Rio de Janeiro’s favela assemblage: Accounting for the durability of an unstable object

What are assemblages?

At the most basic level, assemblage simply denotes “the ‘holding together’ of heterogeneous elements” (Deleuze and Guattari, 1987: 323). However, the varied ways in which Deleuze and Guattari invoked the concept at different points across their diverse works has left considerable doubt about how they understood assemblages as emerging, transforming and disbanding over time. For example, the symbiotic evolution of the wasp and the orchid through a mutual “capture of code” (Deleuze and Guattari, 1987: 11), appears to occur contingently and without external impulse in a process analogous to the notion of “entrainment” in complexity theory (Bonta and Protevi, 2004: 404). By contrast, the “man-horse-bow” assemblage (Deleuze and Guattari, 1977), formed by the “nomadic war machine” of the Asian Steppe, suggests assemblages are over-determined responses to external structuring conditions. In the words of Deleuze and Guattari: “there are always machines that precede tools, always phyla that determine at a given moment which tools, which men will enter as machine components in the social system being considered” (Deleuze and Guattari, 1977: 119). This distinction underpins some recent “discontent” (Buchanan, 2015) regarding DeLanda’s (2006, 2016) efforts to establish a formalised post-Deleuzoguattarian “assemblage theory”. These debates also have important implications for thinking about the importance of scale, power and capital in the formation of socio-spatial assemblages, and links to recent debates in urban theory (see below). ²

Given the disagreements about the precise way in which Deleuze and Guattari intended assemblage to be used and how best to proceed with the concept today, it may be more helpful to use the term “assemblage thinking”, to denote a set of core insights and analytical orientations, as opposed to a formal “assemblage theory” that would need to be adopted in its entirety. Though far from exhaustive, I will identify three core features of assemblage thinking that I think those who use the term would largely agree upon and which forms the basis for the analysis that follows. These are: (1) “relations of exteriority”; (2) anti-essentialism; and (3) the notions of (de/re)territorialisation.

DeLanda (2006) helpfully pins down Deleuze and Guattari’s machinic – as opposed to organismic – vision of society as being constituted by “relations of exteriority”. Organismic metaphors, characteristic of functionalism, and in a different way the notion of “totality”, derived from Hegel and central to Marxian theory, rest on an imaginary of “relations of interiority”, that postulates the social world as “a seamless web of reciprocal action, or as an integrated totality of functional interdependencies, or as a block of unlimited universal interconnections” (DeLanda, 2006: 19). DeLanda argues that such “macro-reductionism”, denies the possibility of emergence, because if the “role” of a part within the functioning of a whole is determined a priori, it cannot be seen to possess the capacity for generative interaction with other entities. By contrast, a component of a machine, depending on the capacities it possesses, may be detached and reconnected to other machines. In this process, although the meaning of the part is transformed, it preserves its own autonomy and its potential for at some point becoming part of different assemblage (Nail, 2017: 23). This does not mean that processes of integration between multiple entities may not occur in assemblages, even to the extent that individual parts lose most of their independence. However, even in these cases the relations between them must be seen as, “not logically necessary but only contingently obligatory: the historical result of their close coevolution” (DeLanda, 2006: 11–12).

A second key feature of assemblage thinking is its opposition to “essentialism” (DeLanda, 2006: 26–47; Nail, 2017: 23–24). Aristotelian “taxonomic essentialism” proposed a three-level ontological hierarchy of the genus, the species and the individual, with each constituting a kind of universal template (DeLanda, 2006: 26–29). Any given entity, at any of these levels, could be regarded as possessing an “essence”, or particular set of distinguishing characteristics that marked it out as unique. By implication, any internal diversity within a group would be understood as random variation around a central prototype. This understanding is belied by evolutionary theory, however. Species do not have timeless essences, but are instead assemblages of components that have been temporarily stabilised through reproductive isolation and are still liable to change through the mechanisms of genetic mutation and natural selection. This means that not only are they constantly undergoing “morphogenetic” transformation, but that this process integrates both the constraints imposed by inheritance and the often highly contingent influences of the environment and the other entities with which it interacts (as in the example of the wasp and the orchid). As such, a species should be seen as “an individual entity, as unique and singular as the organisms that compose it, but larger in spatio-temporal scale” (DeLanda, 2006: 27). Unlike taxonomic essentialism then, “the ontology of assemblages is flat since it contains nothing but differently scaled individual singularities” (DeLanda, 2006: 28).

