Transnational productions of remoteness : building onshore and offshore carceral regimes across borders

Abstract. This article examines transnational framings of domestic carceral landscapes to better understand the relationship between offshore and onshore enforcement and detention regimes. US detention on mainland territory and interception and detention in the Caribbean serves as a case study. While the US domestic carceral regime is a subject of intense political debate, research, and activism, it is not often analyzed in relation to the development and expansion of an offshore "buffer zone" to intercept and detain migrants and asylum seekers. Yet the US federal government has also used offshore interception and detention as a way of controlling migration and mobility to its shores. This article traces a Cold War history of offshore US interception and detention of migrants from and in the Caribbean. We discuss how racialized crises related to Cuban and Haitian migrations by sea led to the expansion of an intertwined offshore and onshore carceral regime. Tracing these carceral geographies offers a more transnational understanding of contemporary domestic landscapes of detention of foreign nationals in the United States. It advances the argument that the conditions of remoteness ascribed frequently to US detention sites must be understood in more transnational perspective.


Introduction
Western governments employ interception and offshore "processing" of asylum claims that lengthen the distance between migrants and asylum seekers and the boundaries of sovereign territory.These practices began in the United States in the early 1980s in response to Haitian and Cuban migration by sea, and became more extensively used around the globe in the 1990s and 2000s.Forcible confinement is a central element of these efforts to regulate migration both on mainland territory and offshore.In the United States, detention of foreign nationals is a legal form of administrative confinement distinct from the rights and procedures developed under criminal justice law.People who are held under the authority of the Department of Homeland Security's Immigration and Customs Enforcement (ICE) (previously Immigration of Naturalization Service (INS) under the Department of Justice) are confined in an archipelago of local jails and dedicated federal detention facilities (that are publicly and privately owned and operated).Customs and Border Protection also operates facilities at ports of entry for short-term confinement.
The remote geographic locations of many of these facilities and frequent transfers of detainees between facilities within the system have proven significant issues of concern for anyone who has tried to track down a loved one, post bail, or serve as a legal advocate (Hiemstra, 2013).Remoteness compounds the unfreedom and harms of detention, making remoteness itself one of the most frequently criticized conditions of the US detention system (National Immigrant Justice Center, 2010;Human Rights Watch, 2009) 1 .In this paper, we show how the problems associated with remote detention are tied to the policy of mandatory detention and to the historical expansion of the detention system.The expansion of onshore capacity, moreover, is deeply entangled with the rapid expansion of local, state, and federal jail and prison capacity beginning in the 1980s (Simon, 1998).We thus develop transnational framings of domestic carceral landscapes  (Schriro, 2009).
to better understand the relationship between offshore and onshore enforcement and detention regimes.
While US detention is a subject of substantial research, activism, and political debate, it is not often analyzed in relation to the development and expansion of offshore enforcement.By developing "buffer zones" in the Caribbean and the Pacific to intercept and detain migrants and asylum seekers traveling by boat, the US federal government has used interception and detention to control migration and mobility to US shores.In this article, we discuss how racialized crises related to Cuban and Haitian migrations by sea in the 1980s and 1990s led to the expansion of an onshore and offshore carceral regime.We seek to show how policies and infrastructure to detain onshore emerge repeatedly in response to racialized crises offshore, better understood in transnational context (Sudbury, 2005;Martin, 2012).
No one dynamic accounts for the dispersed geography of places where migrants are detained.Rather, this is a story about the confluence of a number of geopolitical and domestic events in the late Cold War and immediate post-Cold War era: implementation of mandatory detention for asylum seekers, the use of detention as deterrent, restructuring of US military spaces, and rapid expansion of US carceral landscapes onshore and offshore.This contingent set of onshore dynamics must be held in tension with offshore detention and interdiction practices deployed simultaneously.Thus, the expansion of federal detention facilities "nationally" is better understood as a federal project of confinement, deterrence, and deportation conducted transnationally.By challenging the "domestic" geographical imagination that limits expansion narratives to the domestic scale, we can better understand the systematic, if locally contingent, state production of remoteness to regulate mobility.
