Repression and Protest: The Limitations of Aggregation

Strategic Insights, Volume VII, Issue 2 (April 2008)

by T.J. O'Connell

Introduction

The relationship between state repression and protest is one of the most perplexing in social mobilization theory. Why in some cases does state repression succeed in suppressing organized protest and social movements, while in other cases it only serves to fan the flames of protest? Through various studies, different scholars have proposed and empirically supported contradictory theories that the relationship between the two is either an “inverted U,” positively linear, inversely linear, or curvilinear with a sharp increase at the extremes. A recent comparison of all approaches produced inconclusive results. In fact one expert in the field is doubtful the relationship will ever be resolved. While this may ultimately be correct, we will at least increases the chances of success and overall understanding of the issue if we eliminate the shortcoming present in the majority of the studies thus far, the imprecise naming and categorizing of variables, namely the aggregation of repression. This paper will attempt to shed some light on the issues in three parts. First I present a brief review of existing curve theories. Second, I proceed with a discussion of excessive aggregation, the common limitation of most of the current approaches. Lastly, I describe a new paradigm that deconstructs repression into more manageable and more useful elements.

Which Curve is Correct, if Any?

There is no consensus with regards to plotting the interaction between these two forces using state repression (alternatively called negative sanctions) as the independent variable on the x-axis and political protest (also measured as political violence, rebellion) as the dependent variable on the y-axis. One school of thought supports a negative linear relationship; as repression increases, political violence decreases, in other words, repression works. Based primarily on resource mobilization theory, collective action, of which political violence is a sub-set, is a function of excess time, money, energy, logistical support, organizational strength, opportunities and restrictions by the state. As repression increases, the opportunities decrease as constraints increase and resources in general are increasingly difficult to acquire and retain. As a result, collective action and political violence decrease.[1]

One study, building on the rational-actor model of social movement theory in general, assumes protestors will also behave rationally when faced with state repression based upon their immediate material costs and benefits. When there is little state repression while peaceful avenues to effect change are available, there will be little incentive to risk political violence because the same rewards can be achieved through actions with little or no cost, or at least much lower cost than violence. Likewise, if there is a high degree of state repression, the cost of political violence is too great and holds little benefit. However, the proponents argue, if there is only a moderate level of repression, the likelihood of political violence is high because there is no peaceful alternative, and the cost of violence is not prohibitive.[2] Plotting the rational-actor model results in an “inverted U” shaped relationship, with the highest level of unrest and political violence in countries with only moderate state repression.

In Why Men Rebel, Ted Gurr also supports the inverted U hypothesis, although for different reasons. Rather than protestors as rational actors, he builds upon the relative deprivation theory of mobilization, where those most likely to protest are not necessarily the poorest, or most deprived, but the ones with the largest gaps between expectations and realities.[3] The inverted “U theory” can also be considered the analogous to the “J-curve,” popularized by Ian Bremmer. In the J-curve hypothesis, states are placed along the x-axis in order of authoritarianism with completely closed police states, such as North Korea, at the very left; open advanced democracies, such as India at the right; and in the middle countries in transition such as Russia and South Africa.[4] Political stability—the lack of violent protest—is plotted on the y-axis. The result is a J-shaped curve, with police states experiencing a moderate amount of stability, states in transition experiencing the lowest level of stability, and fully open (least repressive) states with the greatest stability. If the elements of the J-curve are modified—measure the inverse of political stability—violent political protest, and the inverse of increasing openness—increasing repression, the result is the inverted U, albeit a little lopsided. In a sense, both curves measure the same phenomena using inverse parameters, but with the same general relationships.

