The military before the march: Civil-military grand bargains and the emergence of nonviolent resistance in autocracies

Abstract

This article contributes to growing efforts to explain when nonviolent resistance campaigns emerge in autocratic regimes. Building from a novel framework for distinguishing civil-military relations in autocracies, it contends that regimes in which military and political leaders engage in a ‘grand bargain’ generate opportunity structures that are especially amenable to nonviolent resistance. Militaries in these regimes exhibit distinctive characteristics – they are corporate, cohesive institutions as opposed to fragmented in structure and also wield political influence in regime institutions. Consequently, these militaries are especially inclined to care about their societal reputations and to retain their institutional independence from the regime’s political leaders. These factors together can lessen expectations among activists that the military will repress protests and increase the odds of elite splits in the face of mass movements. They also render the military more receptive to nonviolent protest tactics. We operationalize the concept of grand bargains with indicators from three datasets on civil-military relations and autocratic regimes. We then test the argument quantitatively using data on the onset of nonviolent resistance campaigns, as well as events-level data on nonviolent resistance campaigns. The findings support claims that civil-military grand bargains make nonviolent resistance in autocracies more likely, contributing to scholarship on this vital topic.

Keywords

autocratic regimes civil-military relations contentious politics protest

Scholars have identified a variety of factors that favor the emergence of nonviolent resistance movements, including gender equality (Schaftenaar, 2017), oil and gas rents (Pinckney, 2020), modernization of the economy (Butcher & Svensson, 2016), international diffusion (Gleditsch & Rivera, 2017), and conscription in the armed forces (Cebul & Grewal, 2022). Still relatively unexplored are the effects of relations between the military and political leadership in authoritarian regimes. Given the central role that militaries can play both in the ruling coalitions that sustain autocracy and in repression, we expect that civil-military relations should play an important role in shaping the environment in which resistance movements emerge. Are there particular conditions under which civil-military relations in autocracy render these movements more or less likely?

In this article, we help answer this question, arguing that regimes in which political and military leaders engage in what we term a ‘grand bargain’ – in which the political leader accommodates a politically powerful, corporate military by bargaining with its senior officers – are more likely to experience nonviolent resistance movements. Militaries in these regimes exhibit particular characteristics that render the opportunity structure more amenable to nonviolent resistance relative to other autocracies.

Our argument develops in two steps. First, we highlight two features of militaries in regimes in which political and military leaders are engaged in grand bargains: they are corporate in institutional structure and maintain political roles and influence within key regime institutions. Second, we argue that these features of the military affect two aspects of the opportunity structure for nonviolent resistance: (1) the certainty with which dissidents will face military repression, and (2) whether dissidents will be able to exploit divisions within regime elites in order to win concessions. Together, these factors explain why grand bargain civil-military relations are especially propitious for nonviolent resistance.

Empirical tests provide significant support for this theory. We operationalize grand bargains using three datasets and find that regimes with militaries that are politically influential and corporate in their institutional character are more likely to experience nonviolent resistance. We also find that a military possessing only one aspect of the prerequisite characteristics is not enough to make nonviolent resistance more likely in a non-democracy; rather, both properties contribute to the onset of nonviolent resistance.

In supplemental analysis, we evaluate an additional observable implication of our theory that nonviolent tactics should be more effective against militaries accommodated with grand bargains. Using events data, we compare the interactions between two militaries in 2011 – one that is accommodated with a grand bargain (Egypt) and one that is not (Libya). This analysis reveals more positive and cooperative interactions between the military and protesters in Egypt than in Libya, further supporting our theory.

This article contributes to scholarly research first by exploring how the character of states’ security forces affects regime security. Existing research links variation in the structure of the security sector to the indiscriminateness of state repression, civil war, and coups (Roessler, 2011; Greitens, 2016; De Bruin, 2020). This study explores a different linkage, analyzing how the character of an autocratic military affects the potential that the regime experiences mass nonviolent resistance. The article also helps further bridge the study of nonviolence and social movement theory. Seminal scholarship by Stephan & Chenoweth (2008) shows that security forces’ actions during protests are vital to the efficacy of nonviolent mobilization. Alternatively, scholars of social movements demonstrate that opportunity structures can render a situation more or less ripe for contentious politics (De Bruin, 2018). We link these insights, showing that civil-military relations are a critical feature of an autocracy’s political opportunity structure.

This article first defines the ‘grand bargain’ and discusses its implications for nonviolent resistance. The discussion of our research design and empirical analysis follows, as well as events analyses of grand bargains and repression in Libya and Egypt. A brief discussion of our findings and their implications concludes.

Grand bargains in autocratic civil-military relations

A grand bargain describes the relationship between a political leader and a military that is politically influential in the regime and exhibits institutional corporateness. Corporate militaries are those with a conventional hierarchy with leadership accountable within a pyramidal chain of command (Campbell, 2009: 53–55) and often an organizational ethos in which they exhibit a ‘sense of organic unity and consciousness’ such that officers’ identification with the military organization supersede other identities (Huntington, 1957: 10). This contrasts with militaries in which command is more decentralized and siloed into parallel command structures, often answering directly to a political leader (De Bruin, 2018).

We assume that political leaders seek to retain their positions, while military leaders seek to increase their organizational autonomy and resources (Posen, 1986). Bargaining with the heads of such a military is a way in which political leaders can secure themselves in office (Svolik, 2012). Following our assumptions, military leaders negotiate control over their internal organizational structure and hierarchy, while resisting autocrats’ efforts to subdivide the military into autonomous units that report directly to them (Pion-Berlin, 2010). Political leaders consequently cede these prerogatives, control of resources, and policy domains to senior officers, thereby reinforcing the military’s corporate structure; these officers in turn commit to maintain the leader in office and to enforce the deal within their ranks (Singh, 2014). A military accommodated through a grand bargain also controls key veto points in the regime, which enable it to protect its autonomy, monitor the regime, and control patronage flows, while enhancing the credibility of the autocrat’s commitment to sustain the agreement (Brooks, 2008; Talmadge, 2015).

In this article we do not theorize why militaries first assume a corporate character and have political influence and therefore why grand bargains initially emerge – we focus on their implications for nonviolent movements. Nonetheless, we expect that grand bargains originate from multiple causes unrelated to the emergence of nonviolent resistance (and therefore are not endogenous to them).¹ To be sure, nonviolent resistance and the demand for repression could be a factor that increases the military’s political influence and therefore could help explain why a grand bargain emerges (Svolik, 2012). But factors such as the military’s role in regime formation or prominence gained through warfighting also likely affect its influence. For example, a grand bargain in Egypt emerged from the military’s role in bringing the regime to power in the July 1952 coup, while in Algeria it resulted from the military’s role in the country’s independence movement. Grand bargains may also emerge after periods of military rule in which the military negotiates a substantial role in the succeeding civilian-led regime – examples have included, at various times, Thailand, Brazil, and Pakistan. More broadly, existing scholarship emphasizes the overriding importance of avoiding coups in autocrats’ decisions about how to organize the military and security forces; this suggests that coup-proofing, rather than avoiding mass nonviolent resistance, is the key driver of the armed forces structure in a regime (Greitens, 2016; Roessler, 2016; De Bruin, 2018).