Leading on from this, a third key feature of assemblages is the processes of territorialisation–deterritorialisation–reterritorialisation through which they emerge, hold together, transform and, potentially, decompose. The initial “territorialisation” of assemblages establishes them as relatively stable formations, whose structure is reinforced by its internal and external relationships (Deleuze and Guattari, 1987: 508–510). These assemblages will exhibit perpetual “deterritorialising” tendencies (also called “lines of flight”), whereby components of the assemblage seek to break away from it. However, these will typically be “reterritorialised” by the overriding force of the assemblage, through the power relations and the self-perpetuating routines it contains. These competing tendencies produce feedback loops whereby change continues but without allowing the assemblage to decisively break apart. The latter only occurs in cases of “absolute deterritorialisation”, when deterritorialising tendencies ultimately overwhelm the reterritorialising ones. As Deleuze and Guattari note, it is the varied and complex dynamics of specific assemblages that determine their durability:

There are lines of articulation or segmentarity, strata and territories; but also lines of flight, movements of deterritorialization and destratification. Comparative rates of flow on these lines produce phenomena of relative slowness and viscosity, or, on the contrary, of acceleration and rupture. All this, lines and measurable speeds, constitutes an assemblage. (3–4)

Assemblage and critical urban theory

While assemblage thinking has been taken up enthusiastically in geography and urban studies in recent years, it has also been met with resistance in some quarters. For example, in their instructive exchange, Brenner et al. (2011) rejected McFarlane’s (2011) suggestion that assemblage thinking could help to renew and extend the project of critical urban theory. Brenner et al., argue that an “ontological” (as opposed to “empirical” or “methodological”) use of assemblage is incapable of accounting for the systemic dynamics of spatial development under capitalism. Specifically, they claim that its inability to offer, “mediation or at least animation through theoretical assumptions and interpretive schemata” means that the approach inevitably descends into a form of “naïve objectivism” in which it is impossible to distinguish between the trivial and important elements and processes that make up a given assemblage (Brenner et al., 2011: 233). Instead, they argue that the method of critique, derived from the Frankfurt School and ultimately Marxian dialectics, is best equipped to penetrate surface appearances and access the inner workings of social relations and the production of space under capitalism. This underlying logic provides the “context of contexts” in which the formation of empirical assemblages takes place.

Clearly, this critique highlights a crucial point: that the emphasis assemblage thinking places on materiality and often highly contingent interactions between diverse actors can risk losing sight of aspects of social organisation – like capital accumulation and class struggle (or, indeed, hegemonic gender roles, power-laden racial stereotypes and so on) – that exhibit systemic logics and thus transcend specific empirical contexts. On the other hand, such analyses, and the approach defended by Brenner et al. (2011), appear, effectively, to conform to the description of “relations of interiority” provided by DeLanda. As such, they do not offer a response to what is perhaps the principal critique assemblage thinking makes of dialectical models. Furthermore, it is precisely the attribution of empirical phenomena – understood (and therefore easily dismissed) as “surface manifestations” – to a dominant underlying logic that assemblage thinking seeks to avoid. Even the “empirical” use of assemblage that Brenner et al. (2011) condone, which would reduce it to representing temporary stabilisations of arrangements in which capital(ism) is always the primary “animating” force, precludes us from pursuing this line of thinking. Instead, it seems that other approaches are needed that can allow us to explore potential of assemblage thinking for reimagining relationships between diverse actors and logics on more equal and “distributed” terms, while still keeping in view the importance of capital (as well as other powerful structuring forces). I shall return to this question in the conclusion.

Material–symbolic transformations

If at mid-century the favelas were typified by their location on hillsides and other precarious environments, makeshift housing, and the visible absence of public infrastructure, this material–symbolic component of the favela assemblage would undergo a radical process of deterritorialisation over subsequent decades. After decades of systemic inertia in the face of the “favela problem”, rising tensions generated by the evetual collision of favela and middle-class urbanisation processes in the 1950s and 60s united elite opinion behind an aggressive policy of mass favela removals (see Brum, 2012). The policy failed miserably in its aim of ridding the city – and even central and wealthy areas – of their favelas (Brum, 2012), but it did have the unintended consequence of considerably weakening the link between the urban poor and makeshift, self-constructed housing. The resettlement of evicted favelados in sprawling territories of poorly built and subsequently abandoned conjuntos habitacionais at the urban periphery helped to ensure that the favela lost its status as Rio’s residual category of urban poverty. At the same time, it contributed to the emergence of hybrid formations that complexified the notion of what exactly it was that constituted a favela.

