Patterns of Mobilization and Repression in Iraq’s Tishreen Uprising

Chantal Berman, Georgetown University; Killian Clarke, Harvard University; and Rima Majed, American University of Beirut

In the months before the coronavirus pandemic forced the entire world into lockdown, Iraq was undergoing one of the most sustained periods of anti-government protest since the 2003 US-led invasion. This “Tishreen” uprising – named for the month in which it began (October/Tishreen al-Awl 2019) – gave expression to years of pent up frustrations among Iraqi citizens. The uprising was triggered by two events. On September 25th, 2019 a group of university graduates gathered in front of the Prime Minister’s office, protesting their inability to secure jobs, were met with heavy repression. Then, two days later, on October 27th, the Prime Minister demoted a popular general, Abdul-Wahab al-Saadi, who had helped to defeat the Islamic State. The move was broadly interpreted as a capitulation to corrupt politicians, possibly under pressure from Iran. Together, these two events sparked the Tishreen uprising, which was spearheaded primarily by young Iraqis and condemned the regime’s endemic corruption, sectarian polarization, and slavish obedience to foreign powers like Iran and the United States.

The Tishreen uprising was only the latest manifestation of popular unrest in Iraq. Since the end of the post-invasion civil war, Iraq has experienced multiple waves of mobilization – in 2011, in 2015, and in 2018 – all of which aired a similar constellation of demands. In this sense, the Tishreen uprising was the culmination of a decade of mobilization in which Iraqis denounced, with increasing forcefulness, the dysfunctional political system that was set up following the 2003 invasion.

In this brief memo, we explore the first ten weeks of this revolutionary uprising – from its outbreak on October 1, 2019 to the end of the first week of December. Using an original event catalogue on protests and repression during the uprising, we explore the trajectories of popular mobilization as well as the nature of state-protester interactions. Our data reveal a number of noteworthy patterns. We find that protesters largely resorted to non-violent tactics, though in the early weeks of the uprising there were a considerable number of events in which unarmed protesters mobbed and burned buildings associated with various maligned parties and militias. These early events triggered a massive repressive response from Iraq’s political establishment. We find that the majority of this repression was perpetrated by official state security forces, but that elites also called on non-state repressive agents, like thugs and militias, when undertaking especially violent crackdowns. Finally, we see evidence that these repressive responses backfired. The more the political establishment cracked down on protests, the more outrage it triggered, resulting in fresh rounds of mobilization.

These preliminary findings hold a number of implications for our understanding of popular uprisings, particularly ones that target nominally democratic but highly corrupt and/or sectarianized regimes. Such uprisings have become increasingly common in recent decades,[1] yet they are often analyzed using theories derived from the study of revolutions against closed autocracies. Our research shows that such uprisings have the potential to be just as violent and intense as those against authoritarian regimes, in part because these kinds of divided regimes can rely on thugs and militias tied to particular factions in the government to undertake their most egregious acts of repression.[2] These systems, ultimately, have their own distinct ways of resisting change-from-below, suggesting that the long-running conflict between Iraq’s political elites and its mobilized citizenry may be just as “frozen” as the other conflicts analyzed in this briefing.

Overview of the Data

The data we analyze include 1,191 contentious events from September 1, 2019 to December 7, 2019. This includes the first ten weeks of the Tishreen Uprising, plus the four weeks directly preceding it. A contentious event is a public, collective, and voluntary endeavor to influence the actions or policies of some authority.[3] It therefore includes protests, demonstrations, strikes, marches, sit-ins or occupations, roadblocks or blockades, boycotts, petitions, and mass attacks.

One of the major challenges in assembling protest event datasets is identifying sources that provide detailed coverage of protests in a consistent and unbiased way. In Iraq this challenge was accentuated by a media landscape that is highly fragmented and subject to frequent government suppression.[4] We therefore relied on a variety of sources for our data. First, we draw from the newspaper Al-Mada, which is known for its professionalism and national coverage. Second, we rely on two leftist newspapers, which offer strong coverage of protests, strikes, and other forms of popular resistance: Ila al-Amam and Tareeq al-Shaab. Third, we include all events collected by Liveaumap, which sources data on conflict events and natural disasters in various countries, primarily from social media.[5] Finally, we include all protests identified by the Armed Conflict Location & Event Data Project (ACLED), which relies on a combination of local and international news sources.[6] For each event we coded a number of relevant variables, including the location of the event (with GPS coordinates), the number of participants,[7] the demands, the tactics, numbers killed and injured, and the nature of repression.

