Double games: Success, failure and the relocation of risk in fighting terror, drugs and migration

This paper compares the security paradigms for combating terrorism, drugs and irregular migration and argues that while these have largely failed on their own terms, they have also proven rather successful for the actors shaping them. Through a spatial political economy analysis of systems of intervention, the paper shows how vested interests have helped perpetuate counterproductive approaches, while risks (including that of human suffering) have routinely been ‘exported’ into geographical ‘buffer zones’. In analysing the stakes in such sys- tems, we deploy the metaphor of games. This term allows us to highlight divergences between ‘official’ goals, such as ‘winning the war,’ and unstated aims, such as perpetuating security investments, relocating risk or stoking fear for political gain. Equally important, game terminology helps us highlight the spatial and social dynamics of collaboration, conflict and rule-manipulation within the system. In exploring these dynamics, the paper puts focus empirically on the complex collaborations between Western states instigating intervention and poorer ‘partner states,’ showing how a skewed geopolitical distribution of risk may tilt security interventions in the instigators' favour while maintaining ‘skin in the game’ for less powerful actors.


Introduction
Seventeen years from 9/11 and the start of a full-scale "global war on terror", terrorist attacks have been tearing through communities from Paris and Orlando to Istanbul, Baghdad, Brussels, London and New York. After years of "combating migration" through border patrolling and wall-building, Europe experienced its most dramatic border crisis yet in 2014-15, while 2016 saw some 5000 deaths in the Mediterranean. 1 And in spite of a long-running "war on drugs", the narcotics trade continues to thrive while drug-related offences keep driving mass incarceration in the US in particular and fatalities keep mounting in countries such as Mexico and the Philippines. 2 In each case, expensive security interventions have fallen short in terms of the majority of the most loudly expressed aims. Yet these respective interventions retain enduring appeal for policy-makers. Why so?
To answer this question, we approach these interventions comparatively as systems through which logics of security consolidate, benefits accrue, and costs and risks get generated and displaced among actors positioned differently across political space. 3 Our focus on systems chimes with a broader strand of political analysis that is influenced by complexity theory in the natural sciences (Cairney, 2012; e.g. Levin et al., 2013 for an ecological view). First, we emphasise the limitations of focusing on the individual parts of a system and the need to look holistically at the constant interaction of several different parts. Second, and building on this complexity perspective when considering systems of intervention, we emphasise that policy is not set "at the top" but is shaped by a variety of actors at a variety of levels with a variety of motives (an approach influenced by e.g. Clay &Schaffer, 1984 andMosse, 2004); the point, then, is not to assert that policy "failed" because of "a lack of political will" (Clay & Schaffer, 1984) but to map the various political wills and their complex interaction. Third, we emphasise that the system as a whole is characterised neither by a simple policy/implementation dichotomy nor by pure competition but rather https://doi.org/10.1016/j.polgeo.2018.09.008 Received 30 April 2018; Received in revised form 15 September 2018; Accepted 20 September 2018 by a set of relationships (sometimes conflictual, sometimes symbiotic) that have a certain stability. Fourth, we stress that these complex systems may be subject (like ecological systems) to rapid change: they are not as immutable as they appear.
In putting forward such a systemic analysis, we further build on the political economy of "war systems", an approach that seeks to explain the persistence of violence and to illuminate the aims in war other than "winning" (e.g. Keen, 2012; also Kaldor, 2012;Richani, 2013). Many such analyses have focused on the economic and political rewards of continued violence, while an interesting and less commonly explored strain of analysis has investigated how particular distributions of costs (across social groups and geographical areas) can encourage either the continuation of war or (when this distribution changes) some kind of peace (on benefits, see for example Berdal & Keen, 1997;Le Billon, 2000, 2008Sherman & Ballentine, 2003;Felbab-Brown, 2017; on shifting costs in civil war, see, notably, Venugopal, 2009; on international conflicts, Shaw, 2005 andGregory, 2011a).
There is a distinct spatial dimension to how we understand the three systems of intervention on terror, drugs and migration, as we will see in coming sections. In a well-known contribution, Derek Gregory extended pessimistic analyses of an "unending" war on terror (e.g. Duffield, 2007) into a more explicitly geographical dimension, raising the prospect that the entire world might be considered "fair game" for violent counterterrorism in an "everywhere war" (Gregory, 2011a, p. 242;Gregory, 2011b). Yet in any kind of war or security intervention, the point is not just to understand the ruthlessness, expandability and pervasiveness of the violence but also the limits that are placed upon it (whether through attempts to protect particular places or people from the fallout, or through the self-limitations imposed by instigators and "partners" in their actions). We need to understand the interests served by these limits as well as the interests served by the violence itself. Our paper addresses these questions by building comparatively on our longrunning qualitative fieldwork respectively on warfare, including recently on Syria (Keen, 2017), and on migration controls in the EU-African borderlands (Andersson, 2014). We will examine the spatial and social distribution of risks, costs and benefits of the three systems of intervention on terror, drugs and migration, as well as how certain actors position themselves to "game" -and gain from -the system by playing on the risks.
In considering these dynamics through a systemic lens, we adopt the notion of game for its potential to illuminate the conflictive and sometimes symbiotic relationships among "players" within the three systems of intervention. A game involves a set of shared rules and overt competition among adversaries on a circumscribed "playing field"; it may further involve significant scope for manipulation, collusion and even rigging of rules, and we will spend considerable time on these latter aspects. The notion of a game is of course familiar from a range of scholarly fields, including public choice theory and game theory as applied to security interventions (e.g. Alpern, Morton, & Papadaki, 2011;Bier and Azaiez 2009). We do not approach it through such formal modelling, however. Rather, we use game to understand and describe how security actors position themselves in relation to stated and unstated goals. In this, we consciously limit our attention to a set of key actors (or "players") within the three systems of intervention: governments instigating the security interventions; instigating-state security actors (law enforcement and defence); "partner" governments collaborating in the interventions; and finally, partner governments' security actors. Several caveats can be noted in relation to this rather schematic parsing of key actors or players. First, the systems of intervention, certainly, are broader than these actors alone -including corporations providing security technology and NGOs alternately collaborating with and challenging governments (cf. Andersson, 2014). Second, there are struggles aplenty not just among these core and noncore sectors, but also within governments and state agencies, making these far from unitary actors (a point we return to in our conclusion). And third, we recognise that a distinction between (political) government and (institutional/bureaucratic) "security actors" in enforcement is not without its nuances (as later sections will also allude to). However, for the purpose of clarity, this schema works as a starting point for considering the interests, games and gains with the systems of intervention.
