Military Defection

10 Nov 2016

Why do the members of security forces sometimes fight tenaciously for a person in power and at other times defect? Marianne Dahl believes that defections are more likely to take place in regimes where the incumbent applies a broad set of coup-proofing strategies against his or her own forces. These attempts also explain why regimes that seem extremely stable can rapidly collapse in the face of non-violent uprisings.

This article was external pageoriginally published by the external pagePeace Research Institute Oslo (PRIO) in 2016.

Why do members of the security forces sometimes fight tenaciously for the incumbent while other times defect? When large parts of the security forces defect and side with the protesters, the campaign is much more likely to be successful. This policy brief shows that security force defection is more likely to take place in regimes where the incumbent applied a broad set of coup-proofing strategies as a means of protection from the security forces. This provides insights into the puzzling observation that regimes that appear as extremely stable can rapidly collapse in the face of non-violent uprisings.

Brief Points

  • Military defection increases the likelihood of a successful campaign outcome. Non-violent campaigns that generate military defection are 46 times more likely to succeed than those that do not.
  • Coup-proofing strategies increase soldiers’ willingness and opportunity to defect.
  • Large-scale defection is much more likely to take place in regimes which rely on a heavy use of coup-proofing strategies.

‘No government has ever fallen before attackers until it has lost control over its armed forces or lost the ability to use them effectively’ (Brinton 1965). Without the loyalty of the military, even the most scrupulous dictator is turned into a lone-some man with nothing to defend himself but his own rifle. During the revolution in Iran in 1978–1979, a crucial part of the opposition’s strategy was to win the soldiers’ loyalty. Protesters handed flowers to the soldiers and chanted slogans such as ‘The army is part of the nation’ and ‘Brother soldier, why do you kill your brothers?’ (Kurzman 2005:114). The strategy was successful; soldiers started to defect from all parts of the military. By January 1979, approximately a thousand soldiers defected each day–some soldiers joined the op-position, while others simply went home to their families. To avoid soldiers being further exposed to the protesters’ message, and thus potentially liable to defect, military commanders loyal to the regime eventually decided to keep their troops well away from the protesters’ march routes. In effect, the very fear of military defection created a situation in which the military was reluctant to even engage the non-violent protesters. By the time the Shah decided to leave, it was no longer clear for how long he could count on the protection of the security forces.

Military defection, or security force defection, takes place when members of the security forces desert, instead of fighting for the regime, leaving the military and the conflict site, or leaving the military and joining forces with the opposition. As this has a significant effect on the likelihood of campaign success, we need to understand why the security apparatus sometimes decides to side with the protesters and other times it decides to defect.

This policy brief advances a theory to explain the micro-decisions of the security forces: Why do some individual soldiers decide to defect instead of fighting tenaciously for the incumbent? It is based on the presumption that by explaining the actions of individual soldiers we can also explain mass defection. Mass defection is more likely to take place under structures favouring individual defection. As many individual soldiers decide to defect it aggregates to mass defection. Moreover, as the security force leadership observes that they can no longer control the soldiers, they are more inclined to command the security forces to stop fighting for the incumbent and instead support the campaign.

Military Defection and Campaign Outcomes

The end of the communist regime in DDR (1989), Ferdinand Marcos’ regime in the Philippines (1986), Hosni Mubarak’s 30-year-long rule of Egypt (2011) and Zine El Abidine Ben Ali’s regime in Tunisia (2011) all ended due to a combination of civilian revolt and security force defection. Chenoweth and Stephan (2011) find that non-violent campaigns generating security force defection are 46 times as likely to succeed than those that do not. The relationship, however, is less clear for violent campaigns. Figure 1 shows the distribution of success and failure for mass non-violent and violent campaigns aiming to change the regime, secede or oust a foreign occupant between 1900 and 2006. While only 35 percent of the campaigns in total are successful, the number for those generating defection is 55 percent. Figure 2 shows the distribution only for non-violent campaigns. In total 54 percent of the non-violent campaigns are successful. For the non-violent campaigns that generate defection the number is as high as 91 percent. Indeed, only three non-violent campaigns that generated security force defection did not succeed. Figure 3 shows in which countries there were successful campaigns that generated defection between 1900 and 2006.

Non-violent and violent campaigns 1900-2006

There is a broad consensus about security force defection being a key determinant of campaign outcome. Yet, there are few studies of the causes of defection. Indeed, this is the first study to make a systematic analysis of the causes of military defection. In this policy brief, I ask: why does the military fight tenaciously for the incumbent in some regimes, but defect in others? To this end, I investigate how a set of coup-proofing strategies affects the likelihood of military defection. More specifically, I look at how promotion policies, rotation and reporting policies, and the division of the security forces affect soldiers’ willingness and opportunity to defect.

Coup D’etats and Military Defection

Coup d’etats and military defection constitute the two most severe forms of military disloyalty. On the basis that coup-proofing strategies increase the general level of loyalty, one might presume that coup-proofing strategies not only reduce coup risks, but also enhance the loyalty of the security apparatus during times of internal upheaval, and thereby reduce defection risks. Indeed, parts of the literature treat coup d’etats and defection as similar phenomena. Somewhat surprisingly then, many of the countries where the Arab Spring generated military defection were countries in which a broad set of coup-proofing strategies over decades had secured the loyalty of the armed forces and prevented them from staging a coup.

