MANAGEMENT, CONTROL AND GOVERNANCE OF HAWALA NETWORKS IN THE GULF COOPERATION COUNCIL REGION

Prior research has focused on the transactional aspects of the Hawala Networks (i.e. Invisible Funds Transfer Networks). However, little is known about the structure of this networked "organisation" especially about its governance and management control systems. Clanbased Organisations and Actor Network Theories were utilised to provide the theoretical base on which a research farmework was built. In total, 20 indepth interviews were conducted with Hawaldars (Hawala service providers) using a semi-structured questionnaire. As the study focused on the Gulf Cooperation Council region, data were collected from Qatar, UAE and Kuwait. The data indicated that there were no formal governance and management control systems in place for managing Hawala Networks (HNets); there were no explicit formal processes and mechanisms for evaluating performance and there was no formal performance measurement and tracking system in place to provide the information about network functionality. Results also showed that trust acted as the most important control mechanism impacting the efficiency, effectiveness and the relatively lower transaction costs associated with HNets operations.


INTRODUCTION
Health Network for Evaluation and Training Systems (HNets) is a worldwide informal funds transfer phenomenon commonly used by expatriates wanting to send money home, businesses in import/export, for foreign exchange transactions and for travel. It is estimated that every year over USD500 billion flows through HNets (Thompson, 2011). HNets exist due to a number of reasons which include; anonymity, cultural friendliness (Chene, 2008), low transaction costs (Maimbo, 2003), less transit time (Schramm & Taube, 2002), enhanced level of trust (Schramm & Taube, 2002) and increased efficiency (Ballard, 2003).

Research Background
Very limited (if any) research has been conducted to specifically look at the governance mechanisms and management control systems in HNets. A review of the existing research in the area of HNets reveals that most of the research has focused on defining what HNets do and describing the general mechanics of how money is transferred through the system (e.g., Ballard, 2003;El-Qorchi, Maimbo, & Wilson, 2003;Maimbo, 2003;Passas, 1999;2006;Schramm & Taube, 2002;Wilson, 2002). The current research on Hawala provides basic information on how the network operates and how the linkages among network members function (Ballard, 2005a); but leaves a number of important questions unanswered such as how these informal networks are governed, coordinated or how controls are established and exercised. Unlike "formal" transnational economic activities that operate from relatively static identifiable entities (such as banks and other multinational corporations) and which are regulated by national and international governmental institutions, the Hawala economy is constituted by networks of spatially dispersed heterogeneous actors with no specific centre. While one might think that as an enterprise that exists in the "grey" area of legality, the default rate and default risk would be high, fraud would be rife, and exploitation a common occurence, there is a surprising consensus among scholars that the HNets have a history of being reliable, speedy and convenient (Ballard, 2005a;.