Understanding the domestic dynamics of immigration detention requires tracing the transnational sites of offshore carceral spaces.This paper illustrates transnational productions of remoteness through the development of an interlocking offshore and onshore carceral regime.After reviewing conceptual framings for understanding transnational productions of remoteness, we turn to the underappreciated 1980s history of interception and detention expansion in the Caribbean and US southeast, which set the stage for today's detention landscape.This geography illustrates the close ties between offshore interception and detention, and highlights the onshore detention infrastructures that facilitate deportation.

Producing remoteness through transnational carceral regimes
There are many explanations for the remote locations of detention within the mainland US territorial landscape: from proximity to "the border" (often imagined as a clearly demarcated line dividing the US from Mexico) and associated sites of apprehension, to entrepreneurial county sheriffs and prison corporations, to dislocation from legal jurisdictions historically more favorable to asylum seekers (Masters, 1996;Wood, 2007).While each of these may prove important at certain times and places, no single factor can explain the dynamics of carceral expansion since the US resumed long-term migrant detention in 1981.ICE commissioner Dora Schriro's (2009) study of ICE's detention system includes a series of maps illustrating the conflation between apprehension, detention, and borderlands.The maps depict data on detention "demand" by arrest location, detention capacity, and capacity versus "demand".While the report was completed to inform policy re-form, no methods are provided for how these data were collected or mapped, leaving readers to interpret them visually.In the map in Fig. 1 depicting detention "demands", for example, the blurring of administrative data creates an image of "border-ness" from apprehension and removal data, even though border field offices like San Diego and El Paso are not listed among the top apprehension offices.Read within the context of dominant narratives of out-of-control borders, this map's use of traffic light colors (red symbolizing danger and green safety), signals hysteria and suggests that apprehensions threaten to engulf the interior of the country, an impression strengthened by the imagery of "uncontainable flood" of detention data pouring into Mexico, the Gulf of Mexico, and the Atlantic.
The report's map of the ratio of regional detention capacity to apprehensions, by contrast, suggests that many southern border sectors have greater capacity for detention than that met by local apprehensions.Los Angeles and New York City stand out boldly in red for their limited capacity, followed by Atlanta, San Francisco, and Miami.This map, then, does not so much explain apprehension and detention dynamics as act as what geographer Mark Monmonier (personal communication, 2012) calls a "cartographic fog."The scientific status afforded to maps lends authoritative credence to the impression of endless detention "demands", perpetuating institutional interests to meet quotas that keep demand for "beds" in facilities and deportation numbers at historical highs (New York Times, 2014).
While the US-Mexico border has come to stand in for narratives about US national borders and migration policies, Nevins (2010) convincingly shows how local efforts proved fundamental to fueling fortification along the boundary and driving a national ideological debate.We build on these insights about the historical and geographic specificity of national policy and practice by drawing attention to the important place of the Caribbean in the history of US migration enforcement (Noble, 2011).Our findings offer alternative narratives to explain the geographical complexity and diversity of the US carceral regime.They suggest intricate linkages between the development of offshore interception of Haitians and Cubans and the implementation of mandatory detention and expansion of facilities in the 1980s onward.These developments onshore and offshore did not merely coincide in time, but rather one fueled the other, first under President Jimmy Carter, and continuing under the presidential administrations of Ronald Reagan, George H. W. Bush, and Bill Clinton.
The dominance of southern border crossing narratives in explanations of detention geographies overshadows the proliferation of interior enforcement.Proximity to borders also fails to explain the construction in the mid-1990s of a joint Immigration and Naturalization Service-US Marshals Service facility in Batavia, New York.While close to the boundary between Canada and the United States, the facility was not designed to control Canada-US border crossings.Indeed, many people are transferred by plane from crowded facilities elsewhere (Tryon, 2012).The location of this facility has more to do with its proximity to federal courts in Buffalo, its location on the New York State Thruway, and its position within a historic prison belt.Attica is 11 mi south and Albion 17 mi north, each town with more than one carceral facility.
