 0.196* 0.044 0.250* 0.089 Governing party 0.974* 0.218 1.23* 0.486 Gov party × military casualties −0.423* 0.070 −0.400* 0.169 Party family (Left wing as ref) Right-wing parties −0.369* 0.121 0.041 0.333 Ecological/socialist 0.033 0.159 1.16* 0.441 Party family × military casualties Right-wing parties −0.044 0.105 Ecological/socialist −0.442* 0.146 Gov party × party family Right-wing parties −0.439 0.560 Ecological/socialist 2.31 3.50 Gov party × party family × military casualties Right-wing parties −0.068 0.192 Ecological/socialist −0.502 1.20 Post-9/11 0.312* 0.106 0.354* 0.102 War deaths −0.063* 0.007 −0.065* 0.007 ∆Seatshare 1.75* 0.517 −1.38 * 0.537 Constant −0.223 0.180 −0.513 0.302 N (obs) 326 326 N (groups) 65 65 Wald 336.02* 371.19* Note: *p < 0.05. Time-series cross-sectional regression analyses of party position changes, using an AR(1) specification. We use observations from sixty-five parties (N [groups]), with a total of 326 party position changes. Open in new tab As for the other effects reported in model 1, the “post-9/11” variable is positive, suggesting that after 9/11 parties on average have become more positive about the military. The war deaths variable is negative; thus, the more battle-related deaths in conflicts over the world, the more negative parties become about the military. Even though empirical studies consistently find the public to deem one's “own” casualties more important than civilian casualties abroad in times of conflict (Gaines et al. 2007; Kahl 2007; Hultman 2013), a higher number of war-related deaths leads nonetheless to a more negative stance on the military in party manifestoes. For party family (with left-wing parties as a reference), the effect is negative and significant for right-wing parties suggesting that on average right-wing parties become less pro-military in our sample. This perhaps is a consequence of the fact that our sample starts shortly after the end of the Cold War. Parties that have increased their seat share in the previous election become more positive about the military in the next election. Now we move to model 2 and the three-way interaction effects with party family to the right of the vertical bar in figure 3. For left-wing parties, we find that opposition parties, like the general trend, become more positive about the military if there are more military casualties. The effect for left-wing governing parties is insignificant. Right-wing parties fully comply with the findings in the general model. Lastly, we find no effect for ecologist/socialist party. Very few of them are governing parties. But also the general effect of ecologist/socialist party is insignificant in model 1. In conclusion, these parties are unresponsive to military casualties. This might fit into the larger phenomenon that such parties (e.g., niche parties) are less responsive than mainstream parties (Ezrow et al. 2011). Robustness Checks We have conducted several robustness checks: (1) excluding the United States, (2) whether parties have changed their strategy before and after 9/11, (3) absolute numbers rather than the log variants of our independent variable of military casualties, (4) differences in composition of the coalition, (5) using a left–right position variable instead of party family, and (6) control for popularity. The Online Appendix 2 contains the full regression tables. We exclude the United States because it is a statistical outlier in our sample. In the analysis excluding the United States, the results are in the same direction, but the effect of military casualties for left-wing, opposition parties disappears. The significant effect in our initial analyses must have been driven by party position changes of the Democratic Party (US). The effects for right-wing parties are stable. Opposition, right-wing parties become more positive, while government, left-wing parties become more negative. This means that only our results for right-wing parties hold on both sides of the pond. We ran a second robustness analysis that compares the period before 9/11 with the one after. There are two important differences: pre-9/11 there were far fewer military casualties, and the missions post-9/11 are about regime change rather than humanitarian missions. Our results from the post-9/11 analysis are similar to our analysis of the general sample, regardless of whether we include or exclude the American parties. Including American parties in our pre-9/11 analysis shows similar results as well. When excluding American parties from our pre-9/11 sample, the effects are no longer significant, which means that the effects before 9/11 hold only for non-US democracies. With the third robustness analysis, which is for absolute numbers rather than the log variants, the effects are the same but become less significant. Only the effects for left-wing and right-wing opposition parties hold, while those for governing parties disappear. Interestingly, we also found different effects between parties that have won or lost the previous election. In our main model, we found winners tend to become more positive about the military in their next