Why Does the Military Stay in Power after Coups?

Written by Maegan Liew Chew Min


In The Republic, Plato likened auxiliaries to watch-dogs, and warned of them turning into wolves that turn upon the sheep they are meant to guard. Indeed, military intervention in politics has been a long-standing concern among philosophers, political scientists and state leaders alike[1]. The intervention of the military in politics, most notably in the form of a coup[2], is regarded as a threat to be guarded against. However, there exists a largely neglected phenomenon within the literature: the democratic coup d’etat[3], which seeks to facilitate civilian rule.

This neglect illuminates a wider paradigm in the study of Civil-Military Relations (CMR): the traditional coup model, with its extensive empirical occurrence, has become paradigmatic of how military intervention has been analysed in the field. This suggests an implicit assumption that military coups necessarily culminate in military rule under military government/dictatorship (Varol 2012).

By studying the research question of ‘Why does the military stay in power after coups?’, this paper seeks to overcome the skewed nature of the field by rejecting the conflation of military coups with military rule. In identifying the factors that are instrumental in delineating a ‘democratic coup’ that is followed by democratic elections from a ‘military coup’ that is followed by military rule, the focus of the study is re-oriented from the study of coups to the identification of the causal factors explaining why the military stays in power after coups, in the form of military rule. This paper posits that a strong civil society, or the absence thereof, has significant implications for post-coup developments and the establishment of military rule.

This paper begins by providing an overview of existing literature on the causes of military intervention in politics. My hypothesis is then introduced to address the gap in existing literature. The paper takes an actor-based approach in analysing alternative explanations. By showing that these alternative factors remain constant across traditional and democratic coups, I argue that the presence or absence of a strong civil society has a causal effect on the establishment of military rule after coups.

Literature Review

The study of CMR can be generally split into two paths: the sociologically-oriented study of the military[4], and the institutionally-oriented examination of post-colonial CMR focusing on coup occurrence in developing states[5] (Feaver, 1999). While CMR is a broad subject that involves a spectrum of relationships between the military and civilian society at various levels, the field has generally focused on the control of the military by civilian authorities[6] (Feaver, 1999).

Coups have been the traditional focus of CMR[7] as they explicitly embody the fundamental problem of the military using its coercive power to displace civilian leaders (Feaver, 1999; Kanchanasuwon, 1989)[8]. Yet, despite (or perhaps due to) the field’s focus on military coups, there has been a dearth in literature analysing post-coup developments (Marinov&Goemans, 2013). This has translated into a lack of distinction in the literature with regards to the types of military coups and variation in coup outcomes. In particular, the democratic coup has been a neglected phenomenon (Varol, 2012). This paper seeks to add to existing literature by distinguishing traditional coups from democratic coups, in explaining the post-coup development of the military’s continued hold on power. 

There exists vast literature examining the explanatory factors of military intervention in politics. These factors can be differentiated according to whether they are external or internal to a country. They may pertain to structural factors, such as the internal/external threat environment (Desch, 1999), or agential factors, such as the role of international actors and great powers in asserting pressure (Masaki, 2016). Internal agential variables are further differentiated according to the civilian/military distinction[9] (Huntington 1957, 1968; Finer, 1962; Welch, 1976). In theorising military intervention, however, there has been a lack of reference to the agential role of the people. Schiff (2009) points out that the current CMR literature focuses on political institutions as the main ‘civil’ component of analysis, failing to consider the role of the citizenry[10]. This paper seeks to contribute to existing literature through the study of the role of the civil society in discouraging military rule.

Civil Society: Theory & Hypothesis

Civil society can generally be defined as sustained, organised social activity, in support of the public good, undertaken by groups formed outside the state, the market, and the family (Tasnim, 2012; Parnini, 2006). Civil society covers a diverse range of activities, from the provision of public goods to advocacy for social reform, which ultimately works towards ensuring political equality, political liberty and popular sovereignty (Huq, 2005).

The drawing of a positive link between civil society and democracy has been a long-standing one (Tasnim, 2012). Huq (2005) operationalised the relationship between civil society and democracy in Fig 1 below.

fig 1
Fig 1 [11]

In particular, civil society is widely deemed to promote democracy by broadening participation among the citizenry (Huq, 2005; Tasnim, 2012). This is not only through the civil society’s provision of channels for the representation of the people – it plays an important role in building confidence and conferring legitimacy to democratic governance. Firstly, civil society acts to monitor and restrain the power of the state, increasing the accountability of the state to the people (Tasnim, 2012). In addition, civil society has been linked to the generation of social capital[12]. Social capital theory associates dense networks of associations to the generation of trust and cooperation among citizens, sustaining high levels of civic engagement (Newton, 2001). Civil society, with its democracy-building characteristics, may thus influence the occurrence of military rule in two ways.[13]

First, civil society alters post-coup developments and outcomes.[14] A strong civil society that cultivates democratic culture and social capital will translate into broad political and civic engagement of citizens that serves to decrease the military’s success in retaining power after coups.[15]

Second, a strong civil society can exert a socialisation effect on the military, influencing its post-coup policy.[16] By socialising the military into reflecting its democratic and civic values, a strong civil society influences the military to facilitate post-coup elections and voluntarily subordinate itself to civilian rule.