This project began as part of a 5-year study on island detentions off the shores of Australia, the European Union, and the United States.In an effort to explain how the US came to operate a dispersed onshore and offshore carceral archipelago, we conducted qualitative research into the political, legal, and economic dynamics responsible for the pace of expansion and locations of confinement developed since the late 1970s.This research involved extensive collection and analysis of archival materials from national and regional archives and Presidential Libraries that provided insight into political negotiations among a range of actors.We also draw on legal scholarship, media coverage, open-ended interviews, and visits to case study sites in the US northeast, southeast, and southwest.Site visits repeatedly revealed the location of detention facilities in economically depressed towns tied to existing or decommissioned carceral and military uses.
These research findings led us to pursue a transnational perspective to better understand the carceral landscape in the United States.In light of frequent discussion, debate, and research on the production of remoteness through detention in the US, we began research in 2010 by asking what is meant by "remote".In the case of the oft-cited study conducted by the National Immigrant Justice Center (2010), remoteness is defined by distance to a major city as a proxy for access to legal advocacy.Elsewhere, remoteness relies on geographical imaginaries that associate isolation with rurality, conservatism, and racism (Bonds, 2009).Similarly, scholars suggest that "remote locations create 'detached geographies' through which detainees are spatially separated from the services that guarantee their rights" (Martin and Mitchelson, 2009:467), including resources endemic to larger cities that offer support, advocacy, interpretation, and information (Mountz, 2011).For example, Mountz (2011b) shows how Fujianese asylum seekers intercepted off the shores of western Canada were detained remotely in the interior of the province, distanced from Vancouver where they would be more likely to have access to advocacy, legal representation, interpreters, and the tribunals of the Immigration and Refugee Board.
Remoteness, thus, is discursively constructed and materially built in opposition to cosmopolitanism, connection, community, rights, freedom, and even an advocacy industry of law firms and NGOs.However, the role of distance in detention is too often simplified to assumptions that detention is located in isolated, subordinated, and exploited places, relying on these same neocolonial understandings of marginal spaces.Gilmore (2007:11), for example, in her analysis of the California prison system, writes, "this apparent marginal-ity is a trick of perspective, because, as every geographer knows, edges are also interfaces" that "connect places into relationships with each other and non-contiguous places".
Far from disconnected from or incidental to vast political and economic changes since the 1970s, prisons and migrant detention prove central features of economic and state restructuring.Gilmore (2007) ties prison expansion in economically dislocated, and often isolated, places in rural California to surpluses of state capacity, finance capital, and labor that have been generated in the course of deindustrialization and political gains of the right.Bonds (2006) examines how reliance on prisons perpetuates poverty in both the rural places where most prisons are constructed and the urban spaces from which most imprisoned people come.Other scholars also point to the increased reliance on imprisonment and policing in shifting forms of governance of racialized groups (Simon, 1998;Coleman, 2007a).Allspach (2010) and Moran et al. (2013) make parallel arguments to Gilmore's observation about prisons as interfaces.They understand prisons not as Erving Goffman's (1961) "total institution" that is fully enclosed and sealed from the outside, but rather as transcarceral spaces traversed by people, information, resources, and transport systems (Moran et al., 2013).Allspach illustrates how carceral spaces extend beyond prison walls following confinement to reinforce the marginalization of formerly incarcerated women.Moran and coauthors focus on moments of state custody before formal incarceration as temporal and spatial liminality to emphasize blurring of inside and outside life.We scale up these arguments to talk about transnational productions of remoteness that cross national borders, in addition to prison walls, creating transnational carceral spaces.
How is remoteness produced transnationally?Much literature about transnational migration emphasizes mobility of people and remittances across distant places (e.g., Rouse, 1991).We add to this discussion the transnational work of the state and the mobility of its infrastructure and operations.States produce remoteness through the development of a transnational infrastructure to contain and disperse migrants in spaces of formal confinement and blocked migration routes.Remoteness is not simply about the isolated location of any one facility.Rather, the onshore-offshore US detention system produces isolation of asylum seekers while connecting them to a transnational infrastructure of interception, apprehension, detention, and removal.
