The two faces of security in hybrid political orders: a framework for analysis and research

This paper reframes the security and development debate through fresh theoretical lenses, which view security as highly contested both in the realm of politics and in the realm of ideas. 1 For some analysts security concerns political power, including the use of organised force to establish and maintain social orders and to protect them from external and inter-nal threats. For others it is about how individuals and communities are protected (or protect themselves) from violence, abuse of power and other existential risks. We integrate both approaches whilst placing our focus on the deep tensions between them. Combining them is especially apposite in the hybrid political orders of conflict-torn regions in the developing world – where the state and its monopoly of violence are contested and diverse state and non-state security actors coexist, collaborate or compete. We ask what security in these hybrid contexts looks like from below, that is from the perspective of “end users”, be these citizens of states, members of local communities or those who are marginalised and insecure. What are their own vernacular understandings of security, and how do these understandings link to wider conceptions of citizen and of human security? Even when security and insecurity are experienced and decided locally, they are at the same time determined nationally and globally. It is at the interfaces between local agency, state power and global order that the most politically salient and analytically challenging issues tend to arise. To analyse these interfaces we focus on three interconnecting political spaces, each characterised by their own forms of hybridity, in which security is negotiated with end-users: (i) “unsecured borderlands” where state authority is suspended or violently challenged by alternative claimants to power or providers of security, including non-state armed groups; (ii) “contested Leviathans”, that is state security structures whose authority and capacity to deliver security


The Changing Landscape of Security Analysis and Policy
The realist conceptualisation of security, which views security largely through the eyes of the state, whilst still enormously powerful, has lost its earlier monopoly over security thinking. Since the end of the Cold War a new liberal security mainstream has emerged whose referent objects -the things to be secured -are not confined solely to the security of states, their institutions and borders. 2 Security is viewed increasingly as an entitlement of citizens and human beings. The aims of security have also been broadened to encompass a more diverse range of dangers and threats, including issues of human rights and emancipation, freedom from want, the prevention of infectious diseases and the management of climate change. Moreover, security is not necessarily obtained even when states consider themselves to be at peace, as in conditions of authoritarian rule, social injustice or structural violence.
The reframing of security fits within the wider paradigm of a liberal world order and a liberal peace. It features bold assertions about the interdependence of security, development and human freedom (United Nations 2004. The United Nations has officially endorsed "human security" and the "responsibility to protect" (ICSS 2001) as guiding principles of international conduct. The World Bank too has brought security firmly into the development policy mainstream. Its World Development Report 2011: Conflict, Security and Development has embraced the concept of "citizen security", linking it to conflict management and poverty reduction (WDR 2011). It focuses not just on armed conflict but also on political and criminal violence, seen as a continuum. And it contends that insecurity and violence are negatively correlated with the institutional capacity and accountability of states and their elites. The fact that the World Bank, as the premier multilateral development institution, has been launching itself 'into the grubby universe of real-world politics' (Watts 2012:116) may be as significant as the substance of the arguments themselves.
Yet the liberal mainstream by no means constitutes as radical a departure from the realist vision as it could. The WDR 2011 remains firmly within a state-centred policy framework. It insists on building legitimate, inclusive and capable institutions, with "good enough" governance as the preferred policy solutions to citizen insecurity, although it sees them as a long-term exercise, which needs not follow Western models.
Even if states are no longer seen as the sole guarantors of security, political stability and international order, they are still given centre stage. Since 9/11 the emphasis has turned increasingly to the stabilisation of insecure regions and "fragile states". 3 Human security and the responsibility to protect have been invoked to justify armed interventions and stabilisation missions in countries as diverse as Somalia, Afghanistan, the DRC, Sierra Leone, Timor-Leste and Mali. The intractability of conflicts within such countries has reinforced a long-term shift in security doctrine towards counterinsurgency and asymmetric warfare rather than conventional wars between states. Counterinsurgency in turn has placed development, the protection of civilians, political solutions and the reform of state security sectors at the forefront of military doctrine and practice -further blurring the line between security and development. 4 However, within the liberal policy mainstream one finds surprisingly little serious interrogation of the concept of security itself and of how, by whom and with what political agendas security issues are framed and security functions are exercised. 5 Discussion of human and citizen security has largely research designed both to provide insights into the vernacular understandings, coping strategies and potential agency of end-users and to uncover the informal networks, alliances and covert strategies of the multiple actors determining their security in hybrid political orders.
by passed the problem of political power, including the various ways in which even liberal security paradigms uphold or are upheld by prevailing power relations. Thus, in practice, security still tends to be treated as politically non-controversial, with much research, analysis and policy focused on technical solutions to current security challenges.
This applies even to the World Bank's otherwise promising portrayal of citizen security, as 'both freedom from violence and freedom from fear of violence' -construed very broadly to include 'security at home, in the workplace, and in political, social, and economic interactions with the state and other members of society' (WDR 2011: xvi). Conceptualized in this manner, security is an entitlement of individuals which can in principle be measured in terms of indicators of reduced violence and, under the still wider umbrella of human security, other forms of vulnerability. The gathering and analysis of such evidence is used to facilitate large-n cross-country comparisons as well as to provide evidence-based evaluation of stabilisation, aid and development programmes.
Yet the concepts of human and citizen security struggle to capture security's contextually contingent meanings in fluid and complex multi-levelled regional, national and local contexts. They fail to acknowledge how security arrangements stabilise existing inequalities. And they do not pay enough attention to the ways those who clothe their actions with the mantle of international, human or citizen security may in reality damage the safety, livelihoods and welfare of many poor and vulnerable people.
The silences in the theory and practice of security are particularly problematic in hybrid political orders, where insecurity is unseen, easily hidden or unquantifiable, where the entitlements of citizenship are not extended throughout a population, where the state is not the primary actor mobilising to provide public goods and, furthermore, when governance arrangements benefit some but exclude others, including the poor, vulnerable and marginalised. They are more problematic still when security discourse and practice is monopolised by those with power, be they authoritarian states, insurgent groups, occupying forces or development agencies.
These conceptual gaps easily translate into ambiguity and confusion when researchers and policy-makers try to operationalize human and citizen security. They have resulted in widely divergent vocabularies and aims, even amongst those operating within the same research and policy communities. Vulnerable people and groups tend to be homogenised within weak empirical categorisations, obscuring the political power hierarchies and global processes which make them insecure. 6 In consequence they are liable to be treated as subjects of prevailing security arrangements rather than as agents with varying capacities to influence, respond or resist. 7

Securitisation and the Critical Turn
The elevation of human and citizen security into the security mainstream has come at a time when the wider vision of a liberal world order is increasingly challenged. First, it is being made redundant by the emergence of new global centres of power and profit in East Asia and elsewhere that challenge the very foundations of the post-World War II liberal consensus. Second, the limitations of liberal interventions have been cruelly exposed by events on the ground in a variety of different regional and national contexts, including Iraq, Afghanistan, the DRC, Libya and Syria. Third, human and citizen security have come under scrutiny by a critical security literature, which questions both the theoretical and the policy foundations of liberalisation and, in particular, of liberal peace-building (Selby 2013).
The main thrust of these critiques is encapsulated in the concept of "securitisation": the idea that far from contributing to equitable development and the welfare and safety of vulnerable people, liberal interventions carried out in the name of human or citizen security have merely provided normative and policy cover for new forms of global hegemony. Duffield's (2001) Global Governance and the New Wars: The Merging of Development and Security has been especially seminal. He and other analysts have turned the critical lens of securitisation variously upon the enterprises of international development, humanitarian assistance, governance reform and state-building (Booth 2005;Chandler 2006;Richmond 2009;Mac Ginty 2011). Despite its promising steps, they argue human security is largely silent on the aid and security arrangements which prop up the liberal peace. Indeed, as one of their number suggests, human security can be depicted as the dog that has not barked (Chandler 2008a).
