This is the draft text of an article I completed during our staff/student writing workshop, revised from earlier.
Waiting for the delivery man: Temporalities of addiction, withdrawal and the pleasures of dope time
A range of work published in the drug field has reinscribed drug user experience as embodied – challenging researchers to think of pleasure beyond transgression and leisure, into the routine micro-time, embodied, material, domesticated, habitual, remote intimacy. The paper examines changes in the material culture of illicit drug use and addiction discourse in the light of changed modes of drug distribution due to the expansion of digitally mediated markets. The concept of dope time is introduced as one kind of social time that is a part of how opioid dependent users present and talk about themselves. Dope time is generated by both the supply system and the experience of addiction and drug consumption. It can be time spent waiting for drugs to arrive, time waiting for dopesickness to start, or to stop, or time in the clinic waiting for treatment. It is also used as a mechanism of social control, such as deliberately introducing waiting as part of treatment regime to test if the addict is ‘serious’. This paper examines the changing quality of heroin or ‘dope’ time in the context of the shift towards digitally enabled drug markets using qualitative and ethnographic data on user experiences. It draws on data from darknet cryptomarket users who buy and sell using the hidden digital infrastructure. It argues that dope time in the darknet changes to being much more defined by the infrastructure. There are obdurate times dictated by the delivery infrastructure, such as shipment times. Its salience is defined by the drug user’s sense as to whether the time spend waiting is intentional. If he or she regards delayed shipments as the responsibility of the vendor or due to deliberate indifference on their part then this time is experienced more harshly. Dopesickness becomes more painful, and anxiety grows where that is the case. One reason for that is that the user is concerned that the drug may not arrive at all. That feature of the infrastructure then changes the texture of dope time for the user. This is partly a matter of greater convenience and choice but also reshapes the idea of time into one mediated by systems rather than interpersonally.
Introduction: The Time Infrastructure in Illicit Digital Markets
‘When his junk is cut off, the clock runs down and stops. All he can do is hang on and wait for non-junk time to start’ (Burroughs, Junky, 1977)
Time is culturally significant in popular accounts the life of dependent drug users. Burroughs (1977) wrote of ‘junk time’, describing the drug users as in essence a mechanism for obtaining drugs, dependent on the internal ‘clock’ of addiction and withdrawal. Though this is a reductive and inaccurate framing, it does describe on element of social time and drug user. Junk time is one culturally salient element of the drug chronotype which encompasses a range of rhythms and trajectories embedded in consumption rituals, biographies, treatment systems, criminal justice processes and other structuring factors which shape the time of the drug user (Fraser 2006). Time can be used as punishment. Deliberately producing waits for treatment is one way of inducing dopesickeness as a punishment for wayward addicts in treatment (Bourgois 2000). Waiting can be a clinical imposition and a clinical-treatment construct. Letting dependent users wait means they are ‘proper’ patients rather than manipulative addicts (Bourgois 2000).
In this chapter I intend to examine one particular way in which time is structured for drug users, through the times of the marketplace. I do this through a study of experiences of time among users of cryptomarkets, anonymous digital markets used for drug purchase. It matters because the way in which markets structure drug buying is critical to both the way in which users position and identify themselves and experience their drug use. The paper pulls together the findings of many studies of cryptomarkets that have examined the way in which the buying process shapes buyer identity and behaviour (Aldridge and Décary-Hétu 2016; Barratt, Lenton, et al. 2016).
The cryptomarkets are open, private markets hosted on the Tor darknet which take advantage of some of these features. Combined with distributed accountings systems like bitcoin or monero they allow transactions to take place in relative anonymity, with buyer, seller and host unknown to each other. Drugs are purchased and delivered through courier or postal systems, or less typically left at dead drops. Online drug markets create new contexts for dealing and risk, and Aldridge and Askew (2017) describe illicit transactions in this domain as ‘‘stretched’ across time, virtual and physical space, and handlers.’ The reconfiguring of time and space is a key feature which reshapes the experience of drug users interacting with the market. It changes the time spent obtaining the means to buy drugs, obtaining drugs and consuming drugs (Taylor 1993). For example, drug use patterns are changed by the cryptomarkets’ tendency towards requiring larger and larger purchases, encouraging stockpiling and self-titration over a period of weeks and months.