To develop this argument, we build on scholarship illustrating the racialized geopolitics of asylum (Ashutosh and Mountz, 2012) produced in part through bilateral negotiations between states and their influence on landscapes of detention.As Coleman (2007a, b) argues, immigration has always blurred the domestic and foreign spatial and juridical binary.Coleman traces a history of immigration policing in Mexico intersecting with US wars on drugs and immigration across borders and jurisdiction.These "wars" utilize the same practices, strategies, tactics, and authorities, often tar-geting and criminalizing the same people onshore and offshore (Coutin, 2010).Moreover, these wars are not only connected discursively.As Walters (2008) argues, boats serve as important material -if juridically liminal -sites that connect governance of land and sea.
In researching the history of detention facilities, we found the recurrent use of obsolete and active duty military bases to deter, confine, and remove migrants within mainland territory and in offshore Caribbean locations.Site-specific histories need to be unearthed (see, for example, Kaplan, 2005;Vine, 2009;Lipman, 2013;Paik, 2013).Often, these histories of confinement reveal the operationalization of colonial desires to segregate and contain by targeting and criminalizing racialized groups.It is on these racialized, colonized, and militarized grounds that we locate the US's transnational migration detention and deterrence regime.

Case study: island Caribbean to Louisiana
A series of geopolitical, legal, and electoral year battles during the Carter and Reagan administrations set the stage for contemporary US detention and deterrence policies.The 1980 Mariel crisis unfolded over a 6-month period during which 125 000 Cubans arrived by boat in Florida, some transported by Cuban-Americans (Noble, 2011;Lipman, 2013).The Cold War politics of the highly visible "boat lift" proved essential to the response (Stepick, 1982).Cubans had long enjoyed special treatment as people fleeing Fidel Castro's communist regime, but this situation was different.There was deepening concern about the admissibility of this group, some of whom did not have family ties.Rumored to have released people with criminal records, mental health issues, and gay people, Castro bet that President Carter would not risk the image problem of rejecting refugees fleeing communist oppression.Castro bet wrong, and Carter closed the door to Cubans.He then reopened and partially closed entrance again, all within the span of weeks.Rather than enjoying blanket admission as refugees (technically through parole), these Cubans were defined as entrants whose cases would be adjudicated individually.As they arrived in Florida, a civilian-military task force established temporary "processing centers" in civilian spaces in Miami, such as a football stadium, and on military bases as far away as Wisconsin.
With this operation underway, another crisis brewed that would highlight the racialized and geopolitical boundaries of asylum.Late that same summer, some 15 000 Haitian asylum seekers arrived by boat in south Florida.The United States has a long history of neocolonial engagement and military occupation of Haiti (1915Haiti ( -1934)), and maintained relationships favorable to US economic and geopolitical interests, including support of the Duvalier regimes.As a Cold War ally, the US worked to suppress migration from Haiti.During the 1970s, as political and economic conditions in Haiti worsened, over 47 000 Haitian nationals entered the US, most of whom were paroled preceding formal asylum hearings (General Accounting Office, 1983).In 1978, the INS, in consultation with the State Department, established its Haitian Program that mandated confinement for asylum seekers and used mass asylum hearings to reduce the case backlog and speed deportations (Gwynn, 1987;Paik, 2013).None of the 4000 people who went through the program received asylum (Lipman, 2013:121).
In July 1980, a district court ruled that the INS's Haitian Program was unconstitutional because asylum claimants faced a "transparently discriminatory program designed to deport Haitian nationals and no one else" (in Zucker, 1983:157).Asylum seekers arriving that summer nonetheless were confined at Krome, an abandoned Nike missile base some 23 mi from Miami.Lipman (2013:123) explains that Cubans were confined at Krome North and Haitians at Krome South: "Advocates and journalists argued that despite substandard conditions across the board, conditions for Haitians were qualitatively worse than they were for Cubans."The land at Krome South was lower and wetter than Krome North, mosquitoes were relentless, and the barracks designed for 150 operated well beyond capacity with 1000 people (ibid.).
