Introduction

The Arab uprisings of 2011 increased scholarly interestin the relationship between security forces andcitizens in hybrid regimes (Barany 2011; Cook 2011;Haas and Lesch 2013; Kandil 2016; Lutterbeck 2013).On the one hand, deviant practices by members of thesecurity apparatus nourished popular grievances andundermined citizens’ trust in public institutions.While members of the police are supposed to maintainpublic order, police brutality cases have shown thatsecurity agencies could become the primary producersof ‘public disorder’ (Dias Felix 2020). On the otherhand, the way in which security agencies answer thechallenges created by mass mobilisations has adirect effect on the regime's survival as well asthe ability of protesters to achieve their demands(Goodwin 2001; Goldstone 2014; Goldstone and Ritter2019; McAdam, Tarrow and Tilly 2001; Slater2010).

During the Arab uprisings of 2011, the Tunisian andEgyptian security forces were quickly dispatched toprevent the diffusion of demonstrations (Hmed 2015).The strategies employed included the indiscriminatearrest of ordinary citizens, the surveillance andinfiltration of dissident groups and the use of‘thugs’ or agentsprovocateurs to disturb and discredit theprotests. More importantly, security agents usedlethal force against the protesters (Mekouar 2016).In the two cases, senior military members decided todeploy troops in the streets to stop the securityagencies’ exactions and preserve social peace. Theseinterventions precipitated the downfall of Ben Aliand Mubarak.

In Algeria, the 2011 demonstrations and the 2019–20Hirak protests didnot follow the same scenario. Although activists andjournalists faced legal harassment, intimidation andpre-emptive arrests, the use of large andindiscriminate lethal force remained limited. Infact, crowd management strategies have evolved sincethe early 2000s, when episodes of sustained popularmobilisation would still lead security agencies toopen fire on citizens. What made these changes inpolicing strategies possible? Despite the‘theoretical centrality of coercion’ (Greitens 2016:5) to explain the regime's stability, variations inrepressive practices have remained less studied(Chen and Moss 2018; Davenport 2007; Levitsky andWay 2012). We know that rulers can adapt to the newliberal order by diverting institutions to theirbenefit, multiplying channels of patronage andrenewing legitimisation sources (Heydemann 2007),but what about repression?