One influence for our approach is Peter Andreas (2009: xiv), who some years ago treated US-Mexico border enforcement as a "border game" in order to capture "its performative and audience-directed nature" as well as "the strategic interaction between border enforcers and illegal border crossers". Expanding on Andreas' conceptualisation, we may delineate several levels at which the notion of a game is analytically useful. First, consider the enforcer-target interaction. Attempts at combating perceived transnational threats (migration, terrorism, narcotics) are often portrayed as a game (sometimes a "chase") both from practitioner (emic) and academic (etic) viewpoints. At the US-Mexico border, the Border Patrol's crackdowns are frequently referred to as a "cat and mouse game" (Donato, Wagner, & Patterson, 2008), with some Border Patrol agents approaching their task as "a game of tag and catch" (Heyman, 1995, p. 270). In the war on terror, operatives often talk of their interventions as a game of "whack-a-mole", while analysts highlight how drone attacks resemble a chase or hunt (Chamayou, 2015). Many of these idioms -"whack-a-mole" or "cat-and-mouse"hint at the frustrations besetting the quest to chase elusive threats, which have tended to displace, proliferate or change character in response to enforcement. This is a theme we will consider in the first section below.
Second, the game metaphor highlights an important performative dimension in attempts to combat politically "securitised" problems (Andreas, 2009;Buzan, Waever, & de Wilde, 1998), including its staging in highly visible geographical spaces (De Genova, 2002). The game metaphor also points to a clash between the political desire spectacularly and symbolically to "crack down" on a problem on the one hand, and the practical complications in actually curbing it, on the other hand. At the US-Mexico border, the stated aim of "combating migration" is in conflict with the business need for low-skilled migrant labour (though the two aims are partially reconciled in the creation of a "deportable" workforce: see De Genova, 2002;Heyman, 1995). Further, the stated aim works politically, in stoking fears and at the same time holding out the hope of allaying such fears through largely symbolic security intervention -a point we will develop in the second section below.
A third dimension to the game metaphor will form the principal focus of our paper: it helps us to illuminate the relationships and positioning among security actors tasked with carrying out the interventions. Game points towards a set of (unscripted) rules for interaction in a particular "field". 4 In this security field, dominant and dominated "players" take up, defend, and seek recognition for their positions, while safeguarding their interests. This line of analysis is familiar from studies of security professionals (e.g. Bigo, 2014), a line we expand substantially in our third section below by considering how the collaborations among Western and "partner state" security actors often end up undermining the stated rationales and "rules" set by the dominant players.
Again, scholarship on warfare points us towards some of these manipulations under exceptional conditions. In earlier work on Sierra Leone's war, Keen (2005) found that civilians used the term "sell-game" -a football match fixed in advance -to describe a situation of collusion between soldiers and rebels. The "failure" of government troops to defeat rebels could be explained in part by considering how soldiers and their commanders had many priorities (diamond mining, staying alive, supporting an ousted regime) that differed from -and often interfered with -"winning the war". Soldiers sometimes collaborated with the rebels they were supposed to be confronting, even selling them arms. 4 The term is used here in both a Bourdieusian and a geographical sense. D. Keen, R. Andersson Political Geography 67 (2018) 100-110 Rather than understanding such war as a "fight" between two opposing sides, we may be better off thinking of it geographically as a resourcerich zone into which a variety of military formations (rebels, government soldiers, peacekeepers, civil defence forces) are sent, whereupon they get "deflected" from their original -or at least originally expressed -aims. In this sense, the "game" played in this kind of "war system" was what may be termed a double or duplicitous game in which, alongside a highly visible set of rules and an expressed objective of winning, there existed another set of "unscripted" rules and unstated objectives centring on a coveted geographical zone. We see many such examples of collusion, subversion and abetting in the war on terror, the war on drugs and the fight against migration, whose security rationales and funding provisions tend to generate perverse incentives. In particular, "partners" have an incentive to play up the dangers they are claiming to combat while actually neglecting to confront these problems.
With this relational perspective in mind, we also delineate an important fourth dimension of the security games: namely the "game of risk", or the quest to distribute risks to the advantage of instigating players. In section four below, we argue that the negative consequences of our three interventions have been very unevenly distributed, with key instigating countries and actors avoiding some (but not all) of the worst risks and costs -turning the layered game back in the instigators' favour.
In sum, our usage of game allows us to approach jointly several spheres that are frequently considered separately: first, the enforcertarget interaction; second, the performative dimension of security; and third, the conflictive (and sometimes symbiotic) relationship among security actors. In addition, we bring in a spatial and social consideration of risk distribution as a key factor in understanding the dynamics of the three security interventions. This four-dimensional framing of the game helps take us away from the frequently cited gap between practitioner and political readings of security, or between "governance" and "politics" (e.g. Follis, 2016;Walters, 2014). However, we should not fall into the fallacy of seeing different layers of the game as engaged in a zero-sum interaction, nor as a hierarchical layering: rather, again from a systems perspective, an intervention may be "useful" to participants on one or several levels at once, while simultaneously problematic for them on another level. It is in this nonhierarchised sense that the dimensions discussed below need to be understood.

The cat-and-mouse game: chasing the supply
Practitioner metaphors such as "game of tag and catch" or "whacka-mole" are not just sarcastic remarks but point towards a specific logic of intervention: the centrality -and in many ways the futility -of cracking down on the "supply side" of the problem that is ostensibly being "combated". Economic analyses have long noted that the "war on drugs" focuses on cracking down on the supply of narcotics, rather than targeting persistent demand in destination states (LSE IDEAS, 2014). For Western governments targeting drugs, their key metrics centre on levels of drug production, arrests of key drug lords and restriction of particular routes. Similarly "supply-centric" interventions can be observed in our two other security interventions. The war on terror tends to focus on numbers of terrorists killed or neutralized. The fight against irregular migration tends to focus on the number of migrants and smugglers halted or detained (Albahari, 2018). In all three cases, "root causes" have been sidelined, despite frequent lip service to them. If persistent high demand is a key neglected problem in drugs interventions, counterterror efforts have routinely ignored or exacerbated the reasons and desires for engaging in, or supporting, acts of terrorism. In the case of border security the demand for low-cost labour in destination countries has been routinely neglected, as have the causes and motivations for moving in the first place (including global inequality, civil conflict, and unequally distributed refugee protection). The security "game" being played, in other words, has been framed so to exclude certain political lines of action and to give the impression of an arena in which a fearsome adversary is being "fought back", often through military means.