Developing that, this paper uses the concept of social time. It defines time as a multiple set of cultural, structural and disciplinary rhythms (Sorokin and Merton 1937). In that sense, time is created from cultural referents and material determinants (Munn 2020). Social time allow us to understand the combination of drugs as technologies, the market infrastructure, and users’ self and their orientation to a future self. It applies concepts of eventalisation, trajectory and career in order to do this. It tells us how online communities are developing these ways of understanding drug use through processes of asynchronisation and the creation of a community developed drug ontology (Bilgrei 2016). The paper does this in the context of the developing understanding of the interaction of human and non-human in producing drug use contexts (Dennis and Farrugia 2017). A range of work published in the drug field has inscribed drug user experiences as embodied and material. In doing so it challenges researchers to think of pleasure beyond transgression and leisure, into the routine micro-time, embodied, material, domesticated, habitual, remote intimacy times (Dennis 2019; Duff 2011). I aim in my work to examine the connection between materialised market economies and social time, by examining the combination of technologies, practices, users that makes up the illicit economy. Structuring elements are contingent and produce ontological instabilities and reconfigurations of users and contexts as they produce stabilities and configurations (Duff 2014).
Time matters because we are overwhelmed with technical and disciplinary times over other social rhythms. Neoclassical economics and data capitalism both establish conceptions of time that are critical in disciplining society and which create their own socialites, values and hierarchies. Extractive data capitalism presents its own time modes as natural and inevitable (Zuboff 2015). We often encounter this in technical times of nanoseconds and processor theft, the destabilisation of the gig economy and the rapid fire demands it places on labourers within it. Licit and illicit economies of intoxication have grown more like each other in terms of production, distribution and consumption, labour organisation and digital services.
Markets function effectively when they are routinised. As noted by Collier and colleagues (Collier et al. 2020) the infrastructure of illicit digital markets is often hidden from the view of observers and participants. This infrastructure involves a large array of routine labour dependent on shared infrastructure. A wide array of services are provided such as hosting, site design, call centres, franchise management, data analysis and banking/cashing out (Kremez and Carter 2021). The growth of service crime drives down the cost of involvement in illicit digital markets for vendor and buyer. However this initial efficiency can lead to further inefficiencies which then require other services to manage.
To take a simple example, markets are typically thought of as places of instantaneous exchange. When a drug transaction is agreed, cash or electronic currency is transferred immediately. However this is not the case. Bitcoin, the supposedly decentralised currency, is often used for online drug payments. As bitcoin has become more valuable as a commodity more investment has flooded into producing (‘mining’) it electronically. That investment has however not led to the payment procedure becoming more efficient. The system has become bogged down and transactions can take hours to clear without further payment. The response has been to create a technical and social infrastructure to manage that using a combination of exchanges and trusted partners to make the system work. The transactions are stretched over time and embedded in this trust infrastructure (Bancroft et al. 2019).
In this chapter I explore further effects of the cryptomarket infrastructure, largely through the ways in which it structures the time of waiting – waiting for exchange confirmation, and waiting for drug delivery. The chapter began when reviewing data I had collected from a leading cryptomarket I noticed how often concepts of ‘dopesickness’ (heroin withdrawal) were showing up in the same codes as references to time and waiting. That led me to examine how llicit sociality incorporates platform and computational effects in the subjectivity of drug users involved in these platforms. It is central to how they inhabit these digital spaces.
Waiting can be traced through the classic ‘waiting time’ of the dopefiend where time appears as a quality to be manipulated by the dealer. The queue is a typical ordering principle. Market time management practices are observable in face to face markets, for example a free for all can be contrasted with the ordered, governed queuing in a street market (Kleiman 1988). Time can be spent waiting for drugs to arrive through the delivery system, or time waiting for dopesickness to start, or to stop.
Withdrawal is the primary frame for understanding heroin dependence (Walmsley 2016).. It is central to science and subcultural understandings of heroin addiction. Dependent opiate users in the 18th century using substitution with alcohol to cope with withdrawal, or mixed opium with wax as a self-care method. Withdrawal steadily came to be viewed as a process to dangerous to be left to self-care, one that could be traumatic, deadly or just impossible for the user to embark on themselves. In the UK in the early 20th century maintenance was being used but this was strictly class based. The abrupt withdrawal method was used for ‘dangerous’ drug addicts such as prisoners. The Abstinence Syndrome Intensity scale was developed by 1944 to quantity withdrawal. Symptoms of withdrawal were objectified, no longer reliant on subjective self-report. The individual addict could no longer ‘speak’ their condition. Psychological dependence was excluded from the understanding of addiction at the time.
