Tough Enough: How Domestic Pressure to Use Force Affects Leaders Differently

Existing literature assesses how prior military training affects the likelihood that leaders use military force in ofﬁce. One gap in this literature is whether a leader’s domestic environment also shapes individual propensity for use of force. I focus on militarism as a domestic factor that triggers variation in the conﬂict tendencies of leaders and operationalize militarism with a novel metric that compares military to welfare spending. The results show that militarism has an equal effect on the conﬂict propensity of leaders with civilian and military backgrounds. However, comparing military leaders with combat experience to those without combat experience, the results show that militarism produces the strongest effect on the conﬂict propensity of military-no-combat leaders, particularly in dictatorships. I posit that this outcome results from the pressure that these leaders confront to demonstrate toughness. I refer to the pressure to perform toughness as the “demonstration dynamic” and use this concept to explain the dramatic shift in Libyan foreign policy from King Idris to Muammar Qaddaﬁ.


Introduction
In the summer of 1967, the Egyptian regime of President Gamal Abdel Nasser carried out several aggressive international maneuvers that increased the chances of war with Israel (Mor 1991, 359). Nasser mobilized Egyptian troops into the Sinai, expelled UN peacekeepers from the peninsula, and closed the Straits of Tiran. For a fleeting moment, his bellicosity electrified the Egyptian street and demonstrated toughness to his domestic audience. As the host of Radio Cairo attested, "The masses were very excited, sure that their army could win. Feelings were very high among all Egyptians, including us in the media" (Elghdban quoted in Dor-Ner 2007). Nasser was not a reckless military leader unaware of Israel's strength or the human cost of war. His military experience made him wary of confrontation: "Nasser, more than any other Arab, understood how dangerous Israel was and knew the full extent of its power" (Abu Odeh quoted in Mutawi 1987, 51). Interviews with confidants suggest Nasser was aware of Israel's material capabilities (Sharaf 2014, 302). What triggered Nasser's belligerence was the domestic political environment that he faced.
Foreign policy restraint was uniquely difficult for the Egyptian president given the domestic expectations underpinning his rule. Nasser's regime cultivated militarism and linked its legitimacy to Egypt's military prowess. Newspaper clippings from the period featured Nasser's visage aside antiaircraft artillery operated by children. 1 When tensions with Israel escalated in the summer of 1967, Nasser's status as the military strongman of a highly militarized state made foreign policy restraint tricky. Backing down from a conflict with Israel would have discredited Nasser's public image and upset the military officials that he relied on for control. 2 Archival data show the internal pressure Nasser faced to confront Israel. Weeks before the war, the Commander of 1 Pro-Nasser propaganda. Al-Jamhouria. December 24, 1964, Weekly Edition, No. 4021. Retrieved by author from the National Library of Jordan on July 1, 2019 (see the online appendix). 2 A cable from US diplomat William R. Polk, dated November 20, 1968, sug-gests Nasser was more cautious and aware of the costs of war with Israel than his advisors.
the Egyptian Air Force's Gaza Division wrote Nasser urging him to prepare Egyptian forces for a strike against Israel. 3 Before his domestic audience, previous episodes of aggressive foreign policy set a precedent for how Nasser was expected to handle tension with the Israelis. Prior to 1967, Nasser stood up to European powers during the Suez episode, assisted allies in Algeria and the Congo, and sent a third of the Egyptian army to fight in North Yemen's civil war. Headlines from the period suggest aggressive foreign policy was domestically advantageous to Nasser. 4 Though costly, these adventures demonstrated his commitment to restore Egypt's "moral energies" (Nasser 1955). The case of Egypt's belligerence in 1967 captures the intuition behind this article: leaders with military experience making foreign policy decisions in militarized states face heightened pressure to prove their willingness to use force. Other examples, including the foreign policy behavior of Saddam Hussein and Muammar Qaddafi further highlight this point.
This article complements a growing body of literature on the link between prior military experience and use of force in office (Feaver and Gelpi 2004;Weeks 2012;Horowitz and Stam 2014). The motivation for this project is to show that militarism is a state factor that shapes public expectations toward leaders and their use of force. I argue that leaders with military backgrounds face heightened domestic pressure to use military force to project toughness, particularly when militarism inside their states is high. This is because leaders with prior military training frequently acquire political power because of their familiarity with military force. In turn, this creates the social expectation that these leaders will demonstrate a willingness to use force as a foreign policy tool. This project operates at the leader level of analysis but focuses on how militarism at the state level creates pressure on leaders.
To begin, I summarize the existing literature on the link between prior military service and the use of force in office. I then highlight how an environmental factor, whether a leader's state is militarized or not, conditions this relationship between prior military service and use of force. Next, I put forth two hypotheses and outline a research design. In doing so, I put forth the notion of the "demonstration dynamic" to theorize why military-trained leaders face intensified pressure to demonstrate toughness through foreign policy. After discussing empirical findings, I carry out a case study analysis on Libya during the mid-twentieth century. The goal is to explore the demonstration dynamic's validity as a causal explanation for the radical shift in Libyan foreign policy from King Idris to Muammar Qaddafi.