As Brum (2012) has carefully reconstructed, processes of “favelisation” occurred within many of these conjuntos habitacionais in the decades following their construction. In Cidade Alta, in Rio’s North Zone, the inflexible design of the apartment blocks could not accommodate growing families or the entrepreneurial activities that had animated favela life. This led many residents to construct informal extensions and shop fronts on their apartments, giving them the ad hoc appearance traditionally associated with favelas. Meanwhile, the onerous requirement for meeting monthly mortgage payments and market demand from outside of the original rehoused population led to the both the emergence of a black market in apartment sales and sub-letting as well as the establishment of entirely new favelas in the surrounding area. By contrast, in some blocks residents managed to organise collectively to prevent informal construction and preserve the integrity of the original buildings and public spaces.

These diverse processes have produced a scenario in which different observers categorise these micro-territories in different ways (Brum, 2012). Neither the better preserved, nor the “favelised”, apartment blocks are officially considered by the state to be favelas. However, Cidade Alta, like other conjuntos habitacionais built to house evicted favelados, is regarded as such by many outsiders, suggesting they have carried the stigma of the favela with them. Residents themselves, meanwhile, express ambiguity, exaggerating micro-territorial distinctions and invoking hegemonic constructions of what constitutes a “favela”, in which the aesthetic appearance of a neighbourhoods is presumed to reflect the social condition of its residents, despite being a very poor indicator of this (see also Richmond, 2015: 260–280).

At the same time that many conjuntos habitacionais were undergoing processes of favelisation, the favela assemblage also underwent deterritorialisation from within as a result of transformations that reduced homogeneity between settlements that more clearly belonged within the favela category. For starters, most of the new favelas that appeared during this period were built on flat land at the urban periphery, rather than inner-city hillsides. The emergence of both new social movements and new forms of political clientelism that accompanied Brazil’s redemocratisation during the 1980s meant that these areas tended to benefit from more centralised co-ordination than had older favelas (Do Lago, 2003). As a result, these “favela-loteamentos”, as Do Lago (2003) describes them, tended to exhibit a radically different morphological form, with more orderly street patterns and important infrastructure components like drainage and water systems often built in from the start. Although still characterised by legal exclusion and neglect by the state – indeed, usually more so than more centrally located favelas – they were thus able to avert some major infrastructural challenges. Residents of many of these favela-loteamentos refer to them as “vilas” (workers’ quarters) or even “condomínios” (condominiums), both to register their aesthetic distinctiveness and to distance themselves from the entrenched stigma associated with the favela (Richmond, 2015: 270–276).

During the same period, older favelas also experienced important physical transformations. The return of democratic politics to Rio de Janeiro came in the form of the populist State Governor Leonel Brizola and a radical urban reform agenda, including favela upgrading projects and an ambitious land titling programme (McCann, 2014). While the latter largely failed in its aims (see below), the general climate surrounding redemocratisation seemed to signal a broad shift in policy towards favelas, now prioritising on-site upgrading over removal. This greater sense of security fuelled a construction boom as favela residents began to invest in upgrading their homes, rebuilding them with bricks and adding additional floors and aesthetic embellishments (Cavalcanti, 2008).

The following decade saw a deepening of these processes of physical consolidation via the larger and more comprehensive Favela Bairro (“Favela Neighbourhoood”) urban upgrading programme (Burgos, 1998). This brought paved streets, street lighting, landslide defences, recreational areas, and various social services into a large portion of the city’s favelas. It is important to note that after upgrading these areas did not cease to be “favelas” in either official or popular discourse. However, they were left more consolidated and better serviced than many peripheral loteamentos and conjuntos habitacionais, thus challenging both the favela’s status as a residual socio-spatial category and its association with state absence.

Social-economic diversification and the end of favela residualism

Alongside these material–symbolic transformations, the favela assemblage has also been deterritorialised by socio-economic diversification within and between favelas, and by trends towards convergence with non-favela areas on a range of social indicators. This appears to result from long-term shifts in favela residents’ insertion into the urban economy along with new patterns of residential mobility. As was pointed out by Preteceille and Valladares (1999), already in the late 1990s favelas had ceased to be “a locus of poverty” in the city with a majority of Rio’s urban poor living in other kinds of settlement. ³ Subsequently, Neri (2010) has suggested that a long-term process of convergence in poverty levels is underway, with poverty falling in favelas and rising in non-favela areas.