Critically, we anticipate that grand bargains exist across different autocratic regime types.² Authoritarian civil-military relations vary significantly across regimes (Albrecht & Ohl, 2016; Greitens 2016; Bou Nassif, 2021) and within them over time (Brooks, 2008; Talmadge, 2015). In personalist regimes, for example, political leaders may sideline potential political rivals, and avoid significant power-sharing with civilian elites, but still need to maintain some form of accommodation with the military, in which the leader cedes institutional prerogatives. This was, for example, the strategy pursued by Alberto Fujimori in Peru, who agreed to cede various policy domains to the military and incorporated it into his cabinet in order to shore up military support for his regime in the absence of a civilian basis of support (Obando, 1996; Gales, 2000).

Similarly, in military-personalist hybrids, dictators may find it necessary to accommodate the military – one that was powerful enough to bring them to power – by incorporating military leaders into the regime elite. This pattern has been evident among military-personalist leaders such as Idi Amin in Uganda, Pervez Musharraf in Pakistan, and Hugo Banzer in Bolivia. Alternatively, in dominant party regimes in which political leaders engage in a grand bargain, the power of the party is preserved by incorporating the corporate military into its political institutions. This was the case in Zambia under Kenneth Kaunda, who created a single-party system in 1973. Kaunda shored up his civilian base by appointing military officers to cabinet positions (Lindemann, 2011: 26). In short, as these examples highlight and as our empirical analysis underscores, a grand bargain with a military that is politically influential and corporate in character can occur in different autocratic regime types.

The opportunity structure for nonviolent resistance campaigns

We build our argument around the concept of opportunity structure in social movements research, which refers to factors that shape an environment for mobilization and protest. As Tarrow & Tilly (2009: 440) define it, an opportunity structure refers to ‘features of regimes and institutions that facilitate or inhibit a political actor’s collective action’. Our argument assumes that dissidents are sensitive to the structural conditions in their political environments.³ Whether or not movement leaders explicitly frame their decisions in the terminology of our theory, we expect that in authoritarian regimes in which coercion plays a key role, many citizens are likely to be broadly aware of the military’s position within the regime and general organizational character. Specifically, we contend that the grand bargain affects the opportunity structure for protest in three ways.

The (un)certainty of repression

First, the military’s corporateness influences the relative certainty with which protesters will face repression.⁴ The shadow of repression is a central factor shaping the opportunity structure of nonviolent resistance (Chenoweth & Stephan, 2011).

The regular military’s role in administering repression plays a critical role here – if a less immediate one than the role played by police (Nepstad, 2013). In most regimes, police are the first to respond to domestic disturbances, and protesters may be deterred when they anticipate a violent police response. While police matter for dissident action, the anticipated military reaction also constrains nonviolent resistance (Cebul & Grewal, 2022), especially as activists hope to build large movements that might exceed the police’s repression capacity.

The military’s corporate character is one reason why we think activists might anticipate some reticence on the part of the military to repress protests. Corporate militaries experience greater costs from mass repression. Corporate militaries prioritize cohesion (Thompson, 1973) and ordering repression can divide the military by causing junior officers and rank and file to defect (Albrecht & Ohl, 2016; Bou Nassif, 2020). In principle the military’s corporateness might allow it to enforce discipline and ensure repression, as in Burma in 2007 (Lee, 2014), and militaries still repress citizen movements, especially when they can effectively portray them as not representative of society and as part of a marginal community, as did Egypt’s military against the Muslim Brotherhood in 2013. Nonetheless, in many cases senior officers of corporate militaries are reticent to risk dividing the military by ordering soldiers to repress their fellow citizens. Concerns about cohesion, for example, were a primary factor driving the Algerian military’s reluctance to repress protests in 2019 and the Egyptian military’s decisionmaking during the 2011 protests (Bou Nassif, 2020; Ghanem, 2019).

These militaries may also emulate global norms of military professionalism, viewing regime security functions as a distraction from their external security function (Hunter, 1998). Consequently, as in cases such as Egypt, the military may negotiate its mandate to keep itself out of the daily administration of repression. Because these militaries in turn are not identified by the population as purveyors of repression, they may retain their popularity, rendering them in turn reluctant to sacrifice that esteem by repressing protesters (Bellin, 2012).

Exerting political pressure

Second, grand bargains affect the opposition’s ability to divide the military from the political leadership and get the former to pressure the latter. A key feature of the conduciveness of a regime to social movements is ‘the stability or instability of elite alignments within the regime’ (McAdam, 1996). Social movements are empowered by divisions within the elite (Tilly, 1978).

In addition, their corporateness and autonomy means they can separate from the political leadership without threats to their organizational integrity (Bellin, 2012; Stacher, 2012). The military can pressure the political leadership to make concessions to the protesters, or to step down, which increases the chances a protest might succeed, creating greater incentives to carry it out. Nonviolent campaigns are 46 times more likely to be effective if security force defections of this kind occur (Stephan & Chenoweth, 2008: 22–23).

At the same time, because the military is so deeply in control of levers of power in the regime’s institutions, it has less to fear from allowing protests to proceed without repressing them. Its officers know that they are in a good position to protect the military’s interests even if a political leader is ousted. In fact, militaries accommodated with a grand bargain may see some benefit in allowing protests to proceed, because it may strengthen the military’s hand in the bargain with the leader.

The importance of exploiting divisions between the military and political leadership was evident in Algeria in 2019, where the military was both corporate and politically influential (Serrano, 2019). As one protester put it, referring to the country’s president, ‘Boutef was a facade president’ and the military operated autonomously from him.⁵ Similarly, in Egypt in 2011, one of the top activists, Ahmed Maher, told journalists that the protest strategy was built around appealing to the military because, ‘[n]umber one, we consider the army to be different from the Mubarak regime’ (All Things Considered, 2011). Three days after the protests started, many Egyptians were speculating whether the ‘well-financed but politicized military would stay loyal to [Mubarak]’ (Kirkpatrick, 2011). Some protesters were unsure that the military would abandon Mubarak, but important for our argument is that they recognized the distinction between the military and the regime, speculating that it [the military] could ‘take over and potentially oust Mr. Mubarak […] unlike the relatively small and apolitical army in Tunisia’.⁶ It is striking that at least some leading Egyptian activists explicitly linked the character of the military to their incentives to protest. Indeed, the popular slogans of ‘The people and the army are one hand’ can be seen as evidence that the Egyptian protest organizers were aware of the critical role that the military would play in the success of their movement. Similarly, the popular understanding of the institutionally corporate, but politically marginalized, Tunisian army as La Grande Muette (the Big Silent One), indicated a knowledge among dissidents of the political marginalization of the military in that country (Sayigh, 2011; Brooks, 2013). Critically, our theory does not anticipate that every protester will have an in-depth understanding of the military, or that activists will explicitly unpack civil-military relations in the terms of our theory. Yet, following the scholarship that emphasizes the importance of structural factors on social movement behavior (McAdams, 1996; Tilly, 1978), we anticipate that activists often will perceive the military’s potential actions as important to the protesting environment.