The Arc of Protest in the Tishreen Uprising

Figure 1 displays the number of daily events, organized by the start date of the protest. As the figure reveals, the uprising can broadly be divided into three phases. September represents the pre-uprising period, when overall levels of contention in Iraq were fairly low.

Figure 1: Daily Count of Protests

Most of the events that occurred during these months involved labor strikes or student sit-ins related to unemployment. Then, on September 27, General al-Saadi was demoted, and the first major day of protesting occurred four days later, on October 1. Mobilization was intense for the following week, and was met with fierce repression (as we show below), before eventually subsiding by the middle of October. Phase 3 of the uprising then began on October 25. For weeks activists had been calling on social media and other forums for a resumption of protest on this day. These organizing efforts yielded 21 protests in 11 different governorates, and these events in turn catalyzed weeks of elevated and sustained protests, which ultimately forced Prime Minister Adel Abdul-Mahdi to resign on November 29. But our data show that the Prime Minister’s resignation did little to mollify protesters’ anger, which by then had broadened to include not just the corruption and ineptitude of the government but also its brutal response to the mobilization.

Figure 2 tells a similar story, this time showing the geographic location of protests. In the month preceding the uprising protests were scattered across Iraq, with no particularly focal point or area of concentration. In contrast, we see that the initial outburst of protest in early October occurred mainly in Baghdad, though it did spread in limited numbers to some southern cities like Najaf and Basra. But by phase 3 of the uprising, revolutionary mobilization had engulfed virtually the whole country (though they were most concentrated in Baghdad and the Shia-dominated southern governorates).

Though the Tishreen uprising channeled a broad range of social and economic discontents that had been percolating in Iraqi society for years, it was still mainly a political revolution demanding the end of Iraq’s corrupt, patronage-based system of rule.

Figure 2: Geographic Location of Protests

The nature of these demands is reflected in Figure 3, which shows the distribution of demands for each week of the uprising.[8] The month before the uprising was dominated by low-level protests airing social and labor demand, which had grown increasingly common in Iraq since 2018.[9] But from week 5, when the uprising began, to week 13, when the Prime Minister resigned, most events (82%) were staged in overt opposition to the government. Interestingly, even though the uprising was also a rejection of various Shia-dominated political parties and their affiliated militias, which had come to dominate Iraqi politics and were widely blamed for its dysfunction, we see very few events where protesters explicitly condemn these parties. Instead, protesters chose to broadly denounce “the government” or “the system;” essentially lumping the entire political elite together as part of a single maligned establishment.

We begin to see a shift in protester demands during the first week of December, when only 19% of events aired opposition to the government, whereas 66% focused on human rights abuses. In part this shift was a response to the government’s resignation, which essentially met one of the protesters’ main demands. But, as we discuss below, it also reflected a deepening and expansion of grievances, as protesters turned their attention to government abuses, and specifically to several major acts of repression that had been committed in the previous weeks.[10]

Violence and Repression

For an uprising against a nominally democratic government, the Tishreen uprising was met by a shocking amount of violence.

Figure 3: Distribution of Demands by Week

We document nearly 700 deaths and 11,500 injuries during the ten weeks under analysis, nearly all of which were tied to state or non-state repression.[11] But the Iraqi political elites’ vicious reaction to the mobilization makes more sense when we bear in mind its entrenched corruption and extreme fragmentation. A movement demanding wholesale political change represented a real threat to the system of cronyism and rapaciousness that has enriched Iraq’s politicians over the last two decades, and these elites quickly mobilized an array of state and non-state security agents in an attempt to quash this challenge.

Figure 4 plots the number of deaths that occurred on each day of the uprising. While these patterns broadly map onto the protest patterns represented in Figure 1, there are also clear differences, with far more jumps and spikes in violence. The first week of October saw a total of 110 deaths – at the time this was considered a shocking number, though these casualties would soon be greatly surpassed. The most deadly day of the uprising was October 25, the day on which the uprising resumed, when 82 protesters were killed. Though we document only 21 events on this day, the amount of violence inflicted on protesters was extreme: 14 were killed at a protest in Maysan outside the headquarters of the Iranian-backed Asaeb Ahl al-Haq militia; a total of 24 protesters were killed in Thiqar province when they attacked the headquarters of the Badr Organization, another Iranian-backed militia; 12 more were killed in Diwaniya after setting several party buildings on fire.

