Peace Science Made Accessible, Understandable, and Useful.

Does Security Force Fragmentation Make Defections More or Less Likely?

Does Security Force Fragmentation Make Defections More or Less Likely?

Photo credit: Georgia National Guard 

This analysis summarizes and reflects on the following research: Dworschak, C. (2020). Jumping on the bandwagon: Differentiation and security defection during conflict. Journal of Conflict Resolution, 64(7-8), 1335-1357.

Talking Points

  • The kind of security force fragmentation—whether it is “for the purposes of coup-proofing” or “for the purposes of security specialization”—matters when it comes to how likely security forces are to defect in the face of a violent or nonviolent uprising, as it affects their perceptions of who is likely to win and therefore whether they should “bandwagon” with the uprising or not.
  • Whereas previous research has found greater security force fragmentation to correspond with a higher likelihood of defection, this finding applies only to counterbalancing (fragmentation “for the purposes of coup-proofing”) but not to differentiation (fragmentation “for purposes of security specialization”).
  • The greater the differentiation (fragmentation “for the purposes of security specialization”) of a country’s security forces, the less likely those security forces are to defect in a given year of an opposition campaign, whether that campaign is violent or nonviolent.
  • The likelihood of security force defection falls from 76% to 14% when an average country has moderate (as opposed to minimal) levels of security force differentiation (fragmentation “for the purposes of security specialization”).


Security defections in the face of uprisings are important to understand from a range of perspectives. While governments may want to understand what measures they can take to prevent security defections, resistance movements may want to understand how they can more effectively precipitate them to increase their chances of success. Previous research has examined, among other conditions, the extent to which the fragmentation of the security forces may influence the likelihood of defection. In particular, “counterbalancing”—”fragmentation for the purposes of coup-proofing”—appears to make defections more likely due to greater grievances among soldiers and less effective monitoring of them. Christoph Dworschak posits, however, that there is a useful distinction to be drawn between different forms of security force fragmentation, especially as these relate to defections. Counterbalancing is not the only form of fragmentation; there is also what he calls “differentiation”—the fragmentation of security forces not to ward off coups but rather “for the purpose of security specialization.” Curious about the implications of this different form of fragmentation, the author asks, “how does military differentiation influence the likelihood of defections during (non)violent uprisings?”

Security defection: when “part of the security apparatus [either individual soldiers or officers or entire units] breaks away from the government and actively sides with the opposition movement.”

Counterbalancing: a move undertaken by governments “at risk of experiencing a coup d’état,” whereby they purposefully fragment “the military so that power is not concentrated within a single command structure, and individual segments can be used against each other in case of one staging a coup.”

Differentiation: security force “fragmentation for the purpose of security specialization,” such that “multiple parallel forces” have their “training and equipment [tailored] to fulfill specific security tasks.”

Bandwagoning: “when segments of the military recognize that they are fighting on the ‘losing side’” and thereby decide to throw their support behind the “winning side,” in this case the opposition movement.

The author’s hypothesis that differentiation, unlike counterbalancing, makes defections less likely stems from his thinking about the motivations behind the decision to defect. One such motivation is the desire of security forces to be on the winning side of a conflict—a move called “bandwagoning.” If individual soldiers or military leaders sense that an uprising will succeed and that the government’s survival is at risk, they are more likely to defect to the opposition. With this in mind, the author sees the key distinction between counterbalancing and differentiation as being their impact on military “effectiveness”: whereas counterbalancing essentially makes security forces less effective (in an attempt to weaken their capacity to achieve a coup), differentiation makes security forces more effective, insofar as different units are specialized for different tasks and given the tools they need for those tasks. According to this logic, when security forces under either scenario are facing an uprising (whether violent or nonviolent), they make an assessment of their chances of winning against that uprising. Since “counterbalanced” security forces are apt to feel ineffective against an uprising, they are more likely to defect. Since “differentiated” security forces are apt to feel more effective against an uprising, they are more likely not to defect but rather to stick with the government, whom they think will ultimately prevail. Additionally, the author notes that security forces are also more likely to feel loyal to a regime under differentiation, due to higher levels of satisfaction, than under counterbalancing, associated with higher grievances, again making defections less likely in the case of the former.

To test his hypothesis, the author uses statistical analysis to determine if there is a relationship between security force differentiation and defection over multiple countries and decades. Security defection is considered to have occurred in a given campaign-year if “segments of the state security apparatus (elements of the military, paramilitaries, or police) switch[ ] sides and openly support[ ] the (violent or nonviolent) opposition campaign.”