In contrast to the negatively linear, inverted U, and initially positive but ultimately negative relationships discussed above, Marwan Khawaja’s study of the West Bank was the first to support a generalized positive relationship, where increased repression simply encourages more political violence.[5] By measuring elements critical to both resource mobilization theory and relative deprivation, both of which would have predicted, ultimately a decrease in political violence at the extreme levels of repression, he found violence actually increased. However, this positive relationship is also in doubt, as Khawaja acknowledges, if one considers the levels of repression experienced as only moderate. If Israel’s actions were categorized as only moderate, then the increased levels of violence could be interpreted to support the inverted U theory. Although, even this must be caveated by the finding that very low level of arrests resulted in sharp increases in protests.[6]

Another study, by Douglas Hibbs, concludes that repression and protest have both a positive and negative relationship by introducing time as a variable in the analysis. While trying to analyze Gurr’s relationship between deprivation (relative or absolute) and political violence, Hibbs discovered the outcomes varied as he altered the time-frame of the study. The result was an alternative repression-protest hypothesis where increased levels of state repression actually increased political violence in the short term, but in the long term increased repression actually reduces protests.[7] Lastly, Mason and Krane, found that while repression may serve to quiet protests, particularly harsh repression can incite protest.[8] This can either be interpreted as repression initially reduces protest, but over time increases protest (just the opposite finding to Hibbs’ positive, then negative relationships), or it can be used by some to support the inverted-U hypothesis, with a sharp rise at the very end. This has become known as the “backlash” hypothesis.

In contrast to the majority of the empirical studies, completed by researchers testing specific theories, usually their own, Ronald Francisco's study represents analysis by a disinterested party. The work’s primary utility to the field, according to the author, is his successful application of the predator-prey model. However, in the process, he impartially tests the validity of several repression-protest curve theories on three case studies.[9] The results are inconclusive. While “the inverted-U hypothesis receives less support than its “backlash” alternative,” the backlash hypothesis is only “supported weakly” in one of the cases.[10] Additionally, statistically speaking, repression and protests were only “significantly interdependent” in one of the three cases, with the other two showing “little statistical significance in the interactive terms.”[11]

With each scholar “successfully” proving a different curve that represents a particular hypothesis on the relationship between repression and mobilization, and “unbiased” scholars, who are more focused on analytical testing models, unable to determine the best curve, or even the statistical relevance of the relationship, there is clearly a lack of consensus and comprehensive delineation of the forces at work. Charles Tilly does not offer any optimism at the chances of resolving the confusion; “repression and mobilization regularly interact…but those interactions do not conform to covering laws.”[12] In fact he supports the idea that the interactions can be adequately explained at the level of individual events, but doubts that any new overarching theory will be developed to describe either multiple events or the grouping of like events.[13]

Common Error

One basic tenet of system dynamics modeling is that if one variable can both negatively or positively affect another, then it is not properly defined; there must be other intermediate variables that are at work and need to be addressed. This can easily lead to a model that is so complex and full of caveats that its utility is diminished. There needs to be a balance between on one hand, aggregated, generalized and simplistic models, and on the other, complex models that have little application beyond the studies they were built upon. One of the reasons the relationship between repression and protest is so muddied is that studies have erred too far on the aggregated, generalized and simplistic side. While all acknowledge the fact that there are more forces at play, few fully examine how including some of these additional elements might clarify the results.

In the struggle for a theory with universal application, as an increasing number of these variables move from an integral part of the analysis to footnotes, the accuracy and relevance of the theory decreases. Mark Lichbach addresses one half of the aggregation issue, primarily focusing on inaccuracies of aggregating levels of protests in the face of repression.[14] Such techniques he warns do not fully account for the shifts between violent and non-violent protests, adoption methods in light of government responses, and learned repertoires of the protestors as they gain experience. Likewise, other researchers have voiced concerns over excessive aggregation with respect to the type of repression and coercion used against protestors. In Francisco’s predator-pray study where three case studies were used to analyze multiple curve hypotheses, he admits “no single indicator of coercion can capture the full measure of its application to a population.”[15] Despite these recognized shortcomings, however, he relies solely on arrests, injuries and deaths in two cases, and deaths alone in the third case as his measures of coercion.[16] The more nuanced methods of state repression such as punishing protestors by taking away their children, burning down their houses, detention without trial, demotion to manual labor, restriction from job advancement, etc. are acknowledged to have taken place, but are not included in the analysis, instead they are relegated to footnote four.[17] How can we expect to accurately understand the mind, motivations, and influences on protestors when significant acts such as taking away their children or burning down their houses are ignored? While most researchers acknowledge the limitations of aggregation and generalization of repression and protest, few offer a framework for addressing the problem.