Income data tell a similar story. National-level data show that between 2003 and 2013 average incomes rose by 38%, while for the favela population it was 55% (Meirelles and Athayde, 2014). This was primarily the result of rising levels of formal employment, which in 2013 was the condition of more than half of working-age favela residents (Meirelles and Athayde, 2014: 53–61). These trends led  Meirelles and Athayde (2014: 30) to make the highly dubious claim that 65% of favela residents had become “middle class”, using an income-based definition that not only excludes other important aspects of social class (Scalon and Salata, 2012), but is also highly sensitive to cyclical economic trends. Nonetheless, deeper changes to the educational and occupational structure in favelas suggest that the end of favela residualism at least, if not the absolute upward mobility of large numbers of favela residents, is likely to be an irreversible process (Perlman, 2010a).

Part of the explanation for this concerns endogenous socio-economic change within the existing favela population, primarily due to geographic variations in labour market conditions. For example, favela residents in the wealthy South Zone tend to earn significantly more than their counterparts in the poorer North and West Zones (Pero et al., 2005). However, there is evidence that the progressive marketisation of favela housing may be also acting as a sorting mechanism, leading some higher-earning groups coming from both within and from outside the informal housing sector to relocate to favelas with more favourable conditions.

Again, this relates to changes associated with the redemocratisation process. The enshrinement of substantial squatters’ rights in the 1988 Constitution and the creation of a new layer of intermediate legal tools during the 1990s – such as “Habite-se” documents that certify the structural integrity of favela homes – instilled greater confidence, and have led to rising property sales and the emergence of a vibrant rental sector (Perlman, 2010b). As a result, Abramo (2003) suggests that the favela housing market came to resemble a “single market” across the city, with house prices varying according to factors like location, infrastructure, and levels of violence. At the height of the house-price boom of the late 2000s, the difference between the cost of the average home in Vidigal, the most expensive favela, and a typical peripheral favela was as much as ten-to-one (Perlman, 2010b: 16).

These processes certainly should not be exaggerated. The lack of land titles in most favelas, the informal and highly localised channels through which sale and rental opportunities typically circulate, and continuing stigma and security concerns all still represent substantial barriers to entry to the favela housing market (Abramo, 2003; Perlman, 2010b). Recent examples of “favela gentrification”, for example, are limited to a very small number of favelas where specific conditions prevail. Nonetheless, between the endogenous transformation of the social structure within favelas and intensified spatial sorting between them and to a lesser extent from outside, it is clear that favelas have ceased to occupy their former residual position. In terms of social structure, then, the favela assemblage has undergone extensive deterritorialisation meaning that the formal–informal divide no longer neatly maps on to patterns of socio-economic inequality in the city.

Violence and urban fragmentation

It has so far been argued that over recent decades Rio’s favela assemblage has been substantially deterritorialised through internal diversification and the blurring of boundaries – both material–symbolic and socio-economic – that had previously distinguished it from other types of neighbourhood. However, this clearly has not resulted in a waning of the favela as an object of reference, identification and various forms of action. This section argues that just as it was experiencing those processes of deterritorialisation, it was reterritorialising along other lines.

The process by which urban militarisation and spiralling violence came to be organised around Rio de Janeiro’s formal–informal divide has been extensively covered in the literature (e.g. Gay, 2010; Lopes de Souza, 2000) and so requires only a brief recap here. From the early 1980s onwards, Rio de Janeiro became integrated into the emerging global cocaine trade, first as a transit point and later as a major market in its own right. Part of this process resembled the experience of other major cities, in that the more lucrative wholesale stage was controlled by shadowy figures financing and co-ordinating international supply chains, while a part of the retail stage was carried out by small dealers operating in the night-time economy with little attendant violence. However, another part of the retail market took on rather distinct dynamics, becoming highly territorial, extremely violent and almost exclusively focussed on the city’s favelas.