Although offering varying takes on securitisation, critical analysts see it as an important aspect of an overarching hegemonic enterprise instituted after the Cold War and suffused by ideas of Western economic and political liberalism. However they contend that Western humanitarian intervention, statebuilding and development initiatives differ significantly from previous imperialisms in being conducted in the name of the international community. This includes not only powerful states but also the entire panoply of international organisations, international financial institutions, aid agencies and global civil society organisations. The "new aid programme" promoted by international coalitions, it is argued, is an innovative and subtle form of "power-knowledge" paying lip-service to development as progress and to universal entitlements, including human and citizen security. The international actors at the forefront of this agenda have tended increasingly to favour indirect engagement (Chandler 2008b;Veit 2010). At the same time the discourse of "local ownership" has allowed them to deny formal responsibility, especially when interventions generate more insecurity than they prevent or bear heavily upon vulnerable people (Chandler 2000(Chandler , 2006. Analysts utilising the lens of securitisation have offered a useful antidote to liberal understandings of security but have not for the most part provided their own alternative definitions of security. They have tended to discount the motives and values of international actors, as well as to underestimate the extent to which values of democracy, human rights and humanitarianism have gained real traction in parts of the developing world (Weiss 2000;Davidson 2012). They have overplayed the coherence of the securitisation project and underestimated the conflicts and tensions between the major political, military and humanitarian players in global and national security marketplaces (Selby 2013). In so doing they have come perilously close to reducing national and local actors to bit-players in a global game. They have also suffered somewhat from the Cassandra syndrome, tending to downplay empirical evidence of improvements in peace and security globally, as well as in individual national contexts like Mozambique, Sierra Leone or Liberia (HSR 2011;Pinker 2011). Thus far they have been much more effective as critics of the current liberal orthodoxy than in proposing credible alternatives.
Nevertheless, critical security analysis has inspired a new stream of analysis and research. It offers a nuanced view of violent conflict, seeing it not only as an obstacle to development but also a potential site of social and political innovation (Cramer 2006). Moreover it explores the hybrid processes that characterize governance in many supposedly fragile (Rotberg 2002) states and suggests that the meaning of security is contingent upon the contexts within which it is constructed. Hence it challenges researchers to empirically investigate security from the perspective of end-users, including those who are most vulnerable and insecure.

Rethinking the Two Faces of Security
To summarise the discussion above, although security remains a highly contentious concept, it is also a highly necessary one. More rigorous definitions of security are needed than those currently on offer by the realist and liberal mainstreams or by most of their critics. These definitions should help unpeel security's multiple layers of meaning and grasp its contested character, notably in the complex political terrains of supposedly fragile or conflict-affected political spaces. 8 Our definitions start from the assertion that security derives its normative force from the idea that public power is used to protect not just the state but also its citizens along with those deprived of the benefits of citizenship, including marginalised minorities, refugees and displaced persons. In other words, the bedrock of state security is citizen and more widely human security. Accordingly we distinguish between two competing yet interlinked conceptualisations of security.
The first of these is on the supply-side, having long historical roots in the theory and practice of the modern nation state. According to this conceptualisation security can be seen as a process of political and social ordering established and maintained through authoritative discourses and practises of power, including but not confined to organised force. 9 Although security thus involves 'seeing like a state' (Scott 1998), in the hybrid political contexts of the modern world security resides, along with political authority, within a much wider array of global, national and local power structures and security arrangements. Security in this view is achieved through the exercise of power, especially but not solely military power. In an informationrich world it also depends increasingly on surveillance and on control of new media. That is, the processes of social ordering that produce security can operate in parallel with states as well as within them, in some instances complementing state authority while in others competing with it.
However, reducing security to the creation and maintenance of political and social order is analytically incomplete. It risks identifying security with 'the imposed silence and normalised quiet of power' common to authoritarian regimes or criminal orders -that is, with enforced stability rather than true security. 10 Accordingly, a second demand-side conception of security holds that security is an entitlement of citizens and more widely human beings to protection from violence and other existential risks including their capacity in practice to exercise this entitlement. 11 As such it is dependent upon the social contexts, cultural repertoires and vernacular understandings of those who are secured.
This vision of security stems from contemporary international development, humanitarian and human rights concerns. Re-conceptualising security as an entitlement opens the way to challenges to the state's power and monopoly of security provision in those instances where it fails to protect or indeed actively harms its own citizens. It focuses assessments of security provision squarely on ideas of legitimacy, popular consent and political authority. We also see it as related to but somewhat distinct from existing formulations of human and citizen security in focusing on the vernacular understandings of the people and groups who are securedthat is, how they experience, understand and respond to their own security and insecurity.
In broad terms one can say that supply-side security provision, including state security, is needed to assure political order; but in doing so it often stands in conflict with the views and entitlements of those who are secured. Analysis, therefore, should focus simultaneously on the web of relationships between political and social orders, and on the entitlements between individuals and groups.
How the tensions between the two are negotiated is central to empirical enquiry. This distinguishes our own approach from both the realist and the liberal traditions of security thinking. It encourages further analysis of the many ways security is created and maintained through authoritative discourses and practices of power; together with the great variety of actors and organisations contending and cooperating to establish structures of public authority. And it is especially pertinent in situations of on-going conflict, political violence or difficult democratic or post-conflict transitions, where several actors, some state, some non-state and some global, struggle to appropriate and in some cases change the definition, ownership and distribution of security.
Our dual definitions are meant to prompt questioning of the assumption that security is a public good necessary for development. Security, unlike development, tends to be a discourse of order and risk-avoidance rather than of change and transformation. Yet it is also commonplace to argue that domestic security and political order are prerequisites for development. Our understanding allows for deep tensions between the two normative goals. Where indeed security has become the sole governing principle of state policy, as in the national security states of mid-20 th century Latin America or until recently most states of the contemporary Middle East, it may actually harm development and reduce the safety and welfare of citizens (Sayigh 2003;Imbusch 2011).
We thus see security and insecurity as inseparable from the exercise of political power; security is itself politically contested, sometimes violently. Because security like other public goods has an almost unmatched symbolic prestige, the power to create new and shape existing security and justice institutions is intimately bound up with the political processes central to state making and state breaking (Tamanaha 1992: 205). 12 Thus security provision tends to be fought over by those wishing to gain or establish political authority -especially in fragmented states such as Afghanistan, where regional strongmen compete amongst themselves and with the state to attract national and international resources. Those who deliver security assert and protect their mandates to use force or threats of force, as well as to maintain surveillance and gather intelligence about those considered a risk to public order. As a consequence, security provision easily merges into the deep state, becoming hostage to parallel political agendas and establishing for-cibly defended "states of exception" -even within institutions like refugee camps supposedly devoted to protecting those at most risk from insecurity (Hanafi and Long 2010). For this reason the mandates of security and justice institutions, their accountability, their observance of the rule of law and their respect for the rights of citizens have tended to become paramount policy issues in all political orders.
However, when we emphasise that security is an entitlement of end-users we share common ground with current conceptions of human and citizen security. 13 Despite their flaws, these conceptions have a genuine cutting edge, especially when security arrangements protect inequalities of power, status and wealth, be they global, within states or inside and between local communities. We argue that the benefits of security provision tend to accrue disproportionately to wealthy and powerful individuals, institutions and states and that its cost tends to be largely borne by the poor, vulnerable and excluded (Korf 2004;Fluri 2011). By emphasising that security is an entitlement we also stress the potential agency and voice of end-users, including their capacity to protest, mobilise around their rights and hold accountable those responsible for delivering security.