When Reagan entered office in 1981, Krome remained severely overcrowded and there were some 11 000 asylum claims from Haitians to process.Before Reagan was in office, in September 1980, the Dade County Department of Public Health issued a close order for the Krome South facility, citing over two pages of health violations, including inadequate sewage and water systems, dirty portable toilets, severe overcrowding in residential facilities, signs of rodent infestations, locked fire exits, and inadequate laundry facilities2 .In response, the INS transferred Haitians to federal prisons in Otisville, NY, and Lexington, KY, and to inactive military sites, including the Brooklyn Naval Yard and Fort Allen, Puerto Rico.Also considered was the active Fort Drum military base in upstate New York, close to the Canada-US border.

Contingency planning and deterrence onshore
Rather than remedy the discriminatory and dangerous conditions of detention for Haitians by closing Krome and ending their detention, the Reagan administration made detention mandatory for all asylum seekers.This decision and desire to prevent another mass migration proved pivotal in the expansion of detention and the use of detention as deterrent in the years to come3 .The INS's new detention policy aimed to create a "fear of being detained" in order to prevent another mass migration like Mariel (Helton, 1986:374).Thus, detention and interdiction became two prongs of the Reagan administration's deterrence policy.Remoteness and even climate were cited as important aspects of the deterrence strategy.After a downturn in the numbers of people seeking entry to the US, an INS spokesperson said, "a side effect may well be that there has been so much press about it [Fort Drum] being Siberia that it may have caused people to think twice" (Jaynes, 1981).
The Reagan administration committed to a course of "contingency planning" in the case of another "immigration emergency".In April 1981, the INS surveyed seven sites for such use and to meet its long-term detention needs (GAO, 1983).Sites under consideration included Krome North and South; Eglin AFB near Fort Walton, Florida; Ellington AFB near Houston, Texas; Hamilton AFB near San Francisco, California; Craig AFB near Selma, Alabama; a military radar site near Roanoke Rapids, North Carolina; and jails in Florida, and Bureau of Prisons space.After this initial survey, Fort Allen, near Ponce, Puerto Rico, was added to the list.The site planning team recommended using Fort Allen even though there were considerable problems, including the "possibility of terrorist acts" by Puerto Rican nationalists, and the fact that Fort Allen "would not provide easy access for attorney advocates and voluntary agencies" (GAO, 1983:26).
Under pressure from the state of Florida, the Reagan administration announced that it would transfer Haitian asylum seekers to Fort Allen, Puerto Rico (a plan begun under Carter's administration that had failed to get off the ground), and began mass asylum hearings there (Gwynn, 1987:343).Costs were considerably higher (USD 65/day) than they were at either Krome (USD 49) or federal corrections institutions (USD 35), primarily because of the large numbers of security personnel employed (GAO, 1983:29-30).Less than 1 year after it opened, the INS concluded that the "Fort Allen facility is not best suited for current or future detention space needs"4 .
In addition to confining Haitians and Cubans whom Cuba would not accept returned, the US government responded to the influx of Salvadoran refugees in 1983 once again using mandatory detention.The US government's insistence on mandatory detention of asylum seekers created a jurisdictional and logistical crisis, providing conditions for the push to construct additional detention space.The INS did not have the experience or infrastructure to carry out its new commitment to long-term detention.The Bureau of Prisons (also within the Department of Justice) was deeply involved with the INS in planning for the latter's new long-term de-tention facility.The search ensued for a site that was "ideally located in the south-central United States, on a generally direct route to South and Central America, and [would] be able to accommodate the longer term cases from Krome, Brooklyn, and the border facilities" (GAO, 1983:26).Fierce lobbying led by the mayor of an economically depressed, central Louisiana town resulted in the construction of the INS's first new facility, jointly run with the Bureau of Prisons.The new facility in Oakdale more than doubled the INS's capacity nationwide, and residents agreed to accommodate an additional short-term population of 10 000 people in case of an immigration emergency.Local parish jails in the state also pursued contracts for federal detainees.Local agencies provided additional capacity near a federal facility carrying out immigration hearings and processing removals.This institutional landscape would be repeated in jurisdictions across the country.