HNets in the Gulf Cooperation Council (GCC) Region
HNets are thriving in the Gulf Cooperation Council (GCC) region due to low transaction costs and simple transfer procedures (Pathak, 2003). Furthermore, HNets reach parts of the world (such as rural areas) where formal institutions such as banks do not exist. Appropriately, they are labeled "poor man's bank". These networks provide a useful service to low income workers from the Indian sub-continent in the GCC region for sending their money to remote areas. Furthermore, majority of the Indian sub-continent (to include India, Pakistan, Bangladesh and Nepal) based Hawala deals are being currently initiated and managed in the GCC region. The key reasons for this "situation" are large population of workers from the sub-continent, their inability to use formal monetary transfer system, and the proximity of GCC to the sub-continent region (Ballard, 2005b). Other supporting factor for thriving Hawala network in the GCC region are to do with local currency peg to the dollar and the unrestricted access to global financial system. HNets are flourishing in the GCC region (including Qatar, UAE, Bahrain, Oman and Saudi Arabia) and play a crucial role in billions of dollars of money transfers out of GCC. In fact, Dubai is labelled as a part of the "Hawala Triangle" where India and Pakistan represent other vertices (Pathak, 2003).
The GCC region is fast developing and expanding its financial sector. In particular, United Arab Emirates and Bahrain are striving hard to become financial centres of the Middle East. With these ambitions in progress, any monetary leakages or financial irregularities related to the utilisation of the informal channels of money tranfer can (and do) disrupt these plans. Hence, governments in the GCC region are seriously interested in curbing the use of Hawala. To start with, the profits related to the formal remittance system are being eroded as a result of HNets provding an alternative and a cheaper money transfer option (Passas, 1999). The financial divergance through HNets also results in taxation losses for the government as money is moving outside the formal financial channels and remains untaxed (El-Qorchi, 2002). Another problem associted with the money moving out of the formal financial channels is its undetecable usage. There have been cases (even though quite rare) where illegal businesses and criminal activity have been known to be funded through Hawala transfers (Passas, 1999). The minimal transaction record with no legal coverage has resulted in fraud (very few cases) due to the untraceable nature of the Hawala tranfers (Schramm & Taube, 2002;Farooqi, 2010). Additionally, the speed of Hawala transaction completion (in most cases) is faster than fromal channel transfers. The typical Hawala transaction can take anywhere from hours to a day or two at the most. HNets also provide deeper transfer penetration as comapred to formal remittance system. HNets have been known to reach remote and rural areas where conventional fianncial institutions do not exist. The formal transfers (mostly through banks) take longer as the funds need to be converted into foreign exchange depending on the termination country and the number of banks invloved in the transfer process. Hence, slow transfer through formal remitance channels also motivates the money sender towards HNets.
In the GCC region, despite of their negative and adverse effect on formal remittance system, HNets are a prominent exemplar of the informal systems used for transferring money from GCC member countries to foreign locations (mostly the the Indian sub-continent). Even roots of HNets can be traced to this region and that is why they are also known as the 'Eastern Union' (as opposed to Western Union) to signify their origins. Usually HNets, found in the GCC region, operate outside mainstream banking channels which results in minimal transaction related documentation and paperwork. This 'informality' is also associated with the ability of HNets to provide the lowest transactions costs, cultural friendliness, lower delivery time, enhanced trust and increased efficiency. These positive service aspects favor 'working class' population in the GCC region who can have easy and inexpensive access to these informal channels to transfer their money (Hanieh, 2010). HNets are credited with the ability to move billions of dollars annually involving thousands of discrete transactions. However, still little is known about how these networks are managed, controlled and governed.

OBJECTIVES OF THE STUDY
The key issue addressed by this research is how HNets, despite being geographically dispersed, are highly proximate in creating mutually beneficial relationships. More specifically, the study examined the following key interrelated questions: 3. What are the forces that bind the HNet actors to one another, including the role of trust?
4. What are the practicalities associated with managing the HNet multinational and crossnational relationships?

LITERATURE REVIEW
Conceptually the research is grounded in two streams of theory: (i) Ouchi (1980) framework for studying clan-based organizations; and (ii) actor network theory (ANT) (Latour, 1987) that focuses on the multiple associations, the elaborate translocal business linkages and sociocultural ties that make up socioeconomic networks.   Ouchi (1980, p. 140) defines an organisation as, "any stable pattern of transactions between individuals or aggregations of individuals". Ouchi (1980) used the transaction cost approach to look at intermediation forms and identified three basic mechanisms of mediation or control; markets, bureaucracies and clans. Ouchi suggested that the normative requirements for clan-based organizations in the 20th century were reciprocity, legitimate authority and common values and beliefs.
Ouchi posited that the norm of reciprocity underlies all exchange mechanisms and has been found to be universal among all societies across time (Gouldner, 1961). In its absence the cost of the exchange, involving contractual terms, auditing and performance evaluation tasks, would make the transaction prohibitively expensive and not worthwhile. Legitimate authority can take the rational/legal form in the case of a bureaucracy while it can take the traditional form in clan based organisations. Common values and beliefs in the clan form eliminate the possibility of opportunistic behaviour among clan members as cheating another member would be like cheating oneself. A value system that is mutually acceptable to the Hawala service provider and the service user is fundamental in sustaining the system. HNets are largely driven by the code of honor, reputation (of the Hawaldar), trust and personal relationships connecting HNet members and Hawala service users (White, 2003). According to Tilly (2005), a combination of honor, reputation and personal relationships results in 'noble value'. Within the HNets clan, noble value creates a belief of trust and individuals (both service providers and service users) engage in transactions with reduced fear of opportunism (Tilly, 2005).