party manifesto. The same effect of increased security is seen in coalition governments in general: the larger the majority enjoyed by parties in government, the more positive they become about the military. This, in combination with the mixed effects found on a party family level, suggests that future research could benefit from looking deeper into the mechanisms driving these results, and establishes in which cases position in office (and how strong that position is) trumps ideological differences. This allows researchers to disentangle position in office from ideology on a more detailed level and identify the preconditions for blame avoidance strategies. In our fifth robustness check, we omitted the party family variable and replaced it with a variable that indicates the left–right position of the party. For this, we used the rile variable from the MP database (Volkens et al. 2015). This is a composite variable that includes party stances on a range of different issues, which can be aligned along a left–right dimension. The results are highly similar to the results using the party family measure: we find positive effects for right-wing opposition parties, negative effects for right-wing governing parties, positive effects for left-wing opposition parties, but no effect for left-wing governing parties. And finally extreme values on the left end of the left–right scale are associated with a negative effect but only for opposition parties. Arguably, this is the area in which the socialist and green parties are placed. In conclusion, we have controlled for popularity and found parties that are losing in the polls to become (on average) more pro-military. This could indicate that parties that are the victim of being blamed by voters for bad (national) policies shift their attention to foreign policy in an attempt to avoid blame. Since we do not have enough observations for each party family, it is difficult to entangle other effects that might be at play in the background. This might be, however, interesting for future research. Conclusion and Discussion To summarize our findings (also see table 3), we reject our commitment hypothesis. Governing parties do not increase their commitment to the military when military casualties increase. We accept our don't mention the war hypothesis. Governing parties reduce their commitment to the military when military casualties increase. However, this effect is strongly driven by the American parties and the post-9/11 period. For the opposition parties, we reject H3 (they do not reduce their commitment); instead, we accept H4 (opposition parties increase their commitment to the military). Again, this effect is strongest when including the American parties and in the post-9/11 period. As for ideological differences, the general findings are the clearest for right-wing parties. For left-wing parties we have more null findings. Table 3. Summary of effects of the number of military casualties . Including United States . Excluding United States . . Opposition . In office . Opposition . In office . All parties + − 0 0 Left wing + 0 0 0 Right wing + − + − Ecologist/socialist 0 0 0 0 . Including United States . Excluding United States . . Opposition . In office . Opposition . In office . All parties + − 0 0 Left wing + 0 0 0 Right wing + − + − Ecologist/socialist 0 0 0 0 Open in new tab Table 3. Summary of effects of the number of military casualties . Including United States . Excluding United States . . Opposition . In office . Opposition . In office . All parties + − 0 0 Left wing + 0 0 0 Right wing + − + − Ecologist/socialist 0 0 0 0 . Including United States . Excluding United States . . Opposition . In office . Opposition . In office . All parties + − 0 0 Left wing + 0 0 0 Right wing + − + − Ecologist/socialist 0 0 0 0 Open in new tab How do political parties respond to increasing military casualties in military interventions? Several studies demonstrate that with increasing military casualties, public support for the military presence in that conflict declines. In line with Budge and Farlie (1983), opposition parties respond to this by becoming more positive on the military, while governing parties nuance their position, and pay less attention to it. For opposition parties, increasing military casualties are a good opportunity to talk about the military and the failure of the government to use the military effectively. But they are not critical about the military itself. For governing parties, it is better not to talk about the military at all. In contrast to studies of decision-making, we do not find that governments escalate their commitment in their communication. They do not mention the military in a more positive way in an attempt to maintain that the intervention was “the right thing” to do. This suggests that governing parties try to avoid blame for the casualties primarily by mentioning the war. Opposition parties changing their tone after 9/11 could be explained by either the fact that the nature of military operations changed after 9/11, or that these missions were far more deadly (with higher soldiers-killed-to-troop