2 1
Fig 2

My Hypothesis is thus as follows: The absence of a strong civil society is likely to lead to the establishment of military rule following a coup.


The argument I present is developed in the empirical context of Bangladesh, using a within case comparison design[17]. Bangladesh presents an interesting case study and an ideal setting for the study of both military intervention in politics and the role of civil society.

Firstly, the Bangladesh political trajectory has been marked by coups and military intervention in politics. Bangladesh gained independence from Pakistan following the Liberation War which had not only been a struggle for independence, but also a struggle for democracy (Khondker, 1986). Yet, Bangladesh would find itself following in the footsteps of Pakistan’s military regime with the assassination of Sheikh Mujib in 1975, shortly after its hard-fought independence. Amongst numerous coup attempts made against civilian and military regimes alike, the successful coups of 1975 and 1982 led to the rise of the military regimes under General Ziaur Rahmen (1978-1981) and General Ershad (1978-1981) that would define the first few decades of Bangladeshi independence.

It was not until widespread opposition brought down Ershad’s military regime in 1990 that democracy in Bangladesh was restored. However, the competing legacies of democratic inclinations and the legacy of Pakistani military regime would again find themselves in conflict when in 2007, the military-backed caretaker government (CTG)[18] interrupted Bangladesh’s democracy[19]. This covert takeover by the military has been likened to a military coup (Lorch, 2017). Yet, in this 2007 military takeover, the military asserted its commitment to restoring democracy and, unlike the 1975 and 1982 coups, delivered on its promises of returning power back to the civilians in the democratic elections in December 2008 (Ahmed, 2010)[20].

Secondly, Bangladesh’s vibrant and strong civil society is widely-recognised[21]. At the same time, Bangladesh presents an interesting case of a strong civil society existing in a weak state (Lorch, 2017)[22]. The Bangladesh case captures the nuances of democratic coups and of a strong, vibrant but non-vigilant civil society (Lorch, 2017; Tasnim, 2012)[23], presenting a good case for the study of civil society’s role in guarding against military rule.

In addition, Bangladesh, being a post-colonial, developing state, faces similar challenges as the rest of the Third world. The study of Bangladesh’s transition from a military regime to a democratic one today may contribute to existing literature as well as inform policies in addressing military coups that continue to plague parts of the post-colonial world.

In view of Bangladesh’s unique political trajectory, this paper adopts a within-case comparison of the 1975 and 1982 military coups that led to establishment of military regimes, as opposed to the 2007 military-backed CTG that was more akin to a democratic coup characterised by post-coup elections and the voluntary transfer of power to civilians.

Measure of Dependent Variable: Military Rule

The presence/absence of military rule is contingent on whether the civilians or the military are in power. The absence of military rule after coups is operationalised as the presence of free and fair democratic elections that transfers power to civilian leaders within 5 years[24].

In Bangladesh, military rule was observed to follow the 1975 and 1982 coups[25], while military rule was absent in the 2007 coup[26].

Measurement of Explanatory Variable: Civil Society

In view of the lack of civil society index data dating back to the 1970s, the measurement of the explanatory variable will be operationalised in three ways for construct validation purposes. 

  • Literacy Rate

The correlation between literacy rate and political mobilisation/participation has been well-established (Ganguly, 1996). An increase in literacy rate will thus likely be correlated to a stronger civil society.

As literacy rate in Bangladesh has been increasing steadily[27] (Fig 3), it can be inferred that the civil society in 2007 was stronger than the civil society during the 1975 and 1982 coups.

3 1
Fig 3[28]
  • Hartal Occurrence

Hartals are an interesting phenomenon of political strikes in Bangladesh[29]. Strikes occurrence reflect civic/political engagement, revealing the strength of the civil society. A rise in occurrence of hartals will reflect a corresponding increase in the strength of civil society.

The occurrence of hartals in 1990s-2000s was higher than in the 1970s-1980s[30], as a negative correlation was found to exist between hartal occurrence and the time-periods under military dictatorship (Azmon&Salmon, 2004)(Fig 4).