There was a change in the 1960s and 70s away from complete abstinence as an aim – critique derived from recognition that detoxing just led to addicts reusing immediately on discharge. No permanent cure was being offered. Two truth producing mechanisms were worked into diagnosis: urinanalysis and the Opiate withdrawal scale. At this point treatment was moving to exclude ‘pseudo-addicts’ and the reverse, addicts who claimed to be clean but were not. These truth mechanisms framed the patient as untrustworthy, a framing many addicts accepted. From the 1980s on – heroin withdrawal was defined as destablisiing risk management strategies, making the subject resistant to rational, looking ahead decision making. Language is now changing, influenced by neuroscience, which now reunifies addiction as a set of neurobiological mechanisms at work.
When buying drugs for delivery waiting is part of the experience. In the Global Drug Survey 2014, 50.9% of users who had drugs delivered mentioned waiting a long time for the product (Barratt, Ferris, and Winstock 2016). What that meant is subjective but this is something where the face to face and local delivery markets have an advantage. The explosion in drug delivery is partly a matter of greater convenience and choice but also reshapes the idea of time into one mediated by systems rather than personally.
The first time concept I identify is the bio-time of heroin. It is a combination of the bio-pharmacokinetic qualities of heroin which become known through the pharmacological repertoire and the embodied experience of dependence and withdrawal.
Users described a journey towards heroin use, in this account the use of heroin consumes the self in both the ‘nod’, the happy state, and in experiencing withdrawal:
Basically I’m trying to say everyone is different, some are more hedonistic than others, some can simply handle the anxiety, some maybe even enjoy the lifestyle, everyone is different in their use so it really all depends on the type of person. Basically I use because I used to be depressed all the time and hated my medication, weed wasn’t too helpful either (sorry weed, still luv u), but once I started messing around with dope I realized I was either blissfully ignorant and happy all the time, or just going mad with cravings.
The experiencing of heroin dependence was an encounter of the body as obdurate and wilful. Many heroin users combined their use with benzodiazepines, as in this account:
‘Also for me after dose adjustmens it takes about 3 days for my brain to adjust and that is after taking it recreationally for about 6 weeks. In the beginning its a lot of fun but you sadly rapidly develop a tolerance to the hypnotic/sedative effect of benzos. After I successfully tapered off I will take a month brake and then only use it on weekends to smooth out comedowns from opioids or stimulants. It is a lot of fun on its own or as a little helper if you need to catch some sleep after a stimulant binge but taking it recreationally for more then a couple of days in succession is a waste.’
This was part of their pharmacological repertoire. I continue with a typical account of a pharmacological repertoire from a thread reviewing opiate vendors. A developed pharmacological repertoire is typical for experienced users who will have elaborate sets of drugs and drug use practices to manage.
“So if you run out of heroin or just decide to stop using, you will want to drink kratom during the period you would normally be [dope]sick (about 10 days for me). After that, you can keep drinking it (because why not… it’s awesome) or you can stop at any time because unlike suboxone, it doesn’t cause withdrawal or dependence — I would know, I’ve been drinking it with friends for 3 years and we can (and do) stop anytime, with no negative side effects. I can’t believe every junkie doesn’t keep a stash of it, its so cheap.. and legal!
In this example the legal high kratom is being used to manage dopesickness. In this account it is preferred to the prescribed substitution therapy suboxone as it is perceives to not carry a risk of dependence itself. The time horizon is three years long. The extended time horizon illustrates how addiction experiences are built up and change over the long term. The experience is varied and reflected in interactions and social relations with others. The ‘we can stop at any time’ claim is somewhat belied by the fact that they have continued to use it through this period. The stopping criteria are relevant in evaluating the decision to move from one drug to another. Users of methadone mentioned in the forum that they found stopping difficult and cautioned others against it, similarly to this user’s comment on suboxone.