A New Look at Repression

As early as 1976 Snyder called for a more nuanced approach to repression addressing both violent and non-violent forms, and the timing of such actions.[18] When examining the impact of repression on protests, he argued, repression must be analyzed with these various criteria, rather than dealing with it at the aggregate. He did not, however, propose specific relationships between the type and timing of repression and its effect of protests.[19] Mohammed Hafez’s work Why Muslim’s Rebel answers Snyder’s call and provides a specific framework for the conceptualization and analysis of repression. Combining various frameworks from multiple scholars he devolves repression into two parts—lack of political access and physical repression—and breaks down physical repression further into timing and targeting. He concludes that it is the combination of lack of political access and state repression methods that are both reactionary and indiscriminate that will result in the greatest increase in protest, to the point of rebellion.[20]

Adopting the political process approach of Tilly, Tarrow and McAdam, Hafez takes the category of authoritarian state and breaks it into two separate components, political access and state repression which are not necessarily interconnected. States can independently chose to vary the level of political openness and the level of state repression, resulting in very repressive states that allow a specific level of participation in the process, and completely closed states that do not allow any participation in the political process. Goodwin has conducted detailed studies on the idea that the level of inclusiveness can be a determining factor in the success of revolutionary movements. He grades states on three levels: type of state organization (patrimonial/clientalistic to bureaucratic/rational), type of political regime (liberal/inclusive to repressive/exclusive) and infrastructural power (weak to strong).[21] He concludes that the states most vulnerable to political violence are exclusive/repressive regimes with patrimonial/clientalistic organization and weak infrastructural power.

Hafez borrows and simplifies Goodwin’s findings, content to limit the analysis of state regimes to inclusive/exclusive spectrum. The effects of political inclusion are opposition groups are well documented. Hafez summaries that “the more accessible the system the more likely Islamist movements will adopt accommodative strategies and shun violence over time,” likewise, even a repressive authoritarian state that allows some degree of inclusion is “less likely to unify opponents behind a revolutionary strategy or witness internal wars.”[22]

Previous studies attempted to measure the “level” of repression, or put a numerical value on the “harshness” which is inherently problematic. While most would agree that killing 1,000 people is harsh, and killing is harsher than arrests, one only has to try to determine the “harsher” state action between killing 1 person and arresting 1,000 to reach a “harshness” dilemma. Rather than harshness, Hafez adopts the concepts of other scholars and considers the timing and targeting of repression. With regards to timing, repression can be judged on a scale between preemptive or reactionary, while targeting varies between precise and indiscriminate.[23] Summarizing the previous works in these areas, he concludes that preemptive and precise repression by the state will likely succeed in reducing political violence, while reactive and indiscriminate repression will likely incite more political violence.[24] When combined with the analysis on political access, we can say that governments that allow no political access and practice reactive and indiscriminate repression are at an extreme risk for violent political protest and rebellion. Likewise, states that allow greater degree of political access can exert some degree of repression without fear of increased protests.

Instead of tracing a curve along the x- and y-axis, as proposed in over-aggregated studies, this new approach can be visualized as a cube with degrees of timing, targeting, and political access as the variables, where one corner representing the dangerous nexus timing targeting and political access. (See Figure 1) This treatment of repression goes a long way to explaining why repression can sometimes lead to rebellion and in other cases quiets the masses; it also helps explain why various researchers have reached different conclusions when analyzing the problem in a more generalized and aggregated context.

 Figure 1:

Conclusion

The existing theories on the relationship between repression and political mobilization are equally contradictory and inadequate. Independent tests of the various theories proposed offer little clarity as the results are often inclusive. This paper proposes that the common problem shared by most of the previous work on the subject is the oversimplification, or aggregation, of the repression variable. While others have promoted a further refinement of the “political protest” side of the relationship, only Snyder proposed refinement of repression into multilevel elements. Hafez, building on previous works involving the political process model, political inclusion, repression timing and repression targeting, deconstructs the repression variable as it was used in previous studies into three elements. Using the elements of timing, targeting and access, we can more accurately predict violent reactions to repression in specific situations and start to resolve the seemingly contradictory results of previous studies that analyzed repression as a one dimensional variable.