Interdiction, interception, and deterrence
The second prong of Reagan's onshore-offshore deterrence policy aimed to prevent arrival of unauthorized migrants on US territory by effectively creating a buffer zone.This policy issued from the desire to forestall another Mariel, but also from high legal hurdles and political costs of removing asylum seekers once present in US territory.President Reagan issued the offshore deterrence policy on 29 September 1981.Executive Order 12324 directed the Secretary of State to enter into bilateral agreements with foreign governments that would allow US Coast Guard vessels to stop, board, and return foreign vessels (Frenzen, 2010:378).INS officials began conducting credible fear interviews onboard Coast Guard ships to decide whether a person would be allowed to enter the United States to make an asylum claim or be returned to Haiti, the only nation-state that entered into the agreement with the United States at the time. 5 Following the 1991 military coup against popularly elected Jean-Bertrand Aristide, President George H. W. Bush initially indicated that Haitians should not be returned to Haiti.Instead, he offered "safe haven" on the naval base at Guantánamo and sought promises of additional "safe havens" from other nation-states in the Caribbean.With more than 12 500 asylum seekers already confined on Guantánamo, and another exodus expected (Paik, 2013:150), Bush reversed this commitment.In an executive order issued from his Maine summer home, the Kennebunkport Order declared that US obligations to international obligations of non-5 Interception generally refers to the practice of stopping of boats at sea before they reach land.Interdiction is a legal term referring to the practice of stopping potential asylum seekers before they reach sovereign territory where they may seek asylum.We find it significant that the US government routinely referred to its offshore operations at sea as interdiction rather than interception.Most governments today would label the practice of stopping boats at sea interception.
refoulement applied only within US territorial waters, a declaration that led to a sharp decline in the numbers of Haitians attempting to depart (Frenzen, 2010:383).
On the presidential campaign trail, Bill Clinton vocally opposed Bush's policy, but once in office, he continued the Reagan and Bush administrations' claim that practices of interdiction do not violate the principles of freedom of movement in international waters and non-return under the UN Convention Relating to the Status of Refugees or its 1967 Protocol (Paik, 2013:149).The Supreme Court ruling on Sale v. HCC, which challenged Bush's Kennebunkport order, essentially shrunk the space of US obligations to its land boundaries (Frenzen, 2010:386).By setting its own legal geographic terms, this decision provided the political and legal basis for the US to continue its regional deterrence strategy, while ostensibly upholding principles of non-refoulement (Koh, 1994:139).
Reagan's creation of a maritime buffer zone entailed deployment of the Coast Guard to intercept boats at sea.Under the Bush and then Clinton administrations, the confinement of asylum seekers offshore was added to this practice.Moreover, the US government signed agreements with governments in the Caribbean and Central and South America to enhance their migration control efforts (Frenzen, 2010;Paik, 2013).By July 1994, over 16 000 Haitians were held in "safe haven" at Guantánamo, and the US began to construct additional "safe haven" camps in Antigua, Dominica, St. Lucia, Suriname, and the Turks and Caicos (Koh, 1994:154).In the midst of this offshore operation, approximately 30 000 Cubans departed the island in July 1994, after Castro announced that he would allow them to leave the country.Diverging from its historical blanket admission of Cuban refugees, the US began intercepting boats carrying Cubans.By August 1994, 23 000 Cubans were confined on Guantánamo.Crowding there led the US to enter into an agreement with Panama to use space on a US military base in Panama to confine 9000 Cubans (Koh, 1994:155).

Understanding remoteness in transnational context
The turn to offshore migration control efforts in the early 1980s created a buffer zone that would prevent arrivals and the lengthy legal battles over asylum, parole, and detention that had developed around the cases of Haitian and Cuban people.Walters (2008) argues that scholars need to pay more attention to sites where land meets sea.Attention to the buffer zone as one such site challenges "landed" ontologies of geography, shifting attention to fights over jurisdiction, access to rights, and constructions of jurisdiction at sea (Steinberg, 2001;Benton, 2005).