Avoiding dopesickness was one motivation that structured users interactions with the market. It was not completely dominant however. Experiencing withdrawal was not wholly disastrous, as long as there was a sense that it would happen within a manageable time heroin. A user on the heroin thread described this process of controlled waiting
“ alright cool so I’m gonna try to snag a bundle from [vendor] in the morning before he sells out.. idk [I don’t know] why I always wait till the last minute to get more. I’m totally gonna be sick till I get something lol. such an idiot.. i had the cash days ago but im trying to not spend so much right now. ”
Information was relevant to managing dope time. Tracking information gave some predictability to deliveries, but was often not offered for reasons of security. This example from a discussion of heroin vendors describes the work need to make the drug useable which adds to the waiting time.
“ My last order with [vendor] took a total of 8 days, which set a new record for me. … I would not have cared so much about the 8 days if it was fire, but it sadly was not … know I did not feel so great the next day, wasn’t sure if it was the dope or lack of sleep, but usually lack of sleep just makes me a zombie. I eventually cleaned it with some dry acetone and it took out all the brownish tint and somewhat of the iodine smell. I will probably do that for every order here on out. My delima now is i’m out and I am eyeballing the leftovers from my cleaning that I let dry in a jar, my mind says no but my heart says yes! Uggh I should just toss it. ”
This malleability and uncertainty about the drug as an object is common to many users. The individualised context of use was apparent in this account and more typical of darknet market users. They tended to buy and consume individually, and were attracted to the darknet because it allowed them to do that.
Rhythms and stretches
Time terminology appears a lot in drug users accounts in the form of rushes, lost time, blackouts. There is a close relationship with pharmaceutical quality and pharmacokinetics:
‘As I mentioned earlier, I started off with a 25mg shot. I had done no opiates for the past week or two, and the 25mg shot was the perfect amount of dope to get me where I wanted to be. I booted it up, and a few seconds later I had a very familiar warm wave of relief and euphoria slowly creep from my head down to my toes… It’s not a strong rush compared to other opiates like BTH that’s high in 6-‐MAM or hydro/oxymorphone, but it is a very comfortable rush and the way it makes you feel is just plain fantastic.’
Users combine this with understanding of the platform qualities – the times bitcoin takes to clear, escrow to be completed, and deliveries to happen.
It also has very good legs, my first shot lasted me over 4 hours before I decided to add about 10mg more to my bloodstream, and I was still feeling good before I did the second shot. I’m currently 3 hours into my morning dose and I still feel like it’s at full strength. It also may be worth pointing out that I did a speedball shot this morning, and it was just plain amazing. I’ve got some weak coke, but it’s not cut with any active ingredients like amphetamines so with a good little line of that mixed up in the shot with the CW, it was the best breakfast shot ever. 100/5’
The delivery system also provided sensory pleasures. One heroin user described the anticipation and excitement of a delivery and the smell and sight of high quality heroin:
“ Within 5 minutes of receiving I anxiously ripped the letter apart (I’ve been sick all morning so this couldn’t have come at a better time) was going to test it, but opened my paraphernalia container and realized I’m out! I was not too concerned however, because of other user referrals. So I get [vendors] gear open (all shakily from w/d) and the oh so lovely smell hits me! WHACK! I put about ~50mg out, tooted and waited a few to see if I needed more… NOPE! The product looks exactly as the picture. … Best bang for your buck EASY!
The system was not just about avoiding withdrawal. The production of withdrawal could be sought, but also warned against as dangerous as in this thread on naloxone where a user discussing using nalaoxone to shorten the withdrawal period:
‘“ this is a BAD idea:My understanding is forced precipitated withdrawals is not just a short cut to like day 2, it’s a short cut to the most intense hell on earth because the rate at which opioids leave your receptors has an effect on the severity of withdrawals. Meaning, it’s going to be way way worse to force precipitated withdrawals then if you let yourself go into withdrawals naturally. Like way worse. There is a reason ‘rapid detox’ is usually done under anesthesia and under medical supervision.I am not a pro so I don’t know any of this for sure, but please wait until somebody who knows for sure responds. ”
Alluding to a later topic there was a sense of the body being vulnerable and punished by the administration of naloxone. Naloxone could be a serious risk. On the other hand many users described using naloxone in order to avoid dangerous overdose and in that sense it could be a safety valve. Another use of the drug was as a time skip
“ Hey again! So bit of a funny one. I was given a naloxone injector as part of a drugs training thing (with 5 doses in it). I’ve always thought that could be a great way of skipping ahead to day 2 of withdrawals. well, 1= is that possible? The other thing is I can’t inject myself and my wife flat out refuses, thinking it’s a bad idea. She always says I have bad ideas and she’s always right, so I’m inclined to side with her a bit. Sooo… 2= what else can I do with this liquid? Put it on my tongue, mix it with a drink and swallow, put it up my ass? I’m guessing it’s a full on NO. ” (Bluelight)
Naloxone could be used to manage the stretches of waiting time that users were wary of. There was a sense of time being stretched in these accounts, as naloxone and fentanyl being used to manage this empty time.