Fig 4[31]
  • Voter Turnout Rates

Voter turnout rates not only reflects civic/political participation but is also correlated with the strength of social capital (Putnam, 2000). An increase in voter turnout will thus likely reflect a corresponding increase in the strength of civil society.

From Fig 5, there has been a rise in trend of voter turnout rates from the 1970s-80s to the 1990s-2008[32]. This reflects an increase in civic/political participation as well as social capital, indicating the presence of a stronger civil society during the period of the 2007 coup as opposed to the 1975 and 1982 coups.


Fig 5[33]

Alternative Explanations

As noted in the literature review, there is a rich body of scholarship that points to a number of factors that may explain military intervention in politics. This paper takes an actor-based approach and examines alternative explanations involving the role of the civilian government, the military and the international community.

AE1: The absence of legitimacy of civilian government is likely to lead to the establishment of military rule following a coup.[34] 

The absence of civilian legitimacy is operationalised by:

  • Lack of resistance or Support from the citizenry for military coup

The lack of resistance, or even support, from the people for military intervention indicates weak civilian authority and popularity. In both the coups of 1975 and 1982, no demonstrable resistance or opposition was observed (Khondker, 1986). In the 2007 coup, the military had received substantial support from the civil society, including approval from NGOs and the media (Lorch, 2017).

  • Corruption Rankings

While mismanagement of civilian governments can be in various forms, the level of corruption in government often has direct implication for its legitimacy. In view of its consistent ranking amongst the world’s most corrupt countries since 1990s[35] as well as past records of widespread corruption under the Mujib and Sattar civilian governments (Rahman, 1983; Blair, 2013), it can be inferred that civilian governments in Bangladesh have largely been deemed corrupt and lacking in legitimacy.

AE2: The military’s self-perception of its custodianship responsibility is likely to lead to the establishment of military rule following a coup.[36] 

The presence of the military’s self-perceived custodianship responsibility is operationalised by:

  • Occurrence of military coup amidst a state crisis

A state crisis may include political instability, widespread disillusionment or abuse of state power. A military coup executed in such a backdrop may be reflective of the military’s attempt to fulfil its manifest destiny. The coups of 1975, 1982 and 2007 had invariably been executed in the context of state crisis[37].

  • Promise of political reforms and democratic elections

Attempts of military to justify coup by promising to improve and remedy the country’s political scene could be reflective of the military’s self-perception as saviours of the country. All three coups were followed with criticism of the previous government as well as promises of political reform and elections (Khondker, 1986; Blair, 2010).

AE3: High dependence on international actors for disbursement of financial resources is less likely to lead to the establishment of military rule following a coup.

Dependence on outside actors for resources, in particular Western aid, can make the military vulnerable to external pressure. Outsiders’ demands for elections can be potentially decisive in determining post-coup developments when external actors exercise substantial control over the country’s financial resources (Marinov&Goemans, 2013).

  • Aid Dependence

Dependence on international donors has been a traditional form of external dependence of a country. In Bangladesh, aid dependence (as % of GDP) has been on a downward trend (Fig 5). At its peak in the 1970s-1980s, military rule had persisted[38]. With a decreased dependence on aid in 2007, it is thus unlikely that external pressure would affect the occurrence of military rule following the 2007 coup. 

Fig 6[39]
  • UN Peacekeeping Compensation

The participation of Bangladesh Armed Forces in UN Peacekeeping Missions since the 1990s has emerged as another noteworthy form of external financial dependence.

According to Bangladesh’s internal documents, gross compensation from 2001–2010 amounted to US$1.28b (Cunliffe, 2017). This translates into slightly less than 0.2% of Bangladesh’s total GDP of $761.245b during the same period (World Bank, 2018). It appears that contribution to UN Peacekeeping Mission had not substantially increased Bangladesh’s dependence on external actors for financial resources[40]. Since Bangladesh’s dependence on external finance disbursement is low, the impact of pressure from international actors on the military’s establishment of military rule should be limited.


The results from the operationalisation of my hypothesis and AE1-3 are summarised in the table below. 

Coup (Year)  Strong Civil Society Civilian Legitimacy Military Custodianship External Dependence Military Rule
1975 Absent Absent Present Present [Moderate (~0.05)] Present
1982 Absent Absent Present Present [Moderate (~0.07)] Present
2007 Present Absent Present Present [Low (~0.02)] Absent

This demonstrates that the presence of a strong civil society is both sufficient and necessary for the absence of military rule[41].