At the same time, by focusing upon authoritative discourses and practices of power, we emphasise that security is tied up with deep issues about political rights, entitlements and obligations. In principle, those who deliver security should have credible and legitimate mandates based ultimately upon the consent of those secured (i.e., end-users) or their political representatives. In practice, however, end-users tend to regard the claims made upon them by states and social orders as authoritative to the extent that they deliver tangible benefits on an everyday basis. Thus it is especially important to restore providers' legitimacy in situations where the mantle of national security has been used to justify repression and rights abuses committed by authoritarian regimes facing fundamental challenges to their authority, as in Burma, Libya or Syria. Questions of legitimacy also compromise security where governments are too weak or captured by special interests to be able to protect all or some of their citizens, as in Mali or, in the context of sectarian conflict, as in Pakistan. They are also highly salient where the supposed requirements of regional or international security are used to paper over local systems of occupation and repression, as in Afghanistan or Palestine (Hanafi 2010). They can even arise in established democracies and quasi-democracies when the requirements of national security are wheeled in to justify harsh security measures in insecure regions like Northern Ireland, Kashmir or Chechnya.
Finally, in rejecting a purely state-centred approach we focus instead on the real nature of security arrangements, including their varied links to political authority below, within and above the state. 14 The governance of (in)security, especially in conflict prone and post-conflict states is "multileveled" (Cawthra and Luckham 2003;Baker 2010;Leonard 2013). It involves complex arrays of international, state and non-state actors who variously cooperate and compete for power and resources and who determine patterns of security and insecurity. Thus the actors and institutions mobilised to deliver security (and sometimes insecurity) range from (a) the primarily global, such as the United Nations, international peacekeepers, private military companies or transnational militant movements, to (b) the mainly national, including national security and justice institutions as well as the national governments and legislatures to which they are accountable and to (c) the mostly local, including civil society, traditional leaders, business communities and community security and justice bodies. Even more controversially, actors that operate beyond the rule of law and are often considered illiberal by Western paradigms, such as warlords, cartels, paramilitaries, millenarian cults, anti-globalisation movements and mafias, cannot be omitted since they can in some situations be considered agents of security as well as of insecurity (Reno 1999;Goodhand and Mansfield 2010).

Why "Hybrid" Security Arrangements?
The analysis of security outlined above calls into question what Foucault (1980: 78-133) terms the 'sovereign' view of power that lies behind state-centric analysis including much security thinking (von Torotha 2009). 15 Instead we find it more helpful to place our focus on the multiple sites of political authority and governance where security is enacted and negotiated. Here we draw upon the emerging literature on hybrid political orders (HPOs), which highlights the varied and contextually contingent nature of political power and security arrangements, especially in conflict-affected, transitioning and post-conflict contexts ( While previous accounts of state fragility have concentrated on how predation, clientelism, and neopatrimonialism weaken public authority and "hollow out" or "criminalise" the state (Rhodes 1994;Bayart, Ellis, and Hibou, 1999;Chabal and Daloz 1999;Rotberg 2010), analyses of hybridity are less inclined to leap to conclusions about state fragility or failure, and less inclined to see this as a one-way process (Call 2008). Rather they focus upon the multiple ways traditional, personal, kin-based or clientelistic logics interact with modern, imported or rational actor logics in the shifting historical conditions of particular national and local contexts.
Though the processes they focus upon may be presented in terms of the interactions between the formal and informal actors, analyses of HPOs seldom restrict hybridity to this distinction alone (Kraushaar and Lambach 2009). Rather HPOs tend to be characterised by multiple providers of security, welfare and representation, as the state shares authority, legitimacy and capacity with many other actors, networks and institutions (Lam-bach 2007;Wennmann 2011). Indeed the real holders of political power and providers of public goods, including security, may have little allegiance to the state or may not even subscribe to the "idea of the state" itself (Abrams 1988;Hansen and Stepputat 2001).
The HPOs literature has developed a complex and sometimes confusing vocabulary, which characterises hybridity in terms such as "legal pluralism", "twilight institutions", "mediated" and"negotiated" states (Griffiths 1986;Menkhaus 2006/7;Lund 2007;Hagmann and Peclard 2010). These conceptualisations suggest that public authorities in fragile regions, including institutions of the state, wax and wane as governance arrangements are never definitely formed but are in a constant process of reproduction, negotiation and flux (Lund 2007). Moreover they argue that the providers of public goods, including security, enjoy different levels of access to power and authority, and, in some cases, occupy positions in multiple political orders; be they local, national or international.
This broad understanding of hybridity de-naturalises stereotypes of the state and, by extension, other forms of public authority. It proposes that governance should be empirically investigated as a collection of loosely coordinated and constantly changing processes. However, given its intellectual roots in subaltern history and anthropology, hybridity does not denote the mere grafting together of separate actors and institutions to make new entities (Mac Ginty 2011: 8).
Rather it directs attention towards the '(re) negotiation and transformation' or 'unmaking' and 'remaking' of political orders (Mallett 2010: 67-72). Such an approach repositions state-building paradigms within 'transition logics' that focus on the processes and interactions creating and sustaining functioning public institutions, be they of the state or of other political entities (Wennmann 2011: 4).
Concerns with the variety of ways public authority is negotiated in hybrid contexts are also beginning to influence the policy literature on state-building and security provision, focusing it upon inclusivity (DFID 2010;OECD 2011;Carpenter, Slater and Mallett 2012). For instance the WDR 2011(xvii) stresses the importance of close understanding of particular national and regional contexts. It draws upon the literature on "political settlements" and "limited access orders" to argue for ' collaborative, inclusive-enough coalitions' which 'restore confidence and transform institutions and help create continued momentum for positive change' (Di John and Putzel 2009;North et al 2012). Nevertheless the WDR 2011 is somewhat reticent about the political processes through which these coalitions might be formed; nor does it lay down clear criteria by which one might decide empirically that such coalitions are inclusive enough to ensure broad security provision. Its attempts to pull together a wide array of empirical scholarship to support its assertions about the links between narrow elite pacts, weak institutions and cycles of violence do not completely convince, even at times doing 'violence to the scholarship itself' (Watts 2012: 120).
Moreover governance practitioners are enjoined to seek 'local legitimacy' and involve 'broader segments of society-local governments, business, labour, civil society movements, [and] in some cases opposition parties' (WDR 2011: xvii). Donors are left to consider 'how best to manage, exploit, and coexist with [HPOs]', and help public authorities 'to provide human and national security to their populations' (Clunan and Trinkunas 2010: 12). In a similar manner, the OECD's International Network on Conflict and Fragility (INCAF) has called for deeper understanding of HPOs (OECD 2011: 25), arguing that such societies: continue to function, to form institutions, to negotiate politically, and to set and meet expectations. Traditional forms of authority are not necessarily inimical to the development of rules-based political systems […] In fact, the challenge is to understand how traditional and formal systems interact in any particular context, and to look for ways of constructively combining them.
In many respects working with hybridity, like decentralisation before it, appeals to actors from diverse ends of the political spectrum (Lutz and Linder 2004). For the left it accords with the embedded, participatory and communitarian aspects of governance assistance programmes. For neo-liberals it suggests state functions can be outsourced, moving power away from inefficient or corrupt central governments. It also attracts those frustrated by the failure of numerous post-Cold War state-building and security sector reform projects in conflict affected countries, particularly where the centralisation of security repeatedly leads to predation. A hybridity lens has even been adopted by some security analysts to interpret the blurring of irregular and regular threats they encounter in ' asymmetric warfare' on contemporary as well as ancient battlefields (Killcullen 2009;Murray and Mansoor 2012).
In sum the notion of "working with the grain" of hybrid processes has been harnessed for quite different normative goals, be they basic service provision, liberal statebuilding, the promotion of democratic processes, the free market or the defeat of an adversary (Booth 2012: 84-86). Unsurprisingly this apparent looseness has opened the concept of HPOs, and in particular hybrid security arrangements, to a number of critiques.