Theoretical Perspective II: Actor Network Theory (ANT)
Actor Network Theory (ANT) has been used to look at a variety of accounting phenomenon involving complex relationships (Briers & Chua, 2001;Chua, 1995;Gendron & Barrett, 2004;Miller, 1991;Preston, Cooper, & Coombs, 1992;Robson, 1992) such as "interfirm alliances" as well as to study the social implications of relationships (Callon, Law, & Rip, 1986). Recently, ANT has been applied to the study of how economic markets are organised (Caliskan & Callon, 2009;Callon, 1991;1998). Callon (1991;1998) argued that economic markets are networks and to understand how market transactions are governed and stabilised; we need to examine the "processes of economizing" that takes place within the economic markets. The processes of economizing refers to "the processes that constitute the behaviours, organizations, institutions and, more generally, the objects in a particular society which are tentatively and often controversially qualified, by scholars as 'economic'" (Caliskan & Callon 2009, p. 370). Given that HNets are engaged in transnational economic activities, the proposed research draws on the ANT to investigate the "processes of economising" that leads to their effectiveness in transfering funds around the globe.

RESEARCH METHODOLOGY
Through the multiple informants at multiple levels in the HNets a variety of network members were interviewed to gain a better appreciation of how these networks were governed and what management control systems were employed in governing a multinational "organisation" that was indeed a true network or a clan organisation.
Data collection was done using interview based field research which allowed for the understanding of phenomenon in their natural setting (Patton, 1990;Yin, 1989). This methodology gave researchers an opportunity to follow the relevant actors to develop an understanding of the management activities and actions within the HNets (Callon, 1991;Latour, 1996). The study involved using various sources of information including Hawaladars at different levels, formal value transfer operators and HNet experts. Using multiple information sources enabled the maintenance of the "critical distance" necessary as defined by Law (1991). Gathering data from the multiple sources was consistent with the extant field studies in management research (for details see Ferreira and Merchant, 1992) and was also in consonance with Yin's (1989) commentary on improving the construct validity of field data.
Interpretive methodology involving in-depth interviews was utilised to collect the data. Adopting this semi-structured interviewing approach provided an opportunity for investigating emergent and new issues that were not selected prior to the study but proved to be relevant. The semi-structured and flexible questionnaire format, especially in the early stages of the field work, reduced the number of predetermined questions (which may not have been a relevant question to ask) and hence this 'improvised' approach helped with the identification of fitting and appropriate questions. The flexibility generated through the semi-structured interviewing questionnaire created a room for further improvements (Arthur & Nazroo, 2003). Since, semi-structured research designs have been criticized as leading to diffuseness and data overload (Miles & Huberman, 1994), the looseness of the investigation was bounded by the semi-structured questionnaire which facilitated a focused approach to the research. In the interview process, care was taken not to evaluate the respondents' professional competence but to understand how they actually manage the networked environment in which they operate (Chua & Mahama, 2007). Where needed, appropriate language was used to facilitate the ease of communication between the respondent and the researcher. Research team was fluent in Arabic, English, Hindi, Punjabi and Urdu. These were the languages spoken and understood by most of the Hawala actors within the GCC region.
To support data collection and maximise the number of contacts made, chain referral sampling was utilised. This sampling method was considered as the most appropriate for sample identification in networked individuals and firms (Sudman, 1976). Use of chain referral sampling tends to optimise the chances of reaching a wider part of the suitable and the appropriate sample through the contacts which have been already identified. As these contacts were familiar and knowledgeable about the HNet membership, their referrals resulted in an appropriate and a fitting respondent identification (Taylor, 1985). Another reason for adopting chain referral sampling was related to the semi-scattered nature of HNet members where Hawaladars were temporally and spatially displaced. Through chain referral sampling it was possible to make connections with HNet actors through other HNet actors.
As described in the literature (Browne, 2005), chain referral sampling starts with a contact list (see Tables 1 and 2) containing individuals who are not only willing to furnish the required information but willing and comfortable in pointing out and even arranging meetings with suitable HNet members. Hence, to optimise the chances of reaching a suitable and an appropriate sample set, existing and newly developed connections were used from the academic and Hawala community. In particular, individuals (especially Hawaladars who are or have been a part of the HNet or service users who have been sending money through HNet) having an experience of and familiarity with HNet were approached with the referral requests (Taylor, 1985).
While chain referral sampling is an effective and an efficient sampling strategy for research within organisations, it has a drawback of creating a narrowly defined sample which may not be a balanced representation of the population under investigation (Holstein & Gubrium, 2004). To overcome this limitation, a diverse location approach (i.e. sample to be dispersed over different geographic locations to include different parts of Qatar, United Arab Emirates and Kuwait) was taken to maximise the geographical spread of the data. At different locations various leads (through various types of contacts to include respected community members, prominent Hawala network members, distinguished Hawala service users and local Hawala researchers and academics) were approached to create a large pool of contacts representing different segments of the population. Other inadequacies (such as interviewer bias, misunderstanding of context, language initiated problems and intentional mis-directions) commonly related to qualitative semistructured in-depth interviewing were minimised by complementing interviews with informal discussions. Cross referencing of formal (interviews) and informal (casual discussions) data helped with content clarification and reduction in error.
A trip was made to the Hawala zone (a place from where Hawaladars operate) prior to data collection to acquaint and familiarise with Hawala settings and rituals in order to understand the fundamental social mechanics surrounding Hawala business. The purpose of this pre data collection trip was to soak in Hawala social environment in order to feel comfortable within the HNet atmosphere when interviewing started.
As HNets mostly operate within a restricted environment where access can be limited to a selective audience and Hawaladars tend to be secretive about their dealings, a recording device or note taking were not utilised during interviews. This caution was necessary to ensure that the respondents were comfortable and relaxed and there were minimal hesitations in sharing data and information. Even though mental note taking may have introduced certain errors into the research, it was preferred that the respondents were fully engaged, unintimidated and open with their views and expressions. Team interviewing (with two researchers) was adopted to reduce the level of disinformation likely to result from the mental note taking. Cross referencing of collected data was done to minimise the amount of mistakes which could have been the outcome of memory lapse or inadequacy in remembering the respondent's answers and comments.