While the violence on October 25 was widespread, stemming from multiple events that were repressed, other spikes in violence can be tied to the brutal repression of a single event.

Figure 4: Daily Count of Deaths

The first of these massacres, as they came to be known, occurred on October 28 in Kerbala, when security forces used live ammunition to disperse a sit-in at the Tarbiya roundabout, killing 18 and injuring more than 800.[12] Another pair of massacres occurred at the end of November, in the cities of Najaf and Nasiriya, killing a combined 91 people and injuring more than 700. These two massacres were, notably, perpetrated not only by the police but also by non-state militias, like the al-Abbas Brigade, and by anonymous thugs (beltegeyya).[13] We further find evidence that these attacks triggered fierce backlash among protesters, catalyzing new waves of mobilization over the government’s disregard for human rights.[14] In Figure 3 above we saw some evidence of these backlash dynamics with the high share of human rights related protests in the first week of December. Even though the government had resigned the week before, protesters remained in the streets to denounce the atrocities committed in the Najaf and Nasiriya massacres.

While the most extreme violence of the uprising was clearly committed by political elites, protesters themselves did resort to certain limited tactics of violence, particularly at the beginning of the uprising. Figure 5 plots the percent of events per week in which protesters used some form of violence.

The figure reveals that the initial outburst of contention during early October was actually marked by considerable violence – 70% of events in week 5 and 60% of events in week 6 of the uprising involved protestor violence. Week 8, when the uprising resumed, was also marked by more violence.

Figure 5: Share of Events Involving Protester Violence, by Week

But then, as mobilization escalated and broadened in the subsequent weeks, the proportion of events involving protester violence declined, occurring at between 15% and 30% of events.

The use of violence by protesters somewhat cuts against media representations of the revolt as largely non-violent.[15] But it is important to correctly characterize the nature of this violence. First, it was almost entirely unarmed. We documented only three events in which protesters resorted to armed combat. Moreover, scholars have shown that unarmed violence can be a highly effective strategy of resistance, especially when confronting extreme levels of state repression and when combined with non-violent tactics.[16] Second, most of the violence was either undertaken in self-defense or it took the form of mobbing buildings, particularly the offices and headquarters of key parties and militias that protesters blamed for essentially hijacking the Iraqi state. Even though, as noted above, protesters rarely invoked the names of these groups as they aired their demands, they seem to have had no problem physically targeting their properties. In early October, and again on October 25, protesters mobbed, attacked, and burned at least one hundred militia and party buildings, including those belonging to the Badr, Asaib Ahl Al-Haq, and Sayyidd al-Shuhada militias, and the Daawa and Hikma parties.

As we have seen, the Iraqi security forces and their non-state affiliates responded to the protests with disproportionate violence. But when did most of this repression occur, and who was responsible? Figure 6 plots the share of events that were repressed, respectively, by the police and by non-state actors like thugs, militias, and party members for each week of the uprising.

Figure 6: Share of Events Repressed, by Week

The figure reveals that the period of greatest repression was at the initial outburst of protest, when 69% of events elicited some kind of repressive response. Week 8 was also a period of heightened repression. These trends correspond closely to the periods of greater protester violence, suggesting that the early period of the uprising was marked by a bloody cycle of protesters’ mobbing physical properties, followed by disproportionate violent crackdowns.

Much of the coverage of the Tishreen uprising has emphasized the participation of non-state militias and thugs in attacking protesters.[17] But our data reveal that, in fact, most of the repression during the first ten weeks was committed by the police and other state security forces. In total; thugs, militias, or parties were involved in repressing 29 events in our dataset, whereas the police repressed a total of 298. However, further analysis reveals that while thugs and militias may have been used sparingly to put down contention, they may also have been used in a highly strategic way.[18] Figure 7 shows the average number of people killed and injured in protests repressed by the police and in protests repressed by thugs, militias, or parties. It suggests that political elites may have deployed these non-state repressive agents only for the most brutal and egregious forms of repression, when they wanted to have a degree of plausible deniability about their knowledge of or participation in the attack. For example, these types of agents were involved in all three of the massacres noted above, in Kerbala, Najaf, and Nasiriya. We also encountered evidence of these two types of repressive agents hybridizing, for example, when state security forces were commanded or led by a representative from a militia or a political party.