Existing Research on Military Experience and Use of Force
Military experience is a foundational life episode that affects an individual's development (Caspi, Roberts, and Shiner 2005). Prior military experience shapes how leaders think about risk and use of force (Brunk, Secrest, and Tamashiro 1990). Feaver and Gelpi (2004, 65) argue that military elites are more selective about the use of force than career civilians but more willing to escalate the amount of force used once the decision to engage is made. Leaders with military training are not a monolith and recent scholarship shows ingroup variation. For example, leaders with military training but no combat experience are more restrained about using force (Horowitz and Stam 2014, 527) and less selective in the states they target (Horowitz et al. 2018, 1). Despite evidence of a link between personal traits and proclivity for use of force, leaders' conflict management choices do not unfold in a vacuum. Calling for psychological theories of leader behavior to pay more attention to environment, Goldgeier and Tetlock (2001, 88) discuss how environment and cognition interact to produce variation in the behavior of decision-makers. With respect to the question of war initiation, there is long-running interest in both individual traits and environmental factors. War proneness is attributed to both a leader's social conditions and his or her innate attributes (Skjelsbaek 1979, 220).
For example, analysis of female heads of government reveals that under certain political conditions they face pressure to initiate interstate conflicts to project competence (Schramm and Stark 2020, 524). Leaders bring into office differing assumptions about the nature of global threats that affect their propensity for use of force (Whitlark 2017, 546). Life experiences such as military training (Feaver and Gelpi 2004, 67;Horowitz and Stam 2014, 527), travel (Saunders 2009, 139), and revolutionary politics (Colgan and Weeks 2015, 163) shape personal beliefs toward the use of force. Building from prior work that frames war proneness as both the product of personal traits and social factors (Skjelsbaek 1979;Hagan 1994), the motivation of this study is to explore how militarism at the state level affects leaders differently.

Leader Type, Militarism, and Use of Force
While prior military experience can afford leaders credibility (Horowitz and Stam 2014, 541), it may also amplify domestic pressure to demonstrate strength. Particularly in autocracies, leaders with military backgrounds may rise the ranks because their personal backgrounds demonstrate a willingness to use force (Horowitz and Stam 2014, 535). Military dictators often legitimate their rule by asserting that their military training positions them to lead. In civilian dictatorships and monarchies, the heir to power often undergoes military instruction in preparation for ascension to the office or the throne. Because leaders with prior military ex- perience carry these reputations of toughness into office, I predict that they confront heightened domestic pressure to exude strength in their management of foreign policy. Do certain domestic conditions amplify the pressure leaders face to use foreign policy to signal toughness at home? Assessing this question, I focus on domestic militarism as a factor that shapes the utility that leaders assign to the use of force. The focus here centers on how domestic militarism conditions the relationship between prior military experience and the use of force. What is militarism and how does the degree of militarism in a state affect the likelihood that a leader uses force? I draw on Radway (1968, 300) to define militarism as "a doctrine or system that values war and accords primacy in state and society to the armed forces. It exalts a function-the application of violence-and an institutional structure-the military establishment." Militarism mobilizes the masses around the belief that war preparation is a tool for national advancement (Naidu 1985, 3). State militarism manifests itself in symbols, language, clothing, propaganda, and sports (Sabbagh Fajardo and Miller 2017). Militarism can imbue society with an alternative consciousness that relishes war and alters perceptions of reality (Ben-Eliezer 1998, xi).
I make two predictions. First, I predict that militarism has a stronger effect on leaders with prior military training than career civilians. Second, among military-trained leaders, I predict militarism has a stronger effect on the behavior of those without combat experience than those with combat experience. All else equal, I assume that military leaders with combat experience can rely on the gravitas afforded to them from their pre-political involvement in battle to resist pressure to use force once they take office. 5 The theory that I put forth has two main variables of interest: leader type and state setting. Leader type consists of three categories: civilian, military-combat, military-no-combat. State setting consists of two categories: nonmilitarized and militarized. I classify militarized state settings as "Sparta" and nonmilitarized state settings as "Athens." In Thucydides' account of the Peloponnesian War, Pericles describes the culture of the Athenian city-state as enlightened and intellectual in juxtaposition to the war-obsessed Spartans (Strassler 1996, 113). Across all leader types, I predict that militarism triggers a higher probability of conflict provocation (see table 1).