Yet this security game remains riddled with "own goals". In the drug wars, "supply-centric" interventions have repeatedly been shown to reproduce the problem ostensibly being fought. Spending on drug control has increased massively -reaching an estimated $50bn a year from state and federal budgets in the US alone, with global spending estimated at about twice that amount (Castillo, Mejía, & Restrepo, 2014, p. 70). Yet as the LSE's Expert Group on drugs control noted: [T]he pursuit of a militarised and enforcement-led global "war on drugs" strategy has produced enormous negative outcomes and collateral damage … [including] mass incarceration in the US, highly repressive policies in Asia, vast corruption and political destabilisation in Afghanistan and West Africa, immense violence in Latin America, an HIV epidemic in Russia, an acute global shortage of pain medication and the propagation of systematic human rights abuses around the world. (LSE IDEAS, 2014:3).
The US currently houses one-quarter of the global imprisoned population, and every year from 1993 to 2009 saw more people being jailed in the US for drug crimes than for violent crimes, with African Americans three to four times more likely than whites to be arrested for drug crimes (Drug Policy Alliance, 2016; Rothwell, 2015). Meanwhile, the RAND Drug Policy Research Centre noted in 2005 that "[t]he overall trend in cocaine and heroin retail prices during most of the past two decades has been downward (after adjusting for potency)" (Caulkins, Reuter, Martin, & James, 2005, p. 7). In general, even when drug prices have risen, a relatively low price elasticity for addictive substances tends to sustain consumption (LSE IDEAS, 2014: 3, 9), while consumers have often resorted to dangerous means of consumption under the auspices of strengthened criminal networks (Carpenter, 2014). Further, in what is sometimes referred to as a "balloon effect", crackdowns in one place tend to push trade elsewhere. Such crackdowns tend to be associated with rising violence in the new sites of transit, as evidenced by the rising homicide rate in Mexico following the large-scale interdiction in Colombia from 2007 (LSE IDEAS, 2014:29;Castillo et al., 2014).
Similar startling failures have been seen in the war on terror. For all the huge effort at counterterrorism (and the associated invasions) since September 2001, the number of terrorist attacks in the world has skyrocketed, rising from 3329 in 2000 to 29,376 in 2015(Global Terrorism Index, 2015. Meanwhile, the number of fighters in Islamist-inspired terrorist organisations more than tripled from 32,200 in 2000 to in excess of 110,000 in 2013 (Goepner, 2016, p. 113). The greatest expenditure of Western effort and money -and the greatest loss of life for Western soldiers -has been in Afghanistan and Iraq. Yet both countries have been beset by large numbers of terrorist attacks since invasion. Frequently, such violence is taken to justify the intervention, and even to necessitate more of it. As John Kerry said when debating with George W. Bush in 2004: "The President just talked about Iraq as a centre of the war on terror. Iraq was not even close to the centre of the war on terror before the President invaded it." 5 Meanwhile, attacks in Western cities have spread fear well beyond the initial theatres of operations. In many countries, al Qaida and ISIS have successfully exploited Western military interventions for propaganda purposes (Cronin, 2015;Gerges, 2005). Indeed, terrorists have sometimes been explicit in seeking a heavy-handed military response that will win them more recruits (Bolt, 2012;Gerges, 2005;Juergensmeyer, 2002). Civilian deaths arising from such military responses may attract support for terrorist groups, while indiscriminate military responses also tend to 5 First pre-election debate, 2004. See Robert Scheer, "US Is Its Own Worst Enemy in Iraq", Los Angeles Times, accessed at www.globalpolicy.org). D. Keen, R. Andersson Political Geography 67 (2018) [100][101][102][103][104][105][106][107][108][109][110] remove civilians' incentives to distance themselves from terrorists (cf Kalyvas, 2004). Although the "fight against migration" has sometimes had a rather less belligerent framing, nevertheless disastrous results from supplycentric interventions have been evident. In Europe, maritime migration hardly existed before a shift to stringent visa rules for North Africans in southern EU countries in the early 1990s. Since then, a series of highly politicised "border crises" of escalating severity has unfolded alongside tighter border security. In a context where people lack safe and legal pathways, border security initiatives have produced recurrent "balloon" effects -pushing migrants towards dangerous routes and more precarious entry methods while feeding the smuggling business along these new, riskier crossings. This has led to thousands of fatalities and evermounting chaos (Andersson, 2016;Cosgrave, Hargrave, Foresti, & Massa, 2016). While the spike in arrivals of 2015 came to an end after an EU-Turkey accord, the underlying destructive dynamics remain the same, as seen in the escalating fatality ratio and dramatic suffering on the central Mediterranean route. The fight against migration, and its associated "war on smuggling", have here led to stronger demand for smugglers (Albahari, 2018). In the United States, besides the diffusion of violence (Squire, 2014) and increasingly powerful smuggling networks, one principal effect of border security has been a growing undocumented population inland: millions of Mexicans have stayed on as irregular migrants rather than circulating back from seasonal US-based work owing to border security obstacles (Massey, Pren, & Jorge, 2016).
In short, by visibly targeting the "supply" of drugs, terrorists or migrants/smugglers while largely leaving demand unaddressed, the three security interventions have often worsened the problem they ostensibly combat by generating more distressing dynamics, including criminal innovation, chaotic scenes and rising fatalities. Why, then, do governments continue down the same path?
As suggested in our introduction, even if an intervention is failing to achieve the expressed goals, participants may still accrue benefits and succeed in unstated or less-advertised goals. The system may also adjust to "failure", notably by putting some spatial, social and chronological limits on the fallout as problems get loaded onto peripheral or buffer zones, onto politically marginal groups, and even onto the future. Meanwhile, delimited and short-term political and financial gains may swiftly multiply in the instigating state, as the next section will show.

The high political game: fuelling fears and purveying solutions
The framing of our three security games as a "fight" against a particular phenomenon already constitutes a political win, even if the practical results prove disappointing. The economic benefits in core countries, and for key security actors, are also considerable. These two dimensions -the political and the economic -need to be considered together, as they are fundamentally entangled. We will do so in this section by, in tandem, considering the political and financial/institutional gains from the war on drugs, the war on terror and the fight against migration.