Figure 7: Average Deaths and Injuries per Event Repressed, State and Non-State Agents


The picture of violence and repression painted by these data is rather grim. Protesters during the Tishreen uprising have not been protected by the type of liberal institutions and norms of restraint that normally characterize state-challenger interactions in democracies. Instead, anti-protester violence seems to have been at least partly facilitated by the Iraqi regime’s fragmented and sectarianized structure. While the police have engaged in more routine repression, factions and parties who felt particularly threatened by the uprising have deployed their agents at strategic moments to engage in brutal crackdowns. Moreover, the unofficial nature of these assailants has given their sponsors a degree of anonymity and plausible deniability. Nevertheless, Iraq’s protesters have been savvy in their responses. They have targeted the physical properties associated with these pseudo-state and non-state groups, even if they have rarely invoked their names as they aired their grievances. And they have mobilized in response to these massacres, generating further cycles of contention.

Ultimately, then, the conflict between state and society in Iraq, like so many others in the Middle East today, appears to be frozen. On the one hand, Iraq’s political system remains surprisingly resistant to reform, relying on the increasingly sophisticated strategies of violence discussed in this essay, as well as other mechanisms of reproduction and resistance (see, for example, Sara Kayyali’s essay in this briefing, which discusses similar mechanisms in Syria and Lebanon). But, on the other hand, without major reforms Iraqis’ unmet grievances will continue to percolate, and will surely, once the pandemic ends, prompt them to take to the streets once again.


[1] Mark R Beissinger, The Urban Advantage in Revolution (Princeton University Press, Forthcoming).

[2] For a similar argument regarding Lebanon’s 2019 uprising see: Chantal Berman and Killian Clarke, “Deposing a Democracy: Lebanon’s 2019-20 Thawra in Comparative Perspective,” Bobst-AUB Research Note, February 17, 2020.

[3] This definition is derived from: Charles Tilly, The Contentious French (Harvard University Press, 1986).

[4] Katya Schwenk, “The Iraqi Media Landscape in Uncertain Times,” Inside Arabia, January 6, 2020.

[5] The Livemap for Iraq is available at:

[6] In its methodological reports, ACLED often emphasizes the comprehensive and unbiased nature of its coverage. But for this period in Iraq, we found that ACLED had captured only about half of all events, and that these tended to skew heavily toward major cities like Baghdad, and larger, more violent events. For further discussion of bias in ACLED’s data see: Killian Clarke, “Overthrowing Revolution: The Emergence and Success of Counterrevolution, 1900-2015,” PhD Dissertation, Princeton University.

[7] Information on participants was unfortunately unavailable for roughly two-thirds of events.

[8] The figure excludes Week 7, in mid October, when there was only one event.

[9] Hashim al-Rikabi, “The Rising Tide of Change in Iraq: An Assessment of the 2018 and 2019 Protests,” Arab Reform Initiative, 25 November 2019.

[10] On the escalation of demands during the uprising see: Zahra Ali, “Iraqis Demand a Country,” Middle East Report 292/3 (Fall/Winter 2019).

[11] These numbers broadly align with other accounts of violence during the uprising. See, for example: Amnesty International, “Iraq: Protest death toll surges as security forces resume brutal repression,” January 23, 2020.

[12] Amnesty International, “Iraq: Horrific scenes as security forces resort to lethal force to disperse Karbala protests,” October 29, 2020

[13] On the Nasiriya massacre see: Amnesty International, “Iraq: Eyewitness describes ‘street filled with blood’ as at least 25 protesters killed in security force onslaught,” November 28, 2019.

[14] On the potential for repression to trigger backlash, see: Ronald A. Francisco, “Coercion and Protest: An Empirical Test in Two Democratic States,” American Journal of Political Science 40, no. 4 (1996): 1179–1204; Karen Rasler, “Concessions, Repression, and Political Protest in the Iranian Revolution,” American Sociological Review 61, no. 1 (1996): 132–52.

[15] Although for some further discussion of protester violence see: Benedict Robin-D’Cruz, “Violence and Protest in South Iraq,” LSE Middle East Centre Blog, 18 August, 2020.

[16] Mohammad Ali Kadivar and Neil Ketchley, “Sticks, Stones, and Molotov Cocktails: Unarmed Collective Violence and Democratization,” Socius 4 (January 1, 2018).

[17] Omar Sirri, “Revolutionary Protests Spread Across Central and Southern Iraq and Are Ongoing,” Jadaliyya, October 26, 2019; Maria Fantappie, “Widespread Protests Point to Iraq’s Cycle of Social Crisis,” International Crisis Group Commentary, October 10, 2019.

[18] See also D’Cruz 2020.