Causal Logic: The Demonstration Dynamic
I refer to the causal mechanism that links military leaders in militarized states to international aggression as the demonstration dynamic. The demonstration dynamic captures the notion that military leaders in militarized settings perceive political gain in demonstrating toughness through foreign policy. 6 Conversely, these leaders may face political costs, such as damage to their public standing, if their conflict decision-making fails to sufficiently demonstrate strength. Domestic publics have expectations about how willing leaders should be to employ the nation's military resources. All else equal, I posit military leaders are more likely than civilian leaders to face domestic publics who expect them to show a willingness to use force in foreign affairs.
How does domestic militarism affect this relationship between prior military experience and foreign policy decisionmaking? The demonstration dynamic theorizes that militarism increases the likelihood that domestic publics will expect leaders with military training to demonstrate their willingness to use force. Because militarism as a value system assigns utility to preparation for war (Radway 1968, 300;Naidu 1985, 2;Schofield 2007, 1), leaders who pursue diplomatic solutions to interstate strife over military options are more likely to be perceived as shirking a responsibility to utilize the nation's force resources. When militarism is high, not making use of available military resources can prove politically costly to leaders. This argument draws on research that shows the mere opportunity to make use of military means triggers conflict (Schofield 2007, 150) and causes societies to perceive war as inevitable (Bremer 1992, 318). Militaristic attitudes at the societal level affect elite decisionmaking toward the use of force (Van Evera 1984, 62).
I theorize that domestic militarism is an external factor that shapes how much pressure leaders confront to demonstrate their willingness to use force. Militarism is not only a structural factor, but also one that leaders partially control. Particularly in nondemocracies with minimal constraints on the executive, leaders can funnel national wealth toward weapons procurement, defense research, and military education and planning. Existing research shows that leaders' causal beliefs about threats in world politics shape how they budget, staff, and strategize before they make decisions about the use of force (Saunders 2011, 36).
Are leaders with military backgrounds more likely to invest in militarizing their states than those with civilian backgrounds? Military personnel do not necessarily score higher than civilians on survey indicators of militaristic sentiments (Skjelsbaek 1979, 221). Further, I carry out a means comparison (see the online appendix) that shows civilian leaders spend a slightly higher percentage of national gross domestic product (GDP) on their militaries than do military leaders. Most relevant to this study is how militarism acts back on leaders, particularly those who publicly promote military buildup. I posit that when leaders link their political profiles to national military development, they are more likely to face pressure to project toughness by utilizing the military resources they champion.
In performing toughness, the leader's audience is his or her domestic public. However, the effect of the domestic public on the leader's political survival is shaped by regime type. Applying the language of selectorate theory (Bueno de Mesquita et al. 1999, 793), members of the "winning coalition" can easily defect and back political opponents with the institutional means to challenge the leader's authority. In democracies, this occurs through mechanisms like elections and votes of no confidence. In autocracies, the loyalty of regime elites can be undermined by the leader's diminished public standing. This draws on research suggesting that in autocracies, anti-government protests (Casper and Tyson 2014, 548) and elections that reveal a leader's weak marks a social performance of toughness, whereas diversionary arguments focus on using war to distract the public from domestic ills (Oakes 2012; Jung 2014) and rally-'round-the-flag theories link war provocation to nationalist fervor (Lee 1977;Parker 1995). social support (Wig and Rød 2016, 787) frequently precede coup attempts.
I draw on Cheibub, Gandhi, and Vreeland's (2010) six regime types. These include three types of democracy (parliamentary, mixed, and presidential) as well as three types of dictatorship (civilian, military, and royal). In democracies, leaders who fail to meet public expectations of toughness in the foreign policy sphere can be punished through votes of no confidence and elections. A subcase of Margaret Thatcher's handling of the Falkland Islands crisis in the online appendix illustrates the link between toughness, foreign policy, and political survival in a democracy. The Thatcher case suggests that in a democracy, effective use of force may increase public confidence in the leader's capacity to tackle domestic challenges like unemployment (Jenkins 1983).
Dictators are also responsive to public pressure to project toughness. Failure to do so can diminish the dictator's public standing and trigger competition from political elites who see the moment as opportune to grab power. In civilian dictatorships, "insiders" have the power to remove the dictator. In military dictatorships, "officers" possess this option. In royal dictatorships, kings may be confronted by "heirs." During the 1960s, Iraqi Brigadier Abd al-Karim Qasim and Saudi Arabia's King Saud both confronted foreign policy crises related to neighboring states (Kuwait and North Yemen). Both pursued noninterventionist policies that were criticized as weak. Subsequently, both were ousted by elite challengers when public support for their authority was shaky. To summarize, the demonstration dynamic unpacks how leaders confront pressure to meet public sentiments surrounding their expected willingness to use force. Failure to do so may create windows of opportunity for competing elites to challenge the leader's authority, as diminished public standing often triggers elite contestation. Based on this logic, I generate the following two hypotheses.

Hypothesis 1 (Military versus Civilian): The conflict propensity of military leaders is most affected by militarism.
• A military leader in a militarized state setting (Sparta) is significantly more likely to provoke conflict than a military leader in a nonmilitarized state setting (Athens). • A civilian leader in a militarized state setting (Sparta) is only slightly more likely to provoke conflict than a civilian leader in a nonmilitarized state setting (Athens).

Hypothesis 2 (Military-No-Combat versus Military-Combat):
Militarism has the strongest effect on the conflict propensity of military-no-combat leaders.
• A military-no-combat leader in a militarized state setting (Sparta) is significantly more likely to provoke conflict than a military-no-combat leader in a nonmilitarized state setting (Athens) • A military-combat leader in a militarized state setting (Sparta) is only slightly more likely to provoke conflict than a military-combat leader in a nonmilitarized state setting (Athens).