Efficacy of tactics of nonviolence

Third, while nonviolent tactics have been shown to be extremely successful in general (Stephan & Chenoweth, 2008), militaries accommodated with a grand bargain may be especially receptive to them. Nonviolent tactics, such as fraternization and symbolic acts (e.g. flowers being given to soldiers), test the loyalties of rank and file and junior officers, threatening cohesion if they are ordered to fire on the country’s citizens – costs to which corporate militaries are especially sensitive (Bellin, 2012). Appeals to the military’s social prestige – as in ‘The Army, The People, One Hand’ refrain – should be more effective against militaries that care about their reputation.⁷ In sum, we expect that in the aggregate, the opportunity structure is more amenable to nonviolent resistance in regimes with civil-military relations characterized by grand bargains. This leads to our core hypothesis:

H1: Autocracies with civil-military grand bargains are more likely to see the onset of nonviolent resistance campaigns relative to autocracies without these civil-military relations.

The importance of both corporateness and influence in regime institutions

In order to demonstrate that both a military’s corporateness and its influence in regime institutions are important determinants of nonviolent resistance campaigns, we also examine how nonviolent campaign onset is influenced by militaries that have one, but not both properties of a grand bargain.

Corporate militaries can be found in a variety of regimes (Campbell, 2009: 14). Not all militaries that exhibit a corporate structure exercise significant political influence in regime institutions. Corporateness can emerge when the military is marginalized in a regime, as occurs when a leader isolates it from regime politics (Brooks, 2013; Bou Nassif, 2020). Tunisia under Ben Ali and Morocco under Mohammed VI both exhibit corporate militaries, but neither was a pivotal player in intraregime decisionmaking (Willis, 2012).

In turn, where militaries are corporate in nature but lack political influence, the efficacy of protests is less clear for dissidents. While corporateness still renders repression costly to the military as under a grand bargain, the military’s role in the inner circle of decisionmaking and its political influence is more ambiguous, or absent (Cook, 2007). Whether and how the military might be willing and able to engage in insubordination and exert political pressure to force change is more uncertain, as is its propensity to act as an ally. Hence, in the case of a corporate military without political influence, any positive effect on the opportunity structure generated by a corporate military’s disincentives to repress protesters is offset by uncertainty over whether it can and will resist a political leader’s orders to do so and exert pressure on the regime to elicit concessions. See Table I for a summary.

H2: The positive effect of military corporate structure on the onset of nonviolent resistance campaigns is conditional on the military having a political role in regime institutions.

Yet another class of states may lack the second property associated with militaries accommodated with grand bargains: they may be politically influential but lack a corporate structure. In this case, military authority is decentralized into competing units that operate in parallel command structures and often directly report to the political leadership (De Bruin, 2018). These militaries lack the identifying characteristics of corporate militaries, including a traditional hierarchy, an overarching identification with the institution as a whole, and high levels of institutionalization (Campbell, 2009: 35). For example, in Syria under Bashar al-Assad the military has had several independent commands that technically answer to the army’s Chief of Staff, but in practice operate through parallel command structures and report to the president (Al Hendi, 2011).

This arrangement is less amenable to protest for three reasons. First, whether protesters will experience

Table I.

Summary of expectations

		Military political influence in regime institutions
		Significant	Marginal
Military corporateness	Corporate	Increased likelihood of nonviolent campaign onset	No effect
Military corporateness	Not corporate	No effect	Decreased likelihood of nonviolent campaign onset

repression by military units is hard to discern when the military is so fractionalized. This uncertainty mitigates incentives to protest. Heads of units in divided security sectors may not prioritize the institutional integrity of the military as a whole, as in a corporate military. Hence, they may not worry if some soldiers and junior officers defect when ordered to fire. In addition, these chiefs face potentially greater risks if they do not defend the regime, given the co-mingling and ties between the political leadership and unit chiefs that often occurs in fractionalized security sectors (Stacher, 2012). This dynamic was apparent also in Syria during the 2011 uprisings. Direct ties to Bashar al-Assad and sectarian alliances among Alawi in elite units like the 4th Armored Division, Republican Guard, and Special Forces overrode concerns about the institutional integrity of the army, which was largely comprised of Sunni conscripts (Al Ayed, 2017; and also Koren, 2013).

Second, even if some military chiefs refrain from repression, they may have less control over chiefs’ actions in these decentralized security sectors. Research has shown that under such conditions, an inconsistent repressive response by the armed forces is more likely (Cunningham & Beaulieu, 2010: 173–195; Greitens, 2016).

Finally, the potential benefits of protest are less certain when the military is not corporate. Non-corporate militaries often have more personalistic ties between the political and military leadership (Campbell, 2009), making it more risky to turn against the regime without compromising its entire structure (Stacher, 2012).⁸ For example, in early March 2011, one expert anticipated Syria would not experience large nonviolent resistance campaigns such as those that occurred elsewhere in the Arab world because in those places the ‘state, regime and government did not overlap as much as those of Syria do’.⁹ Of course, large protests did occur in Syria, but this was not something either the protesters or outside analysts anticipated prior to the torture of several teenagers in Dara’a in March 2011.¹⁰ In other words, protests occurred despite Syrians’ expectations that protests were unlikely, and not through the instigation of activists, weeks after they initially occurred in other parts of the Arab world. This leads to our final hypothesis:

H3: The positive effect of a military political role on the onset of nonviolent resistance campaigns is conditional on the military being corporate in structure.

Table I summarizes our expectations.

Research design and data

The empirical analysis centers on several attributes of the military that capture the core features of a grand bargain and combines these into a single indicator. We rely on three datasets: from Geddes, Wright & Frantz (2018, hereafter ‘GWF’), De Bruin (2018), and White (2017). Drawing on multiple datasets allows us to capture the different ways in which military corporateness and political participation manifest.

We measure political influence first by analyzing the military’s role in national cabinets, especially in holding portfolios that are not security related. Militaries accommodated with grand bargains are likely to hold these positions because they render the bargain with the political leader more credible (Boix & Svolik, 2013). White (2017) has collected data on the number and type of positions held by military officers in national cabinets for all countries from 1964 to 2008. The Military Participation in Government (MPG) dataset captures how many military officers served in a national cabinet (or equivalent) and whether they occupied a security or a non-security role.

Table II.