The author’s statistical results confirm his hypothesis, as he finds that the greater the differentiation of a country’s security forces, the less likely those security forces are to defect in a given year of an opposition campaign. Furthermore, this inverse relationship is especially observed in countries with low coup risk—the most likely sites of “pure” differentiation (rather than possible counterbalancing, more likely in countries with high coup risk). This finding affirms his core conclusion that differentiation (as opposed to counterbalancing) of security forces makes them less likely to defect to the opposition in conflict situations. To illustrate this relationship, he finds that the likelihood of security force defection falls from 76% to 14% when an average country has even just moderate (as opposed to minimal) levels of security force differentiation.

The author concludes, therefore, that previous research finding a correlation between greater security force fragmentation and higher likelihood of defection applies only to counterbalancing but not to differentiation. He suggests that this distinction emerges from the different effects these forms of fragmentation may have on security forces’ actual and perceived military effectiveness against an uprising and therefore on their propensity to bandwagon with the opposition movement.  

Informing Practice

The implications of this research are different depending on whether one interprets them from a state perspective, hoping to prevent defections, or from an opposition movement perspective, hoping to elicit them. The variables explored here—different forms of security force fragmentation—certainly privilege a statist reading, as these are variables within a government’s control, not within a resistance movement’s control. That said, those with an interest in the success of nonviolent resistance movements can still find something of value here for thinking through their strategy and how they might facilitate security force defections.

Most notably, the author highlights the importance of bandwagoning as a determinant of security force behavior when faced with an uprising. It is useful to remember, in other words, that it is not only ethical concerns about being ordered to fire upon unarmed, nonviolent activists that may motivate security force defections—as central as these may be—but also pragmatic assessments about which side security forces think is going to prevail in the conflict. The trouble is that security forces may simply assume that what counts as effectiveness against an armed uprising also counts as effectiveness against an unarmed, nonviolent uprising—causing them to miscalculate the chances of a nonviolent campaign’s success. With this in mind, the task facing nonviolent activists—but also public figures in academia and the media who have the capacity to influence narratives about movement effectiveness—is how to better inform security forces as to what it looks like when a nonviolent campaign is winning. If security forces perceive decisive military repression against an unarmed crowd of people as evidence of a nonviolent movement’s weakness and eventual defeat, they may decide not to bandwagon with the movement. Conversely, if they perceive armed elements within an opposition movement as indicative of that movement’s strength and eventual victory, they may decide to bandwagon with it. Both decisions, however, would be based on faulty assessments about the actual relationships between nonviolent and violent resistance and the likely success of such movements: the former scenario often provokes a dynamic called political jiu-jitsu, or backfire, that can ultimately shift public support and power towards the movement, and the latter scenario often weakens a nonviolent campaign by depleting broad-based support for it and making it easier to quell with traditional military force employed by the security forces who remain loyal and ever more steadfast in their commitment to defense against an armed adversary.

Ensuring that security forces have accurate information about the dynamics of nonviolent resistance is especially important, as security force defections themselves actually contribute to a nonviolent uprising’s success insofar as they are part of the power shift that ultimately makes it impossible for a regime to maintain control; defections, then, are both a reaction to the uprising’s likely success and a move that may further facilitate this success. Bandwagoning with a nonviolent movement (or not doing so) can, therefore, to some extent (although not always) be a self-fulfilling prophecy, making security force perceptions of their or a nonviolent movement’s effectiveness (whether accurate or not) a crucial factor to consider and address.

It is also crucial to address security force perceptions of the best way to support a nonviolent uprising if they do end up defecting, disabusing them of the assumption that they can best “help” by adding firepower to the movement. This move towards armed “defense” or “protection” of a nonviolent uprising again reveals a fundamental misunderstanding of the dynamics of nonviolent resistance and ultimately weakens a nonviolent campaign by taking away the moral high ground and the potential for political jiu-jitsu, at the same time making it easier for remaining security forces to justify violent repression against it, potentially causing a descent into civil war and massive civilian casualties. Rather, defected security forces can be more helpful to a nonviolent movement by publicizing their stories in a way that encourages and facilitates further defections, while also informing those in a position to help about the challenges or barriers facing those soldiers or officers who may still wish to defect. [MW]

Continued Reading

Bartkowski, M. (2020, September 29). For members of security forces: A guide to supporting pro-democracy movements. Retrieved on December 29, 2020, from

Chenoweth, E., & Stephan, M. J. (2014, July/August). Drop your weapons: When and why civil resistance works. Foreign Affairs. Retrieved on December 29, 2020, from

Neu, K. K. (2018, October 4). Do military defections help or hinder pro-democracy civil resistance? Retrieved on December 29, 2020, from

Nepstad, S. E. (2013, November 13). Civil resistance and military dynamics: Examining security force defections in the Arab Spring. ICNC Webinar. Retrieved on December 29, 2020, from


International Center on Nonviolent Conflict:

Key Words: nonviolent/civil resistance, security forces, defections

Next article The Status of Civil Society Influences Whether Democracy Aid Spurs Civil War
Previous article Experiences of Structural Violence in the Stories of Undocumented Latinas in the U.S.