The Growth of Civil Society

What explains the ability of the civil society to influence the military to pursue the democratic coup policy? This can be attributed to the growth of the Bangladesh civil society in both its vertical (increase in number of educated/professional citizens) and horizontal (increase in broad participation) strength[42].

The 1971 Liberation War and its aftermath had necessitated and marked the birth of the work of NGOs. The beginning of Bangladesh’s civil society was motivated by the need for non-state actors to perform functions normally ascribed to the state[43] (Lorch, 2016). However, it would take nearly 2 decades before the impact of Bangladesh’s civil society can be felt in the political sphere, when the people’s demonstrations brought down its last military regime under General Ershad[44].

It is notable that this 1990 movement had been largely driven by student groups and professional associations within the civil society. Huq (2005) notes the abstention of organised labourers in the tertiary sector (agriculture, trade and industry) of the civil society from the 1990 mass upsurge. This phenomenon is supported by the association between education and literacy with political/civic mobilisation (Gunguly, 1996). As literacy rates increased from the 1970s to 1990s, civil society had gained in vertical strength, and was able to drive the downfall of Ershad’s military regime.

By the 2000s, civil society has strengthened in terms of broad citizen participation as well.  This growth can be charted by the voter turnout rates (Fig7)[45]. This may be attributed to the role of the civil society in increasing the confidence of the citizenry in democracy through its building of social capital and monitoring of the state to increase accountability to the people. The caretaker government concept made possible by the civil society was one such example[46].

In view of the high voter turnouts (above 70%) in the decade preceding the 2007 military takeover, on top of the country’s past experience of mass demonstrations against Ershad’s military regime[47], it is likely that the military’s policy of facilitating post-coup elections had took into consideration the civil society’s values and position − which was ultimately reflected in the 2008 voter turnout, the highest in the country’s history.

The absence of military rule after a coup only occurred in 2007, following the growth of civil society in vertical strength by the 1990s and horizontal strength by the 2000s. Along with the causal mechanism illustrated in Fig 2, we can conclude that the absence of military rule occurred after and as a consequence of the growth of Bangladesh’s civil society.

Fig 7[48]


To overcome the conflation of military coups with military rule and address the lack of research into post-coup developments, this paper has sought to explain the occurrence of military rule after coups through a within-case analysis of Bangladesh. In addition, this paper looked at an often-neglected actor in CMR − the citizenry. I argue that the strength of a civil society can influence the military’s coup policy through the process of socialisation as well as by acting as an obstacle to the establishment of military rule.

More analysis into the role of the different actors of civil society[49] may be helpful to clarify the causal mechanism between civil society and military rule. The study of agential factors can also be better informed by considering the changes in structural factors/conditions[50].

Through its findings, this paper hopes to highlight that the civil society, or “society of ‘men without arms’” (Huq,2005,p.89), can play an important role in shaping the CMR dynamics, and is an area worthy of greater attention by political scientists and practitioners alike.


Ahmed, N. (2010). Party politics under a non-party caretaker government in Bangladesh: the Fakhruddin interregnum (2007–09), Commonwealth & Comparative Politics, 48:1, 23-47, DOI: 10.1080/14662040903444491

Azam, J., & Salmon, C. (2004). Strikes and political activism of trade unions: Theory and application to bangladesh. Public Choice, 119(3/4), 311-334. doi:10.1023/B:PUCH.0000033323.54694.90

Ahsan, R. and Iqbal, K. (2016). “How Do Exporters Cope With Violence? Evidence from Political Strikes in Bangladesh,” Department of Economics – Working Papers Series 2025, The University of Melbourne.

Blair, D. C. (2013). Military Engagement: Influencing Armed Forces Worldwide to Support Democratic Transitions. Washington: Brookings Institution Press. Retrieved from https://muse-jhu-edu.libproxy1.nus.edu.sg/book/27280

Cunliffe, P. (2018). From peacekeepers to praetorians – how participating in peacekeeping operations may subvert democracy. International Relations, 32(2), 218-239. doi:10.1177/0047117817740728

Khan, Zillur R. “Politicization of the Bangladesh Military: A Response to Perceived Shortcomings of Civilian Government.” Asian Survey 21, no. 5 (1981): 551-564.

Kuehn, D. (2017). Midwives or gravediggers of democracy? The military’s impact on democratic development. Democratization, 24:5, 783-800, DOI: 10.1080/13510347.2017.1324421 Retrieved from https://www-tandfonline-com.libproxy1.nus.edu.sg/doi/pdf/10.1080/13510347.2017.1324421?needAccess=true

Desch, M. C. (1999). Civilian control of the military: The changing security environment. Baltimore: Johns Hopkins University Press.