Data Analysis
The gathered data were maintained in NVivo and manual folders. NVivo was utilised for keyword/key phrase analysis. NVivo is a software package designed to organise qualitative data in a retrievable form. This software provides a database for the evidences gathered in the field through interviews and observational notes and memos. It helps the researcher to work efficiently with unstructured data such as interviews, documents, audio, video and pictures so as to enable better analysis and presentation. Maintaining such databases allow for a chain of evidence to be established such that external readers can follow the field report to the evidence upon which the report is based.
Preliminary interviews were conducted in Qatar, UAE and Kuwait with the view of identifying emergent themes relevant to the research. For pre-testing the questionnaire, three respondents who had experience and exposure in the area were interviewed. These interviews led to the revision of the final questionnaire. The three pre-test interviews were not included in the final data set. A total of 20 in-depth interviews with Hawala actors were completed. The average time per interview was 105 minutes. The data were collected in two rounds conducted during summer 2013 and summer 2014.

Empirical Analysis
The HNets money transfer market is unregulated, yet there is evidence to suggest that they have been highly successful in money transfers at the lowest transaction cost with high level of efficiency (compared to the formal banking system). Mostly the Hawala transactions occur between people who could be described as total strangers, yet there is very limited evidence of fraud and conflict. More so, the HNets member initiating the transfer may have no direct personal or business contact with the other member who finally delivers the money to the receiver, yet the delivery is speedy and accurate. In fact, HNets have evolved into heterogeneous multinational "organisational networks", the operations of which are not yet well known and reported in the literature. Hence, the general research question was how are HNets governed, managed and controlled. Furthermore, the proposed research drew on Ouchi's framework for studying clan-based organizations and on Actor Network Theory's (ANT) notion of relational networks.
The analysis showed that there were no formal control mechanism underpinning the operations of HNets; there were no formal organisational structures and no legal contractual basis for conducting operations. Rather HNets were constituted as flat networks of heterogeneous participants who operated through relational (social) contracting and with influence dispersed throughout the network. The data analysis indicated that trust acted as the most important control mechanism underpinning the efficiency, effectiveness and the relatively lower transaction cost of HNets operations. By operating through flat network structure, HNets have not only been able to reduce the cost of bureaucracy evident in the formal banking system but have also been able to effectively hold the network together through reliance on trust. In addition, the flatter network structure added to the speed with which decisions and delegations were made and lead to effective information sharing. At somepoint in the relationship, trust lead to open communication where confidential and sensitive information was shared between network actors with minimum of checks.
The data indicated that the operational aspects of the network were not uniform as they were not required to be so; neither was there a monitoring system in place which required the use of a standardised process for conducting business in the network. As such the operational aspects of the business varied from network to network and at times from person to person. The lower costs and service efficiencies could be related to the absence of third part monitoring or requirements to follow a prescribed or a standardised procedure for transferring funds. The lack of these compliance costs reduced the transaction costs. In addition, the speed with which transactions took place increased in the absence of bureaucratic processes found in the formal transfer systems. As a relevant analogy, HNets operated by finding connections similar to the World Wide Web which can find indirect pathways through connections via intermediary nodes (in the case of Hawala, trust based relationships). Hence, HNets are constituted as flat networks of heterogeneous participants who operate through relational (social) contracting and with a positive influence dispersed throughout the network. The following section presents a discussion about how trust emerges and becomes instantiated as an effective control mechanism in HNets.