Some analysts contend that the HPOs approach risks swapping the state fragility literature's essentialist focus on the deficiencies of the state with a ' celebration' of potentially chaotic, regressive and violent forms of governance beyond the margins of the state (Meagher 2012(Meagher : 1077. In particular, it is suggested that a rush to embrace the traditional can blur the true nature of the relationship between localised orders and legitimacy, or, even more dangerously, deploy Tillyan notions of state formation to apologise for violence. 16 Moreover a concentration on hybridity may lead to an overemphasis upon the negotiability of governance arrangements and foreclose robust empirical investigation of existing power structures (Doornbus 2010). Critics even fear that analysts, who see hybridity everywhere, may disregard the existing tensions and divisions in fragmented governance contexts more accurately described by the concept of "institutional multiplicity' (Goodfellow and Lindemann 2013).
These are useful warnings. Like all openended analytical concepts (as we have seen security is another), hybridity is vulnerable to a variety of constructions, critiques and misinterpretations. 17 Furthermore, it is useful only if it raises fruitful questions for analysis and empirical investigation. Indeed we prefer it over alternatives such as institutional multiplicity precisely because it highlights the complex interplay among multiple and often contradictory forms of social ordering, each having their own sources of power, distinct organisational logics and sources of legitimacy. 18 We also argue that hybridity should not be seen as a concept in search of a pleasing theory of the traditional. Rather it is an analytical lens that explicitly challenges reductionist positions by focusing on the interactions that make talking of, let alone reverting to, supposedly traditional governance arrangements impossible. Later in this paper we turn the same analytical lenses not only upon local level security arrangements but also upon the political spaces controlled by national elites and global security actors, both of which are far from homogeneous, being tugged in different directions by a diversity of political and security actors, each with distinct agendas, ways of operating and political alliances.
To be sure, we argue that hybrid political contexts tend to be permeated by deep contradictions and clashes between different ways of organising security and political power. These clashes may become violent, especially in situations of major insecurity, contested political authority or armed international intervention. But violence is neither inherent nor necessary, either for security or in order to drive the slow, locally embedded processes of political centralisation that some see as essential to state-building. Rather, the presence or absence of violence, and how it is organised or resisted, remain empirical issues for investigation in particular national and local contexts.
In sum, the main value of a hybrid approach to security is its emphasis on empirically grounded investigations that uncover how, and for whom, security is determined in complex, multilayered political contexts. Indeed it brings together a supply-side approach to the determination of security by a variety of security actors, with a demand-side emphasis upon inclusive security based upon the agency of end-users. Thus the approach can help analysts understand the informal networks and political spaces in which end-users voice protest against or withhold cooperation from illegitimate institutions, at the same time as it reveals networks which threaten their rights or worsen insecurity. Through such a lens realistic appraisals of the "weapons of the weak" become every bit as essential for understanding security provision as analysis of the political powers of the strong (Scott 1985).

Webs of (In)Security: A Multi-Level Approach
As already suggested, hybrid interactions between actors and institutions occur across as well as within national and local boundaries. Security arrangements tend to be determined at multiple levels both on the demand side and on the supply side. Webs or chains of security and of (in)security stretch from the global to the national to the most local levels and back. 19 The lives and survival strategies of end-users, in particular poor and vulnerable people, often depend upon remote national and global processes over which they have no control and upon powerful actors who are in no way accountable for the misery and insecurity they may cause. Conversely both global and national decision-makers often find themselves disconcerted by seemingly local upheavals, which generate wider conflicts and insecurities: what some security analysts term "blowback".
Below we turn our focus upon three types of political space, within which these webs of (in)security tend to interconnect, namely: "unsecured borderlands", "contested Leviathans" and "securitised policy spaces". Although they span the local, the national and the global, we see them as being mutually constituted, not merely as separate levels of analysis. However we are deliberately selective in our focus on these particular political spaces rather than others and, furthermore, in confining our discussion to the primary decision-makers active within them (rather than all possible actors determining local (in)security). As our dual definition of security attests, this focus should not be taken to imply that end-users have no agency. Rather it acknowledges that their agency and experience of security in each space may be constrained by other more obviously powerful actors and dynamics.
While our treatment of the global and the national may seem to take us rather far from our original concern with the vernacular understandings and lived experience of end-users, we think it can be justified: endusers are not only entangled in networks across local and national boundaries, they sometimes have a surprisingly acute understanding of how they are put at risk by wider national and global insecurities. Inevitably trying to look at security "from below" (Luckham 2009) is something of a thought experiment. It is not made any the easier by the fact that so much of the literature sees security through the eyes of states and powerful global actors. One way of turning the tables upon the latter is to scrutinise their policies and programmes through the same hybridity lenses that researchers turn upon local actors in the developing world.

Unsecured borderlands are spaces where state authority is suspended or violently challenged by alternative claimants to power and providers of security, including non-state armed groups.
These borderlands are unsecured since they fall outside the security umbrella of the state; but they are not necessarily insecure. They can be configured around both social exclusions and geographical divisions ). Often they traverse established state boundaries. But they may also take the form of unsecured spaces inside existing states.
Examples are the borderlands between Pakistan and Afghanistan (White 2008) and between Indian and Pakistani-controlled Kashmir (Aggarwal and Bhan 2009), the troubled peripheries between eastern DRC and its Great Lakes neighbours (Raeymaekers 2010), the porous border areas between Sudan and Northern Uganda, the Somalispeaking region that traverses Somalia, Somaliland, Ethiopia and Northern Kenya (Simonse 2011) and the militarized border areas between Lebanon, Syria, Israel and Palestine. Sometimes the entire de jure territories of states like Somalia (Menkhaus 2006/7), Haiti, Central African Republic or, arguably, Yemen can be seen as unsecured borderlands. National parks too can share a number of the characteristics of borderlands, in some instances functioning as safe areas, but in others becoming unpoliced or unpoliceable spaces, as in rebel-penetrated parks in the Great Lakes region of Africa (Dunn 2009).
However there are important differences among borderlands, including distinctions between those where borders themselves are porous and unpoliceable, as between Pakistan and Afghanistan, and those where boundaries within peripheral regions are militarily enforced but where state authority is routinely ignored or contested, as in Indian-controlled Kashmir or Palestine. In neither case can regional borderlands be regarded as political voids. Often there exist well-organised security links between states and non-state actors across national boundaries, such as those between Iran, Syria and Hezbollah in Lebanon or those between Paki-stan's military intelligence apparatus and the Taliban in Afghanistan.
Not all unsecured borderlands are situated across or adjacent to national boundaries. The salient borders may be largely or wholly interior and may be characterised by deepseated horizontal inequalities, ethnically or religiously polarized identities, or geographical patterns of urban and rural exclusion. Examples are Northern and Eastern Sri Lanka (Goodhand, Klem and Korf 2009), parts of Gujarat in India (Berenschot 2009), the Niger Delta in Nigeria, areas exposed to Maoist violence in Nepal (Bohara, Mitchell and Nepal 2006), Baluchistan in Pakistan or the Oromo and Somali regions in federal Ethiopia.
Exclusions may be even more local still, including political and social spaces where the writ of the state does not extend or which suffer significant state and non-state violence, for instance Palestinian refugee camps in Lebanon (Hanafi and Long 2010), urban slums like the Cape Flats in Cape Town (Burr 2008), the outskirts of Karachi (Khan 2010) or favelas in Rio de Janeiro (Dowdney 2003). Indeed securitized border spaces can be found even in otherwise stable and wellgoverned countries as in India's Jammu and Kashmir (Aggarwal and Bhan 2009) or Northern Ireland.
Yet far from being ungoverned, such border spaces tend to have their own hybrid forms of political regulation, often involving violence alongside complex interactions among various armed groups. There can also be multiple articulations with the ' absent' state, among neighbouring states and with global players. Certainly one cannot in truth say states are absent in places like Pakistan's Federally-Administered Tribal Areas, Darfur or Kashmir (on either side of the Line of Control); they are very coercively present.