 About the Author

Maj Thomas J. O'Connell, USAF is a student in the National Security Affairs department at the Naval Postgraduate School. Maj O'Connell received his B.A. in Political Science from the U.S. Air Force Academy and M.S. in International Relations from Webster University. A career military pilot, he has flown the C-17 and KC-135 on combat missions in the Balkans, Afghanistan, and Iraq.

The views expressed are solely those of the authors and do not reflect the Naval Postgraduate School or the U.S. military .

 References

1. David Snyder and Charles Tilly, “Hardship and collective violence in France 1830—1960” American Sociological Review No 37 (1972): 527, and Charles Tilly, From Mobilization to Revolution (Reading MA, Addison Wesley, 1978) as cited in Marwan Khawaja, "Repression and Popular Collective Action: Evidence from the West Bank" Sociological Forum 8, No. 1 (1993): 47-49.

2. E. N. Muller and E. Weede "Cross-National Variation in Political Violence: A Rational Action Approach." Journal of Conflict Resolution 34, No. 4 (1990): 625.

3. Ted Gurr Why Men Rebel. (Princeton NJ: Princeton University Press, 1970). as cited in Khawaja, 48 and Muller, 626.

4. Ian Bremmer, The J-Curve—A New Way to Understand Why Nations Rise and Fall (New York: Simon and Schuster, 2006): 1-2.

5. Khawaja, 47.

6. Khawaja 62, 64-65.

7. Douglas A. Hibbs. Mass Political Violence. (New York: Wiley, 1973): 88-91, as cited in Muller, 626, 635.

8. Robert Francisco, "The Relationship between Coercion and Protest: An Empirical Evaluation in Three Coercive States." Journal of Conflict Resolution 2, No. 39 (1995): 265

9. Ibid., 263, 280.

10. Ibid., 263, 277.

11. Ibid., 277.

12. Charles Tilly. “Repression, Mobilization, and Explanation” in Repression and Mobilization, Christian Davenport, Hank Johnston, and Carol Mueller eds. (Minneapolis: University of Minnesota Press, 2005): 218.

13. Ibid., 211.

14. M. I. Lichbach, "Deterrence or Escalation? The Puzzle of Aggregate Studies of Repression and Dissent." Journal of Conflict Resolution 2, No. 31 (1987): 266.

15. Francisco, 270.

16. Ibid., 271.

17. Ibid., 270.

18. David Snyder “Theoretical and methodological problems in the analysis of government coercion and collective violence” Journal of Political and Military Sociology, No. 4 (1976): 288-90 as cited in Khawaja, 50.

19. Khawaja, 50.

20. Mohammed Hafez. Why Muslims Rebel (Boulder: Lynne Reiner Publishers, 2003), 72-76.

21. Jeff Goodwin. No Other Way Out. (Cambridge: Cambridge University Press, 2001). 27-29.

22. Hafez, 28-29. Other studies supporting this conclusion, specifically in the Middle East include: Jillian Schwedler, Faith in Moderation: Islamist Parties in Jordan and Yemen (New York: Cambridge, 2006), and Nathan J. Brown, Amr Hamzawy, and Marina Ottaway, “Islamist Movements and the Democratic Process in the Arab World: Exploring the Gray Zones,” Carnegie Papers: Middle East Series (67), and Carrie Wickham, “The Path to Moderation: strategy and learning in the formation of Egypt’s Wasat Party,” Comparative Politics, January, 2004.

23. For a more detailed study regarding the timing of repression see Charles D. Brockett, “A Protest-Cycle Resolution of the Repression/Popular-Protest Paradox” in Mark Traugott, ed, Repertoires and Cycles of Collective Action (Durham, NC: Duke University Press, 1995). For earlier studies on the targeting aspects of repression see T. Mason and D Krane, “The Political Economy of Death Squads: Towards a Theory of the Impact of State-Sanctioned Terror,” International Studies Quarterly 33 (1989): 175-198.

24. Hafez, 72-76.