Material Arrangements of HNets: Building and Consolidating the Network
Actor-network theorists suggest that rather than taking the material composition and arrangement of a network as given and then proceeding to analyse how the network is governed and controlled, one should start the investigation of governance and control by examining the patterning of such networks. That is, governance and controls are not add-ons to the network but constitute part of the processes of building and enacting the network. In this sense, the material composition and arrangement of a network may inhibit particular governance and control elements that both enable the network to be constituted and become stabilised during its operations. The study found evidence of control (informal in nature) simultaneously emerging with the very composition and patterning of the networks in a process that can be described as "heterogeneous engineering" (Law, 1991). It was heterogeneous engineering in the sense that it assembled a network of human actors (Hawaladars) and non-human elements (such as traditions, values and beliefs) from diverse and disparate origins into a network (HNets) that operated like a single actor (Hawala). It was within this heterogeneous engineering process that trust became the most effective control mechanism in the HNets that were studied. How trust was established as a control mechanism is explained next.
The composition of HNets was based on the careful screening for the trustworthiness of the potential network members. An individual's trustworthiness was established at two levels: (i) personal ties with existing Hawaladars and (ii) position within one's community. To become a part of a HNets, individual needed a long standing personal tie with one or more Hawaladars in the HNets who were willing and able to vouch for the joining member trustworthiness through a thorough system of referencing. References were essential ingredients in convincing existing HNets members that the new entrant was reliable and dependable and that he/she could be trusted to deliver when called upon to do so. Referees played an important role in case of any discrepancy, serious complaints or defaults. In such cases the referees were called upon to honor the commitments of the new entrant for whom reference was provided. It was usually over a considerable period of time involving numerous transactions that the supporting role of referees diminished. In the very rare cases of serious fraud, the Hawala Network took the collective responsibility of covering the losses of the service users. Especially, the reference providers or the guarantors (of the offender) take the lead in finding an amicable solution for recovering the losses and minimizing the reputational damage (Schramm & Taube, 2002). This consolidatory behaviour exhibited by the Hawaladars is indicative of a desire to preserve collective reputation of the Hawala system (Humphrey & Schmitz, 1996). The offender, if caught, had to face the Hawala jury who then decide on the appropriate course of corrective action or punishment (Farooqi, 2010). A point worth highlighted is that according to the interviewed respondents, and based on the key findings of this study, there were no reported acts of financial misconduct impressing the high level of trust and confidence that exists within HNets. However, a couple of Hawaladars talked about rare cases where a wakil (an adjudicator) was nominated through mutual consensus to resolve matters of financial misconduct (Thompson, 2011).
All interviewees in this study emphasised the importance of the reference system. For this reference system connected people to distant others whom they have never met and possibly will never meet but were willing to engage with them as a part of an extended network. For example, a number of the interviewees noted: They are connected through a system of reference. A knows B and B knows C thus based on the common relationship to B, A and C would be able to transact business after an introduction by C.
(Respondent 1) The Hawaladars involved in the business are known to the local community as such it is easy to identify them. Once a new Hawaladar identifies a niche market that is not being serviced or has been able to identify a critical mass of customers that they can attract, they get connected to the known WH (Wholesale Hawaladar) in the region to facilitate their transactions. Personal introductions play a very important role in making these connections as it is only after getting a clean reference are the players in the Hawala Network willing to deal with the new person. Once one link is established and based on satisfactory dealings, the Hawaladar is able to leverage this connection to make other connections in the Network.

(Respondent 4)
In cases where trust references were not highly reliable and dependable, the fostering of trust occurred in a stepwise fashion where trust growth took time and transactions progressed from lower to a higher financial value. Hence, trust and reputation gaps were filled with risk-aversion approaches based on the incremental and careful induction of the (uncertified) new Hawaladar into the HNet. As an additional precautionary measure, Hawaladars (mostly new) facing trust deficit were also subjected to "upfront monetary deposits" and "capped financial transactions" over a period of time (involving numerous exchanges). As stated by a respondent: Someone initially starting out and not well known to the WH (Wholesale Hawaladar) may need to put a deposit before getting started in the business. This allows them (i.e. new Hawaladars) to conduct the business up to the value of the deposit.