Research Design
The dependent variable in the dataset is the initiation of a militarized interstate dispute (MID). Initiator MIDs represent a standard metric for measuring the conflict propensity of leaders (Peceny and Beer 2003;Weeks 2012;Colgan and Lucas 2017). MIDs are defined as "historical cases of conflict in which the threat, display or use of military force short of war by one member state is explicitly directed towards the government, official representatives, official forces, property, or territory of another state" (Jones, Bremer, and Singer 1996, 163). MIDs occur across 19 percent of the leader-years in the dataset. I focus on initiator MIDs (rather than defender MIDs) because they generally represent conflicts of choice rather than necessity. Existing research conceptualizes initiator MIDs as the preferable proxy indicator for measuring a leader's proclivity for use of force (Sechser 2004;Stein 2015;Schramm and Stark 2020). I focus on two explanatory variables in the analysis: leader type and domestic militarism. These variables stem from the logic behind the demonstration dynamic.

Leader Type
With the Why Leaders Fight replication data , I categorize leaders into three groups: (1) civilian, (2) military-combat, and (3) military-no-combat. Horowitz and Stam (2014, 527) equate combat experience with deployment to a war zone and their statistical results suggest that it causes more conservative beliefs toward the use of force later in life. While the combat versus no-combat binary produces in-group variation across military leaders, it is a coding dichotomy that faces limitations. 7 I proceed to use the combat binary over other metrics such as military rank or years served. The rationale for this stems from the argument. I expect domestic audiences to be more familiar with whether the leader's prior military service includes combat than alternative designations, such as the leader's final military rank or number of years served. This is important as the argument that I advance centers on how the public's familiarity with the leader's prior military service shapes expectations surrounding the use of force.

Militarism
Militarism captures the degree to which a population embraces preparation for war as a tool for national advancement. As an attitudinal factor, operationalizing militarism as a cross-national variable presents challenges. This is solved by turning to the related concept of militarization: observable expenditures on arms, troops, defense research, and the expansion of the military sector (Naidu 1985, 2;Ross 1987, 8). Conventional metrics for militarization include military spending as a percentage of GDP and military personnel as a percentage of the population (Bowman 1996, 29). One shortcoming of these metrics is they fail to capture the conceptual trade-off that militarism may entail between investment in military goals versus other national priorities, like education and healthcare (Hansen 1982, 11). I create a new cross-national time-series metric that measures military spending as a percentage of GDP relative to welfare spending as a percentage of GDP. This ratio serves as a proxy indicator for whether a nation prioritizes investment in preparation for war or investment in the public standard of living. Military spending prioritizes future readiness for war. Welfare spending prioritizes the immediate material needs of citizens (Snyder and Bhavnani 2005, 571;Taydas and Peksen 2012, 276). Governments that invest in welfare initiatives over other priorities tend to maintain their credibility through the provision of public goods that raise the social safety net (Taydas and Peksen 2012, 284). Conversely, 7 Thanks to a colleague of mine, a military officer, for pointing out the limitations of coding combat experience as a binary variable. No two individuals experience combat identically. Combat produces divergent experiences and varying types of risk. This variation is not captured in the simple combat versus no-combat binary. Military-no-combat leader in militarized state (Sparta) N = 401 I assume that when military spending is prioritized over welfare spending, a government's credibility is linked to performance in the defense and foreign policy spheres.
To create the metric, I merge Correlates of War (COW) military expenditure data with Gleditsch's (2002) expanded GDP data to measure military spending as a proportion of GDP. I then merge this military spending data with Taydas and Peksen's (2012) imputed welfare spending data to create a military to welfare spending ratio covering the period 1974 to 2000. I test the metric's correlation with World Value Survey (WVS) cross-national data assessing public confidence in the armed forces. Comparing figures from 1989 to 1999, there is a positive linear relationship between this militarism metric and confidence in the armed forces. 8 This demonstrates the metric's conceptual validity as an indicator of domestic public support for national defense mobilization. I classify militarized states as those above the median and nonmilitarized states as those below. 9 The highest militarization score in the dataset is that of Iraq in 1991 with a value of 10.35. In 1991, Saddam Hussein's government allocated 32 percent of Iraq's GDP to military spending and roughly 3 percent of GDP to welfare spending. Combining leader type and state setting, I create six leader-setting categories (see table 2).
If the hypotheses are accurate, these different leadersetting categories should correspond with different levels of conflict initiation. Figure 1 offers tentative support for my first hypothesis that the conflict propensity of military leaders is more affected by domestic militarism than the conflict propensity of civilian leaders. The Sparta-Athens distinction (militarized state setting versus nonmilitarized state setting) triggers greater variation in the conflict propensity of military leaders relative to civilians. Figure 2 shows that all three leader types (civilian, military-combat, military-no-combat) are more prone to provoke conflict when situated in Sparta versus Athens. It also confirms that leaders with prior military training but no combat experience in Sparta are most prone to provoke conflict.
These results offer tentative support for the hypotheses. Based on these preliminary findings, I proceed to test the hypotheses using multivariate regression analysis. I use 8 Pearson's r is 0.38. I use time series WVS Data from Wave 2 (1989)(1990)(1991)(1992)(1993) and Wave 4 (1999Wave 4 ( -2004 to assess change in national perceptions of the armed forces over a decade and compare the results against change in national levels of militarism for the same period (see the online appendix). 9 A robustness check (see the online appendix) confirms that increasing the threshold for militarism from the median to the upper quartile produces the same finding: militarism has the strongest effect on military-no-combat leaders. Clarify software (Tomz, Wittenberg, and King 2003) to estimate the probability of conflict initiation and include in the models a battery of control variables. My focus centers on relative increases to the respective probabilities of conflict initiation when different types of leaders are situated in nonmilitarized versus militarized state settings. If Hypothesis 1 is accurate, militarism should trigger a relative increase in the conflict propensity of military leaders that is greater than that for civilians. If Hypothesis 2 is accurate, militarism should trigger a relative increase in the conflict propensity of military-no-combat leaders that is greater than that for military-combat leaders.