Consider, first, the war on drugs. While moral panics around narcotics significantly preceded President Nixon's launch of the "war" in the 1970s (Nadelmann, 1990), a profitable system of intervention grew swiftly in the United States from this time onwards. Anti-drug funding further rocketed in the 1980s under President Reagan, as did the "prison-industrial complex" (Alexander, 2012, pp. 49-50). Meanwhile, programmes for prevention of drug addiction, treatment and for relevant education were severely scaled back. One study of the burgeoning US prison system noted that Reagan's large cuts to welfare programs exempted law enforcement agencies in the context of a relaunched "war on crime", including especially the drug wars. Alexander (2012:335) notes: When it came to the bellicose branches of government … -what Reagan called the "legitimate functions" -he countenanced bureaucratic bloat at every turn. The crime war became his domestic equivalent of the cold war.
Meanwhile, targeting "drug criminals" seems to have been part of a wider project of intimidation with a distinctly racial dimension. African Americans were targeted and incarcerated disproportionately (Parenti, 2008;Perkinson, 2010). Crack users were treated far more harshly than (richer) cocaine users, and by the early 1990s nearly 90 per cent of crack prosecutions targeted African Americans (ibid.). Significantly, Nixon had originally launched the domestic war on drugs in the context of widespread social unrest and anti-war protests. This "war" sent out an intimidating message to the black population and anti-war protesters (Parenti, 2008). John Erlichman, a former White House Counsel, said in 1994 that Nixon "had two enemies: the antiwar left and black people": We knew we couldn't make it illegal to be either against the [Vietnam] war or black, but by getting the public to associate the hippies and marijuana with African Americans and heroin, and then criminalizing both heavily, we could disrupt those communities. We could arrest their leaders, raid their homes, break up their meetings, vilify them night after night on the evening news. Did we know we were lying about the drugs? Of course we did (Baum, 2017, np).
Nixon was well aware of the electoral benefits from getting tough in the wars on crime and drugs. As Parenti (2008) notes, surveillance increased sharply, and Nixon's Organized Crime Control Bill was used to summon protesters before secret "special grand juries" for interrogation.
To move on to the war on terror, gains within the West from its "failures" have also been significant. As the old Soviet enemy receded, the spectre of terror presented an opportunity for vast continued investments in the military-industrial complex, which since then has remained very close to post-Cold War levels (Keen, 2012): according to one careful calculation in 2006, the Pentagon was absorbing fully 42 per cent of US tax dollars (Hossein-zadeh, 2006). One of the main arguments against the war on terror is the extremely high cost to the taxpayer: one tally puts the cost of the war efforts in Iraq, Afghanistan, Pakistan and Syria as well as Homeland Security at $3.6 trillion in current dollars between 2001 and 2016 (Crawford, 2016). But this astronomical cost may also help to explain the persistence of the war on terror despite all its counterproductive effects: every purchase, after all, is also a sale.
Political gains have also been substantial: to mention but one wellknown example, President George W. Bush's waning popular support rallied strongly with the declaration of a "war on terror" after 9/11, while the "tough" stance also seems to have helped his re-election in 2004 (Hetherington and Nelson 2003). The fear of terrorism is politically effective (Robin, 2004), and neoconservatives -and more recently President Trump -have routinely exploited it.
In international migration, fear and an "emergency" framing have similarly been useful for powerful Western actors, both economically and politically. On the economic side, corporate lobbying has helped convince governments to increase spending on punitive migrant detention, border barriers, surveillance technology, and military hardware (Bigo & Jeandesboz, 2010). In the US, the fight against migration has multiplied US Border Patrol personnel and the budget of Customs and Border Protection under the post-9/11 Department for Homeland Security (Massey et al., 2016). The same holds true in Europe, as southern European border, military and law enforcement forces have had their resources increased via the fight against migration, while the EU border agency Frontex has seen strong growth to deal with its everlarger agenda. (Figs. 1 and 2).
The fuelling of fears around migration has also often conferred significant political benefits -especially when combined with the "wars" on drugs and terror in a discrete political-physical space such as the US-Mexico border or the Mediterranean.
In the US, the post-9/11 period saw the southwest border becoming heavily securitised, with migration interdiction framed as halting potential terrorists (Chebel d'Appollonia 2012). In Europe, while the securitisation of migration may in the past have taken less extreme forms (Huysmans, 2006), since 2015 the external borders have served as a lightning rod for hard-right denunciations of an "invasion". Here, the drama at the borders is deployed politically to silence alternatives and to present border enforcement as the only feasible solution to the "emergency".
As in the war on terror, with its frequent exaggeration of the actual threat (Lemieux, 2016), the fight against migration depends on a deceptive numbers game that inflates the fears of "invasion": the vast majority of migrants (legal or otherwise) arrive by air, not land and sea (Andersson, 2016) -a trend that is even clearer once we take the one-  . This is merely illustrative: most border security spending occurs on member state level. Source: https:// frontex.europa.eu/about-frontex/key-documents/ (Andersson, 2014); by Hungary's right-wing government in 2015 (Haraszti, 2015); and by the UK in Calais, with the "emergency" there featuring heavily in Britain's right-wing newspapers in the run-up to the 2016 Brexit vote.
To conclude, while the three security interventions on terror, migration and drugs may be counterproductive and ill-conceived in relation to their most widely stated aims, they also present remarkable "successes" for powerful actors involved in propagating them. Politically, they fuel citizens' fears that (when amplified by the media) allow governments to impose drastic measures that seem to assuage such fears while also providing an effective diversionary tactic in terms of rallying anxious voters and shifting blame for societal problems elsewhere. Economically, the various security actors involved in implementing the interventions expand their operations and prestige not simply in spite of the systems' failings but also because of them. Given the political and economic functions of the phenomena ostensibly being confronted, even an "own goal" has many uses.

Gaming the system: partnership and subversion
So far, we have considered the enforcer-target "game" and the political-economic dimensions of "gaming the problem" within instigating Western states. Next, we will consider the double games played by non-Western partner countries. Here we find substantial scope for ostensible partners to manipulate the rules and even set some of their own. Both instigating and partner governments have often paid lip-service to a "joint" endeavour that few actors believe is actually going to work, while perverse incentives keep generating more of the problem ostensibly being combated.
We identify three overlapping ways in which the official game may be subverted. First, by appearing to collaborate, something that occurs when partner states that are known to be involved with criminals, terrorists and so on readily trumpet short-term "successes" to the instigating/donor state. Second, by stoking the threat, something that occurs when partner states actively seek to worsen high-profile problems or even collaborate with enemies. Third, by seeking impunity, which occurs when partner states take advantage of the political goodwill and resources obtained from signing up to a global struggle by engaging in (and sometimes escalating) domestic repression. In all cases, as we shall see, the political and financial importance that instigating states give to "combating the threat" in the official game is what allows these double games to develop.