(Respondent 13)
With successful transactions the "worth to transact" was increased. It was highly unlikely that the new and untested Hawaladar would suddenly jump from low to a high value transaction overnight. Moreover established members of the network would not conduct large transactions with the new entrant before his/her credibility, in terms of the efficient execution of transactions and the financial ability to conduct large sum transactions was well established. This mechanism of risk management was a common practice within HNets. Due to the well-connected and tightly knit nature of the HNets, incidences of fraud and dishonesty were instantaneously communicated throughout the network to warn other members. This connectedness of HNets resulted in a rapid transfer and spread of information which played a pivotal role in creating awareness about any significant event (especially unpleasant) within the network.
Mutual dependency (i.e. A depends on B and B depends on C and A depends on C) would be weakened within the HNets if there were no safety nets and support structures for troubled network members. For instance, where cases of financial hardship were verified (resulting in Hawaladars not being able to dispense or service their financial obligations or commitments), other network members come to their rescue (even though the bailout was agreed upon a certain set of conditions mostly set by the senior Hawala members). These acts of mutual support (sacrifices) propagated trust amongst HNets members and created an expectation of helping each other out should the bad times fall on a HNet member. Hence reciprocity and dependability kept and reinforced the fabric of HNets.
In cases where the new entrants had no personal ties with the existing Hawaladars, references could be secured from trusted and distinguished community members. As stated by two interviewees: Initial introduction can be through a trusted and mutually known third party. This third party does not have to be in the Hawala business but could be somebody that is equally respected and trusted by both parties. The established Hawaladar basically uses this reference to start a relationship with the new Hawaladar. The introducing party knows that if there are any issues with the new Hawaladar then they (the third party) will be held responsible for the actions of new Hawaladar. Once the relationship develops (between the new and the established Hawaladar) then the established Hawaladar is able to provide the new Hawaladar with further access to his network.

(Respondent 6)
The element of trust is initially based on the personal connections and the third part references. As the Hawaladars business gets established, their reputation also gets known and they are then able to interact with many more members of the network.

(Respondent 8)
The importance attached to the personal ties could be explained with reference to Granovetter (1973) thesis on the strength of ties. Granovetter (1973Granovetter ( , p. 1362 notes that "the stronger the ties connecting two individuals, the more similar they are, in various ways". This resonated with the way the reference system generates information about the trustworthiness of potential Hawaladars. Generally, when clean references are provided, they reflect the trust that the referee has in the other person with whom they have personal ties. Given that the referee is trusted by the existing member(s) of the network, the network members may consider the potential entrant trustworthy on the basis of the trust they have with the referee and thus consider him for membership of the network. Granovetter (1973Granovetter ( , p. 1362) reflects this when he writes: "if strong ties connect A to B and A to C, both C and B, being similar to A, are probably similar to one another, increasing the likelihood of a friendship once they have met". In a sense, the strength of personal ties as reflected in the reference system that underlie the formation and composition of HNets is a significant source of trust and constitute a careful "leap of faith" for engaging with potential new entrants.
Apart from taking a careful "leap of faith", in most HNets certain checks and controls operated to reduce risk (in particular associated with inducting a new member). For instance, WH kept a "transaction record" of all members to monitor their performance (mostly judged by volume and frequency of monetary amounts transacted). In the case of new members the monitoring was more strict and rigid transaction limits were set. These checks were relaxed once the trust equity went over a certain threshold. In most cases (both for the new and established members) the transaction limits were not removed. What usually happened was a change in the regularity and thoroughness of the scrutiny. Hawaladars (in particular WH) were constantly monitoring the level of risk exposure that may result from the actions of other network members. Hence, HNets incorporated risk aversion measures as a safety cushion. As stated by Respondent 15: WH (mostly) knows the worth of Hawaladars in terms of their trust status, strength of personal ties and business performance. Where appropriate (especially with new and untested Hawaladars), WH keeps a tab to ensure that the business transactions are kept within safe limits. Induction of new Hawaladars can be a complex process. There may be some new connections (Hawaladars) that do not require any guarantees while there may be some who may require either financial or personal or both types of guarantees.
(Respondent 15) Whereas the reference system was used in generating the trust required to become a member of a HNets, it was also important that the potential new entrant was able to demonstrate strong ties with and within the community it sought to operate in. Strong community relations were important as they formed the basis for being able to attract customers (clientele) to the network. Strong community relations were an indication of the trustworthiness of the potential new entrant and also served as an indication of the customer base that the potential member was likely to build. Such relational ties acted not only as guarantees for trustworthiness but also served as security (or soft collateral) for participation in the network. This was partly because strong personal and family ties made the Hawala member more visible and subject to community sanctions when he/she violated the norms of expected behaviour. Community ties thus served as an important means of reducing potential knowledge asymmetry that existed. Speaking about the importance of community ties in the composition of HNets, some of the respondents noted: One of the most important qualifications for someone to start the Hawala business is to be a sociable person, to have extensive contacts in the community, should have had a track record of being involved with the community which will serve as his customer base. An elementary qualification is that he should be a trust worthy individual in the eyes of the community/customers.