Comparing the micro-mechanisms of coup d’etats and military defection shows that these are fundamentally different phenomena, and military defection should indeed be more likely to take place in regimes that apply a broad set of coup-proofing strategies. The main difference is that for coups, the destiny of the coup plotters hinges entirely on whether the rest of (or most of) the security apparatus follows their lead. In contrast, the destiny of defecting soldiers does not to the same degree depend on whether the rest of the security apparatus follows. This implies that coups need to be coordinated while defection does not. As the logic of coup-proofing strategies depends on the need for coordination, it follows that they will not deter defection. To the contrary, as coup-proofing deteriorates the military institution, increases personal as well as group grievances and reduces the likelihood that defecting soldiers will be detected, and subsequently punished, it increases the likelihood of military defection.

The two main differences between coup d’etats and military defection are that (1): Coups take place at the centre of political power. Thus, if the coup attempt fails, few if any escape opportunities exist. Hence it is almost a given that coup plotters will be punished. Normally, the punishment for a coup attempt varies from exile, at best, and execution at worst. Only rarely are coup plotters granted amnesty. For defection the opposite is the case: Defection can happen anywhere, and often takes place far away from the centre of power. Escape opportunities are therefore much easier to find. And (2): For military defection during civil resistance, there exists a third party defectors can join. Defecting soldiers can join the civil society campaign, potentially being offered protection, food and shelter by members of the campaign. This is seldom if ever the case for coup plotters. As a result, while it might be preferable if the rest of the security forces join the defectors, it is not a necessity as it is for the coup plotters. Indeed, no sound man would ever try to commit a coup that he believed no other soldier would support, but a soldier could switch sides and defect, even if he believed that no one would follow his actions.

The absolute need for coordination in a coup has consequences which coup-proofing strategies rely on. Only a few people are likely to be in the position where they can effectively initiate and organize a coup. Being able to do so depends on having an extensive network. While individuals in possession of this can exist in all branches of the security apparatus, they will tend to cluster in top positions. As a result, a common group of coup-proofing strategies aim to make key security force personnel unwilling to stage a coup. This involves recruiting people to top positions on the basis of ethnicity and kinship, as well providing material benefits exclusive to those in top positions. This should not only reduce coup risks, but also defection risks amongst key security personnel, as the strategy is likely to increase their general level of loyalty to the regime. However, as defection does not hinge on coordination anyone can choose to defect. As a result, as these strategies do not increase the general level of loyalty amongst the rest of the security personnel, the majority of the security apparatus can be both willing and able to defect. As I argue below, these very strategies should increase the likelihood that non-top security personnel decide to defect.

The second type of coup-proofing strategies are those that aim to make it more difficult to coordinate a coup. One way to do so is by setting up parallel military units and thereby increasing coordination problems. By doing so, the incumbent can play the different units up against each other. Should one unit decide to plot a coup, the dictator can potentially use one of the other security units to defend his position. Another strategy that rulers employ is the frequent replacement and rotation of commanders of the armed forces. Again, as defection does not hinge on coordination, this is unlikely to deter defection risks.

Non-violent campaigns 1900-2006

Explaining Military Defection

So what determines whether an individual soldier decides to fight for the regime or defect? Following a rational choice model (see Gates 2002), the decision should be a function of (1) the utility of cooperation, (2) the utility associated with defection and (3) the likelihood of being captured and punished.

Successful campaigns that generated defection 1900-2006

The utility of cooperation is a function of pecuniary and non-pecuniary rewards. Pecuniary rewards consist of wages, access to housing, education etc. Non-pecuniary rewards relate to the joy of fighting the “good fight” and being part of the military organization. The opportunity costs are related to the benefits a soldier could gain if choosing to defect. The punishment for defection is plausibly execution, but this depends on the soldier being captured. If the defecting soldier is not captured, he or she will not be punished. Hence, the security apparatus’ ability to monitor the actions of each soldier is decisive for decisions regarding whether or not to defect. Thus, any structure that (1) lowers the likelihood that defecting soldiers are detected, (2) reduces the benefits of being part of the security apparatus or (3) increases the opportunity cost of being a member should cause a higher likelihood of military defection.

Promotion Policies aimed at making those in top positions unwilling to carry out a coup comes at the cost of favouring one group over another. As such, officers are systematically promoted to key positions based on loyalty instead of skills. One disadvantage of such a promotion strategy is that it creates a large share of out-group soldiers, who face few or no promotion opportunities. One consequence of such a promotion strategy is that both skilled and lesser skilled soldiers who do not belong to the group from which officers are promoted face few or no opportunities within the system. While it strengthens the preference for the current regime at the top of the organization, loyalty does not trickle down the system. To the contrary, being deprived of promotion possibilities tends to increase grievances, and reduce soldiers’ preference for the sitting regime. As a consequence, the utility of continued cooperation amongst those who belong to the out-group should be lower than it would have been had it not been for this coup-proofing strategy. Moreover, as top personnel is less competent under such a regime, they are also less likely to efficiently monitor the actions of soldiers. As a result, the likelihood of defecting soldiers being captured is reduced. This should increase defection risks.