Feaver, P. (1999). “Civil-Military Relations,” Annual Review of Political Science, Vol. 2, pp. 211-241.

Finer, S. E. (Samuel Edward). (1962). The man on horseback; the role of the military in politics

Huntington, S. P. (1968). Political order in changing societies. New Haven: Yale University Press.

Huq, P. A. (2005). Civil society and democracy in bangladesh. Social Change, 35(2), 85-100. doi:10.1177/004908570503500206

International Institute for Democracy and Electoral Assistance. (2018). Bangladesh, Voter Turnout. Retrieved from https://www.idea.int/data-tools/question-countries-view/521/59/ctr

Istiak, K. M. (2012). foreign aid to bangladesh: Some iconoclastic issues. The Journal of Developing Areas, 46(1), 331-343. doi:10.1353/jda.2012.0005

Khondker, H. H. (1986). Bangladesh: Anatomy of an unsuccessful military coup. Armed Forces & Society, 13(1), 125-143. doi:10.1177/0095327X8601300106

Lewis, D. (2011). State, politics and institutions. Cambridge: Cambridge University Press.10.1017/CBO9781139017138.005

Lorch, J. (2016). Civil society and mirror images of weak states: Bangladesh and the philippines. London: Palgrave Macmillan UK. doi:10.1057/978-1-137-55462-8

Lorch, J. (2017). Civil society support for military coups: Bangladesh and the philippines. Journal of Civil Society, 13(2), 184-201. doi:10.1080/17448689.2017.1312790

Masaki, T. (2016). Coups d’État and foreign aid. World Development, 79, 51-68. doi:10.1016/j.worlddev.2015.11.004

Newton, K. (2001). Trust, social capital, civil society, and democracy. International Political Science Review / Revue Internationale De Science Politique, 22(2), 201-214. doi:10.1177/0192512101222004

Parnini, S. (2006). Civil society and good governance in bangladesh. Asian Journal of Political Science, 14(2), 189-211. doi:10.1080/02185370601063191

Pattanaik, S. S. (2008). Re-emergence of the military and the future of democracy in bangladesh. Strategic Analysis, 32(6), 975. doi:10.1080/09700160802404521

Putnam, R. D. (2000). Bowling alone: The collapse and revival of american community. New York: Simon & Schuster.

Rahman, A. (1983). Bangladesh in 1982: “beginnings of the second decade”. Asian Survey, 23(2), 149.

Schiff, R. L. (2009). The military and domestic politics: A concordance theory of civil-military relations. London: Routledge.

Tasnim, F. (2012). How vigilant is the vibrant civil society in bangladesh? A survey-based analysis. Journal of Civil Society, 8(2), 155-183. doi:10.1080/17448689.2012.726548

Trading Economics. (2018). Bangladesh Corruption Rank. Retrieved from https://tradingeconomics.com/bangladesh/corruption-rank

UNESCO Institute of Statistics. (2018). Bangladesh. Retrieved from http://uis.unesco.org/country/BD 

Varol, O. (2012). The democratic coup d’etat. Harvard International Law Journal, 53(2), 291.

Welch, C. E. (1976). Civilian control of the military: Theory and cases from developing countries. Albany: State University of New York Press.World Bank. (2018). Bangladesh GDP (current US$). Retrieved from https://data.worldbank.org/indicator/NY.GDP.MKTP.CD?end=2010&locations=BD&start=2001&view=chart


[1] There exists a common belief, especially in the democratic world, that it is ‘natural’ for the military to obey civilian power (Finer, 1962; Feaver, 1999).

[2] In this paper, a coup d’etat is defined as the seizure of effective executive authority through the threat or use of force (Marinov&Goemans, 2013).

[3] Also known as a guardian coup, the democratic coup is characterised by efforts to facilitate free and fair elections, ultimately bringing about the transfer of power to democratically elected civilian leaders (Varol, 2012). The democratic coup, being an exception rather than the norm of coup models, has largely been neglected by existing literature (Varol, 2012).

[4] Such as Janowitz’s (1960) The Professional Soldier

[5] Such as Huntington’s (1968) Political Order in Changing Societies

[6] This has its roots in the civil-military problematique, whereby “the very institution created to protect the polity is given sufficient power to become a threat to the polity” (Feaver, 1999, p214). This also explains the preoccupation of the CMR field with the ‘coup’ phenomenon.

[7] It needs to be noted that the study of CMR is not limited to the area of coup occurrence as an indicator of military intervention in politics. In fact, scholars have argued that the coup/no coup dichotomy ignores the nuances of CMR. CMR can also be studied by examining military influence, civil-military friction, military compliance, delegation and monitoring (Feaver, 1999).