To start again with the war on drugs, the benefits of at least appearing to collaborate have often proven substantial for partner states. US aid to Mexico has been designed, in part, to encourage a clampdown on drugs: between financial years 2008 and 2017, Congress appropriated nearly $2.8 billion for Mexico under the Mérida Initiative for combating drug trafficking and organised crime (Ribando Seelke & Finklea, 2017, p. 11). Yet such programmes may create a perverse incentive to promote the persistence of the problem. These also rarely acknowledge how, within "buffer" states such as Mexico, there have been longstanding symbiotic relationships between senior military officers, law enforcement and the drug cartels. Such is the corruption in the Mexican police that the entire police force of Veracruz, a major drugs conduit, was dissolved in December 2011 (Keen, 2012). In Colombia, the pre-eminent US partner, a pre-and post-9/11 war on terror combined damagingly with a war on drugs, propping up abuses by counterinsurgency actors while at the same time minimising the attention paid to grievances that helped inform rebellion (Fajardo, 2003;Restrepo & Spagat, 2005). 6 In many ways, the repressive apparatus of the state has depended on positioning itself as an ally of the US in one or more supply-centred security intervention.
An example of both appearing to collaborate and of seeking impunity comes from further south, in Guatemala. With the end of Guatemala's civil war in 1996, a "war on crime/drugs" came to provide a rationale for high military spending at a point when the country's swollen military risked major downsizing. During our fieldwork in 2002, Edgar Gutierrez, head of civil intelligence, stressed that shadowy structures linked to the old counterinsurgency were now involved not only in the "war on crime" but in many of the criminal enterprises that they claimed to be combating. Clandestine intelligence organisations were being used to fight organised crime, and "this same apparatus and method have been used to intimidate the human rights organisations." As in Colombia, the "criminal" enemy helped sustain flows of external aid as well as continued impunity for official and semi-official attacks on "subversives". 7 Katherine Saunders-Hastings (2015) has shown how the Guatemalan army has continued to recover from its vulnerable situation at the end of the civil war by positioning itself as the most effective and least corrupt of the institutions involved in combating crime and drugs, while politicians have benefited from being seen as "tough on crime". Yet crime has routinely been shunted into other parts of Guatemala City and beyond, in another predictable "balloon effect". Turning to counterterror interventions, we not only see attempts at appearing to collaborate and seeking impunity, but also a more active stoking of the threat. Indeed, the global war on terror has since 2001 created perverse incentives for non-Western governments to fuel -or at least allow -the threat of terrorism, attracting international approval and resources in the process. At the same time, government security structures (and security organisations linked to government) have frequently engaged in limited (and usually well-publicised) actions to show their value as partners, switching back and forth between appearing to collaborate and stoking the threat. Scholars and reporters have observed this pattern in Afghanistan, where an anti-Taliban agenda helped cement a government that itself included abusive warlords and where a symbiotic relationship grew up between the Taliban and security firms that were paid large sums to provide protection against them (Galbraith, 2010;Rashid, 2007;Wilder, 2009). In Yemen, the government's desire to eliminate its ostensible enemies (notably al-Qaeda in the Arabian Peninsula) has repeatedly been questioned in a context where the terrorist threat was serving to justify both aid and repression (Attree, 2016;Phillips, 2011;Whitaker, 2010). In conflict-hit northern Mali, meanwhile, evidence suggests that the presence of a jihadist threat has been actively abetted by an Algerian regime that saw the benefits of participating in a war on terror (e.g. Keenan, 2013). Meanwhile, the Malian government has itself benefited from close ties to criminal groups that it has ostensibly been opposing.
Let us consider two cases, Sri Lanka and Syria, in a little more detail, as they illustrate well the three forms of double-gaming (collaborating, stoking and seeking impunity) delineated above, as well as the overlaps between them. In Sri Lanka, the government's avowed commitment to a "war on terror" helped secure a significant degree of international quietude in relation to escalating official abuses against the Tamils that culminated in the killing of possibly 40,000 or more Tamil civilians trapped alongside abusive Tamil Tiger rebels in northeastern Sri Lanka in 2009 (United Nations (UN), 2011). The "war on terror" discourse proved useful for the Rajapaksa government not only in legitimising attacks on Tamils but also in intimidating journalists, lawyers, aid workers and human rights groups, and in providing cover for economic exploitation, including the forcible takeover of a large number of companies by the Rajapaksa family and its associates (Weiss, 2012; Keen's fieldwork). Although nominally "weak" relative to Western powers, Sri Lanka was able to use the fear of terrorism, the fear of a Sinhalese nationalist backlash, and rivalry between donors to manoeuvre itself into a position where it was effectively dictating terms to aid organisations and to donor governments (which one might have expected to shape "the rules of the game": Goodhand, 2010; Keen's fieldwork). While the Rajapaksa regime was quick to trumpet victory in 2009, it was also anxious to stress that the struggle against terror -and the "state of emergency" -was ongoing.
In Syria, amid a "global war on terror", the highly abusive government of Bashar al-Assad created a degree of ambivalence in the international community (and a heightened level of impunity) by presenting itself as a bulwark against "Islamist extremism" and terrorism. Yet behind the scenes the Assad regime has also engaged in actions that helped nurture extremist groups, notably ISIS/Daesh. Reports abound of collusion by government actors in terrorist attacks that were blamed on Islamist terrorists, while the regime released violent jihadists from prison and exempted ISIS from the bulk of regime military actions (Khalaf, 2015;Keen, 2017;cf Weiss and Hassan (2015) on Assad's strategy in Iraq). In December 2014, open-source data showed the regime's "counterterrorism operations -more than two-thirds of which were airstrikes -skew heavily towards groups whose names aren't ISIS. Of 982 counterterrorism operations for the year up through Nov. 21 [2014], just 6 percent directly targeted ISIS." 8 By understanding the geographical limits to violence, we get a much better sense of the role of the "global war on terror" in incentivising these kinds of collaborative and duplicitous relationships.
Turning finally to the fight against irregular migration, we should first note how signing up to this "fight" has given impunity to governments from Morocco to Mauritania and Turkey -and this impunity has been notably bolstered by active stoking of the threat. Morocco has obtained substantial diplomatic leeway with Europe (notably in reference to occupied Western Sahara) by presenting itself as a bulwark against migration, a bulwark that may at any time stop functioning. To take but one of many recurring examples of such a withdrawal of controls, in early 2017, after a European Court of Justice ruling excluded Western Sahara from EU-Morocco trade deals, Moroccan authorities suggested they might stop playing the "gendarme" of Europe. Soon a border crisis ensued at the Spanish enclave of Ceuta, where sub-Saharan migrants yet again climbed the fortified fences with minimal Moroccan control (El Confidencial, 2017).