Splitting up the security apparatus should both reduce the utility of continued cooperation and the likelihood of being punished if choosing to defect. The division of the security apparatus tends to generate grievances and competition between the different sections of the security forces. As a result, marginalized parts of the security apparatus should be less willing to fight for the incumbent, as their benefit from being part of the security system is relatively low. Moreover, it makes it harder for information to flow freely. As a result, information about soldiers who plan to defect is more likely to not reach those who could potentially stop this. This also lowers the risks taken by the individual soldier when deciding to defect. In fact, it might make the military unable to crack down on incidents of defection. In addition, other soldiers could potentially observe this, in turn concluding that defection is relatively risk free.

A last method of lowering coup risks is to inhibit alternative power bases from developing. This is done by rotating and dismissing top officers. However, similar to the argument above, this strategy reduces the ability of the leaders of security apparatus to exercise control over the organization, and should accordingly lower the risk each defecting soldier takes on. As a consequence, the likelihood of defection increases.

As there only exists one measure on coup-proofing, the division of the security forces, I use regime type as a proxy for the broader set of coup-proofing strategies that I look at here. A broad set of scholars have documented that personalist regimes tend to make the most extensive use of coup-proofing strategies, followed by monarchies, militaries and lastly single-party regimes. Amongst authoritarian regimes, single-party regimes are known to have the most professionalized security apparatus. Hence, military defection should be most likely to take place in personalist regimes and least likely to take place in single party regimes.

Coup-Proofing and Military Defection

Analysing all maximalist non-violent and violent campaigns shows that military defection is more likely to take place when the security apparatus is divided. The variable Division of the Security Forces is an index of the ratio of troop members in the paramilitary vs. the regular military, and is provided by Pilster and Bömelt (2011). Figure 4 shows the effect of having a divided security apparatus on the likelihood of military defection. The figure shows that when all other variables are set to their mean, if a regime scores 1 on Divided Security Apparatus, this is one with an undivided security apparatus, the likelihood of defection is 15 percent. If instead a regime scores 3, the likelihood increases to almost 40 percent. For a regime with a score of 4 the likelihood is above 50 percent. The average score for the variable Divided Security Apparatus in the Philippines during the non-violent campaign against Ferdinand Marcos (1983–1986) was 4.15. That is the highest score observed among countries included in the NAVCO dataset. This campaign did indeed successfully generate military defection among the security apparatus.

The division of the security forces and military defection

The analysis also shows support for an effect of regime type on military defection. The regime data is provided by Geddes, Wright and Franz (2013), and divides authoritarian regimes into militarist, single-party, personalist regimes and monarchies. All models support the theory that military defection is far more likely to take place in personalist regimes than in single-party regimes. The estimates indicate that military defection is about 86 percent less likely to take place in single-party regimes than in personalist regimes. This indicates that there is a positive relationship between applying the coup-proofing strategies studied here and the likelihood of military defection.

Conclusion

This analysis suggests that while the incumbent can protect himself from coup risks by applying coup-proofing measures, this comes at a cost. One cost of coup-proofing measures is that it erodes the organization of the security forces, and increases both the willingness and opportunity for individual soldiers to defect. From the perspective of civil society, this might constitute an opportunity. Regimes that appear as stable, where the incumbent has consolidated power and faces a very low risk of a coup, might not be at all immune to defection. As civil society is able to mount a campaign (non-violent or violent) against the regime, thousands of individuals’ decisions to defect may spell the end of the regime–the dictator is overthrown by a thousand cuts.

References

Brinton, Crane (1965). The Anatomy of Revolution. 3rd ed. New York: Vintage Books.

Chenoweth, Erica & Maria J. Stephan (2011). Why Civil Resistance Works: The Strategic Logic of Nonviolent Conflict. New York: Colombia University Press.

Gates, Scott (2002). ‘Recruitment and Allegiance: The Microfoundations of Rebellion’, Journal of Conflict Resolution 46(1): 111–130.

Geddes, Barbara; Joseph Wright & Erica Frantz (2013). ‘New Data on Autocratic Breakdown and Regime Transitions’.

Kurzman, Charles (2005). The Unthinkable Revolution in Iran. Cambridge, MA: Harvard University Press.

Pilster, Ulrich & Tobias Bohmelt (2011). ‘Coup-Proofing and Military Effectiveness in Interstate Wars, 1967–99’, Conflict Management and Peace Science 28(4): 331–350.

About the Author

Marianne Dahl is Doctoral Researcher at PRIO. Her research interests includes: Non-Violence, Civil War Theory, Post-Conflict Stability, Political Economy and Statistical Modeling. Twitter external page@climatesecurityexternal page 

For more information on issues and events that shape our world, please visit the CSS Blog Network or browse our Digital Library.

JavaScript has been disabled in your browser