[8] In particular, the study of CMR in the post-colonial world has been dominated by the phenomenon of coups (Feaver, 1999; Kanchanasuwon, 1989).

[9] Huntington’s two classic works (1957, 1968) delineate this debate; in his earlier work, he emphasises the military’s professionalism and subsequently turns his attention to the degree of political institutionalisation by civilians in society. Likewise, Finer (1962) ties in the civilian/military factors through his disposition/opportunity framework, in which disposition to intervene encompasses factors like the military’s manifest destiny and corporate interests while opportunity to intervene may exist under weak civilian leadership. For example, Welch (1976) highlights the role of legitimacy and efficiency of civilian governments in deterring coups.

[10] Schiff (2009) attempts to fill this gap through the concordance theory which emphasises the achievement of concordance via a cooperative relationship among the three partners: the military, political elites and citizenry, which will work to decrease the likelihood of a military intervention. However, Schiff’s argument fails to account for democratic coups in which the military stands in direct opposition to political elites.

[11] Retrieved from Huq (2005), p.90

[12] Social capital, found in civic communities and associational culture, refers to the features of social organisations facilitating coordinated action and has been found to be linked to the success of democracy (Tasnim, 2012; Putnam, 2000).

[13] Fig 2 summarises how a strong civil society may lead to post-coup elections, thereby guarding against the occurrence of military rule.

[14] Marinov and Goemans (2013) demonstrate that post-coup dynamics can affect the military’s ability to intervene in politics. In particular, shared democratic norms and ideals among people in established democracies will lead to popular pressures for elections. Even in cases where the military seized power on account of civilian government mismanagement, the military will ultimately face “mounting pressures to return to the barracks once their job is done” (Marinov&Goemans, 2013, p.803).

[15] This is also likely to be factored into the military’s calculus in deciding their post-coup policy, thereby orientating the military towards relinquishing power to civilian leaders through facilitating democratic elections.

[16] This is related to the concept of subjective control put forward by Janowitz (1960) in which the interconnectedness of the military with society results in the military reflecting the core values of society, as embodied by the civic virtue of the citizen-soldier ideal (Schiff, 2009).

[17] The within-case comparison of Bangladesh across time is an ideal model for the comparative case-study method which overcomes the many variables small-N conundrum for the testing of my hypothesis, as it enables for most other variables to be controlled for.

[18] CTGs have been a practice relatively unique to Bangladesh. The concept of a caretaker government, composed fully of non-political personages and set up with the sole purpose of holding a national election, was an innovative idea made possible by the relentless efforts of the civil society (Huq, 2005). In the 2007 CTG, however, the role of the military as the mastermind behind the installation of the 2007 military-backed CTG installation distinguishes it from other CTGs in previous years (Blair, 2013). This has been interpreted as the military’s intervention into politics not dissimilar to a guardian coup (Marinov&Goemans, 2013).

[19] On 11 January 2007, the military had forced President Iajuddin Ahmed to postpone elections and declare a state of emergency (Lorch, 2017). He also announced his resignation from the position of chief adviser to the caretaker government (Blair, 2010). The military would proceed to successfully unseat the incumbent executive, installing a new military-backed CTG the very next day (Lorch, 2017).

[20] Having assumed power in the backdrop of widespread political violence, the military-backed CTG committed itself to introducing political reforms to ‘cleanse’ the political scene before returning to the barracks in December 2008.  

[21] In 2006, a Nobel Peace Prize was notably awarded to Professor Yunus and the Grameen Bank, signifying the rising world recognition for the achievements of the Bangladeshi civil society (Tasnim, 2012).

[22] For example, the country has consistently been ranked as one of the world’s most corrupt countries (Tasnim, 2012).

[23] The literature has featured several studies regarding the paradox of the Bangladesh civil society’s inability to strengthen democracy in the country despite its vibrant nature. In particular, its non-vigilant nature has seemingly challenged the widely-assumed democracy-building function of civil society (Tasnim, 2012). The role that the strong but unvigilant Bangladesh civil society can play in a weak state has thus remained a gap in existing literature. This paper posits that the Bangladeshi case fits perfectly into the nuances of a democratic coup, which while at the one hand indicates poor democratic governance and on the other a strong democratic culture that the military subordinates itself to after usurping power. As is widely criticised, Bangladesh’s unvigilant and politicised civil society has diminished its surveillance role in keeping the democratic state in check, contributing to political disorder as part of the country’s confrontational politics and bitter rivalry between the two main political parties: the Awami League and the Bangladesh Nationalist Party. This motivated the 2007 military CTG to take over in a covert coup. At the same time, the strong, vibrant civil society worked to guard against the establishment of military rule, shaping the military policy into a democratic coup.