There were various reasons why the drug trafficking factions that emerged from this context came to base their operations in favelas. Not least among these was the favelas’ geography and morphology. Many were favourably located near to key transit points, like the city port, train stations, and major highways, as well as the prime consumer markets of the wealthy South Zone (Gay, 2010: 206–07). Meanwhile their narrow streets and complex layouts facilitated the discrete storage of drugs and arms, and the ability for gangs to defend territory from rivals and evade capture by police (Gay, 2010). Social conditions also contributed, as poverty and high levels of unemployment among young men in favelas provided a steady flow of recruits for the gangs. Each of these factors suggests that the rise of the drug trade acted as an “assemblage converter” (Deleuze and Guattari, 1987: 324–325) for the favelas, by taking existing, contingent features of their structure and injecting them with new, emergent content.

Another important contributing factor, also fitting such an interpretation, was the somewhat distinctive pattern of sociability in favelas, which combined high levels of internal social cohesion with intense suspicion of the police – both legacies of the historic failure of the state to provide for and protect favela populations. These dynamics combined with the traffickers’ capacity for violence, persuaded residents to observe the so-called “lei do morro” (“law of the hill”) by turning a blind eye to illegal activities (see Penglase, 2009). In exchange for this, the traffickers, who typically hailed from the same favelas and were themselves embedded in local social networks, maintained security in the neighbourhood by outlawing petty crime and violence committed by non-traffickers. Although broadly beneficial to the maintenance of local order within favelas, this came at a high cost as residents became vulnerable to abuses by the traffickers themselves. At the same time, the onus on controlling territory frequently turned many favelas into battle zones between rival gangs. Military-style policing by the state intensified in response, treating favelas as enemy territories where the human rights of residents could be disregarded with effective impunity (Gay, 2010; Machado da Silva, 2010).

This account suggests that favela militarisation was largely the result of dynamics within the favelas themselves (notwithstanding the market demand for cocaine coming from outside). However, these developments must be placed in the context of socio-spatial transformations occurring on a larger scale. Over the same period, elite neighbourhoods were also being fortified with walls, CCTV cameras and private security, while urban areas that were controlled by no hegemonic armed group came to be seen as insecure “neutral territories” to be avoided after dark (Lopes de Souza, 2000). These interlinked, centrifugal processes related to a more general transformation of social relations as an increasing sense of insecurity fuelled what Machado da Silva (2010) describes as “violent sociability”, characterised by an individualisation of demands for security and a greater willingness to condone extra-legal means of attaining it. It was within this context that, encouraged by sensationalistic media portrayals, favelas became a universal scapegoat for rising violence in the city and the police were handed increasing powers for repressing them (Machado da Silva, 2010).

This scenario of favelas as exceptional territories within a fragmenting urban landscape also gave rise to two further important developments from the 2000s onwards, both of which have tended to reinforce favela territorialisation and exceptionalism in new ways. The first of these was the emergence of so-called “militias”, primarily in favelas in the city’s western periphery (see Zaluar and Conceição, 2007). ⁴ Mainly consisting of off-duty police officers, these groups arose to prevent the emergence of drug trafficking and violent conflict that had become endemic elsewhere in the city. However, many have subsequently morphed into large, organisationally sophisticated and heavily armed mafias, running lucrative protection rackets and exercising monopoly control over local utilities and services. While their activities tend to be accompanied by lower levels of physical violence than is found in areas controlled by drug traffickers, they share with them the ability to operate in favelas with impunity.

The second major shift began in 2008 as Rio de Janeiro’s military police began a steadily expanding programme of favela “pacification” (Cano, 2012). In stark contrast to the hitherto dominant model of favela policing, based on mobile, and frequently bloody, capture-and-kill operations, Unidades de Polícia Pacificadora (Police Pacification Units, UPPs) were designed to establish a permanent presence in favelas and build relationships and trust with local organisations and residents. While the programme had some initial success in reducing violence, it did so in ways that strongly reinforced favela exceptionalism. UPPs militarised territory in much the same way as drug traffickers, reterritorialising favela borders through patrolling and surveillance (Fleury, 2012). While open conflict between police and traffickers in pacified favelas declined – despite traffickers continuing to operate in these areas – high-profile cases of abuse against residents indicated that police could continue to act with impunity. Indeed, individual UPP commanders were formally granted special powers within their territories, for example in the adoption of tactics like curfews and blanket stop-and-search (Fleury, 2012). Much like the traffickers and militias, then, pacification reproduced the favelas as militarised spaces of exception.