In the 1990s, this strategy would develop into a "safe haven" strategy wherein Haitian and Cuban asylum seekers were confined on US military bases and in spaces that the US funded in other nation-states.The maritime spaces of US territorial and international waters functioned as a buffer zone deflecting people from US shores.This carceral space at sea grew with the proliferation of spaces of confinement, from Coast Guard cutters to so-called safe havens.Therefore, the expansion of the US carceral regime involved a transnational proliferation of spaces of confinement on land and sea, on mainland territory and offshore: the buffer zone at sea, the safe haven on military bases, and the construction and repurposing of spaces on military bases on the mainland.To return to our use of the term remoteness, it is impossible to ascribe any one definition or narrative to remoteness, and helpful instead to look to the transnational, interlocking networks of spaces of confinement and associated policies.This panoply of sites resembles Agamben's (1998) space of exception: zones in the borderlands where people are at once inside and outside of sovereign territory.In this case, the US government exercises power to control territory, mobility, and access while operating beyond its jurisdiction.Boats function as islands of law (Benton, 2005) and carceral spaces (Pickering, 2014) as they move through these jurisdictionally ambiguous spaces of exception.
Most discussions of transnationalism focus on mobility of people and remittances across borders.In this analysis, we have added the mobility of state enforcement infrastructure (Mountz, 2011b), focusing on creation of a transnational regime of confinement in the Caribbean and on US territory.The US established bilateral agreements with Haiti and Cuba to return migrants, monitored Bahamian and Dominican expulsions of Haitians to anticipate potential arrivals in the US, and partnered with other states to carry out its safe haven policy.These geopolitical negotiations built on domestic detention infrastructure designed to facilitate removal.These efforts illustrate Rodríguez's (2008) contention: "The U.S. prison is a global statecraft, an arrangement and mobilization of violence that is, from its very inception, already unhinged from the delimiting 'domestic' (or 'national') sites to which it is presumptively tethered." The history we trace suggests that the globality of US statecraft could be achieved only through its negotiations with other nation-states.Transnational analysis of the US carceral regime thus extends the existing conceptualization of transcarcerality in carceral geographies literature to additional spaces of liminality (Walters, 2008) and where maritime spaces are constructed as sites of policing (Pickering, 2014).
The concerted and creative lengths that the US went to to remove and prevent the arrivals of Haitian people repeatedly led to the invocation of exceptional categories, creation of new laws, and consolidation of relations between Blackness and excludability, detainability, and criminality (Loyd, 2015).As Miller writes, "the legislative will to criminalize certain kinds of immigration-related conduct correlates closely to a crisis of legitimacy that immigration policy experienced after 1975 -acutely so after the 1980 Mariel Boatlift -as well as the popularity of 'tough on crime' measures already well underway in the same legislative arena" (Miller, 2003:627).Yet the imposition of long-term, mandatory detention of Haitian asylum seekers is consistently overlooked as the origins of contemporary migration detention.The persistent carceral treatment of Haitian asylees and refugees in detention and safe haven spaces -the "ceaseless struggle to expeditiously return as many as possible to Haiti" (Stepick, 1982:163) -reveals how carceral logics of humanitarian migration control fueled detainability, setting the stage for subsequent criminalization.
The use of the buffer zone strategy was repeated by Western governments in borderlands that migrants and asylum seekers cross frequently.These spatial patterns to offshore, distance, and other asylum seeking and seekers through remote control, interception, and detention became wellrehearsed tactics (Mountz, 2010(Mountz, , 2011a)).Australia and the European Union set up their own buffer zones with interception at sea and use of islands to detain people closer to regions of origin and sites of interception.Tracing this offshoring from our historical analysis to the present shows that transnational carceral regimes are neither new nor isolated.Rather, state authorities think and act transnationally to control human mobility, and carceral regimes must be understood in these broader geopolitical contexts.