Material and governance structure
There are obdurate times dictated by the delivery infrastructure, such as shipment times. These become part of the dope time. Its salience is defined by the drug user’s sense as to whether the time spend waiting is intentional. If he or she regards delayed shipments as the responsibility of the vendor or due to deliberate indifference on their part then this time is experienced more harshly. Dopesickness becomes more painful, and anxiety grows where that is the case. One reason for that is that the user is concerned that the drug may not arrive at all. That feature of the infrastructure then changes the texture of dope time for the user. It reminds them that the power in the relationship fundamentally lies with the vendor. The user worries that they may be thrown back on an unreliable face to face market, or have to go without. Dope time becomes upended.
This user describes the anxiety induced by these platform limits
Placed a little order and will report back, but interested to hear others thoughts. I ordered and supposedly the order was “shipped” from HIDETHEDRUGS 9 days ago. In resolution now. Told him it was going there if he did start communicating about the location. 4 days ago he asked for the zipcode. Then nothing until I went to resolution. Then he asked for the zip again and said he would get back after shipping everything out. Sent it with privnote this time and then he said he couldnt open it (though I got the “note has been read” notice. He wanted it again in PGP [encrypted]. Sent it AGAIN 9 hours ago. He has been on 6 hours ago. No answer. Still waiting.
The market’s escrow system, encryption and bitcoin’s gumminess all add to this wait.
Time can be perceived as being as a punishment, by enforcing waiting: waiting to be admitted to a programme, for treatment to start, for doses to become effective. In this example of prescription methadone
‘At the clinic (the one I went to) the lowest they start you is 15mg/day. But if .5mg of Buprenorphine keeps you good I would imagine 5mgs of methadone would be along the same lines… I was in the clinic for 3 years on 140mg/day, lost my job, and had to quit cold turkey from 100 mg/day. I was sick for a month. Methadone is a road best treaded carefully.’
Waiting time became more salient in the context of medicalised methadone and suboxone prescription compared to waiting for a delivery of heroin. Waiting for medicalised substitution drugs was different in two ways. It was mediated by a clinic or other service which required the user to adapt to the service’s timescale. It also was not going to produce a pleasurable experience but solely allay withdrawal symptoms.
Heroin’s bio time appears as a comprehensible, graspable form of sociality which is articulated through the infrastructure of the cryptomarket. It is one way in which a recognised shared cultural understanding of time in the context of drug use is reproduced and transformed through the digital market. The experience of time waiting for the drug is now distributed through several novel systems: the market infrastructure itself, the associated payment systems, the discussion boards, and the postal/courier system. As described at the start of the chapter, many heroin users have experienced institutional time as a sometimes supportive, sometimes alienating experience. Clinical time might be perceived as protective or as deliberately hostile depending on the context, prison time as explicitly punishing. The cryptomarket infrastructure allows users to share social time constructs and to some extent manage time themselves without being subject to the will of dealers or the governance processes of the treatment and criminal justice worlds. The market itself however is not purely experienced as a convenient, consumer focused infrastructure. It produces its own governance systems and demands which users also have to adapt to and incorporate into their novel sense of dope time materialised through a digital market.
Aldridge, J., and D. Décary-Hétu. 2016. ‘Cryptomarkets and the Future of Illicit Drug Markets’. Pp. 23–30 in Internet and drug markets, EMCDDA insights. Publications Office of the European Union, Luxembourg.
Aldridge, Judith, and Rebecca Askew. 2017. ‘Delivery Dilemmas: How Drug Cryptomarket Users Identify and Seek to Reduce Their Risk of Detection by Law Enforcement’. International Journal of Drug Policy 41(Supplement C):101–9. doi: 10.1016/j.drugpo.2016.10.010.