[24] This 5-year criteria was adopted from Marinov and Goemans (2013). In this paper, elections staged by the military or which involve the participation of ‘civilianised’ military coup-plotters are not considered as free and fair elections. The outcome of elections must lead to the transfer of power to civilian leaders.

[25] Following the 1975 and 1982 coups, performative elections were conducted to enable the military and coup leaders to come to power. As power was not transferred to civilians, the tenures of General Zia Rahman (1977-1981) and General Ershad (1983-1990) as the President of Bangladesh are considered as military regimes in this paper.

It is also notable that on top of the participation of military leaders-turned politicians in these elections, the ‘free and fair’ nature of elections was also questionable. For example, Zia’s 1977 referendum, which saw him secure an impressive 98.89% of the popular vote, featured problematic wording of the referendum question: “Do you have confidence in President Major General Ziaur Rahman and in his policies and programs enunciated by him?”. The referendum question was worded such that voters were neither aware of nor given a choice of an alternative to the proposed military regime (Khan, p.559). The 1986 and 1988 elections held during the Ershad regime had been essentially uncontested as the major opposition parties (the Awami League and BNP) had boycotted the elections (Blair, 2013).

[26] In contrast, the 2007 military-backed CTG ushered in democratic elections within 2 years of its covert takeover in December 2008. The 2008 elections had largely been considered as free and fair, and brought the presently-incumbent party, the Awami League, to power. The Bertelsmann Transformation Index’s Free and Fair Elections component scored Bangladesh’s 2007 elections at 7, which was in line with world median and above the threshold score of 6 (below which is classified as an autocracy).

[27] As data on pre-1980s literacy rate is not available, extrapolation is necessary. In 2007, literacy rate among adult population stands at 46.7%, a substantial increase from 29.2% in 1981.

[28] UNESCO Institute of Statistics. (2018). Bangladesh. Retrieved from http://uis.unesco.org/country/BD 

[29] This observable phenomenon was chosen in view of Bangladesh’s unique political culture of confrontational politics. The use of political strikes, known as hartals, is a particularly interesting political performance in Bangladesh. Hartals have traditionally symbolised protest against misrule and had also played a significant role in the demise of the country’s last military regime in 1990 (Ahsan&Iqbal, 2016). Today, the main grievance behind hartals has been political, such as concerns about election fairness (Ahsan&Iqbal, 2016). Participation in hartals thus signifies political and civic participation.

[30] Azman&Salmon (2004, p.327) observed that the “military dictatorship period witnessed a lower level of strike activity than the subsequent democratic one”.

[31] Retrieved from Azam, J., & Salmon, C. (2004). Strikes and political activism of trade unions: Theory and application to bangladesh. Public Choice, 119(3/4), 311-334.

From their results, it can be understood that the occurrence of strikes is negatively correlated with the military dictatorship. They found that this is not statistically significant – this can thus be interpreted as an anomaly. Since strikes during the military dictatorships of 1970s-80 is lower than under the subsequent democratic governments, the civil society is likely to be stronger during the 2000s.

[32] The voter turnout rates in 1970s-80s hovered at 50-60% before shooting up to 75% by mid-1990s, with the voter turnout in 2008 hitting a historic high of 85.26%. The 2014 dip in voter turnout is an interesting phenomenon but is outside the scope of this paper. It is however likely to be associated with the BNP’s boycott of the 2014 elections, rather than a reflection of civil society strength. The 2018 elections voter turnout data is not yet available.

[33] International Institute for Democracy and Electoral Assistance. (2018). Bangladesh, Voter Turnout. Retrieved from https://www.idea.int/data-tools/question-countries-view/521/59/ctr

[34] The strength and legitimacy of the civilian government has been found to be important in military intervention in politics (Welch, 1976). In his disposition/opportunity framework, Finer (1962) categorises weak civilian leadership as providing the opportunity for the military to interfere with civilian politics. The lack of civilian legitimacy discredits the civilian government, weakening the authority and popularity of the incumbent civilian leaders. This provides the stage for military intervention and also enhances the support for the military.

[35] From 2001-2005, Bangladesh was ranked by Transparency International as the world’s most corrupt country for five successive years (Tasnim, 2012). Data on Transparency International Corruption Index began in the 1990s. The corruption levels in the 1970s-80s have to be extrapolated. In 1996, the first year when Bangladesh was included in the Transparency International corruption index, it was ranked the 4th most corrupt country. In 2018, Bangladesh was ranked 149 out of 180 countries in terms of perceived level of public sector corruption (Transparency International, 2019).