Properties of civil-military grand bargains

	Grand bargain
Non-security MPG (White)	Yes
Military in party executive committee (GWF)	Yes
Military consultation (GWF)	Yes
Party interference in military (GWF)	No
Joint party-military influence (GWF)	No
Executive-reporting armed orgs (De Bruin)	No
Executive-affiliated armed orgs (De Bruin)	No

An additional indicator of political influence is the likelihood that the military will participate in informal consultative bodies with the political leadership, such as advisory councils. For these aspects of military regime participation, we draw from the GWF dataset on Autocratic Regime Features, which contains indicators for military participation in political party executive committees and other modes of informal consultation with the military.

We use several indicators to measure corporateness. First, we focus on the absence of intrusion and politicization by a political party. While a corporate military may coexist with an institutionalized civilian party, we expect that a politically powerful corporate military will not tolerate intrusive party oversight or a reciprocal relationship where there is significant party influence on the military. To measure this, we use a variable from the GWF data that captures the nature of the relationship between a dominant political party and the military in an autocracy. Second, to measure corporateness, we introduce indicators related to the absence of fractionalization in the security sector. We also expect that a cohesive, corporate military should resist efforts to create militarized, independent security forces directly answerable to the political leadership that would infringe on the military’s autonomy. Hence, with the absence of competing units in a decentralized security sector, we can infer that the military has a corporate structure. De Bruin’s (2018) dataset provides a count of the number of security forces (fractionalization), and importantly accounts for their relationship with both the military – that is, is the security force a counterbalance to the regular military – and the regime – that is, does the force have ties to the regime that circumvent the regular military chain-of-command?

Table II summarizes the decision rules that indicate a grand bargain. A ‘Yes’ indicates that civil-military relations must exhibit at least one, if not more, of these features to qualify as a grand bargain, while the items bolded as ‘No’ must be absent in a grand bargain. That

Table III.

Distribution of nonviolent resistance campaign onsets by strategy type

	No grand bargain	Grand bargain	Totals
No campaign onset	1,664 (98.00%)	605 (97.11%)	2,269 (97.76%)
Campaign onset	34 (2.00%)	18 (2.89%)	52 (2.24%)
Totals	1,698	623	2,321

is, if any one of these features is present, we do not classify it as a grand bargain.

With these decision rules, we generated an indicator for the presence of a grand bargain using the GWF regime-year data as a basis. We coded our indicators for a subset of non-democratic regimes from 1964 to 2008 dictated by data availability. As the MPG data are only available for 1964–2008, this dictated the temporal limitation. De Bruin’s data cover a random sample of 111 countries. Accordingly, our indicator is coded for the same random sample of non-democracies. If a country became a democracy or for some other reason exits the component data, it was not coded in that particular year. In the 2,321 regime-years in our sample, grand bargain occurs in 605 – or 26.07%. The remaining 73.93% comprise regimes where the military lacks the corporateness or political influence indicative of militaries accommodated with grand bargains – that is, it has only one dimension of a grand bargain – or lacks both dimensions.

Our dependent variable is the onset of a nonviolent resistance campaign, as indicated by the NAVCO 2.0 data. In line with NAVCO’s inclusion criteria, we examine the beginning of a nonviolent campaign with at least 1,000 participants and some underlying organization. We do not examine ongoing nonviolent campaigns, focusing rather only on the first year in which they emerge. Table III shows the bivariate relationship between grand bargains and nonviolent resistance campaign onset. Campaign onset is a rare event – in our sample, slightly more than 2% of regime-years see the onset of a nonviolent campaign. In terms of the independent variable, we see that approximately 2% of regime-years without a grand bargain see a nonviolent resistance onset, while nearly 3% do for those with a grand bargain. In this basic examination, nonviolent resistance campaign onset is nearly 50% more likely under grand bargain than a different civil-military arrangement.

We control for factors that are likely to be correlated with both civil-military relations and the emergence of nonviolent resistance campaigns, including armed rebellion – whether or not the state is facing an internal armed conflict according to the UCDP/PRIO Armed Conflict Dataset (Gleditsch et al., 2002; Themnér & Wallensteen, 2014)¹¹ – as well as the number of coup attempts that the state has experienced since 1950 from Powell & Thyne’s (2011) data. We also differentiate regimes with a grand bargain from overall regime-type and include controls for different non-democratic regimes using the GWF data: military regimes, dominant party systems, and personalist regimes. We also address the relative openness of the non-democratic regime by including the country’s Electoral Democracy (Polyarchy index) from the V-DEM project (Lindberg et al, 2014), and control for the duration of the regime (from GWF) – given that how old a regime is may affect both its civil-military relations and the likelihood of a nonviolent resistance campaign. Also included are controls for both overall economic health (GDP per capita) and the change (change in GDP per capita) – from Gleditsch (2002) – as well as an indicator for the Cold War, coded as 1 for years prior to 1991.

We used a logistic regression, where the dependent variable is the onset of a nonviolent campaign. As such, country-years where there is an ongoing nonviolent campaign – that is, subsequent years in a multiyear campaign – are excluded from the analyses. To address temporal dependence on the dependent variable, we include also a cubic polynomial of the number of years since the last nonviolent campaign onset within the country (Carter & Signorino, 2010). Robust standard errors were calculated by clustering on country. All right-hand side variables – except for the polynomial of the number of years since nonviolent campaign onset – were lagged by one year to address potential reverse causality. Here we exclude also military juntas from the analysis of regimes, as juntas are likely to resemble grand bargains in our component datasets¹² yet involve the military as an institution controlling the executive rather than bargaining with it. However, in robustness checks, we do add juntas to the analyses, and it does not change the results substantively.

Table IV.

Civil-military grand bargains and the onset of a nonviolent resistance campaign

	Basic	Controls	Full controls
Grand bargain	0.655*	1.050**	0.886**
	(0.280)	(0.300)	(0.283)
Military regime			1.968†
			(1.124)
Single party			0.899
			(0.989)
Personalist			1.906†
			(1.101)
Polyarchy index		–0.386	–0.356
		(1.212)	(1.166)
Civil war		–0.447	–0.483
		(0.459)	(0.467)
No. of coups		0.042	0.005
		(0.066)	(0.064)
GDPPC (log)		0.146	0.218
		(0.157)	(0.162)
Δ GDPPC (log)		0.097	0.186
		(0.699)	(0.836)
Cold War		0.499	0.479
		(0.668)	(0.672)
Regime age		0.023**	0.036**
		(0.008)	(0.010)
Years since last campaign onset	–0.332**	–0.356**	–0.332**
Years since last campaign onset	(0.112)	(0.104)	(0.111)
Constant	–1.777*	–3.745**	–5.838**
	(0.743)	(1.368)	(2.042)
Observations	1,947	1,889	1,886

Robust standard errors in parentheses. **p < 0.01, *p < 0.05, †p < 0.1. Polynomial of years since nonviolent campaign onset not shown.