Conclusions
The US strategy of relying on detention as deterrent clashed with its capacity to detain.Crises of capacity would recur as this commitment grew more entrenched institutionally.For example, the lengthy detention of asylum seekers, which varied by nationality, contributed to the space the government needed for confinement.In 1990, people were detained for an average of 23 days, but Haitians were held at Krome for an average of 101 days, and Salvadorans between 21 and 92 days, depending on the facility (GAO, 1992:25, 27).This disparity reflects the intertwining of racial and geopolitics.Moreover, the Anti-Drug Abuse Act of 1988 defined a new category of offenses called "aggravated felonies", and mandated the deportation of non-citizens.Thus, in addition to extending periods of detention for asylum seekers, this legislation increased categories of people subjected to detention.Overcrowding in jail and prison space that the INS relied upon meant that much of it was unavailable for contract use.By the 1990s, as a consequence of all of these developments, coupled with renewed asylum movements from Haiti, Cuba, and Central America, the US commitment to detention as deterrent was in crisis.The General Accounting Office concluded that it was not "feasible to expand INS' detention capabilities sufficiently to solve the problems" (GAO, 1992:43).Its report focused on capacity onshore, but as we have shown in the controversy over safe haven, the depths of the crisis must also be considered offshore.
A. Mountz and J. Loyd: Transnational productions of remoteness This brief sketch of the Caribbean roots of the contemporary US detention system sheds light on three dynamics.First, the prevailing understanding of detention as a response to undocumented entry across the US-Mexico border conflates detention dynamics of different agencies (INS and Border Patrol), the role of interior enforcement, and the policies for mandatory detention developed to deter and remove Haitian asylum seekers unwanted by power brokers in Florida (Stepick, 1982).The dominance of the southern border narrative thus obscures different racialized dynamics of migration and Caribbean Cold War dynamics that set the institutional stage for deterrence policy on that boundary (Loyd, 2014).This late Cold War story is one in which remoteness was produced transnationally through geopolitical, carceral, and racialized crises.Preventing Haitian arrivals expanded the government's reliance on detention for all groups, extending racialized carceral responses to mobility.Migrant detention, then, is built not only on the racism of the criminal justice system but also on the racism of asylum policy.This finding contributes to interrelated histories of racism and xenophobia that have implications for building interracial coalitions against criminalization (Simon, 1998;Loyd et al., 2012).
Second, ahistorical and placeless explanations obscure the importance of regionally specific detention dynamics such as local military and prison geographies.The new strategy of deterrence relied on existing military facilities and repurposed military bases to respond to mass spontaneous migration.The current system continues to rely on some of these same places, such as Krome.Furthermore, the ways in which local communities responded to the restructuring of the military itself played an important role in the expansion of detention and prison capacity.Leaders and residents near Oakdale, for example, organized to transform the area's new surplus in WWII military bases into new federal revenue streams in the form of immigrant detention, the Bureau of Prisons, and a regional prisoner and alien transport hub.Further attention to the repurposing of military capacity for the operation of detention and removal is needed.
Third, the problem of remote conditions of detention is tied up with mandatory detention, and the very expansion of the detention system.Over this period, each of our case study sites either developed a detention and prison economy or deepened longstanding reliance on prison revenues.Our research shows that while legally distinct, in practice civil detention is built upon the material and ideological infrastructure of the criminal justice system and military infrastructures.
This history and the spatial patterns deployed to mediate access to rights and sovereign territory have been repeated many times, globally.In the US, the pattern of dispersed and fraught site selection was repeated in protests over federal decisions to detain and process some 59 000 Central American minors who crossed the US-Mexico border in 2014.Once again, the implementation of national policy hinged on lo-cally contingent histories of economic dependence and resistance.As we write, the Obama administration has invited leaders of Honduras, Guatemala, and Nicaragua to a meeting to discuss implementation of processing of children's asylum claims in their countries.
We join other scholars in arguing for transnational framings of domestic carceral regimes (Sudbury, 2005;Rodríguez, 2008;Loyd et al., 2012), geopolitical understandings of racialized struggles over asylum (Ashutosh and Mountz, 2012), and deeper analysis of the implications of immigration policies that blur domestic and foreign spaces and enforcement policies (Coleman, 2010).These geographical moves to geopoliticized and racialized analyses of domestic carceral regimes require sustained attention to both transnational perspectives and site-specific histories and contingencies.The past must be examined in order to understand contemporary landscapes of detention and their symbiotic onshore and offshore dynamics.

Figure 1 .
Figure 1.Hysteria along the US-Mexico border: map showing relationship between detention capacity and demand(Schriro, 2009).