Unsecured borderlands are particularly salient contexts to examine how poor and vulnerable people themselves think about and experience security (Richards 1996;Lorenco-Lindell 2002;Allen 2012). They poignantly, sometimes brutally, expose the vast gaps between the way academic and policy analy-sis frames security and lived experience of those at the receiving end of insecurity. Even the concepts of human and citizen security do not come close to conveying the vernacular understandings and hidden "transcripts of resistance" of people at grass-roots (Scott 1985). Indeed it is doubtful if the term "security" itself can be translated. Approximations can be found; Acholi people in war-torn Northern Uganda distinguish between 'bad surroundings' (piny marac) and 'good surroundings' (piny maber) (Finnström 2008). However such terms are imbued with cultural resonances all of their own, as well as being open to change and reinterpretation as violent events transform local realities.
Using different methodologies, Uvin and Finnstrom study grass roots perceptions of security in Burundi (Uvin 2009) and Northern Uganda (Finnström 2008). Both conclude that most people do not make sharp distinctions between freedom from violence, social peace and the ability to meet basic needs, including the ability to move freely from place to place. Both suggest that local people tend to have more complex, less judgemental understandings of armed groups than national elites or international actors. And both seem to endorse the sentiments of a respondent in Finnstrom's (2008: 12) study, that 'the silence of guns does not mean peace'.
At the same time unsecured borderlands tend to be highly gendered spaces, in which gender subordination interlinks with other exclusions (Saigol 2010). Coulter challenges stereotypical accounts of women as victims in her study of girl soldiers and bush wives in Sierra Leone, as do Abdullah, Ibrahim and King in their analysis of women as civil society activists and peacemakers (Coulter 2009;Abdullah, Ibrahim and King 2010). Nevertheless Coulter (2009: 10) observes 'women's choices in times of conflict and war are at best circumscribed, at worst non-existent'.
Hybrid or informal security and justice institutions in unsecured borderlands are sometimes regarded as credible alternatives to failing, corrupt or oppressive state secu-rity provision. They are, however, incredibly diverse, including traditional justice institutions, local defence forces, community policing bodies (Baker 2008), paramilitaries, private security companies, assorted vigilante groups (Buur 2008;Meagher 2007) and community-led peace initiatives (Colak and Pearce 2009). Isima's (2007) study of nonstate security provision in Nigeria and South Africa suggests there may be contradictory relationships between informal security provision and "civil militarism", when providers alternate between being protectors and oppressors of poor and vulnerable people. Indeed informal institutions are seldom impartial, having their own political and economic agendas, some pursued through violence. Those relying on traditional authority, like the Arbakai in Afghanistan, can be patriarchal and reinforce local inequalities (Tariq 2009), while privatised security provision may be market-driven and biased towards those with wealth and power (Isima 2009). Furthermore, where outlawed groups provide much needed security and justice, they often do so at the expense of due process and respect for rights and the rule of law. Conversely, in some situations, as with paramilitary formations in Darfur and Southern Sudan, they may act in collusion with the state in repressing both armed resistance and unarmed protest (Ylonen 2005).
Yet in different circumstances narratives of subaltern resistance to economic exploitation, state repression or foreign occupation may be harnessed to wider transformative agendas. Wood's historical and ethnographic analysis of peasant revolt in El Salvador offers a persuasive account of insurgent collective action within struggles for land and social justice (Wood 2003). Others explore how popular movements and civil-society groups can pose alternatives to violence, as with civil society organisations and agrarian conflict in Guatemala (Van Leeuween 2010), Afghanistan's often overlooked activists (Theros and Kaldor 2011) and the resistance of Colombian communities to guerrilla as well as state violence (Alther 2006).
Yet one cannot assume that the hidden transcripts of dissent, still less armed resistance, are progressive. Indeed, where insurgents mobilise submerged ethnic, national or religious identities, even in support of wider goals of national liberation or social justice, they often create new forms of exclusion and violence, as for instance studies of the Taliban have shown (Fleischner 2011;Rangelov and Theros 2012). Moreover, a well-recognised feature of successful armed struggles is the betrayal of the hopes of many of those who supported them: the 2006 violence in East Timor (Nixon 2012) and the well-documented peacetime reversals of gains made by women in armed insurrections are cases in point (Coulter 2006;Hale 2008).
In sum, whilst it is critical to understand hybrid security provision and the ways it can fulfil unmet security and justice needs, a strong dose of realism -based on detailed empirical research into how informal mechanisms function, for whom they work and what tangible benefits, if any, they provide to end-users -is necessary.

Contested Leviathans are states and state security structures whose authority and capacity to deliver security are weak, disputed or compromised by the special interests that tend to predominate in HPOs.
They are so termed in order to capture the contingent and disputed nature of state authority in many national contexts. To call such states "fragile" can be something of a misnomer, however, since even the most challenged retain considerable powers to coerce or watch over their citizens. Moreover, their primary security institutions (the armed forces, intelligence apparatuses, police and judicial systems) matter enormously for the rights and security of citizens, even when in the main they act as agents of insecurity rather than security.
At the extreme end of the spectrum stand states whose capacity to exercise any form of legitimate authority, nationally or locally, is severely diminished or non-existent. In large parts of Somalia or eastern DRC (Renders 2007; Kaiser and Wolters 2012), for instance, almost all semblance of public authority has vanished; state power is highly contested and geographically fragmented in Lebanon (Mac Ginty 2011); and in Afghanistan it faces prolonged armed resistance (Goodhand and Sedra 2010). Yet none of these countries, even Somalia, can be written off as ungoverned and ungovernable political spaces (Leonard and Samantar 2011).
Empirical analysis is not best served by hammering such states into a single theoretical mould of state fragility. Their categorisation as "fragile" or "failing" has often been after the event, only following major state crises or outbreaks of violence. Seldom has there been much serious ex ante analysis of their susceptibility to breakdown. Moreover it is striking that some states currently considered fragile were once considered developmental success stories, like the Ivory Coast, Zimbabwe or (before the genocide) Rwanda. 20 The Arab Springs seemed to challenge what we knew, or thought we knew, about the closed or oligopolistic political marketplaces of many authoritarian and quasi-democratic regimes. While such regimes deploy impressive capabilities for political coercion and surveillance of their citizens, sometimes penetrating deep into civil society as in states like Syria, Burma, Sudan, Yemen or Pakistan, the apparent centralisation of power has not been all that it seems. Rather as the former President of Yemen expressed it ruling 'is like dancing on the heads of snakes' (ex-President Ali Abdullah Salih quoted in Clark 2010: xi). Hence we suggest turning the analytical lens of hybridity upon apparently more durable state Leviathans as well.
First, state power and security may in practice be negotiated with major independent social sectors, like Islamist groups in Pakistan, Lebanon, Sudan or Egypt, or cartels in Mexico, Colombia and Peru. Second, the central institutions of the state itself may in reality be hybrid in their own right, with formal chains of command and accountability penetrated and even superseded by informal patronage networks and systema-tised corruption. Indeed, as in Syria, Iraq, Libya, ex-Yugoslavia, Zimbabwe and others, it is often state elites themselves who deploy personal, ethnic or religious ties to cement their regimes and control their military and security apparatuses.
This informalisation of power tends to be a double-edged sword, consolidating the power of state elites within patronage structures but also weakening public authority and the state's capacity to deliver security and other public goods. In times of crisis it can also generate acute political tensions in the heart of the state itself, rendering apparently fearsome state machines and their security apparatuses suddenly vulnerable. This is especially so where, as during the Arab Springs, state elites lose control over information to new media and where their monopoly of force is challenged by new forms of popular unarmed and armed resistance on the streets.