Table IV shows the results of logistic regression analysis of the relationship between grand bargains and nonviolent resistance campaign onset.¹³ We present a model first with no control variables, then with all control variables excluding autocratic regime type, and in the final specification we add indicators for regime type. To understand substantive effects, we calculated predicted probabilities, setting all other covariates at their observed values in line with Hanmer & Kalkan (2013). Focusing on the model with all control variables, on average and holding other factors constant, the probability of nonviolent resistance campaign emergence in a regime that exhibits a grand bargain is 3.97%, whereas the

Figure 1.

Conditional effect of grand bargain components

probability for those that do not is 1.73%.¹⁴ Autocracies with grand bargains are more than twice as likely to see the emergence of nonviolent resistance as those with different civil-military relations.

To test H2 and H3 we separated grand bargain into its two dimensions – corporateness and political influence. We then duplicated our main analyses, but rather than including a single grand bargain indicator as an independent variable, we included indicators for whether or not a regime featured a corporate military and whether or not the military was a regime participant. These were coded in line with the decision rules outlined in Table II. The intersection of these two dimensions – that is, when both the corporate and regime participation indicators were coded as 1 – indicates the presence of a grand bargain. This was modeled in the analyses by interacting the two indicators.

Figure 1 shows the predicted effect of each grand bargain component by themselves and interacted with each other generated from the coefficients of a logit model (Table A1 in Online appendix). Here we see that when each component occurs without the other – for example, a politically active military is embedded in a fractionalized security sector or has a high level of party intrusion into the military (i.e. is not corporate) – the effect on campaign onset is greatly reduced and is not significant for either component. In contrast, when each component is interacted with the other, the effect of each component is much greater, and is statistically significant in the case of political participation and marginally significant in the case of corporateness (p = .058).¹⁵

These analyses provide strong support for H3 and more modest support for H2 – the two dimensions that make up a grand bargain in autocratic civil-military relations do not have an effect on the onset of a nonviolent resistance campaign by themselves – but when the two dimensions combine, they exert a stronger effect. However, only in the case of political participation does this joint effect reach statistical significance. The joint effect of the corporateness component is significant only at the .10 level.

A further illustration of the effect of the components of grand bargain – alone and combined – can be seen in Figure 2, which shows the predicted probability of a nonviolent resistance campaign onset when a regime’s military has: (a) neither corporateness nor regime participation, (b) corporateness but not participation, (c)

Figure 2.

Predicted probability of nonviolent resistance campaign onset

participation but not corporateness, and (d) corporateness and participation together (grand bargain).

Reflecting the expectations in Table I, in Figure 2 we see that the predicted probability of campaign onset with a grand bargain is four times that when the autocratic military is neither a participant in government nor characterized by a high degree of corporateness (.010 versus .043). Critically, these predictions are statistically distinguishable from each other. In contrast, in line with Table I, both predictions for when a regime’s military has one, but not the other, component of a grand bargain are statistically indistinguishable from the complete absence of any component of grand bargain.

Grand bargain and tactics of nonviolence

We also investigate another observable implication of our theory: that nonviolent tactics should be more effective under grand bargains. We examine the interaction between protesters and the military using the NAVCO 3.0 data (Chenoweth, Pinckney & Lewis, 2018). The NAVCO 3.0 data are at the events level, allowing researchers to observe the mechanisms of campaign onset and evolution over time. The data are available for 21 countries across Africa, Asia, Europe, Latin America, and the Middle East.¹⁶

The NAVCO 3.0 data are coded using the CAMEO structure (Schrodt, 2012). The CAMEO framework lists for each event a location, the main actor in the event, a ‘verb’ for the type of action by the actor, and the ‘target’. We anticipate that countries characterized by a grand bargain should see a lower proportion of hostile military–opposition interactions and a greater proportion of cooperative or conciliatory events. We coded two indicators for each opposition–military interaction, differentiating between (1) hostile and (2) cooperative or conciliatory interactions. We considered, for example, military repression to be hostile. We considered an event indicative of military repression if the event had the actor ‘military’; the target as either ‘opposition’ or ‘activists’;¹⁷ and the verb as ‘repress and abuse’, ‘engage in violent combat’, ‘use unconventional mass violence’, or threaten some other type of violence or repression. We similarly considered these types of actions initiated by the ‘opposition’ or ‘activists’ and specifically directed at the military to be indicative of a hostile military–opposition interaction.

Cooperative or conciliatory military actions towards the opposition were coded if the actor was the military and the verb was ‘defect’ or taking a position as ‘neutral arbiter’ (NAVCO 3.0 codebook). We also included additional cooperative interactions, such as the military ‘yielding’ to protesters, ‘providing aid’, ‘engaging in material cooperation’, ‘engaging in political cooperation’, ‘consultation’, or ‘express intent to cooperate’ –

Figure 3.

Military–opposition interactions during the 2011 Arab Spring

as well as instances where the military used violence against or expressed dissent with other aspects of the state, such as the government, police, or intelligence services. Opposition–initiated interactions directed at the military were similarly coded.

As an illustration, Figure 3 shows the sequence of military–opposition interactions in two countries during the Arab Spring – Egypt and Libya. Note that while Egypt is a grand bargain, Libya has a non-corporate military.¹⁸ The figure shows the occurrence of these interactions across 18 days. In the case of Egypt, this was the time from the onset of protest on 25 January 2011 to Hosni Mubarak’s resignation on 11 February. For ease of comparison, the same length of time is shown for Libya – from 18 February to 7 March 2011. Blue symbols indicate a conciliatory/cooperative military-opposition interaction on that day while red symbols indicate that a hostile event took place.¹⁹

While both countries saw protest during this period, the interactions are exclusively cooperative for Egypt – which affirms accounts of fraternization between protesters and the military and explicit attempts by the protesters to align themselves with the military. In contrast, military-opposition interactions in Libya see a mixture of hostile and cooperative actions on the part of both the opposition and the military. This is likely a reflection of the fragmentation of the Libyan military at the outset of the uprising, which itself stemmed from the historical fractionalization of the military by the regime (Bou Nassif, 2015; Gaub, 2019). This does not lead to exclusively cooperative or conciliatory actions, but high variability in opposition–military interactions as the fragmentation of the military leads to a greater diversity of responses to protesters and increased uncertainty regarding how different units of the military will act. As we argue above, we would expect fraternization and other nonviolent tactics to be far more effective against militaries with a strong corporate ethos than one divided and whose commanders had strong personalistic ties to the political leader as in Libya. To more directly examine how military–opposition interactions vary between autocracies with grand bargains and those without, we merged the NAVCO 3.0 events data with our regime-year data. The sample of events was limited to those that were (a) military–opposition/activist interactions, (b) covered by the range of our data on autocracies and grand bargains, and (c) included in the 21 countries in NAVCO 3.0 – yielding a total of 60 events in 11 countries between 1990 and 2008.²⁰ Despite these limitations, the data display patterns suggestive of support for our expectations. More than 21% of military–opposition interactions under grand bargain could be characterized as conciliatory or cooperative (in contrast to ∼4% for interactions not under grand bargain). Also, approximately 5% of military–opposition interactions under grand bargain were hostile, whereas more than 13% were hostile in its absence.