If these are examples of political gains, there are also -as in the wars on drugs and terror -substantial economic gains to be had for key "partner" governments and underfunded state agencies in both appearing to collaborate and in selective stoking. Turkey is the starkest example today: the country was promised billions of euros in aid as well as free movement for its citizens into the EU on the back of a deal in which it agreed to stem the flow of Syrian migrants, setting the stage for further demands (Greenhill, 2016). Not only have agencies (besides governments) in "partner states" benefited from such largesse, but they have in parallel kept tapping into the cross-border smuggling business, where (as in the war on drugs) evidence shows they often play a fundamental role as facilitators (see, e.g., Bradol, 2018). Besides state agencies, in some countries paramilitary groups have also been seeking to tap into the promise of funds, exemplified by the efforts of Libyan militias and of Sudan's Rapid Support Forces (which grew out of Janjaweed militias responsible for genocide in Darfur) to bring crackdowns on migrants and smugglers to international attention at a time when the EU is committing large funds to halting irregular migration in collaboration with African regimes through its "Partnership Framework", an "Emergency Trust Fund for Africa", and associated initiatives. 9 Often the double games around terrorism, drugs and migration have combined, multiplying the potential gains, as seen from Mali to Mexico and Turkey. The most effective player on these overlapping fields was long Libya's Muammar Gaddafi. Having been ostracised and labelled a sponsor of terrorism, Gaddafi strategically clawed his way back onto the international stage in part by offering to help "rein in" international terrorism, and in part by offering to control migration flows. In return, Gaddafi asked for -and to a large extent received -political and economic favours, including the lifting of the embargo; the expensive Libya-Italy "Friendship Pact" of 2008; and the transfer of resources for controlling migration (Rinelli, 2016). As relationships with Western countries eroded and NATO launched its air strikes in 2011, Gaddafi shifted fully from "collaboration" to stoking, as he threatened to unleash an "unprecedented wave of illegal immigration" on southern European shores. 10 In sum, in the complex and layered games occurring among ostensible partners in fighting drugs, terror or migration, it is no longer clear that the instigating, dominant players are in control. Instead, the rules are being actively subverted and remade by ostensibly weaker partners. Governments and other influential actors within non-Western countries are often seen as subjected to one-way "policy transfers" -in other words, as rule-or regime-takers rather than "regime makers" (Ruggie, 1982) -yet as we have seen, they have frequently been able to barter their apparent cooperation for political favours, economic resources and impunity for their own abuses.
These dynamics have tended towards escalation within the systems of intervention, as especially evident in strategies of stoking. We may draw on Byman's (2006) work on perverse incentives in counterinsurgency, as well as on de Waal's (2015) notion of the "political marketplace", to suggest that, once perceived (security) threats are accorded supreme value, the temptation to stoke the relevant threat increases. As with any "price" or "barter", the terms of trade are influenced by the perceived desperation of the relevant parties. In fact, the more Western governments have made explicit the extreme priority they are giving to combating a given phenomenon (notably by placing it within an existential "security" and/or "war" paradigm), the more difficult it has been for the instigators to threaten to withdraw assistance from cooperating "partners", and the worse the terms of trade are likely to be for the instigators. All this means that apparent "regime takers" are best conceived not as passive recipients of "Western policy", but as active participants in a complex game where the official rules are complemented by unscripted, tacit and partially hidden ones (cf Scott, 1992).

Shifting the goalposts: games of risk
Returning to the game as seen by the instigators, this section introduces further complexity. Even amid failure in terms of most of the loudly expressed goals and in terms of dealing with these various problems, we are not witnessing total failure when it comes to reducing -or one might better say distributing, delimiting and relocating -the various risks that these wars or fights claim to address.
Nothwithstanding the numerous problems with 'supply-sided' intervention in the war on drugs, it could be argued that consumption might have been even higher if prices had been lower (Caulkin in LSE IDEAS, 2014). Meanwhile, a large proportion of the collateral costs of prohibition has been displaced onto producer and transit countries (LSE IDEAS, 2014:6). States collaborating in drug control have suffered from growing usage among their own citizens on top of severe criminality and violence, as seen in Mexico.
Notwithstanding the numerous terror attacks in the West, Lemieux (2016) notes only 3 percent of deaths caused by a growing number of terrorist attacks occurred in Western countries in the period 2000-2014, with the bulk of terror attacks concentrated in a handful of mainly Muslim countries. In counterterror operations, further, the use of drones and "vertical" forms of power (Elden, 2013) minimises the risk to Western combatants, as does the focus on "hard security" for the reduced numbers of personnel deployed on the ground (Chamayou, 2015). By contrast, countries that collaborate in the war on terror have often found themselves on the frontline of violence and even (as with Turkey after its strikes on ISIS) retaliation.
These dynamics reflect what Shaw (2005) calls "risk-transfer war"a reaction, in part, to American casualties during the Vietnam war. At the same time, the contemporary "export" of war-on-terror casualties to poorer countries represents in many ways an extension of a Cold War pattern (not to mention of the colonialism that predated it), as the Cold War saw "superpower" conflicts being fought out mostly in "Third World" countries such as Vietnam, Guatemala and Afghanistan. Significantly, today's radically skewed distribution of costs and risks also feeds a common perception that "only Western lives count", a complaint used for mobilisation purposes by some terrorists (for example bin Laden: Slim, 2007, p. 150). In our fieldwork on the Syria conflict, many Syrians stressed that terrorist groups have benefited from a perception that the West (because of a preoccupation with Western casualties) is hostile to ISIS but not to Assad (Keen, 2017).
The concept of containment -another inheritance from the Cold War and, in different guises, the colonial period -has, as Duffield (2001) argues, been put to new uses in today's era of proliferating transnational threats. On one interpretation, we now have a three-tier zonal system of unevenly distributed security risk. In the "outermost" tier, we find regions of the world that resemble "kill zones" (in US counterterror parlance) where the dangers resulting from wrongheaded wars proliferate. Barriers of different kinds (military, legal, border security) go up to limit the fallout as Western countries attempt to confine risk to the region -in effect establishing a middle tier of "buffer zones" in the global borderlands that have a key role in "absorbing" refugees (86 per cent of refugees are today hosted by low-and middle-income countries 11 ). In the innermost tier, finally, we find Western countries, which (while failing in many of their most loudly expressed goals) are in some respects protected from some of the gravest risks arising in large part from their own interventions.