ratios). Before 9/11, a large part of the military interventions undertaken by Western democracies were third party, humanitarian interventions and consisted of a mandate to intervene between two (or more) already existing striving parties. After 9/11, Western democracies themselves became one of the striving parties by initiating conventional military conflict, by invading Afghanistan and Iraq leading to far higher death tolls. Here, political elites seem to agree (despite their ideological orientation) on their strategy: avoid blame once in government, and escalate commitment once in opposition. Future research could benefit from further unpacking these findings, by looking into the composition of government coalition. This can be done by looking at ideology, or (as our results indicate) by position in office. It would be interesting to test whether secure governments (with large coalition majorities and public support) use different tactics when faced with military casualties than insecure governments. One would expect (based on findings from the diversionary use of force literature and prospect theory; see, e.g., Farnham 1992; Levy 1996; Mercer 2005; Brulé and Williams 2009) that insecure governments are more prone to escalatory policies, and thus might become more positive about the military. There is evidence that public opinion is not always responsive to military casualties (Boettcher and Cobb 2006; Gelpi and Feaver 2006; Berinsky 2009). This has consequences for the strategies of political parties. Still, we find that political parties do respond to military casualties even if public opinion does not. Even though we hypothesized active damage control from governing parties when confronted with military casualties, we found no evidence to support this. The most interesting finding is the consistent finding of governing parties becoming more negative, and indications of parties from different party families adopting similar strategies as opposition or governing party. However, even though party manifestoes can reveal underlying policy preferences over time, they only provide 4-year-interval snapshots and thus capture only part of the dynamics between government and opposition, namely, close before the elections. In addition, by design, our study does not include a content analysis on the tone, therefore limiting the outcomes only to positive or negative references to the military in general. Another point worth making in this context is the fact that the decision to intervene is largely a co-decision between government and opposition parties in multiparty systems and it has two consequences. First, it does not allow for our models to adequately capture strategic interaction on the domestic level between political parties. That is, the strategies of government and opposition parties are likely dependent not only upon events such as military casualties but also on the behavior of opponents. Second, it is hard to specify to what extent the conflict is inherited (or not). This is especially the case for the post-9/11 large-scale interventions in Afghanistan (average duration of deployment 60 months with a maximum of 225 months) and Iraq (average duration of deployment 46 months with a maximum of 180 months). Since governments fear public backlash over military casualties, one would expect a higher incentive for shifting the blame to the previous government who decided to intervene in the first place. For instance, after being confronted with two increasingly unpopular and costly wars (Afghanistan and Iraq) President Obama promised during his campaign (in both 2008 and 2012) to end the war in Afghanistan, and took credit in 2012 for retreating American troops from Iraq (which already began under President Bush since September 2007). Unfortunately, our data do not allow to differentiate between “inherited” and “chosen” conflicts (see the Online Appendix 3). Still, despite the high amount of military casualties that are only to be found in the United States, we were able to systematically test and compare party strategies and conclude that political strategy for right-wing parties does not differ much on both sides of the pond. Footnotes 1 Even though it is not possible to control for this within the European context, it is worth noting that US studies that controlled for the number of soldiers killed per state found casualty tolerance to be lower among voters in those particular states that endured heavy military losses in wartime (Hayes and Myers 2009; Kriner and Shen 2014), and could vary over time as well (Althaus, Bramlett, and Gimpel 2011). 2 Data gathered through open sources (databases of national departments of Defense, rolls of honor, veteran websites etc.); for more information see the Online Appendix. 