These results are intended to be suggestive – the small sample size and selection criteria mean that any inferences must be drawn with extreme caution and that the differences observed are not statistically significant.²¹ However, taken with the strong evidence in our regime-year analyses, we see the patterns evident in the NAVCO 3.0 data as providing additional support for our theory. Not only is nonviolent resistance campaign onset more likely under grand bargain, at the micro-level, opposition–military interactions tend to be more cooperative and conciliatory.

Additional analyses and robustness checks

In additional analyses, we examine grand bargain’s impact on nonviolent campaign onset using more restrictive versions of the dependent variable, according to campaign goal (see Online appendix). In order to ensure that the findings for grand bargain were not merely a function of other factors related to military size or importance, we reran the main analyses with controls for military size and expenditures, as well as an indicator for whether or not the military recruited through conscription (Asal, Conrad & Toronto, 2017) – which could affect the propensity of the military to repress (Posen, 1993: 88). We also added a control for the population of the country in order to capture the influence that a larger population may have on the opportunity structure for protest. In none of these analyses is the effect of grand bargain altered substantively.

Recent research on the onset of nonviolent resistance campaigns has also found that gender equality (Schaftenaar, 2017), oil and gas rents (Pinckney, 2020), the modernization of the economy (Butcher & Svensson, 2016), and the potential for international diffusion of nonviolent resistance campaigns are also potential drivers of the onset of nonviolent resistance (Gleditsch & Rivera, 2017). We added controls from these studies as well and found no substantive change to our main results.

It is possible that a third factor, such as a history of lower-level unrest, may influence regimes to both adopt grand bargains and make nonviolent resistance more likely. Accordingly, we draw on the Governmental Incompatibilities Data (Cunningham et al., 2017), which captures the presence of organized and articulated maximalist claims against the government in a given country – reflecting a basic level of mobilization (i.e. the existence of organized dissidents articulating maximalist claims) – which is a necessary condition for higher level mobilization. We add to our analyses a counter of the number of years that the regime has seen maximalist claims since 1960. The inclusion of this counter does not substantively change the results.

We also consider what factors drive the emergence of grand bargains in the first place – if grand bargains are more likely to emerge where autocrats evaluate the threat of protest as high, then that association may drive the results. We reran our analyses on more limited samples to ameliorate potential reverse causality. First, we examined only those regime-years where the regime had not experienced a prior nonviolent campaign onset, and second, we limited the sample to those regime years where the regime had experienced neither a prior nonviolent resistance campaign nor an organized incompatibility over government. In these subsamples, it is less likely that grand bargain has emerged because the regime leader evaluates the risk of mass mobilization as high. The effect of grand bargain in these subsamples then is on the emergence of nonviolent resistance absent a substantial prior history of nonviolent opposition mobilization in the regime. Figure 4 shows the effect of grand bargain in the main analyses and in these two more restrictive samples.

We also tested for potential bias in our results presented by unobservable factors – for example, the military’s propensity to repress, cultural and historical factors – by employing the test developed by Oster (2019). The results of this test are provided in the Online appendix and provide strong evidence that our results are not driven by unobserved factors. We also reran the analysis including ‘juntas’, which occur when the military as an institution runs the government. Conversely, we also reran the analyses while excluding all subtypes of military regimes, including military-personalist hybrids. In both of these analyses, the grand bargain indicator remains substantively unchanged. Relatedly, we also reran our analyses controlling for whether or not the regime was

Figure 4.

Effect of grand bargain on nonviolent resistance campaigns in subsamples

Figure 5.

Effect of grand bargain index on nonviolent campaign onset

founded in a coup – using an indicator from GWF – and our results do not change substantively.

For analytical clarity and simplicity we model and describe grand bargain as a binary condition (present or not). Yet, as an additional check, we also constructed an additive index of the different components that we use to code the grand bargain indicator and replaced the indicator with this additive index as a linear term in a logistic regression of nonviolent campaign onset. Figure 5 (and Table A7 in the Online appendix) suggests that the more features of grand bargain that are present in a regime, the greater the impact of grand bargain on nonviolent campaign onset.

In addition, in Figure 6 we present the results of analyses that use different versions of the binary grand bargain indicator, which iteratively drop different components we use to characterize grand bargains overall. While Figure 5 suggests an additive effect for different components, the results here demonstrate that some aspects of grand bargain may be more important than others. Critically, these results underscore the importance of corporateness and political influence by showing that the absence of loyalist security organizations and a role for the military in the national cabinet are essential determinants of nonviolent campaign onset. When they

Figure 6.

Versions of grand bargain indicator and nonviolent campaign onset

are dropped from the indicator’s composition, the effect on campaign onset is no longer significant.

In summary, we find strong evidence in support of our expectation that autocracies whose civil-military relations are characterized by a grand bargain between the political and military leadership (H1) are more likely to see the emergence of nonviolent resistance campaigns. We find further support for our expectations that the individual components of a grand bargain – that is, military participation in politics and military corporateness – do not alone exert this effect (H2 and H3). Rather, it is the combination of these two dimensions that yields an increased likelihood of nonviolent campaign onset. Under grand bargain, we anticipate that dissidents will anticipate that mass repression by the military is less likely, and that both the cohesion of the military and its existence as a distinct political actor make it more likely that it can be separated from the regime. Analysis of events data tracking direct interactions between the military and opposition in a subsample of countries from 1990 to 2008 also suggests interactions between the military and opposition are more conciliatory or cooperative when there is a grand bargain.

Conclusion

This article explores how states’ civil-military relations condition the onset of nonviolent resistance campaigns in non-democratic regimes, arguing that those in which the political leadership accommodates the military with a grand bargain are more amenable to protest. The corporate and politically influential militaries in these settings affect the certainty with which protesters will experience repression and the possibility of leveraging divisions between the military and regime to win concessions.

Large-N empirical tests provide supporting evidence that civil-military relations characterized by grand bargains exhibit a higher likelihood of nonviolent resistance campaign onset. Empirical tests confirm that both features of grand bargains – military corporateness and political influence – are necessary to increase the likelihood of nonviolent campaign onset. Additional analyses of nonviolent protest tactics further support the theory, suggesting that when militaries are accommodated by grand bargains, interactions between protesters and the military tend to be relatively cooperative and conciliatory. This article thus contributes to growing literature on the regime security consequences of the structure of armed forces and provides new insight into the scholarship on nonviolent protest.

Future scholarship in turn might study how dissidents evaluate militaries, and whether there is cross-national variation in their perceptions that aligns with differences in civil-military relations. In addition, scholars might examine the role that extra-military security forces play in shaping expectations about the efficacy of protest, or protest onset. In this respect, researchers might explore whether variation in the character of internal security forces affects protest onset and also whether protest onset is affected by whether those forces carry out indiscriminate or selective repression (Greitens, 2016).