Yet the resilience of the deep state should not be underestimated either. Security elites tend to act as power and profit maximizers, translating their control of security and organized violence into personal or institutional gain within national, regional and global political marketplaces (North et al 2012). They may even work to mould democratic governance around their own securitydominated vision of the polity, as in Central America's post-conflict democracies. "Political armies" and security institutions almost invariably continue to be major players, even in transitional or democratic regimes (Luckham 1996;Koonings and Kruijt 2004). A particularly graphic example of the baleful influence of security services under a 'violence called democracy' has been Guatemala (Schirmer 1998;Goldman 2007). Even less explored by researchers has been the political consolidation of unaccountable security sectors under "states of emergency" in the securitised border spaces of Colombia, Sri Lanka or India (not only in Kashmir but also in areas contested by Naxalite insurgents)bringing down the wrath of the state upon embattled minorities and economically and socially excluded regions.
Despite a wealth of useful insights, the policy and academic literatures on the stabilisation of fragile states tend to remain couched in the restricted language of statecraft. Moreover they barely touch upon the deep politics of reform, draw upon the critical literatures on the state and on HPOs or empirically investigate how security is actually delivered to end-users. Only latterly has more critical and empirically grounded attention been turned upon the real politics of stabilisation (Collinson, Elhawary and Muggah 2010), post-conflict security reform (Hutchful 2009;Sedra 2007;Peake, Scheye and Hills 2009) and day-to-day policing and justice in conflict and post-conflict situations (Baker 2010), including the activities of what Baker (2002) terms 'lawless law enforcers', who may in practice be the only recourse of poor and vulnerable people seeking a modicum of safety and justice.
Over three decades ago a classic study by Enloe (1980) focused upon how colonial and post-colonial elites framed security policy around the manipulation of ethnic and religious identities. More recently attention has turned to how fragile states and predatory elites thrive upon durable disorder (Chabal and Daloz 1999), merging state security institutions with the parallel networks of HPOs. Nevertheless there has not been enough serious empirical investigation of how they and their security apparatuses contribute to, and extract advantage from, such disorder, either nationally (as in Burma, Zimbabwe and arguably now Iraq and Syria) or within marginalized border regions (like Darfur, Abyei and Kofordan in Sudan) -nor of the ways durable disorder touches on the lives of those who are threatened or excluded.
There is also a distinct shortage of detailed micro-analysis of the invisible faces of power and security, including the intelligence and surveillance systems often at the heart of contested Leviathans. As Tadros (2011) has shown for Egypt, state security apparatuses are often parallel powers in their own right, interconnecting with corporate and political interests and penetrating deep into civil society. Research on these apparatuses in Latin America has analysed how they perpetuate legacies of impunity, rights abuses and social exclusion -even in supposedly democratic or 'post-conflict' states (Schirmer 1998).
There have been calls for 'pragmatic realism' both about the prospects of security reform (Scheye 2009) and about post-war stabilisation more generally (Colletta and Muggah 2009). The empirical foundations for such a pragmatic approach are spelt out in case studies of the politics and practice of security sector reform (Cawthra and Luckham 2003;Hendrickson 2008;Peake, Scheye and Hills 2009). Most of these case studies, however, tell the story of reform and the political obstacles it encounters from the viewpoint of the reformers themselves rather than that of the end-users whom the reforms are supposed to benefit.
Existing analysis has been notably weak on how dysfunctional security institutions can be challenged or held accountable by endusers living in conditions of insecurity. Furthermore we have little understanding of the how citizen demands for change emerge and catalyse a reordering of the political and security landscape. For his part, Hattotuwa (2009) has drawn attention to how new media have kept open spaces for political debate in the security-dominated political environment of post-conflict Sri Lanka, as well as engaging with grass roots audiences in vernacular languages. Furthermore, Somaliland is widely cited as a paradigm case of citizen action and peacebuilding from below. Yet it remains a special case and it is far from clear that its lessons are transferable to other conflict-torn or fragile states, where as we see below peacebuilding largely been driven by international rather than local actors and agendas.

Securitized policy spaces are policy arenas in which international actors (peacekeepers, donors, international agencies, INGOs etc) intervene to ensure peace and security, claiming to act for poor and vulnerable end users as well as for the international community.
Our central contention is that interventions by members of the international community are characterised by their own forms of hybrid politics, which warrant similar analytical lenses to those turned upon national and local actors. Scrutinised through these lenses, security provision is globally and historically constituted (Ayers 2010); and the welfare and security of end-users all too often take second place to geopolitical concerns, inter-agency rivalries and patron-client relationships. Even when intervening for humanitarian goals, international actors rapidly become entangled in hybrid relationships with powerful, and sometimes destructive, national and local actors. Good intentions are no protection against the perverse and sometimes violent consequences of international engagement.
As we have seen, the WDR 2011 tackles the political dimensions of peacebuilding and security reform. Yet it remains very difficult to translate its analysis into sound operational guidelines for international engagement in the bad surroundings of fragile states. More broadly the distinct historical trajectories by which states and regions become fragile or insecure, or are 'stabilized' or opened to reform, are all too often glossed over. Nor indeed has there been enough recognition of the major differences in the scale, types and impacts of international engagement, ranging from the relatively limited policy support for security and justice reforms in countries such as Nigeria, Uganda, Rwanda, Ethiopia or Nepal, to the wholesale reordering of entire states and their security institutions under the rubric of stabilisation as in Afghanistan, Iraq, Bosnia, Kosovo, Haiti, East Timor, Liberia or Sierra Leone (see the typology of external post-conflict engagement in Luckham 2011: 98-106).
The policy literature tends to view peacebuilding through the interventionist gaze of the international community, rather than through the lenses of national, let alone grass-roots stakeholders. This top-down perspective is reflected in analyses which attribute the success or failure of peacebuilding variously to the sequencing of reform (Paris 2004), to inability of donors and international agencies to coordinate policies and act with a single voice (Toft 2010), inherent tensions between humanitarian and military action and a lack of local ownership and absence of political will. Although no doubt important, these issues do not sufficiently address deep contradictions inherent in the international enterprise of peacebuilding itself (e.g. Paris and Sisk 2009;Collinson, Elhawary and Muggah 2010;Gordon 2010;Lothe and Peake 2010).
Policy analyses of interventions have tended to edit out the political interests and calculations of the major players, including the international ones. Yet experience shows that stabilisation and the prioritisation of security can easily become counterproductive in situations of highly contested political authority, large-scale violence and external military intervention, especially where stabilisation merges into counter-insurgency as in Afghanistan and Iraq. Thus the international community remains open to the charge that its humanitarian aid and development assistance, and still more its sponsorship of security sector reform (SSR) and stabilisation programmes, have been securitized or driven by foreign policy agendas.
An emerging stream of critical analysis has turned its attention to the links between the liberal peace and stabilisation agendas (Chandler 2006;Pugh, Cooper and Turner 2009;Richmond and Franks 2009), as well as to the cooptation of both by the war on terror (Keen 2006;Howell and Lind 2009). A major contribution of these approaches is to frame external actors as objects of studyrather than taking their policy agendas as the starting point for inquiry. Accordingly the entire assemblage of external actors who are active in unstable regions are characterised as international "regimes" or indeed HPOs in their own right (Mac Ginty 2011; Veit 2010). Veit even characterizes the complex relation-ships in eastern DRC between representatives of the international community, local elites and armed groups as a new trope in the old colonial narrative of indirect rule.
The priorities and animating logics of the different protagonists -peacekeeping forces, aid bureaucracies, humanitarian agencies, international NGOs etc -differ and sometimes clash (Bagayoko and McLean Hilker 2009). Yet one finds little detailed empirical enquiry into how bureaucratic timetables, funding requirements and inter-agency rivalries determine the workings of stabilisation policies or SSR programmes. There is even less empirical study of the relationships of such policies and programmes to the agendas of military alliances, large powers, and global corporations. Moreover relatively few studies have focused upon relationships between international interveners and national and local security actors. Beginning to fill these gaps, Autesserre's (2010) study of peacebuilding in the DRC is one of the first anywhere to approach the messy and violent real politics of relationships peacebuilders and local armed actors.