Control Variables
To control for military strength, I use a state's score in the COW Composite Index of National Capability (CINC) for each leader-year (Singer, Bremer, and Stuckey 1972). States with numerous borders may be more conflict-prone because of the threat perception that geographic contiguity triggers. I use the COW Direct Contiguity dataset (v3.2) to measure the number of borders for each leader-year. I control for leader age  as well as regime type (Cheibub, Gandhi, and Vreeland 2010). To capture tension in the international system, I use the Tau-b coefficient in the analysis, which indicates a state's satisfaction with the leader of the international system (Signorino and Ritter 1999). This variable is provided as a control in the Why Leaders Fight dataset . I create a Cold War dummy variable given the effect of this period's ideological tension on interstate relations and use peace years to control for ways in which the outcome variable (MID initiation) may be dependent on time (Carter and Signorino 2010, 273;Schramm and Stark 2020, 532). I test the hypotheses through monadic analysis using the occurrence of an initiator-MID for each leader-year. I conduct standard logit regressions because the dependent variable is coded as 0 in the absence of a MID and 1 when a MID is initiated for the leader-year. Table 3 presents the results. Model 1 compares civilian and military leaders. The model shows that military leaders in militarized settings are most conflict-prone relative to the reference group: civilian leaders in Athens. Model 2 disaggregates military leaders based on those with combat experience versus those without. Military-trained leaders with no combat experience presiding over militarized states are most prone to conflict. Model 3 again compares civilian leaders against military leaders but restricts the population of cases to dictatorships (civilian, military, royal). Civilian leaders in Athens are the reference category for all three models.