In turn, we hope also that our insights might prove useful to nongovernmental groups and activists seeking to support the growth of nonviolent social movements globally. Our argument suggests that dissidents might pay special attention to their military’s specific roles within regime institutions and its institutional character. These considerations could aid efforts to strategize about how and when to organize protests and the repertoire of tactics of nonviolence employed.

Footnotes

Replication data

The dataset and do-files for the quantitative analysis in this article, along with the Online appendix, can be found at .

Acknowledgements

The authors thank the editors and four anonymous reviewers for their feedback on this article. Earlier versions were presented at the 2019 meetings of the American Political Science Association and International Studies Association. The authors would also like to thank Ryan Grauer, Kristen Harkness, and Kara Kingma Neu for their valuable feedback on earlier versions of this article.

Funding

Data used in this article were collected with support from a National Science Foundation Dissertation Improvement Grant (SES #SES-1424001). This article does not necessarily reflect the views of the National Science Foundation or the US Government.

ORCID iDs

Risa Brooks

Peter B White

Notes

References

Al Hendi

Ahed

(2011) The structure of Syria’s repression: Will the army break with the regime? Foreign Affairs 3 May.

Al-Ayed

Abdel-Nasser

(2017) Can Syria be salvaged. In: Kodmani

Bassma

Moussa

Nayla

(eds) Arab Reform Initiative Report. https://s3.eu-central-1.amazonaws.com/storage.arab-reform.net/ari/2017/08/16145152/Arab_Reform_Initiative_Report_en_OUT_OF_THE_INFERNO_Rebuilding_Security_in_Iraq_Libya_Syria_and_Yemen.pdf.

Albrecht

Holger

Ohl

Dorothy

(2016) Exit, resistance, loyalty: Military behavior during unrest in authoritarian regimes. Perspectives on Politics 14(1): 38–52.

All Things Considered (2011) Founders of April 6 movement weigh in. National Public Radio 14 February. https://www.npr.org/2011/02/14/133756340/Founder-Of-Egypts-April-6-Movement-Weighs-In.

Asal

Victor

Conrad

Justin

Toronto

Nathan

(2017) I want you! The determinants of military conscription. Journal of Conflict Resolution 61(7): 1456–1481.

Bellin

Eva

(2012) Reconsidering the robustness of authoritarianism in the Middle East: Lessons from the Arab Spring. Comparative Politics 44(2): 127–149.

Boix

Carles

Svolik

Milan

(2013) The foundations of limited authoritarian government: Institutions, commitment, and power-sharing in dictatorships. Journal of Politics 75(2): 300–316.

Bou Nassif

Hicham

(2015) ‘Second-class’: The grievances of Sunni officers in the Syrian armed forces. Journal of Strategic Studies 38(5): 626–649.

Bou Nassif

Hicham

(2020) Endgames: Military Response to Protest in Arab Autocracies. Cambridge: Cambridge University Press.

10.

Brooks

Risa

(2008) Shaping Strategy: The Civil-Military Politics of Strategic Assessment. Princeton, NJ: Princeton University Press.

11.

Brooks

Risa

(2013) Abandoned at the palace: Why the Tunisian military defected from the Ben Ali regime in January 2011. Journal of Strategic Studies 36(2): 205–220.

12.

Butcher

Charles

Svensson

Isak

(2016) Manufacturing dissent: Modernization and the onset of major nonviolent resistance campaigns. Journal of Conflict Resolution 60(2): 311–339.

13.

Campbell

Kirk S

(2009) Civil-Military Relations and Political Liberalization: A Comparative Study of the Military’s Corporateness and Political Values in Egypt, Syria, Turkey, and Pakistan. Washington, DC: George Washington University Press.

14.

Carter

David B

Signorino

Curtis S

(2010) Back to the future: Modeling time dependence in binary data. Political Analysis 18(3): 271–292.

15.

Cebul

Matthew D

Grewal

Sharan

(2022) Military conscription and nonviolent resistance. Comparative Political Studies 55(13): 2217–2249.

16.

Chenoweth

Erica

Stephan

Maria J

(2011) Why Civil Resistance Works: The Strategic Logic of Nonviolent Conflict. New York: Columbia University Press.

17.

Chenoweth

Erica

Pinckney

Jonathan

Lewis

Orion

(2018) Days of rage: Introducing the NAVCO 3.0 Dataset. Journal of Peace Research 55(4): 524–534.

18.

Cook

Steven A

(2007) Ruling But Not Governing: The Military and Political Development in Egypt, Algeria, and Turkey. Baltimore, MD: Johns Hopkins University Press.

19.

Cunningham

Kathleen Gallagher

Beaulieu

Emily

(2010) Dissent, repression, and inconsistency. In: Chenoweth

Erica

Lawrence

Adria

(eds) Rethinking Violence: States and Non-State Actors in Conflict. Boston, MA: MIT Press.

20.

Cunningham

David E

Gleditsch

Kristian Skrede

González

Belén

Vidović

Dragana

White

Peter B

(2017) Words and deeds: From incompatibilities to outcomes in anti-government disputes. Journal of Peace Research 54(4): 468–483.

21.

De Bruin

Erica

(2018) Preventing coups d’état: How counterbalancing works Journal of Conflict Resolution 62(7): 1433–1458.

22.

De Bruin

Erica

(2020) How to Prevent Coups D’État: Counterbalancing and Regime Survival. Ithaca, NY: Cornell University Press.

23.

Earl

Jennifer

(2003) Tanks, tear gas and taxes: Toward a theory of movement repression. Sociological Theory 21(1): 63–102.

24.

Gales

Phillip

(2000) The Peruvian Military’s Role in Sustaining Democracy. Montgomery, AL: Air Command and Staff College.

25.

Gaub

Florence

(2019) Like father like son: Libyan civil-military relations before and after 2011. Mediterranean Politics 24(2): 181–195.

26.

Geddes

Barbara

Wright

Joseph

Frantz

Erica

(2018) How Dictatorships Work: Power, Personalization, and Collapse. Cambridge: Cambridge University Press.

27.

Ghanem

Dalia

(2019) What will Algeria’s military do next? Middle East Eye 29 July. https://www.middleeasteye.net/opinion/what-will-algerias-military-do-next.

28.

Gleditsch

Kristian S

(2002) Expanded trade and GDP data. Journal of Conflict Resolution 46(5): 712–724.

29.

Gleditsch

Kristian S

Rivera

Mauricio

(2017) The diffusion of nonviolent campaigns. Journal of Conflict Resolution 61(5): 1120–1145.

30.

Gleditsch

Nils Petter

Wallensteen

Peter

Eriksson

Mikael

Sollenberg

Margareta

Strand

Håvard

(2002) Armed conflict 1946–2001: A new dataset. Journal of Peace Research 39(5): 615–637.