This typology is simply a broad (and state-centric) starting point for considering risk distributions in the three security interventions. The zones can certainly be disaggregated into countries and subnational regions -something that helps to explain political tensions over intervention. The instigators of the war on terror (the US and the UK), for instance, have been among the least affected by its fallout in terms of forced migration from Iraq and Afghanistan; similarly, the sharply rising migratory "risk" stemming from the NATO bombardment of Libya (led by the UK, the US and France) has been weighted towards other European (and non-European) states dealing with the bulk of the influx. The US has been relatively insulated from international terrorism since 9/11 (though less so from US-born terrorists), while countries in the western part of continental Europe have faced multiple acts of terror despite their relatively limited role in the "everywhere war".
Further, within Europe's free-movement Schengen area itself, the EU-wide "fight against migration", with all its associated instruments and rules, has helped push the "problem" towards geographical margins. The strategy, albeit far from failsafe, has consistently been to shift the risks and responsibilities of disastrous border security to "frontline states" -that is, far from core northern European destinations. In addition, pre-emptive crackdowns have been put in place with a view to preventing people from crossing the external EU border, thereby transferring risks onto "transit countries", as in the case of Libya where EU states have supported the Libyan Coast Guard in forcibly returning people to detention centres where they face torture and other horrendous abuse (Amnesty International, 2017).
The distribution of much of the human, social and even financial cost to the West's periphery has further served as a form of politically (if not always practically) efficacious deterrence. Within this system, human suffering -mass mortality in the Mediterranean, humanitarian crisis in Greece, and even the biopolitical exceptionality of Libya's "forced labour camps" and "living hellholes" (UNICEF, 2017;Vaughan-Williams, 2015) -may perversely appear to some dominant players as a "win" in deterrence terms. This is implied when officials suggest rescues in the Mediterranean should be limited since they are said to serve as a "pull factor", or when militarised naval patrols are said (however erroneously) to destroy the smugglers' "business model" (House of Lords, 2017). Given that more "liberal" goals such as reducing global terrorism or meeting humanitarian needs have not been achieved, there may be a temptation retrospectively to redefine goals in more nationalist termsand via a more explicit language of deterrence. Indeed, part of the emerging system involves plucking success from the jaws of failure by abandoning humanitarian values.
In considering the transfers of risk within systems of intervention, we must approach risk as double-edged -as a "bundle" involving a combination of calculated future gains and costs (cf Andersson, in press). The dominant players calculate that the gains of this risk bundle will be enough to keep partner states engaged in distributing the problem in a way favourable to them. They also anticipate that the additional leverage gained by such "partners" will not outweigh these distributive gains. As "buffer state" governments use their status to press for diplomatic or financial favours, they may sometimes successfully undermine the established rules of the official game, while only rarely upsetting the unofficial risk distribution game, at least for any considerable period of time.
In sum, as with the civil wars with which we started our paper, the interlinked disasters occurring in the fight against terror, drugs and migration thus appear to have functions. Through a systemic analysis of the kind proposed here, we may illuminate the spatial and political distribution of risks (in terms of costs and gains) from "failing" interventions in ways that can powerfully help to explain their persistence.

Conclusion: blowing the whistle on double games
The various security games played around terrorism, drugs and migration provide a powerful tool for explaining why destructive interventions persist. Even though each of the three systems of intervention may "fail" on a global level, they "succeed", first, in framing a nebulous issue in narrow political terms; second, in enrolling and rewarding a very large array of actors, "setting the game" for everyone else to follow; and third (to a degree), in making sure that risks are transferred away from "core" to "non-core" players and their territories. As complexity theory would suggest, such systemic mechanisms are not neatly and hierarchically organised: "non-linear" causalities abound, as do negative feedback loops, as seen in the incentives for stoking threats. Moreover, the system has significant elements of instability, as our concluding discussion will explore with reference to possible alternatives.
But first, let us try to obtain a clearer view of how the objectives of different players interact within multi-level security games. We may simplify these interactions between the actors' objectives in the following analytical terms: symbiosis occurs when one actor approaches its relationship with another in terms of a shared endeavour or a "winwin", notwithstanding their differing objectives. This holds when, for instance, Washington obtains substantial political capital from wallbuilding and Border Patrol deployments while the agencies carrying out these interventions see their economic interests defended. In its most extreme case, symbiosis may involve a "sell-game" of the Sierra Leone kind, where supposed adversaries collude under the radar. Next, subversion occurs when a less powerful actor sets out to undermine the official "rules of the game" by fuelling the problem (or fears associated with the problem), as we have seen among partner states in the war on terror and the fight against migration. Finally, redistribution occurs when a dominant player seeks to transfer or relocate risks to dominated 11 UNHCR figures: http://www.unhcr.org/statistics/unhcrstats/576408cd7/ unhcr-global-trends-2015.html. D. Keen, R. Andersson Political Geography 67 (2018) [100][101][102][103][104][105][106][107][108][109][110] players -including those risks generated by the intervention itself. This transfer usually involves both a portion of gains and of costs, in a conscious trade-off with the partner. However, in its most extreme form, such redistribution involves a form of "risk dumping", as seen when the blocking of refugees seeking to leave Greece from 2015 onwards led to perennial encampment with all the risks (political, financial, social and human) that this blockage involved for the Greek state, officials, communities, and for refugees and migrants themselves.
To see these relationships more clearly from each actor's perspective, we may list them as a table (the left-hand column indicates the actors from whose viewpoint we are observing the relationship): The table suggests that Western (instigating) governments as a rule engage in a symbiotic relationship with their security agencies (and contractors), while building a redistributive relationship with poorer "partner" regimes and forces. Meanwhile, partner regimes at times seek a symbiotic relationship with instigators (by at least appearing to crack down on a problem) yet may over time be increasingly inclined to turn this into a subversive relationship, given the incentives to "game the system". This subversive tendency is also pronounced among partnerstate security actors, who may seek gains not only from cracking down on the problem but also from stoking it or simply avoiding dealing with it.
The complex and interrelated gains suggest that it will be very hard to shift away from today's disastrous security interventions. In particular, as long as "core" states send out the message that combating terror, drugs and migration are overriding priorities, potential partners will be incentivised to take advantage of these "imperatives" -firstly, by driving hard bargains that end up damaging human rights and human welfare and, secondly, by tolerating or fuelling the respective "threats".