Results
Because these coefficients can be difficult to interpret, I use Clarify software (Tomz, Wittenberg, and King 2003) to estimate the probability of MID initiation for each of the leader-setting categories. The results in figure 3 allow me to assess Hypothesis 1: militarism has a more pronounced effect on the conflict propensity of military leaders versus civilian leaders.
The percent increase in the probability of MID initiation rises 96 percent for civilian leaders when they preside in Sparta versus Athens. The percent increase in the probability of MID initiation rises 95 percent for military leaders when they preside in Sparta versus Athens. The relative  effect of militarism on civilian and military leaders is essentially equal, which leads me to disconfirm my first hypothesis. However, using Clarify to interpret the results of Model 3, it is worth noting that in dictatorships, militarism has a noticeably stronger effect on the conflict propensity of military leaders than it does on civilians. In dictatorships, the percent increase in the probability of MID initiation is 34 percent for civilian leaders when they are situated in Sparta versus Athens. The percent increase in the probability of MID initiation rises 136 percent for military leaders situated in Sparta versus Athens (see the online appendix). Regime type shapes militarism's effect on leaders. Under dictatorship, military-trained leaders are far more prone to conflict than civilians when the state militarizes.   Figure 4 produces results that allow me to assess Hypothesis 2: militarism has the strongest effect on the conflict propensity of military-no-combat leaders. The results in figure 4 lend support to Hypothesis 2. The percent increase in the conflict propensity of military-no-combat leaders rises by the largest margin when these leaders are situated in militarized settings relative to nonmilitarized settings. The percent increase in the probability of MID initiation rises 151 percent for these military-no-combat leaders situated in Sparta versus Athens. For military-combat leaders, the percent increase in the probability of MID initiation rises just 66 percent for Sparta relative to Athens. For civilian leaders, the percent increase in the probability of MID initiation rises 88 percent for Sparta relative to Athens.
Across the three leader types, domestic militarism has the strongest relative effect on the conflict propensity of military-no-combat leaders and the weakest effect on the conflict propensity of military leaders with combat experience. This may lend support to existing arguments of military conservatism, which posit that exposure to war induces future caution toward the use of force. These results also justify exploring the demonstration dynamic as a potential causal explanation for this variation.
Anecdotally, several international aggressors like Siad Barre, Muammar Qaddafi, and Saddam Hussein point to this link between military-no-combat leaders in militarized states and international conflict provocation. The results in table 4 offer additional evidence of the acute effect of militarism on the relative conflict propensity of military-nocombat leaders. Controlling for the same variables as before, the military-no-combat leader in Sparta category increases the odds of MID initiation by the highest margin.
Separate from the emphasis that the demonstration dynamic places on toughness, what are potential alternative explanations for militarism's effect on military-no-combat leaders? One potential explanation is that these leaders are most inclined to use force to signal resolve. Leaders strive to project themselves as resolute to other world leaders, particularly early in their tenures (Lupton 2020). One limitation in applying the resolve scholarship to the findings here is that it focuses on how leaders respond to targeting (Lupton 2018, 198), more so than why they become unprovoked aggressors. Initiating MIDs, leaders may not achieve the deterrence advantages highlighted in the resolve scholarship. Their provocations may invite rather than discourage future targeting from adversaries.
A second potential explanation focuses on regime type and ideology. Leaders are surrounded by policy elites with their own preferences about the use of force (McManus 2017, 28). Military-no-combat leaders may be prone to use force in militarized states because they preside over the types of national security-oriented military dictatorships linked to global conflict (Peceny, Beer, and Sanchez-Terry 2002, 19). To account for this possibility, I restrict the analysis to dictatorships. Notably, military-no-combat leaders have a higher rate of dispute initiation when they preside over civilian dictatorships rather than military dictatorships. Dictatorship increases the general likelihood that military-trained leaders in militarized states initiate conflict, but the military dictatorship subtype does not have a stronger effect than the civilian dictatorship subtype.
Finally, two concerns are worth discussing with respect to endogeneity. External security threats may cause states to militarize and select into office military leaders who are perceived domestically as best suited to manage defense during periods of state vulnerability. I account for this consideration in the statistical analysis by incorporating control variables to approximate interstate tension. Using Alliance Treaty Obligations and Provisions (ATOP) S-scores that measure foreign policy similarity between states (Chiba, Johnson, and Leeds 2015, 974), I find that civilian and military leaders confront roughly the same level of interstate tension. Civilian and military leaders also preside over militarized states at comparable frequencies (49 percent of civilian leaders preside over militarized states; 59 percent of military leaders preside over militarized states). This suggests that the international system does not disproportionately place certain types of leaders in state settings prone to international conflict. A second concern is that tension in the international system causes states to militarize in advance of looming international conflicts. This possibility implies that tension in the international system drives both domestic militarism and international conflict. To address this concern, I run a robustness check that restricts the dataset to leader-years with Tau-b scores above the median. In theory, this allows me to exclusively examine cases where system tension is low. Military-no-combat leaders in militarized states maintain the highest rate of MID initiation when system tension is low. Further, even when system tension is high, leaders are not guaranteed to initiate conflict at a higher rate. The following case study shows that amid regional tension, King Idris refrained from involving Libya in foreign conflicts.

Case Study: Libya's Transition from Monarchy to Military Rule
To illustrate the causal logic concerning the demonstration of toughness, I focus on Libya during the mid-twentieth century. I identify this case as illustrative because it presents an abrupt transition from a civilian leader (King Idris) to a military leader (Colonel Qaddafi). Analyzing the case, I focus on finding evidence that proves Qaddafi, unlike Idris, confronted domestic pressure to involve Libya in foreign conflicts because of the perceptions surrounding him as the military figurehead of a militarized state. A brief comparison of the two leaders (see table 5) reveals their divergent propensities to initiate MIDs. 10 Libya was less militarized during the tenure of King Idris. During Idris, the percentage of Libyans serving in the armed forces was half the level of that under Qaddafi. Libya also allocated a noticeably higher percentage of its GDP to the military under Qaddafi than Idris. However, the relative increase in military spending (as a percentage of GDP) suggests comparability between Idris's tenure and Qaddafi's. Under Idris, military spending as a percentage of GDP rose 85 percent. Under Qaddafi, this figure rose 95 percent. Also justifying the comparison, Libyan oil was discovered in 1959. Both leaders held power during transformative decades when oil wealth increased the country's per capita GDP. Thus, to properly understand the radical shift in Libyan foreign policy, it is important to examine how militarization affected Idris and Qaddafi differently. If the demonstration dynamic accurately explains the Libyan case, I anticipate evidence that shows Idris and Qaddafi faced unequal levels of public pressure to project toughness and that this variation fueled divergent foreign policies. 10 I create the militarism settings above (Sparta/Athens) using the number of military personnel in relation to the population because data for the military to welfare spending metric do not start until 1974.