In recent years significant shifts in global security marketplaces have had major consequences for the capacity of the international community to ensure security within the contested spaces of fragile or shadow states (Nordstrom 2000). These shifts have included developments in the commerce in weapons and other conflict goods and services (Cooper 2002), including the expansion of privatised security provision (Abrahamsen and Williams 2009), the commercialisation of conflict resources, including the erosion of the distinction between lootable and nonlootable resources like oil (Le Billon 2012; Kaldor, Karl and Said 2007) and, above all, the trade in destructive "illicit" commodities such as drugs in countries like Afghanistan, Colombia, Mexico or Guatemala (Briscoe and Rodriguez 2010). Illicit or shadow markets have even begun to reconfigure entire states, as seen dramatically in eastern DRC or Guinea Bissau, which are transit points in the drugs trade between Latin America and Europe. They have also brought new actors into the security frame, including criminal mafias, diaspora networks, international security firms, natural resource corporations and even international NGOs (Avant 2005), all impacting in various ways upon vulnerable people and communities.
There are many ways in which developments in global and regional security marketplaces have spread violence and undermined peacebuilding. Examples include: (i) the enormous difficulties of regulating poppy production and trade and of breaking its links to warlordism and insurgency in Afghanistan (Goodhand and Mansfield 2010); (ii) the illicit trade in coltan and other high-value commodities in the DRC, which have impeded the creation of a viable national economy and provided incentives to sustains criminal and political violence (Kaiser and Wolters 2012); (iii) the problems of implementing security and justice reforms where resource or drug-induced corruption penetrates deep into security agencies, as in Mexico, Colombia or even Ghana; or (iv) the undue influence that international security firms have in civilian protection programmes or SSR in countries like Iraq, Afghanistan or Liberia (Leander 2001).
Moreover international actors in donorsaturated spaces confront major issues of accountability, above all in relation to endusers in host countries and local communities. Donors, international humanitarian actors and NGOs are accountable primarily to their own governments, agencies, funders and stakeholders, not to end-users in host countries and local communities. In other instances, international actors such as global security firms, dealers in illicit or high-value commodities, arms traders, natural resource corporations, criminal mafias or networks of religious militants may be accountable to non-state authorities such as shareholders or faith communities; in others they may be answerable to no one at all. Yet none of them are accountable in any way to end-users. Indeed the latter not only have few means of redress against those exposing them to violence, exploitation and rights abuses -be they state security agents, warlords, armed militants, criminals or indeed intervention forces; they have almost none against those supposed to protect them, such as peacekeepers, international agencies, donors, humanitarian bodies and NGOs, when they fail to deliver on their responsibility to protect or become complicit in their exploitation and abuse (Mamdani 2009).

By Way of Conclusion: Some Challenges of Research from an End-User Perspective
Our analysis above has presented a framework for evaluating security from an enduser perspective. This demands not only in depth local-level research but also the ability to turn an end-user lens upon the global, regional and national power relations which determine the security of poor and vulnerable people. Below we briefly sketch some challenges of implementing such a research agenda.
The first challenge is to tap end-users own vernacular understandings of how they navigate the terrains of violence and seek security in unsecured borderlands. A few researchers have provided vivid and at the same time analytically focused accounts of how particular groups navigate insecurity. These include, for instance, Coulter's (2009) research on girl soldiers and 'bush wives' in Sierra Leone and Vigh's (2006) account of young urban fighters in Guinea-Bissau, both of which include perceptive accounts of their methods of research. Others have combined more structured surveys with ethnographic techniques or historical sources to good effect, including Uvin's (2009) research on perceptions of war and peace in Burundi and Wood's (2003) analysis of grass-roots insurgency in El Salvador. Some of the best research indeed has not announced itself as being "about" conflict and insecurity per se but has rather generated insights about the latter indirectly, as for instance in Coburn's (2011) ethnographic study of power and the pottery trade in an Afghan town and Hutch-inson's (1996) examination of the Nuer's relationship to cash, guns and the Sudanese state. Paige's (1997) innovative use of ethnography, surveys and historical political economy analysis to interpret relations between coffee elites and violence in three Central America countries may also be put in this category.
Approaching end-users from another perspective, a number of research initiatives challenge the received wisdom that it is impossible to do large scale or rigorous research in conflict affected regions (Justino, Leavy and Valli 2009). 21 Some of these utilise large time series livelihood or household surveys designed to understand endusers perspectives of public goods provision, including security and justice, and the everyday outcomes of conflict for different social groups. Although the drive for large-n data sets is arguably spurred by donors' need to make "business cases" for development interventions, the possibility of combining these with further, historically and contextually nuanced, methodologies presents an exciting prospect for future understandings of security provision and everyday life in HPOs. 22 However we argue in a companion piece on our literature searches (Luckham and Kirk 2013) that, in aggregate, existing studies still tend to be geographically scattered, thematically and methodologically diverse, and in many cases lacking in empirical rigour. One cannot extract firm empirical or policy conclusions without greater conceptual integration and sharper empirical focus on (a) who precisely end-users are, (b) to whom they look for protection, (c) how far they have capacity to influence or indeed frustrate formal policy structures and agendas and (d) when instead they turn to informal security and justice providers, protests or violent revolts. Moreover, vernacular understandings must still be placed in historical context, including the processes of uneven development, social exclusion and political violence, which have rendered unsecured borderlands peripheral and their inhabitants insecure.
In sum, mapping end-users' vernacular understandings of hybrid security arrangements demands a combination of methodological innovation, historical understanding, empirical rigour and willingness to enter their social worlds and respect their agency (on the challenges of research in violent contexts see McGee andCramer, Hammond andPottier 2011). The bulk of existing studies deploy the standard techniques of ethnography and participatory research. They can with profit be supplemented by creative use of a wider repertoire of research techniques, such as: (a) the previously mentioned integration of qualitative methods with quantitative surveys; (b) mathematical modelling of social networks; (c) the use of life histories alongside documentary sources to record local-level social change; (d) accessing the resources of poetry, fiction 23 and the mass media to draw on the popular imagination; (e) deploying the crowd sourcing techniques which have shown their effectiveness in preventing outbreaks of electoral violence in Kenya and elsewhere (Bott, Gigler and Young 2011;Mancini 2013); or (f) using blogs and SMS messages to ensure a voice for the excluded in documenting their own experiences of abuse and insecurity, as in Sri Lanka (Hattotuwa 2009).
Analytical innovation too can sharpen the tools of empirical inquiry to better serve end-users, drawing for instance on the "subaltern" perspective of post-colonial historians, or on Scott's (1985) analysis of the weapons of the weak. Our point is not that these analytical perspectives should be wheeled in simply to place the focus on the agency of those who have been made insecure. It is that the premises as well as the techniques of research on security from below are in need of rethinking.
A second challenge is how to undertake empirical archaeology of the informal security relationships within and around the state. How and for whom do they work or fail to work? We have seen how durable disorder is endemic in HPOs, with diverse forms of political authority coexisting and competing over how and by whom security is defined and provided. As Stepan (1988: ix-xv) argued in a seminal analysis of democratic strategies towards security apparatuses during the Latin American transitions of the 1980s, rendering the deep state transparent is a truly major research challenge. Secrecy is not just endemic, it is the way the state works, as well as the way it shields its beneficiaries from prying eyes. Even in seemingly consolidated autocracies, such as Syria and Libya prior to the current upheavals, patronage networks of family, clan, ethnicity and religious confession criss-crossed state and security structures, holding elites together but also dividing them and linking them to wider political alliances beyond the state, which often descended to grass roots.