Exceptional governance

While policing provides the most visible example of exceptional governance arrangements in favelas, it is hardly unique in this regard. As McCann (2014) has persuasively argued, even as more positive forms of state intervention developed in favelas during the 1980s and 90s it was channelled through structures that served to reinforce, rather than weaken, the division between them and the rest of the city. The evolution of another key political and social institution shaping favela life – the residents’ association – bears this out. Residents’ associations were established in the 1950s, under the tutelage of the Catholic Church, to formalise the representation of favela residents and act as a conservative bulwark against political radicalism (Burgos, 1998). These bodies had a statutory obligation to articulate the relationship between the state and favela residents, were required to hold regular presidential elections, and were afforded collective representation via a citywide federation. They also came to exercise an important social role within communities, for example by mediating conflicts and co-ordinating the subdivision of favela land.

Politically neutered under the military regime, residents’ associations became major drivers of radical urban reform during the 1980s as a new generation of favela activists challenged more pliable older leaders in local elections (McCann, 2014: 960–1110, 1457–1534). However, the residents’ associations’ intermediary position in the context of redemocratisation and rising violence often meant they distorted state interventions to the detriment of long-term reforms. As mainstream democratic institutions struggled to meet the basic demands of favela residents, residents’ associations often found they could wrest more immediate benefits by delivering resident votes en masse to clientelistic politicians in exchange for investments in local infrastructure. In many cases, moreover, residents’ associations were subdued or brought under the direct control by drug traffickers and militias pursuing their own agendas (McCann, 2014: 2545–2754).

Even when they remained independent, residents’ associations sometimes served to stymie radical reforms, or at least to ensure they were implemented in ways that reproduced the favelas’ separation. An exemplary case of this was Brizola’s land titling programme Cada Família Um Lote (“A Plot for Every Family”) (see McCann, 2014: 1534–1621), which the residents' associations ended up opposing because they perceived it would weaken their role as informal planning authorities and threaten their status as intermediaries between favela residents and the state. In this regard, the comparison with semi-formal loteamentos is striking. While conditions for loteamento regularisation were from the outset more straightforward, part of the reason they were invariably successful is that their residents’ associations did not constitute a powerful intermediary layer of governance, allowing land titling to be carried out directly with individual residents.

Despite the distorting influence of criminal groups, clientelist politicians and residents’ associations in articulating the state in favelas, some more direct state interventions have managed to bypass these intermediaries and implement far-reaching reforms. In this respect, the Favela Bairro programme stands out for its comprehensive and holistic, if highly technocratic, approach to carrying out upgrading in a large proportion of the city’s favelas. However, in many respects even these interventions came to reinforce, rather than challenge, divisions between favelas and the wider city. For example, Favela Bairro had no tools for dealing with land titling or trafficker dominance, and thus in these areas tended to simply reproduce existing conditions, for example by bolstering informal real estate markets and providing gangs with new public spaces to dominate (McCann, 2014: 3123–3146).

More recently, the urban reform agenda associated with the 2016 Olympics served to reterritorialise the favela assemblage in both familiar and novel ways. Despite apparent the shift away from favela removals since redemocratisation, large numbers of favela residents have been evicted from their homes in recent years on diverse, and often highly questionable, legal grounds (Richmond and Garmany, 2016). These processes highlight the continued underlying legal insecurity of favelas in spite of the various nominal constitutional and legal protections supposed to avoid such outcomes. They also demonstrated the political weakness of favela populations when faced with a policy programme that can unite the political class, business interests, and much of public opinion (see Brum, 2013).

Meanwhile, other seemingly more benign favela policies have also reasserted favela exceptionalism in recent years (Richmond and Garmany, 2016). For example, large investments in monumental transport and infrastructure projects, such as controversial cable-cars installed in some of Rio’s most visible favelas, have been carried out without meaningful consultation and seemingly designed more for the benefit of tourists than residents themselves. As such, these interventions also suggest that favelas continue to be treated by government as spaces in which normal democratic and legal protocol need not apply. Indeed, this represents perhaps the most consistent feature of state engagement in favelas. From the mass removals of the 1960s to UPPs and cable-cars in the 2010s (though with the partial exception of Favela Bairro in the 1990s), the state has always tended to act as though it were responding to an emergency that could only be addressed through drastic and geographically targeted interventions that bypassed broader democratic procedures. This history “in the present tense”, as Fischer (2014) aptly puts it, has come at the expense of gradual, mainstreamed policies that would probably, over the long term, have had more positive outcomes for favela residents and more extensively deterritorialised the favela assemblage.