King Idris
Despite reigning over a country categorized as militarized for nearly half his term in office, Libya's King Idris never provoked an interstate dispute. Idris was a paradigmatic dove. Secondary sources describe Idris as "shy" (Associated Press 1983) and "gentle" (Lewis 1970, 35). His authority was premised on his position as the successor of a religious movement, the Sanusi brotherhood, which united Libya's tribal society (Lewis 1970, 34). Libyans expected Idris to use his religious authority to resolve tribal disputes. He reportedly enjoyed this role (Golino 1970, 347). It was more important that Idris appear religious than strong. Idris fulfilled this expectation and was recognized publicly as "puritanical" (Associated Press 1983). Libyan society had little exposure to Idris who spent much of his time outside Tripoli in his home province of Cyrenaica surrounded by advisors (Lewis 1970, 35). Some Libyans preferred for Idris to stay above the political fray in accordance with his divine mandate (Anderson 1986, 68).
Although oil wealth gave Idris the resources to pursue a more assertive foreign policy, he was restrained on the international front. Idris used Libya's oil revenues to double the number of embassies abroad (Cecil 1965, 27) not to provoke foreign conflicts. Diplomatic cables reveal Idris's willingness to outsource Libyan defense to Western powers. 11 While Idris increased the size of Libya's federal army by almost 20 percent, he pursued military buildup to improve "national loyalty" 12 among a tribalized society rather than to prepare for conflict abroad.
During Idris's tenure, the rise of Arab nationalism threatened his domestic mandate. 13 Nasser's nationalist regime targeted Idris with transnational radio broadcasts from Cairo. In 1956, British military officials sent to Libya to protect Idris uncovered an Egyptian assassination plot to induce regime change. 14 Still, amid Egypt's hostility, Idris maintained a cautious foreign policy toward Cairo. During the Suez crisis, he rejected Britain's request to send Libyan troops to Egypt. The conditions existed for Idris to conduct a more assertive foreign policy. Instead, he pushed Western powers to guarantee Libyan security in the event of an Egyptian attack and maintained a noninterventionist international posture. 15 Propped up by oil revenues, Libya became a more militarized state during the second half of Idris's tenure. Why did Libya's militarization during this period not push Idris toward a more aggressive foreign policy? The statistical results predict that even as a civilian leader, Idris's foreign policy 11  should have become more conflict-prone during the 1960s given Libya's militarization during this decade. Yet Libyan foreign policy was restrained even amid civil wars in the region that afforded Idris opportunities to join surrounding monarchies in defense of embattled royal regimes. There are multiple reasons for Idris's foreign policy restraint. On a personal level, Idris's disposition was passive and riskaverse. At the state level, the logic behind the demonstration dynamic highlights how Idris's restraint stemmed partially from a domestic reputation that incentivized him to act as a religious mediator rather than a foreign policy hawk. Finally, taking a more forceful position against regional adversaries was at odds with Idris's political interests. The process of militarization during the second half of Idris's rule expanded Libya's officer corps. 16 Ideologically, these junior officers aligned more closely with the Arab nationalism espoused by Egypt's Free Officers than they did with Libyan royalism. Unlike his peers, Idris was not a regional defender of monarchy. Saudi Arabia's King Faisal, Jordan's King Hussein, and Iran's Shah Pahlavi mobilized military interventions in support of royalist governments in North Yemen and Oman. Idris likely lacked the loyalty from his military officers necessary to carry out such interventions and instead debated transforming Libya into a republic as early as 1956. 17
At home, Qaddafi engaged in a concerted effort to militarize Libyan society, much of it focused on the schooling system. Military science became the focus of the educational curriculum, children were mandated to wear military uniforms, and civilian teachers were replaced with military officers. Qaddafi created a people's militia, separate from the Libyan armed forces, which enlisted 40,000 recruits within its first year (Pargeter 2012, 111). In his Green Book, he emphasized the virtues of fighting: "Different types of boxing and wrestling are evidence that mankind has not got rid of all savage behavior" (Qaddafi and Christman 1988, 114). From the outset, this militarization was a tool of political legitimation as Qaddafi championed agriculture, industry, and arms. 18 At a celebration in Tripoli marking his first year in office, Qaddafi described the troops parading before the crowd as "stronger than steel" and prepared to confront an 16  international climate that adhered to "the law of the jungle." 19 Did Libyans expect Qaddafi to initiate conflict abroad because of the image that he cultivated at home? State repression and the absence of independent media make it difficult to assess Libyan public opinion during Qaddafi's tenure. Nonetheless, diplomatic cables and archival content suggest that Qaddafi confronted multiple political crises during his first decade in office. Repeatedly, he relied on public speeches grounded in alarmist rhetoric about external aggression to galvanize domestic support at critical junctures when he was vulnerable to coup and assassination attempts.
The fall of 1971 was particularly perilous. A botched assassination attempt killed several members of Qaddafi's motorcade in September that year. In the aftermath, Qaddafi briefly resigned and spent a short period under house arrest, according to Jordanian officials. 20 French diplomats suspected that he was receiving medical care for nervous breakdowns. 21 The British speculated that a change in leadership was on the horizon as Qaddafi appeared isolated from his top aide in the Revolutionary Command Council (RCC), Abdelssalam Jalloud. 22 After ten days away from the public, Qaddafi stepped before a crowd and claimed that popular protests in support of his authority required he maintained power. He reminded the crowd of his military credentials and success in expelling the Italians. 23 Two months later, Qaddafi nationalized the property of the British Petroleum Company. 24 These maneuvers enhanced Qaddafi's domestic profile and stabilized his rule within the RCC.
A similar dynamic unfolded in the spring of 1974. After the RCC briefly stripped Qaddafi of his duties, he returned to the spotlight three months later to hand out war medals to veterans of Libya's struggle against the Italians as onlookers chanted "the colonel, the colonel." 25 Qaddafi's embrace of international rivalry amid periods of internal instability corresponds with existing work on revolutionary foreign policy (Snyder 2005, 62). Throughout his tenure, Qaddafi tied militarism to anti-imperialism. His vision of an anti-imperialist state hinged on an armed society capable of defending itself. In public speeches, he pushed for private citizens to be armed at a faster pace to fend off a surprise aggressor. 26 As his tenure unfolded, Qaddafi increasingly used his annual address commemorating the September 1 revolution to invoke defense and foreign policy concerns over domestic matters. During his 1978 speech, he championed the nation's military support to revolutionary movements abroad, progress in training citizens in small arms, and the ability of Libyans to pilot advanced Soviet jets. 27 This policy helped stabilize Qaddafi's political position. Diplomats diagnosing Libyan internal affairs during the 1970s described potential challengers as "less visionary" and noted that Qaddafi's identification with the revolution made elite insubordination more difficult. 28 Qaddafi initiated conflicts abroad to signal a willingness to forcibly defend Libyan revolutionary goals. This logic offers insight into both Libya's wars with Egypt and Chad. Qaddafi's decision to send troops to Egypt was linked to his domestic politics. As Arabic writers point out, Qaddafi called for a mass march from Tripoli to Cairo (Badawi 2015;Soleiman and Mostafa 2015) just before Libyan tanks shelled the Egyptian border town of Sallum. It is possible Qaddafi's bluster reached a threshold at which point he anticipated that the domestic reputational cost of not using military force was more worrisome than the cost of sending Libyan units into Egypt. Libyan troops were repelled, but the episode triggered observable public support for Qaddafi. One day after the crisis, a popular Libyan student movement issued a statement expressing its willingness to fight on Qaddafi's behalf against Anwar Sadat. 29 Qaddafi's long-running intervention in Chad killed 7,500 Libyan soldiers (Brooke 1988). Domestic public opinion surrounding the war is hard to assess. However, during the conflict, one analyst described the intervention as "genuinely popular" among Libyan society (Joffe 1981, 95). Even a Libyan opposition newspaper conceded that the war stemmed from grievances predating Qaddafi's reign. 30 Before his domestic audience, Qaddafi stated his primary objective was to reclaim the Aouzou strip: territory in northern Chad claimed by Libya since independence. Idris was criticized for failing to seize an opportunity in the late 1960s to recoup the land (Joffe 1981, 91). Conversely, Qaddafi publicly pitched Libya's intervention as a principled effort to recover the strip (Joffe 1981, 95). Qaddafi also used the conflict with Chad to advance anti-imperialism, which was interwoven with how he projected strength. As an antiimperialist, Qaddafi asserted that Libyan troops would remain in Chad until French forces evacuated. 31 At home, this message likely resonated among some Libyans given lingering anti-French sentiment, as evidenced by a mob attack on the French embassy in Tripoli in 1980.