Such systems will not be uncontested, however. At least three important sources of opposition -both actual and potential -can be identified. First, those who lose out within the current systems, while politically weak in many important respects, can be expected to oppose them (and often to have good ideas about what policies to favour instead). This will include political opposition within "buffer" states (as with the coalition against the "war on drugs" in Colombia); it will also include people within "core" countries (as in the opposition to the mass incarceration within the US that is a consequence in part of the "war on drugs").
Second, besides opening spaces for political contestation, the "loading" of costs and risks onto actors in the non-core and buffer states also tends to generate different kinds of "blowback" that negatively impact even the core states, perhaps after a significant time-lag. For example, where the EU is ruthlessly exporting migration control to a country like Libya, the credibility of the EU as a body that stands for human rights is radically undermined. Or where the costs of war are perceived as being systematically "confined" to poorer countries, this may reinforce an urge among both partner states and armed actors to ensure that some of the "costs" are felt within core countries (even to the extent of acts of terrorism). Another source of opposition arises from the fragile claim to be "winning" against terrorism, drugs and migration, a claim that is vulnerable to the evidence of people's own eyes and ears as well as to the view, still adhered to by many officials, that policy should be based on evidence (Rothstein & Teorell, 2008). Furthermore, whether in "core" on "non-core" countries, the parts of the administration that have to struggle directly with the costs of current interventions (for example, foreign, health and development ministries) will often find themselves in tension with ministries (defence, interior) that are more committed to enforcing "supply side" policies.
Third, there seems to be a growing momentum behind reforms based on harm reduction and addressing demand, with the possibility that hard-won lessons from one setting can be applied to another. In the "war on drugs" in particular, encouraging steps have been taken towards a focus on harm-and/or demand-reduction rather than a nearexclusive focus on supply. In Colombia, amid a large fallout from the US-sponsored war on drugs (in the shape of mass displacement, human lives lost, and political turmoil), Bogotá has began (albeit in stops and starts) to pursue a more inclusive approach to peace and to drug control. Among drug destination regions, meanwhile, Portugal has been exemplary in decriminalising drugs, in treating addiction as a public health problem, and in putting in place policies of harm reduction. As the Colombia and Portugal cases suggest, a wider shift towards harm reduction and a focus on the wider public good may be the best starting points for a shift of approach in all three security interventions -and the instigators of such a shift are likely to be those most badly affected by the current risk distribution, working effectively as a transnational coalition.
If this is the "good news", we must also note some additional "bad news" -namely that the beneficiaries of the current system can themselves be expected to adapt to trends and critiques that call the system into question. Indeed, they are already doing so. To invoke the complex biological systems alluded to in our introduction, we may say that just as medical interventions need to take account of bacterias' resistance to antibiotics, so too attempts to reform our supply-centric approaches will need to track (and adapt to) processes of resistance to criticism and reform. As migration policies fail to prevent major political crises in Greece and Italy, for example, the conclusion is not necessarily that these policies have failed but that a more intensive outsourcing of migration control (and of violence) is necessary -for example, in relation to Libya, Turkey, Niger and Sudan. Meanwhile, the potential political fallout from outsourcing migration control is being met with an intensified "de-responsiblisation" -denying knowledge of what is going on in Libya, Calais, Lesbos or Darfur, for example -and denying one's own responsibilities in relation to this suffering.
Given these shifting parameters, we wish to highlight one way in which academic investigators may help produce positive change: via different parameters of evaluation.
One of the problems with "combating" or "waging war" on something is that the political stakes for success become so high that it may not be possible to admit failure. Indeed, instead of being based on "evidence-based policy", our three security interventions have been beset by under-evidenced official reporting that trumpets short-term "success" without accounting for fundamental "failures" or potentially massive and damaging side effects (e.g Johnson &Tierney, 2006 andGoepner, 2016 on the war on terror; Andersson, 2016 and Albahari, 2018 on migration). One way of bringing this case home analytically is through evaluative parameters that take account of the full risks and costs of business as usual.
Let us return to a key concern of this paper -that is, how risks are should note that the largest transfer of risks and costs does not take place among core security actors within the system but rather towards groups with little or no stake in it. In the war on drugs, ordinary Mexicans have suffered severe violence while public authority has been undermined. In the fight against migration, border communities and bordercrossing migrants and refugees have for years suffered under the chaos and displacement effects caused by business as usual. In the war on terror, sites of displaced fighting have generated downward spirals with severe consequences for local societies, as seen for instance in northern Mali and the Afghan-Pakistan borderlands. These costs are a form of "negative externalities", in the sense familiar from environmental economics: they are not of central concern to the security players perpetuating the "game". In fact, we may say that in this respect, the four types of actors in our table act symbiotically whilst sharing a stake in this particularly skewed distribution of risks towards third parties. Moreover, they also share an interest in misidentifying the reasons behind these costs -that is, in portraying the escalating violence and chaos as disconnected from the functioning of the system of intervention. The game metaphor, on this level of systemic justification, may be quite simply understood in one of its dictionary definitions: as a "secret and clever plan or trick." This is something civil society actors, academics and politicians in countries such as Colombia have started to note in the war on drugs, and in so doing they have helped open political space for holistic assessments of the real costs of intervention, which in turn builds momentum for shifts in policy. In other words, the "fight against the fight" is not lost. We are overdue a more serious look at the underlying functions of various "wars", including their insulation from cost-cutting in times of austerity; their generation of wide-scale impunity; their active fostering of human rights abuses and suffering in "buffer" states and border regions; and their endemic blindness to the destructive consequences that they themselves have induced. In terms of academic efforts, it would also be worth going beyond the high-stakes "fights" against terror, drugs and crime and applying a systemic analysis of "gaming" to other high-profile fields of uneven risk distribution, including climate change, high finance and the "prison-industrial complex". As distinct from a traditional "cost-benefit analysis", and based on our analysis in this paper, we would suggest that a range of systems can be analysed in terms of the social and geographical distribution of costs and benefitswhat we call a "Costs and Benefits of the System" (CBS) analysis. Whether in relation to our three security interventions or other arenas, there may be opportunities to use such a framework in more formal comparative studies as well as in micro or ethnographic studies -particularly when seeking to explain the longevity of systems that are failing in important respects and when exploring opportunities for reform. In any case, our hope is that a more methodical underpinning for comparing such distinct areas of policy and public concern, drawing on an awareness of the politics behind the distribution of risks, costs and benefits, may help academic and policy analysts to ask different questions in relation to current "problem areas" of global concern. This may in turn help shift policy attention towards systemic aspects of intervention and away from the problem "in itself". The latter, after all, has usually been framed by powerful actors with a keen eye on their own interests.

Declaration of interests
None.