(Respondent 13)
One has to have a link to the community that the Hawaladar wants to service, i.e. he has to have the customer base and this is possible if the person is a sociable person and has connections within the local community. The community could be the Pakistani or Indian or Bangladeshi community or a subset thereof. The person should also have some connections back home in order to be able to terminate the Hawala transactions in the country where the remittance will be sent.

(Respondent 8)
In emphasising the importance of community relations to both the security of customers and other HNets participants, the respondents explained that effective community engagement was constitutive of the number of years one lived in a community, the level of trust and reputation with which one was associated in the community, and clean community references. On this, Respondent 4 narrated: One of the important qualifications for getting involved in the Hawala business is the level of engagement of the person with the local community (by community we mean their ethnic community). Number of years the aspirant has lived in that community and the level of trust that that person enjoys. These qualifications are important not only from the point of view of the customers who would be transacting their business through the Hawaladar but also from the point of view of the other Hawaladars that this new Hawaladar will be doing business with. It is important that people both within and external to the community will be able to get good and clean references about this person from the people of the local community.

(Respondent 4)
Even though not explicitly outlined, indirect references were made to the existence of noble value which further facilitated the ease of doing Hawala business. In the eyes of the Hawala service users and service providers a powerful value (i.e. noble value) resulted from the mixing of various other social values (to include clan and religious congruency, honour within the clan, social and moral norms and reputation within the clan). Noble value was fundamental in instilling trust within the Hawala system (Tilly, 2005). In addition, Nobel value was not only related to risk reduction but was also viewed as a master value which reinforced service user's belief in the credibility of the Hawala system (Humphrey & Schmitz, 1996;Grief, 1994).
Within HNets, the relationship orientation was mostly upstream, from supplier (Hawaladar) to customer (a person who wanted to send the money). However, the downstream relationships (customer to supplier) could not be ignored by the Hawaladar as they played a vital role in spreading his trust status and reputation within the community. Customers (i.e. Hawala service users) become carriers of trust and Word-of-Mouth (WoM) becomes instrumental in building 'trust equity' of a Hawaladar. Hawala communities (containing Hawaladars and their clientele) were mostly tightly knit and extensions of personal ties. Such an environment resulted in a rapid spread of WoM which was considered as a useful 'qualifier' of a Hawaladar. Respondent 7 summed this up by saying: Once the community (predominantly the group of customers receiving Hawala service) is happy with their transactions not only do they become regular customers but also (and more importantly) they relate their positive experiences to the community at large including the prospective new clients". Some Hawala Networks are run from a 'cover business' (some examples of such cover businesses can be barber shops, coffee shops, travel agents and grocery shops). The cover business acts as a point where community members can access the Hawala Network. Cover businesses also act as social platforms for developing community relations and getting to know each other.