31.

Greitens

Sheena Chestnut

(2016) Dictators and Their Secret Police: Coercive Institutions and State Violence. Cambridge: Cambridge University Press.

32.

Hanmer

Michael J

Kalkan

Kerem Ozan

(2013) Behind the curve: Clarifying the best approach to calculating predicted probabilities and marginal effects from limited dependent variable models. American Journal of Political Science 57(1): 263–277.

33.

Hunter

Wendy

(1998) Negotiating civil-military relations in post-authoritarian Argentina and Chile. International Studies Quarterly 42(2): 295–317.

34.

Huntington

Samuel P

(1957) The Soldier and the State: The Theory and Politics of Civil-Military Relations. Cambridge, MA: Harvard University Press.

35.

King

Gary

Zeng

Langche

(2001) Logistic regression in rare events data. Political Analysis 9(2): 137–163.

36.

Kirkpatrick

David D

(2011) Mubarak orders crackdown, with revolt sweeping Egypt. New York Times 28 January (https://www.nytimes.com/2011/01/29/world/middleeast/29unrest.html).

37.

Koren

Ore

(2013) Not (only) Assad’s fault: The military effect in Syria. International Affairs Forum 14 May. https://www.ia-forum.org/Content/ViewInternalDocument.cfm?ContentID=8006.

38.

Lee

Terence

(2014) Defect or Defend: Military Responses to Popular Protests in Authoritarian Asia. Baltimore, MD: Johns Hopkins University Press.

39.

Lindberg

Staffan I

Coppedge

Michael

Gerring

John

Teorell

Jan

(2014) V-Dem: A new way to measure democracy. Journal of Democracy 25(3): 159–169.

40.

Lindemann

Stefan

(2011) The ethnic politics of coup avoidance: Evidence from Zambia and Uganda. Africa Spectrum 46(2): 3–41.

41.

McAdam

Doug

(1996) Conceptual origins, current problems, future directions. In: McAdam

Doug

McCarthy

John D

Zald

Mayer N

(eds) Comparative Perspectives on Social Movements: Political Opportunities, Mobilizing Structures, and Cultural Framings. Cambridge: Cambridge University Press, 23–40.

42.

Nepstad

Sharon Erickson

(2013) Mutiny and nonviolence in the Arab Spring: Exploring military defections and loyalty in Egypt, Bahrain, and Syria. Journal of Peace Research 50(3): 337–349.

43.

Nossiter

Adam

(2019) In epic standoff, unarmed Algerians get the army to blink. New York Times 29 July (https://www.nytimes.com/2019/07/29/world/africa/algeria-revolution-standoff.html).

44.

Obando

Enrique

(1996 ) Fujimori and the military: A marriage of convenience. NACLA Report on the Americas 30(1): 31–36.

45.

Oster

Emily

(2019) Unobservable selection and coefficient stability: Theory and evidence. Journal of Business & Economic Statistics 37(2): 187–204.

46.

Pinckney

Jonathan

(2020) Curving the resource curse: Negative effects of oil and gas revenue on nonviolent resistance campaign onset. Research & Politics 7(2). https://doi.org/10.1177/2053168020936890.

47.

Pion-Berlin

David

(2010) Through Corridors of Power: Institutions and Civil-Military Relations in Argentina. University Park, PA: Pennsylvania State University Press.

48.

Posen

Barry

(1986) The Sources of Military Doctrine: France, Britain, and Germany between the World Wars. Ithaca, NY: Cornell University Press.

49.

Posen

Barry R

(1993) Nationalism, the mass army, and military power. International Security 18(2): 80–124.

50.

Powell

Jonathan M

Thyne

Clayton L

(2011) Global instances of coups from 1950 to 2010: A new dataset. Journal of Peace Research 48(2): 249–259.

51.

Roessler

Philip

(2011) The enemy within: Personal rule, coups, and civil war in Africa. World Politics 63(2): 300–346.

52.

Sayigh

Yezid

(2011) Agencies of coercion: Armies and internal security forces. International Journal of Middle East Studies 43(3): 403–405.

53.

Schaftenaar

Susanne

(2017) How (wo)men rebel: Exploring the effect of gender equality on nonviolent and armed conflict onset. Journal of Peace Research 54(6): 762–776.

54.

Schrodt

Philip A

(2012) CAMEO: Conflict and mediation event observations event and actor codebook. Available at: chrome-extension://efaidnbmnnnibpcajpcglclefindmkaj/http://data.gdeltproject.org/documentation/CAMEO.Manual.1.1b3.pdf.

55.

Serrano

Francisco

(2019) Bouteflika may have stepped aside, but the generals really running Algeria won’t. World Politics Review 13 March. https://www.worldpoliticsreview.com/articles/27634/bouteflika-may-have-stepped-aside-but-the-generals-really-running-algeria-won-t.

56.

Singh

Naunihal

(2014) Seizing Power: The Strategic Logic of Military Coups. Baltimore, MD: Johns Hopkins University Press.

57.

Stacher

Joshua

(2012) Adaptable Autocrats: Regime Power in Egypt and Syria. Redwood City, CA: Stanford University Press.

58.

Stephan

Maria J

Chenoweth

Erica

(2008) Why civil resistance works: The strategic logic of nonviolent conflict. International Security 33(1): 7–44.

59.

Svolik

Milan W

(2012) The Politics of Authoritarian Rule. Cambridge: Cambridge University Press.

60.

Talmadge

Caitlin

(2015) The Dictator’s Army: Battlefield Effectiveness in Authoritarian Regimes. Ithaca, NY: Cornell University Press.

61.

Tarrow

Sidney

(1996) States and opportunities: The political structuring of social movements. Comparative Perspectives on Social Movements: Political Opportunities, Mobilizing Structures, and Cultural Framings 90(2): 41–61.

62.

Tarrow

Sidney

Tilly

Charles

(2009) Contentious politics and social movements. In: Boix

Charles

Stokes

Susan C

(eds) Oxford Handbook of Comparative Politics. Oxford: Oxford University Press, 461–496.

63.

Themnér

Lotta

Wallensteen

Peter

(2014) Armed conflicts, 1946–2013. Journal of Peace Research 51(4): 541–554.

64.

Thompson

William R

(1973) The Grievances of Military Coup-Makers. Thousand Oaks, CA: Sage.

65.

Thyne

Clayton

(2017) The impact of coups d’état on civil war duration. Conflict Management and Peace Science 34(3): 287–307.

66.

Tilly

Charles

(1978) From Mobilization to Revolution. New York: McGraw-Hill.

67.

White

Peter B

(2017) Crises and crisis generations: The long-term impact of international crises on military political participation. Security Studies 26(4): 575–605.

68.

Willis

Michael

(2012) Politics and Power in the Maghreb: Algeria, Tunisia and Morocco from Independence to the Arab Spring. Oxford: Oxford University Press.