3 The reason for limiting the sample to these eleven countries is two-fold. First, and for theoretical reasons, full-fledged Western democracies outside the transatlantic geographical area were excluded. Second, for reasons of data availability and time constraints, not all European countries were included in the sample. In our sample, there are countries with presidential systems (e.g., the United States), semi-presidential systems (e.g., France), proportional representation systems (e.g., Denmark and the Netherlands), and majoritarian systems (e.g., the United Kingdom). Despite these differences in specifics of government arrangement and party systems, they share broad similarities in the functioning of the democratic process. 4 Empirical work on public opinion and military casualties has been done on, e.g., Canada, France, the United Kingdom, and Israel (see, for instance, Cornish 2003; Clements 2013; Gribble et al. 2015). However, these empirical tests have mainly been done on high-profile events (Afghanistan, Iraq, Libya) and on—from a European perspective—military outliers. 5 Party position on peace is not taken into consideration since this variable measures to what extent parties are in favor of peace, but do not give any information on the preferred means to achieve this (e.g., diplomacy or the military). 6 Even though different countries measure the number of soldiers killed differently (e.g., whether soldiers who die of a heart attack are considered “killed during a mission” or not), the number of soldiers killed in the dataset is based on the numbers reported by the national governments as such. References Adams J. 2012 . “ Causes and Electoral Consequences of Party Policy Shifts in Multiparty Elections: Theoretical Results and Empirical Evidence .” Annual Review of Political Science 15 : 401 – 19 . Google Scholar Crossref Search ADS WorldCat Adams J. , Clark M., Ezrow L., Glasgow G.. 2004 . “ Understanding Change and Stability in Party Ideologies: Do Parties Respond to Public Opinion or to Past Election Results ?” British Journal of Political Science 34 ( 4 ): 589 – 610 . Google Scholar Crossref Search ADS WorldCat Adams J. , Haupt A. B., Stoll H.. 2008 . “ What Moves Parties? The Role of Public Opinion and Global Economic Conditions in Western Europe .” Comparative Political Studies 42 ( 5 ): 611 – 39 . Google Scholar Crossref Search ADS WorldCat Adams J. , Somer-Topcu Z.. 2009 . “ Policy Adjustment by Parties in Response to Rival Parties’ Policy Shifts: Spatial Theory and the Dynamics of Party Competition in Twenty-Five Post-War Democracies .” British Journal of Political Science 39 ( 4 ): 825 . Google Scholar Crossref Search ADS WorldCat Aday S. , Cluverius J., Livingston S.. 2005 . “ As Goes the Statue, So Goes the War: The Emergence of the Victory Frame in Television Coverage of the Iraq War .” Journal of Broadcasting & Electronic Media 49 ( 3 ): 314 – 31 . Google Scholar Crossref Search ADS WorldCat Althaus S. L. , Bramlett B. H., Gimpel J. G.. 2011 . “ When War Hits Home: The Geography of Military Losses and Support for War in Time and Space .” Journal of Conflict Resolution 56 ( 3 ): 382 – 412 . Google Scholar Crossref Search ADS WorldCat Arena P. 2015 . “ Crisis Bargaining, Domestic Opposition, and Tragic Wars .” Journal of Theoretical Politics 27 ( 1 ): 108 – 31 . Google Scholar Crossref Search ADS WorldCat Bakker R. , De Vries C., Edwards E., Hooghe L., Jolly S., Marks G., Vachudova M. A.. 2015 . “ Measuring Party Positions in Europe: The Chapel Hill Expert Survey Trend File, 1999–2010 .” Party Politics 21 ( 1 ): 143 – 52 . Google Scholar Crossref Search ADS Google Scholar Google Preview WorldCat COPAC Bausch A. W. 2015 . “ Democracy, War Effort, and the Systemic Democratic Peace .” Journal of Peace Research 52 ( 4 ): 435 – 47 . Google Scholar Crossref Search ADS Google Scholar Google Preview WorldCat COPAC Bazerman M. H. , Giuliano T., Appelman A.. 1984 . “ Escalation of Commitment in Individual and Group Decision Making .” Organizational Behavior and Human Performance 33 ( 2 ): 141 – 52 . Google Scholar Crossref Search ADS WorldCat Beck N. , Katz J. N.. 1995 . “ What to Do (and Not to Do) with Time-Series Cross-Section Data .” American Political Science Review 89 ( 3 ): 634 – 47 . Google Scholar Crossref Search ADS WorldCat Bell S. 2007 . “ The UK's Risk Management Approach to National Security .” The RUSI Journal 152 ( 3 ): 18 – 23 . Google Scholar Crossref Search ADS WorldCat Berinsky A. J. 2007 . “ Assuming the Costs of War: Events, Elites, and American Public Support for Military Conflict .” Journal of Politics 69 ( 4 ): 975 – 97 . Google Scholar Crossref Search ADS WorldCat Berinsky A. J. . 2009 . In Times of War: Understanding American Public Opinion from World War II to Iraq . Chicago & London: University of Chicago Press . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Berinsky A. J. , Kinder D. R.. 2006 . “ Making Sense of Issues through Media Frames: Understanding the Kosovo Crisis .” Journal of Politics 68 ( 3 ): 640 – 56 . Google Scholar Crossref Search ADS WorldCat Biddle S. 2005 . American Grand Strategy after 9/11: An Assessment. Challenges ( vol. 1 ). US Army War College: Strategic Studies Institute . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Boettcher W. A. , Cobb M. D.. 2006 . “ Echoes of Vietnam? Casualty Framing and Public Perceptions of Success and Failure in Iraq .” The Journal of Conflict Resolution 50 ( 6 ): 831 – 54 . Google Scholar Crossref Search ADS WorldCat Boettcher W. A. , Cobb M. D.. 2009 . ““ Don't Let Them Die in Vain”: Casualty Frames and Public Tolerance for Escalating Commitment in Iraq .” Journal of Conflict Resolution 53 ( 5 ): 677 – 97 . Google Scholar Crossref Search ADS WorldCat Brambor T. , Clark W. R., Golder M.. 2006 . “ Understanding Interaction Models: Improving Empirical Analyses .” Political Analysis 14 ( 1 ): 63 – 82 . Google Scholar Crossref Search ADS WorldCat Brockner J. 1992 . “ The Escalation of Commitment to a Failing Course of Action: Toward Theoretical Progress .” Academy of Management Review 17 ( 1 ): 39 – 61 . Google Scholar Crossref Search ADS WorldCat Brulé D. J. , Williams L. K.. 2009 . “ Democracy and Diversion: Government Arrangements, the Economy, and Dispute Initiation .” Journal of Peace Research 46 ( 6 ): 777 – 98 . Google Scholar Crossref Search ADS WorldCat Budge I. 1994 . “ A New Spatial Theory of Party Competition : Uncertainty, Ideology and Policy Equilibria Viewed Comparatively and Temporally .” British Journal of Political Science 24 ( 4 ): 443 – 67 . Google Scholar Crossref Search ADS WorldCat Budge I. , Farlie D. J.. 1983 . Explaining and Predicting Elections: Issue-effects and Party Strategies in 23 Democracies . London : Allen & Unwin . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Bueno De Mesquita B. , Downs G. W.. 2006 . “ Intervention and Democracy .” International Organization 60 ( 3 ): 627 – 49 . OpenURL Placeholder Text WorldCat Bueno de Mesquita B. , Morrow J. D., Siverson R. M., Smith A.. 2002 . “ Political Institutions, Policy Choice and the Survival of Leaders .” British Journal of Political Science 32 ( 4 ): 559 – 90 . Google Scholar Crossref Search ADS WorldCat Chesebro J. W. 1997 . Analyzing Media: Communication Technologies as Symbolic and Cognitive Systems. College Composition and Communication (vol. 48 ). New York & London: Routledge . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Clements B. 2013. “ Public Opinion and Military Intervention: Afghanistan, Iraq and Libya .” The Political Quarterly 84 ( 1 ): 119 – 31 . Crossref Search ADS WorldCat Cornish P. 2003 . “ Myth and Reality: US and UK Approaches to Casualty Aversion and Force Protection .” Defence Studies 3 ( 2 ): 121 – 8 . Google Scholar Crossref Search ADS WorldCat Croco S. E. 2011 . “ The Decider's Dilemma: Leader Culpability, War Outcomes, and Domestic Punishment .” American Political Science Review . http://doi.org/10.1017/s0003055411000219 . OpenURL Placeholder Text WorldCat Croco S. E. Weeks J. L. P. 2016 . “ War Outcomes and Leader Tenure .” World Politics . http://doi.org/10.1017/s0043887116000071 . OpenURL Placeholder Text WorldCat Dannenbaum T. 2011 . “ Bombs, Ballots, and Coercion: The Madrid Bombings, Electoral Politics, and Terrorist Strategy .” Security Studies 20 ( 3 ): 303 – 49 . Google Scholar Crossref Search ADS WorldCat Dardis F. E. 2006 . “ Marginalization Devices in U.S. Press Coverage of Iraq War Protest: A Content Analysis .” Mass Communication & Society 9 ( 2 ): 117 – 35 . Google Scholar Crossref Search ADS WorldCat Dimitrova D. V. Connolly-Ahern C.. 2007 . “ A Tale of Two Wars: Framing Analysis of Online News Sites in Coalition Countries and the Arab World During the Iraq War .” Howard Journal of Communications 18 ( 2 ): 153 – 68 . Google Scholar Crossref Search ADS WorldCat Dimitrova D. V. Strömbäck J.. 2005 . “ Mission Accomplished? Framing of the Iraq War in the Elite Newspapers in Sweden and the United States .” International Communication Gazette 67 ( 5 ): 399 – 417 . Google Scholar Crossref Search ADS WorldCat Döring H. Manow P.. 2016 . “ Parliaments and Governments Database (ParlGov): Information on Parties, Elections and Cabinets in Modern Democracies .” Development version. Druckman J. N. Peterson E. Slothuus R.. 2013 . “ How Elite Partisan Polarization Affects Public Opinion Formation .” American Political Science Review 107 ( 1 ): 57 – 79 . Google Scholar Crossref Search ADS WorldCat Dunlap C. J. Jr. 1999 . “ Kosovo, Casualty Aversion, and the American Military Ethos: A Perspective .” Journal of Legal Studies 10 : 95 . OpenURL Placeholder Text WorldCat Dyson S. B. 2006 . “ Personality and Foreign Policy: Tony Blair's Iraq Decisions .” Foreign Policy Analysis 2 ( April 2005 ): 289 – 306 . Google Scholar Crossref Search ADS WorldCat Eichenberg R. C. 2005 . “ Victory Has Many Friends: U.S. Public Opinion and the Use of Military Forde, 1981–2005 .” International Security 30 ( 1 ): 140 – 77 . Google Scholar Crossref Search ADS WorldCat Eichenberg R. C. Stoll R. J. Lebo M.. 2006 . “ War President: The Approval Ratings of George W. Bush .” Journal of Conflict Resolution 50 ( 6 ): 783 – 808 . Google Scholar Crossref Search ADS WorldCat Eriksson M. 2015 . “ UCDP/PRIO Armed Conflict Dataset Codebook .” Upsala Conflict Data Programs (UCDP) ( 4 ): 22 . OpenURL Placeholder Text WorldCat Ezrow L. De Vries C. Steenbergen M. Edwards E.. 2011 . “ Mean Voter Representation and Partisan Constituency Representation: Do Parties Respond to the Mean Voter Position or to Their Supporters ?” Party Politics 17 ( 3 ): 275 – 301 . Google Scholar Crossref Search ADS WorldCat Farnham B. 1992 . “ Roosevelt and the Munich Crisis: Insights from Theory Prospect .” Poitical Psychology 13 ( 2 ): 205 – 35 . Google Scholar Crossref Search ADS WorldCat Fearon J. D. 1994 . “ Domestic Political Audiences and the Escalation of International Disputes .” American Political Science Review 88 ( 3 ): 577 – 92 . Google Scholar Crossref Search ADS WorldCat Foster D. M. Gibler D. M.. 2016 . Is Anyone Listening? Crisis Signaling by the British House of Commons, 1918–2004 . Working Paper, University of Alabama . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Gaines B. J. Kuklinski J. H. Quirk P. J. Peyton B. Verkuilen J.. 2007 . “ Same Facts, Different Interpretations: Partisan Motivation and Opinion on Iraq .” Journal of Politics 69 ( 4 ): 957 – 74 . Google Scholar Crossref Search ADS WorldCat Garland H. 1990 . “ Throwing Good Money After Bad: The Effect of Sunk Costs on the Decision to Escalate Commitment to an Ongoing Project .” Journal of Applied Psychology 75 ( 6 ): 728 – 31 . Google Scholar Crossref Search ADS WorldCat Gartner S. S. 2008 . “ The Multiple Effects of Casualties on Public Support for War: An Experimental Approach .” American Political Science Review 102 ( 1 ): 95 – 106 . Google Scholar Crossref Search ADS WorldCat Gartner S. S. Segura G. M. Barratt B. A.. 2003 . “ War Casualties, Policy Positions, and the Fate of Legislators .” Political Research Quarterly 53 ( 3 ): 467 – 77 . Google Scholar Crossref Search ADS WorldCat Gelpi C. Feaver P. D.. 2006 . “ Success Matters Casualty Sensitivity and the War in Iraq .” International Security 30 ( 3 ): 7 – 46 . Google Scholar Crossref Search ADS WorldCat Gelpi C. Feaver P. D. Reifler J.. 2009 . Paying the Human Costs of War: American Public Opinion and Casualties in Military Conflicts . Princeton: Princeton University Press . Google Scholar Crossref Search ADS Google Scholar Google Preview WorldCat COPAC Gemenis K. 2013 . “ What to Do (and Not to Do) with the Comparative Manifestos Project Data .” Political Studies 61 : 23 – 43 . Google Scholar Crossref Search ADS WorldCat Gennaioli N. Voth H.-J.. 2011 . State Capacity and Military Conflict . Working Paper . Google Scholar Crossref Search ADS Google Scholar Google Preview WorldCat COPAC Gettleman M. E. Young M. Franklin J. Franklin H. B.. 1995 . Vietnam and America: A Documented History . New York : Grove Press. Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Gordon P. H. 2002 . “ NATO and the War on Terrorism a Changing Alliance .” The Brookings Review 20 ( 3 ): 36 – 8 . OpenURL Placeholder Text WorldCat Green-Pedersen C. 2007 . “ The Growing Importance of Issue Competition: The Changing Nature of Party Competition in Western Europe .” Political Studies 55 ( 3 ): 607 – 28 . Google Scholar Crossref Search ADS WorldCat Gribble R. , Wessley S., Klein S., Alexander D. A., Dandeker C., Fear N. T.. 2015 . “ British Public Opinion After a Decade of War: Attitudes to Iraq and Afghanistan .” Politics 35 ( 2 ): 128 – 50 . Google Scholar Crossref Search ADS WorldCat Hayes A. F. Myers T. A.. 2009 . “ Testing the “Proximate Casualties Hypothesis”: Local Troop Loss, Attention to News, and Support for Military Intervention .” Mass Communication and Society 12 ( 4 ): 379 – 402 . Google Scholar Crossref Search ADS WorldCat Hetherington M. J. Nelson M.. 2003 . “ Anatomy of a Rally Effect: George W. Bush and the War on Terrorism .” PS: Political Science & Politics . http://doi.org/10.1017/S1049096503001665 . OpenURL Placeholder Text WorldCat Hobolt S. B. de Vries C. E.. 2015 . “ Issue Entrepreneurship and Multiparty Competition .” Comparative Political Studies 48 ( 9 ): 1159 – 85 . Google Scholar Crossref Search ADS WorldCat Hultman L. 2013 . “ UN Peace Operations and Protection of Civilians: Cheap Talk or Norm Implementation ?” Journal of Peace Research 50 ( 1 ): 59 – 73 . Google Scholar Crossref Search ADS WorldCat Hyde C. K. 2000 . “ Casualty Aversion .” Air & Space Power Journal 14 ( 2 ): 17 . OpenURL Placeholder Text WorldCat Kahl C. H. 2007 . “ In the Crossfire Or the Crosshairs? Norms, Civilian Casualties, and U.S. Conduct in Iraq .” International Security 32 ( 1 ): 7 – 46 . Google Scholar Crossref Search ADS WorldCat Karlsson N. Juliusson Á. Gärling T.. 2005 . “ A Conceptualisation of Task Dimensions Affecting Escalation of Commitment .” European Journal of Cognitive Psychology 17 ( 6 ): 835 – 58 . Google Scholar Crossref Search ADS WorldCat Karol D. Miguel E.. 2007 . “ The Electoral Cost of War: Iraq Casualties and the 2004 U.S. Presidential Election .” Journal of Politics 69 ( 3 ): 633 – 48 . Google Scholar Crossref Search ADS WorldCat Kent Weaver R. 1986 . “ The Politics of Blame Avoidance .” Journal of Public Policy 6 ( 4 ): 371 – 98 . Google Scholar Crossref Search ADS WorldCat Kriner D. L. Shen F. X.. 2012 . “ How Citizens Respond to Combat Casualties .” Public Opinion Quarterly 76 ( 4 ): 761 – 70 . Google Scholar Crossref Search ADS WorldCat Kriner D. L. Shen F.. 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 . Google Scholar Crossref Search ADS WorldCat Kull S. Destler I. M.. 1999 . Misreading the Public: The Myth of a New Isolationism . Washington, DC: Brookings Institution Press . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Lacquement R. A. 2003 . “ The Casualty Aversion Myth .” Naval War College Review LVII ( 1 ): 40 – 57 . OpenURL Placeholder Text WorldCat Larson E. V. Savych B.. 2005 . American Public Support for US Military Operations from Mogadishu to Baghdad: Technical Appendixes. Rand Corporation. Levy J. S. 1996 . “ The Diversionary Theory of War .” In Handbook of War Studies , edited by Midlarsky Manus I., 259 – 88 . Boston: Unwin Hyman . OpenURL Placeholder Text WorldCat Levy J. S. . 1997 . “ Prospect Theory, Rational Choice, and International Relations .” International Studies Quarterly 41 ( 1 ): 87 – 112 . Google Scholar Crossref Search ADS WorldCat Lian B. Oneal J. R.. 1993 . “ Presidents, the Use of Military Force, and Public Opinion .” Journal of Conflict Resolution 37 ( 2 ): 277 – 300 . Google Scholar Crossref Search ADS WorldCat Litwak R. S. Litwak R.. 2008 . Regime Change: U.S. Strategy through the Prism of 9/11. Baltimore: JHU Press. Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC MacAskill E. 2013 . “ Obama Faces Tough Decisions on Syria As Pressure for Intervention Mounts .” The Guardian , p. 2 . OpenURL Placeholder Text WorldCat McCroskey J. C. 2006 . An Introduction to Rhetorical Communication: A Western Rhetorical Perspective . New York : Routledge . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC McDermott R. 2004 . “ Prospect Theory in Political Science: Gains and Losses from the First Decade .” Political Psychology 25 ( 2 ): 289 – 312 . Google Scholar Crossref Search ADS WorldCat Mercer J. 2005 . “ Prospect Theory and Political Science .” Annual Review of Political Science 8 : 1 – 21 . Google Scholar Crossref Search ADS WorldCat Mintz A. De Rouwen jr K. . 2010 . Understanding Foreign Policy Decision Making . New York: Cambridge University Press . Google Scholar Crossref Search ADS Google Scholar Google Preview WorldCat COPAC Mintz A. Yang Y. McDermott R.. 2011 . “ Experimental Approaches to International Relations .” International Studies Quarterly 55 ( 2 ): 493 – 501 . Google Scholar Crossref Search ADS WorldCat Mueller J. E. 1970 . “ Presidential Popularity from Truman to Johnson .” American Political Science Review 64 ( 1 ): 18 – 34 . Google Scholar Crossref Search ADS WorldCat Mueller J. E. . 1971 . “ Trends in Popular Support for the Wars in Korea and Vietnam .” The American Political Science Review 65 ( 2 ): 358 – 75 . Google Scholar Crossref Search ADS WorldCat Mueller J. E. . 2005 . “ The Iraq Syndrome .” Foreign Affairs 84 ( 6 ): 44 – 54 . Google Scholar Crossref Search ADS WorldCat Nincic D. J. Nincic M.. 1995 . “ Commitment to Military Intervention: The Democratic Government As Economic Investor .” Journal of Peace Research 32 ( 4 ): 413 – 26 . Google Scholar Crossref Search ADS WorldCat Noetzel T. Schreer B.. 2009 . “ Does a multi-tier NATO Matter? the Atlantic Alliance and the Process of Strategic Change .” International Affairs 85 ( 2 ): 211 – 26 . Google Scholar Crossref Search ADS WorldCat Norpoth H. Sidman A. H.. 2007 . “ Mission Accomplished: The Wartime Election of 2004 .” Political Behavior 29 ( 2 ): 175 – 95 . Google Scholar Crossref Search ADS WorldCat Pierson P. 1994 . Dismantling the Welfare State: Reagan, Thatcher, and the Politics of Retrenchment in Britain and the United States . Cambridge : Cambridge University Press . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Pierson P. . 1996 . “ The New Politics of the Welfare State .” World Politics 48 ( 2 ): 143 – 79 . Google Scholar Crossref Search ADS WorldCat Plümper T. Troeger V. E. Manow P.. 2005 . “ Panel Data Analysis in Comparative Politics: Linking Method to Theory .” European Journal of Political Research 44 ( 2 ): 327 – 54 . Google Scholar Crossref Search ADS WorldCat Preston T. 2011 . Pandora's Trap: Presidential Decision Making and Blame Avoidance in Vietnam and Iraq . Plymouth : Rowman and Littlefield Publishers. Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Riker W. H. 1988 . “ Liberalism Against Populism: A Confrontation Between the Theory of Democracy and the Theory of Social Choice .” In Liberalism Against Populism: A Confrontation between the Theory of Democracy and the Theory of Social Choice , 1 – 19 . Prospect Heights, Illinois : Waveland Press . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Ripstein A. 2004 . Critical Notice Too Much Invested To Quit. Economics and Philosophy (vol. 20 ). Oxford: Pergamon Press . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Rubin J. Z. Brockner J.. 1985 . “ Social Influence and Entrapment .” In Entrapment in Escalating Conflicts , 57 – 100 . New York, NY: Springer. Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Rupp R. E. 2016 . NATO after 9/11: An alliance in continuing decline. NATO After 9/11: An Alliance in Continuing Decline . New York: Springer . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Schubert J. N. Stewart P. A. Curran M. A.. 2002 . “ A Defining Presidential Moment: 9/11 and the Rally Effect .” Political Psychology 23 ( 3 ): 559 – 83 . Google Scholar Crossref Search ADS WorldCat Schultz K. 2013 . “ Domestic Politics and International Relations .” In Handbook of International Relations , 478 – 502 . London: SAGE Publications Ltd. Google Scholar Crossref Search ADS Google Scholar Google Preview WorldCat COPAC Smith A. 1998 . “ International Crises and Domestic Politics .” The American Political Science Review 92 ( 3 ): 623 – 38 . Google Scholar Crossref Search ADS WorldCat Smith H. 2005 . “ What Costs Will Democracies Bear? A Review of Popular Theories of Casualty Aversion .” Armed Forces & Society 31 ( 4 ): 487 – 512 . Google Scholar Crossref Search ADS WorldCat Somer-Topcu Z. 2009 . “ Timely Decisions: The Effects of Past National Elections on Party Policy Change .” The Journal of Politics 71 ( 1 ): 238 – 48 . Google Scholar Crossref Search ADS WorldCat Sperling J. Webber M.. 2009 . “ NATO: From Kosovo to Kabul .” International Affairs 85 ( 3 ): 491 – 511 . Google Scholar Crossref Search ADS WorldCat Staw B. M. 1976 . “ Knee-Deep in the Big Muddy : A Study of Escalating Commitment to a Chosen Course of Action .” Organizational Behavior and Human Performance 16 ( 1 ): 27 – 44 . Google Scholar Crossref Search ADS WorldCat Staw B. M. . 1997 . “ The Escalation of Commitment: An Update and Appraisal .” In Organizational Decision Making , edited by Shapira Z., 191 – 215 . Cambridge: Cambridge University Press . Google Scholar Crossref Search ADS Google Scholar Google Preview WorldCat COPAC Stimson J. 1985 . “ Regression Models in Space and Time: A Statistical Essay .”American Journal of Political Science 29 ( 4 ): 914 – 47 . OpenURL Placeholder Text WorldCat Stoll R. J. 1987 . “ The Sound of Guns: Is There a Congressional Rally Effect After US Military Action ?”American Politics Research 15 ( 2 ): 223 – 37 . OpenURL Placeholder Text WorldCat Tardy T. 2011 . “ A Critique of Robust Peacekeeping in Contemporary Peace Operations .”International Peacekeeping 18 ( April 2015 ): 152 – 67 . OpenURL Placeholder Text WorldCat Tir J. Singh S. P.. 2013 . “ Is It the Economy Or Foreign Policy, Stupid? The Impact of Foreign Crises on Leader Support .”Comparative Politics 46 ( 1 ): 83 – 101 . OpenURL Placeholder Text WorldCat Tomz M. 2007 . “ Domestic Audience Costs in International Relations: An Experimental Approach .” International Organization 61 ( 4 ): 821 – 40 . Google Scholar Crossref Search ADS WorldCat Valentino B. A. , Huth P. K., Croco S. E.. 2010 . “ Bear Any Burden? How Democracies Minimize the Costs of War .” Journal of Politics . http://doi.org/10.1017/S0022381609990831 . OpenURL Placeholder Text WorldCat Vis B. 2016 . “ Taking Stock of the Comparative Literature on the Role of Blame Avoidance Strategies in Social Policy Reform .” Journal of Comparative Policy Analysis: Research and Practice 18 ( 2 ): 122 – 37 . Google Scholar Crossref Search ADS WorldCat Vis B. , van Kersbergen K. 2007 . “ Why and How Do Political Actors Pursue Risky Reforms? ” Journal of Theoretical Politics 19 ( 2 ): 153 – 72 . Google Scholar Crossref Search ADS WorldCat Volkens A. , Lehmann P., Matthiess T., Merz N., Regel S., Werner A.. 2015 . “ The Manifesto Data Collection. Manifesto Project .” Manifesto Project (MRG/CMP/MARPOR) . Wagner W. , Herranz-Surrallés A., Kaarbo J., Ostermann F.. 2017 . “ Politicization, Party Politics and Military Missions: Deployment Votes in France, Germany, Spain, and the United Kingdom .” Wissenschaftszentrum Berlin Für Sozialforschung , Discussion (January) . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Walgrave S. , Nuytemans M. 2009 . “ Friction and Party Manifesto Change in 25 Countries, 1945–98 .” American Journal of Political Science 53 ( 1 ): 190 – 206 . Google Scholar Crossref Search ADS WorldCat Wenzelburger G. 2014 . “ Blame Avoidance, Electoral Punishment and the Perceptions of Risk .” Journal of European Social Policy 24 ( 1 ): 80 – 91 . Google Scholar Crossref Search ADS WorldCat Whyte G. 1986 . “ Escalating Commitment to a Course of Action: A Reinterpretation .” The Academy of Management Review 11 ( 2 ): 311 – 21 . Google Scholar Crossref Search ADS WorldCat Whyte G. 1993 . “ Escalating Commitment in Individual and Group Decision Making: A Prospect Theory Approach .” Organizational Behavior and Human Decision Processes 54 : 430 – 55 . Google Scholar Crossref Search ADS WorldCat Williams E. 2008 . “ Out of Area and Very Much in Business? NATO, the U.S., and the Post-9/11 International Security Environment .” Comparative Strategy 27 ( 1 ): 65 – 78 . Google Scholar Crossref Search ADS WorldCat Williams L. K. 2015 . “ It's All Relative: Spatial Positioning of Parties and Ideological Shifts .” European Journal of Political Research 54 ( 1 ): 141 – 59 . Google Scholar Crossref Search ADS WorldCat © The Author(s) (2020). Published by Oxford University Press on behalf of the International Studies Association. This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/open_access/funder_policies/chorus/standard_publication_model)

Journal

Foreign Policy AnalysisOxford University Press

Published: Sep 26, 2020

There are no references for this article.