Conclusion
A range of individual traits and experiences affect leaders and their willingness to use force in office. Military experience is one factor that attracts interest with respect to how it shapes the beliefs of leaders toward the use of force. This article unpacks the public pressure to use force that leaders with prior military experience confront when domestic militarism is high. In militaristic states with populations more prone to see war preparation as the primary tool for national advancement, leaders face amplified pressure to demonstrate a willingness to use military force as a tool of foreign policy. The statistical analysis here shows that this effect is strongest on leaders with military training but no combat experience.
Building from core assumptions about how the logic of political survival affects foreign policy decision-making, the demonstration dynamic theorizes why militarism creates more pressure on a certain category of leaders (those with 28  military training but no combat experience) to use force in office. This pressure to project toughness through foreign policy is particularly strong in dictatorships. Focusing on Libya, the demonstration dynamic offers insight into the domestic circumstances that induced foreign policy restraint under Idris and foreign policy belligerence under Qaddafi. Qaddafi's foreign policy advanced his domestic status as a tough-minded, but reckless military strongman amid competition from regime elites. Under certain conditions, such as those in Qaddafi's Libya, the decision to initiate interstate conflict may prove politically advantageous even when it results in defeat. This breaks from conventional logic, which links unsuccessful foreign provocations to domestic political risk (Bueno de Mesquita et al. 1999, 794). The pressure to demonstrate toughness may offer insight into the behavior of other world leaders. For example, the American intelligence community's early psychological profile of Kim Jong-un highlighted the pressure the North Korean leader faced to signal toughness at home. 32 A separate contribution of this article is its development of a crossnational metric to measure militarism. Measuring militarism as a ratio that compares military investments to welfare investments captures the conceptual trade-off that militarism entails and has the potential to advance future quantitative research. 33 Finally, this article highlights how external factors shape individual leader traits and affect the likelihood of international conflict. Future research might explore how factors like authoritarianism or militarism shape the foreign policy behavior of leaders with other relevant traits. 34 The findings could have important implications on international affairs.

Supplementary Information
Supplementary information is available at the ISAGSQ data archive.