Bancroft, Angus, Tim Squirrell, Andreas Zaunseder, and Irene Rafanell,. 2019. ‘Producing Trust among Illicit Actors: A Techno-Social Approach to an Online Illicit Market’. Sociological Research Online.
Barratt, Monica J., Jason A. Ferris, and Adam R. Winstock. 2016. ‘Safer Scoring? Cryptomarkets, Social Supply and Drug Market Violence’. International Journal of Drug Policy 35:24–31. doi: 10.1016/j.drugpo.2016.04.019.
Barratt, Monica J., Simon Lenton, Alexia Maddox, and Matthew Allen. 2016. ‘“What If You Live on Top of a Bakery and You like Cakes?”–Drug Use and Harm Trajectories before, during and after the Emergence of Silk Road’. International Journal of Drug Policy 35:50–57. doi: 10.1016/j.drugpo.2016.04.006.
Bilgrei, Ola Røed. 2016. ‘From “Herbal Highs” to the “Heroin of Cannabis”: Exploring the Evolving Discourse on Synthetic Cannabinoid Use in a Norwegian Internet Drug Forum’. International Journal of Drug Policy 29:1–8. doi: 10.1016/j.drugpo.2016.01.011.
Bourgois, Philippe. 2000. ‘Disciplining Addictions: The Bio-Politics of Methadone and Heroin in the United States’. Culture, Medicine and Psychiatry 24(2):165–95. doi: 10.1023/A:1005574918294.
Burroughs, William S. 1977. Junky. London: Harmondsworth/Penguin.
Collier, Benjamin, Richard Clayton, Alice Hutchings, and Daniel Thomas. 2020. ‘Cybercrime Is (Often) Boring: Maintaining the Infrastructure of Cybercrime Economies’.
Dennis, Fay. 2019. Injecting Bodies in More-than-Human Worlds. Routledge.
Dennis, Fay, and Adrian Farrugia. 2017. ‘Materialising Drugged Pleasures: Practice, Politics, Care’. International Journal of Drug Policy 49:86–91. doi: 10.1016/j.drugpo.2017.10.001.
Duff, Cameron. 2011. ‘Reassembling (Social) Contexts: New Directions for a Sociology of Drugs’. International Journal of Drug Policy 22(6):404–6. doi: 10.1016/j.drugpo.2011.09.005.
Duff, Cameron. 2014. ‘The Place and Time of Drugs’. International Journal of Drug Policy 25(3):633–39. doi: 10.1016/j.drugpo.2013.10.014.
Fraser, Suzanne. 2006. ‘The Chronotype of the Queue: Methadone Maintenance Treatment and the Production of Time, Space and Subjects’. International Journal of Drug Policy 17:192–202.
Keane, Helen. 2002. What’s Wrong with Addiction? New York: New York University Press.
Kleiman, Mark A. R. 1988. ‘Crackdowns: The Effects of Intensive Enforcement on Retail Heroin Dealing’. 17.
Kremez, Vitali, and Brian Carter. 2021. Crime Laundering Primer: Inside Ryuk Crime (Crypto) Ledger & Risky Asian Crypto Traders. Advanced Intel.
Munn, Nancy D. 2020. ‘The Cultural Anthropology of Time: A Critical Essay’. 32.
Sorokin, Pitirim A., and Robert K. Merton. 1937. ‘Social Time: A Methodological and Functional Analysis’. American Journal of Sociology 42(5):615–29. doi: 10.1086/217540.
Taylor, Avril. 1993. Women Drug Users: An Ethnography of a Female Injecting Community. Oxford: Clarendon Press.
Taylor, John R., and John E. Helzer. 1983. ‘The Natural History of Alcoholism’. Pp. 17–65 in The Biology of Alcoholism: Volume 6: The Pathogenesis of Alcoholism Psychosocial Factors, edited by B. Kissin and H. Begleiter. Boston, MA: Springer US.
Walmsley, Ian. 2016. ‘Coming Off Drugs: A Critical History of the Withdrawing Body’. Contemporary Drug Problems 43(4):381–96. doi: 10.1177/0091450916666641.
Zuboff, Shoshana. 2015. ‘Big Other: Surveillance Capitalism and the Prospects of an Information Civilization’. Journal of Information Technology 30(1):75–89. doi: 10.1057/jit.2015.5.