[36] The distinctive corporate identity of the military as the ‘guardians’ of the nation has been a fundamental factor motivating military interventions in countries. Finer (1962) terms this ‘the Manifest Destiny of Soldiers’ as saviours of their countries. He attributes the military’s consciousness of its unique mission and self-sacrificial corporate virtues as providing the basis for its self-perceived duty to save the nation by intervening in politics, shaping the disposition of the military to intervene (Finer, 1962).

[37] The 1975 coup had occurred in the midst of an increasingly authoritarian and corrupt civilian government and declining standards of living in the country as a result of the 1973 oil crisis and 1974 famine (Lewis, 2011). The 1982 coup had been launched in an atmosphere of disillusionment amongst the public and loss of faith in a corrupt government embroiled in party factionalism (Rahman, 1983). The 2007 coup occurred amidst political chaos and violence in the lead up to elections (Blair, 2010).

[38]This appears to challenge AE3. However, a limitation of my analysis is the lack of consideration of the Cold War and post-Cold War dynamics. Marinov&Goemans (2013) argue that greater dependence on Western aid flows can increase the likelihood of post-coup elections, but only in the post-Cold War era. However, as only the 2007 coup happened in the post-Cold War era, it would be hard to ascertain the impact and extent of international donor pressure in post-Cold War. My analysis will thus be based on objective aid amount.

[39] Istiak, K. M. (2012). foreign aid to bangladesh: Some iconoclastic issues. The Journal of Developing Areas, 46(1), 331-343. doi:10.1353/jda.2012.0005 Retrieved from https://muse-jhu-edu.libproxy1.nus.edu.sg/article/470006

[40] Although this paper is unable to prove that Bangladeshi participation in Peacekeeping Missions had not granted external actors leverage to exert pressure on the military , this paper contends that Bangladesh contribution to UN Peacekeeping Mission had not substantially increased the country’s dependence on international actors for financial resources, and that pressure from international actors will thus be limited. Then, AE3 is disproved by the Bangladeshi experience as high aid dependence had not prevented military rule in the 1970s-80s, while post-coup elections were held despite low external financial dependence in 2007.

[41] AE1-3 have been constant factors in all 3 coups, while the presence of a strong civil society is the only varying factor. Since the constant factors cannot explain the divergent post-coup developments, this paper argues that the presence/absence of a strong civil society is a causal factor in explaining the occurrence of military rule after coups. As there is an absence of military rule only when a strong civil society is present, a causal relationship can be drawn between the presence of a strong civil society and the absence of military rule. A strong civil society is thus both necessary and sufficient for an absence of military rule. 

[42] This growth can be attributed to the increase in literacy rate and can be charted to Fig3 which maps the growth in literacy rate in Bangladesh.

[43] Bangladesh’s civil society has been internationally recognized for its contributions and efforts in poverty alleviation and social development, as epitomised by Grameen Bank’s famous micro-credit system (Tasnim, 2012). Other NGOs like the Bangladesh Rural Advancement Committee have also been pivotal in initiating welfare and education programmes throughout the country.

[44] The civil society’s opposition to General Ershad and demands for his resignation had been taken into account by the military in their refusal to back Ershad in imposing martial law (Blair, 2013).  

[45] While the 1970s-80s voter turnout had hovered at 50-60%, by the mid-1990s, voter turnout soared to above 70%. Significantly, voter turnout rate shot up to a historic high of 85% in 2008, reflecting broad civic participation. This suggests a high level of political/civic engagement of the citizenry at the time of the 2007 military coup. This represents a growth in horizontal strength of civil society.

[46] Elections held in 1991, 1996 and 2001 had been considered highly competitive free and fair elections (Huq, 2005). This can be attributed to the holding of elections under caretaker non-political governments. The concept of a caretaker government (CTG) set up with the sole purpose of overseeing national elections was largely the achievement of civil society efforts in Bangladesh (Huq, 2005). In turn, the CTG arrangement driven by the civil society boosted both accountability as well as civic/political participation of the citizens, as reflected by the growth in voter turnout rate from 1996 onwards.

[47] The factoring in of the civil society’s strength and position into the military’s policy had been observed in 1990 when Ershad’s appeal to the military to implement martial law against the people’s demonstration was rebuffed (Blair, 2013).

[48] Graph of voter turnout rate across years, based on data from International Institute for Democracy and Electoral Assistance (2018).

[49] For example, the media or NGOs as opposed to the citizenry

[50] Such as the Cold War/post-Cold War environment