Mapping such networks is crucial to determine if elite and elite-mass coalitions are 'inclusive enough' (in the World Bank's terminology) to manage emergent conflicts and ensure the security of end-users. It may also diagnose the tension points in the edifices of power that can open spaces for change or render state structures vulnerable to major shocks and upheavals. However it is very hard to penetrate the deep state and the clandestine social networks and patronage systems within it. Added to this are the risks to researchers own safety and their responsibilities towards their informants. Yet even the most fearsome state Leviathans are not completely monolithic. Opportunities can open for research in the most unexpected places, especially at moments of political crisis, when the unravelling of political authority opens new windows for inquiry.
Researchers can construct reasonably convincing accounts of the deep state and of its informal networks of power through the gathering and triangulation of scraps of information from a variety of non-obtrusive measures and indirect data sources; even if it is not always easy to ensure that such a bricolage meets rigorous research standards. 24 Researchers can draw upon and learn from organisations, such as the International Crisis Group, which have documented political violence and human rights abuses in many national contexts. Investigative journalists too have much to teach them. Goldman's (2007) The Art of Political Murder, for instance uses the murder of an archbishop in Guatemala as the starting point for a far-reaching inquiry into the activities of the country's security agencies, their links with paramilitary and criminal groups and the struggles of human rights groups and social activists to hold them accountable. Furthermore it is possible to use the resources of new media to gather information about political spaces that are difficult or dangerous to enter. A recent study of internet censorship in China shows how researchers can even extract salient conclusions from the state's own efforts to close down spaces for debate and criticism (King, Pan and Roberts 2013).

A third research challenge is to lay bare the webs of causality and of accountability linking poor and vulnerable people to the international actors who in various ways determine their (in)security.
As we have seen the literature on the securitisation of development focuses in general terms on the globalised nature of the security arrangements entangling the developing world. Critical scholars such as Mamdani (2009) have argued persuasively that the international community bears a major share of the responsibility for the humanitarian disasters, such as that in Darfur. However their case needs to be backed by more rigorous research (especially from an end-user perspective) on how donors, international organisations and international NGOs actually navigate the terrains of war and the shifting security marketplaces of troubled states and unsecured borderlands.
There has been some progress, as we have seen, from the initial normative concerns of the literature on security and development towards more empirically grounded analysis of the real politics of international engagement: for instance (amongst oth-ers) the studies put together by Peake, Scheye and Hills (2009) on field experiences of security reform; the work of Hutchful (2009) on SSR in peace agreements; or Collinson, Elhawary and Muggah's (2011) scrutiny of stabilisation policies. There have also been detailed accounts of the interface between international and local security actors in particular national contexts, for instance the work of Autesserre (2010) and Veit (2010) on the DRC. The latter's analysis of how international actors fall back upon colonial tropes of indirect rule and are thereby drawn into hybrid relationships with politicians and warlords scrutinises both international and local actors through the same analytical lenses.
Policy-makers often grumble that the social research that arrives on their desks does not address their most pressing policy concerns. Researchers for their part complain that policy-makers disregard their findings by pursuing quick policy fixes in situations of great historical and social complexity. Acknowledging this complexity can help international actors to identify and minimise the potentially regressive outcomes of their interventions. More crucially it could potentially provide those at the receiving end of interventions with evidence-based analysis with which to hold international actors accountable. However end-users cannot hope to hold policymakers (or indeed researchers themselves) to account without more access to the research upon which the framing of policy is based. They also need empirical inquiry that reflects their own concerns, including better understanding of how and by whom their security is determined. 25 Despite an expanding research literature, security in HPOs remains such an acutely contested area that firm empirical conclusions are rare, especially about how security touches on the lives of poor and vulnerable people. Research, like policy, has been skewed by the inequalities inherent in the theory and practice of security itself. We hope that this article will encourage researchers to pay greater attention to the capabilities and concerns of those whom the prevailing security architectures have left out or failed.

Notes
1 This paper builds upon a systematic literature search undertaken by the Justice and Security Research Programme (JSRP) at the London School of Economics (LSE) funded by the UK's Department for International Development (DFID). The search itself and its main findings are discussed in Luckham and Kirk (2012) and Luckham and Kirk (2013). 2 The argument that security is not confined to the security of states predated the end of the Cold War (Palme 1982;Buzan 1983;Luckham 1983). During the 1990s it was mainstreamed through the concept of human security (UNDP 1994;Ogata and Sen 2004;Tadjbakhsh and Chenoy 2007;Jolly and Basu Ray 2007;Kaldor 2007). 3 State fragility, state-building and stabilisation are the focus of a substantial official literature; (OECD 2009;DFID 2010).
For an excellent critique and analysis of the implications for development and humanitarian policy see; (Collinson, Elhawary and Muggah 2010). 4 The US Army/Marine Corps' (2009) Counterinsurgency Field Manual remains the most authoritative official statement. 5 There is of course an entire academic subfield of critical security studies, referred to later in this paper. But it has had little impact on mainstream research, still less on policy discussion. The 'Security in Transition' research programme at the LSE is examining the discourses and policy papers of powerful international actors to understand what it terms the 'security gap'. http://www.securityintransition.org/ [Last accessed 17 August 2013] 6 These receive some discussion in the WDR 2011 but are neither confronted head on nor linked to its main policy conclusions and recommendations. 7 Both Ogata and Sen (2004) and Tadjbakhsh and Chenoy (2007) build human capabilities into their analyses, but do not relate them to security governance. 8 The essentially contested nature of security was highlighted by Buzan (1983). See also; (Smith 2005;Luckham 2007Luckham , 2009). 9 This definition is distinct from but consistent with 'realist' accounts of international relations. It draws from analyses of the making, unmaking and remaking of political orders and states; (Huntington 2006;Tilly 1985Tilly , 1990; Bates 2010; North, Wallis and Weingast 2009). 10 Edward Said quoted by Nadine Gordimer (2011). 11 This definition draws heavily upon Sen (1981Sen ( , 1999. See also endnote 2 above. Nevertheless we emphasise the contested and contextually contingent nature of entitlements. 12 See also the sources in endnote 8. 13 See the sources in endnote 2. 14 de Sardan (2009), Researching the Practical Norms of Real Governance in Africa contrasts 'real governance' -how states are really managed, public policies actually implemented and public goods actually delivered to the normative ideals of 'good governance' promoted by development agencies. 15 See Foucault's (1980) 'Two Lectures' and 'Truth and Power' where he argues that power is diffused through multiple sites rather than hierarchically concentrated, enacted rather than possessed and discursively constituted through 'regimes of truth' rather than purely coercive. 16 Tilly's famous notion of 'state formation as organised crime' portrays violence as a tool for populations to hold abusive authorities to account and drive a slow, locally embedded process of political centralisation. See: Tilly (1985). 17 The accusation of privileging violent forms of social ordering simply does not apply to influential formulations of hybridity, such as Boege et al (2009) However Watts' (2011:118-120) critique of the WDR 2011's use of large-n studies serves as a useful reminder about the limitations of such research in politically contested situations.. 23 In societies like Afghanistan and Somalia poetry has enormous cultural resonance, notably in dealing with issues of war and peace: on the political dimensions of Somali poetry and literature see Hoehne and Luling (2010: Part V). Writers like Nuruddin Farah in Somalia or Aminatta Forna in Sierra Leone provide insights into the raw realities of political violence and conflict, which social scientists fail to capture. 24 A good example is Cawthra's (1986) deconstruction of the apartheid war machine put together largely from careful assemblage of existing documentary sources. Others include Luckham (1971) and Schirmer (1998), although not focusing directly on how military and security establishments interface with end-users.. 25 International development organisations and major donors are increasingly publishing the individual research findings which go into major reports. For instance the WDR 2011's background papers can be found online.