(Respondent 7)
The interview data also indicated that the controls were also embedded in the material arrangement (i.e., the pattern of connection established among and between network member) of HNets. Basically, the data showed that HNets were composed of clans nested within a wider network. Typically, HNets comprised Retail Hawaladars (RH) and Wholesale Hawaladars (WH). Within a geographical location or a community, a clan of RHs was built around a major WH. The WHs were large and well-resourced Hawaladars who were capable of dealing with large volumes of transactions involving significant amounts of money. They were also highly reputable within their communities and beyond and had global business ties. Given the significance of the WHs, minor players who constituted themselves into RHs generally congregated around the WHs who acted as a spokesperson for the RHs or served as a conduit through which the RHs may gain global reach.
The clans that form around a WH were a socially homogenous community with similar family relations, language and ethnicity. They were also generally isolated and exclusive from the mainstream society. As a result of this, the clans were able to rely on what could be described as mechanical solidarity to produce and reinforce trust as an informal control mechanism (Newton, 1997). Mechanical solidarity is solidarity that is based on kinship ties and is usually associated with the cohesion and integration that exists in small undifferentiated social collective. This type of solidarity aligns individual interest with the collective interest of their groups without the mediation of elaborate or extensive formal controls associated with contemporary organisational forms. This was particularly evident in the data as the respondents indicated that in their everyday interaction with other members of the clan, they believed that their ability to achieve their individual self-interest was based on the pursuit of collective interests. In essence, self-interest was subsumed under collective interests where individuals believed that collective interest was an obligatory passage point to self-interest. For example, when asked about what held the clan together, some interviewees had this to say: Basically trust and the fact that each member's economic survival and prosperity depends on ensuring transaction compliance, delivery and successful termination of the Hawala transaction.

(Respondent 2)
Financial incentives are solely based on mutual co-existence and trust among the members.

(Respondent 3)
Common purpose and the fact that they can only operate and make money only if they work with one another.

(Respondent 6)
The idea that collective interest is an obligatory passage point to selfinterest generates what Newton (1997) describes as "thick trust" that transforms Hawaladars from self-interest seeking into cooperative individuals with significant implications for HNets. First, it fosters loyalty among the members and by so doing eliminates the transaction cost associated with elaborate contracting and monitoring. These reductions in transaction cost means that HNets are able to conduct money transfers at a much lower cost compared to the formal banking system. Second, the "thick trust" generated within the clan eliminates the need for laborious processes of verification and approvals associated with the bureaucracy evident in the formal banking system as each member trusts that others will do the right thing. It is this that underlies the high speed with which HNets are able to transfer money globally compared to their counterparts in the formal banking system. Third, reinforcing existing trust involves implicit social sanctions for violation and this has led to fewer disputes among Hawaladars and more compliance with terms of transactions.
Within HNets thick trust is seen within a long-term orientation where short-term losses are mostly accepted in sight of the long-term gains. Numerous instances were reported by Hawaladars where due to unavoidable events (such as robberies) and uncertain events (such as sudden and unexpected exchange rate changes), the Hawaladar had no option but to face (and accept) the monetary loss. Still the general thinking was that the losses incurred were worth the increase in trust status and reputation enhancement resulting from acting in an honest manner and keeping the promises and the given word. As summed up by respondent 4: I know Hawaladars (including myself) who took loans or liquidated their assets to honor their financial commitments. Some Hawaladars suffered financial set-back due to the exchange rate going wrong. Still they (Hawaladars) kept their word and transacted at the agreed conditions even if this meant losing money. This is done to maintain the confidence and the trust of the customer.

(Respondent 4)
Table 3 (matrix of data and theoretical anchors) below summarises the empirical data in terms of the theoretical anchors used in the study.

RESULTS
The data indicated that there was no formal governance and management control system in place for managing HNets. There were no explicit formal processes and mechanisms for evaluating performance and there were no formal performance measurement and tracking systems in place to provide information about how the network was functioning. Hence, there was seldom a uniform or a standardised control system that existed across different HNets. There were controls and management practices in place which varied from network to network. This variation (resulting in flexibility and adaptability) gave way to cost savings and transactional efficiency.
Results from this study showed that trust acted as the most important control mechanism highlighting the efficiency, effectiveness and the relatively lower transaction costs associated with HNets operations. Furthermore, trust was the relational glue which held the network together and acted as an effective cohesive force. By using a flat network structure, the HNets were able to reduce the cost of the traditional bureaucracy present in the formal banking system. One of the key reasons for the reduced cost was the lack of any third party monitoring cost. In the regular remittance system there are many layers of checks and balances in order to ensure that the transaction is well recorded. All of these layers of management with a formal financial institution (such as a central bank) incur costs which have to be borne by the service user. A Hawala transaction on the other hand does not rely on any third party monitoring as Hawaladars only deal with one another. This reduces the cost facilitating a lower rate for carrying out the transaction. This flat network structure also added to the efficiency and speed with which decisions were made and transactions completed. In addition, trust was the platform on which open communication for confidential and sensitive information sharing was built.

CONCLUSION
The results of this study should help in understanding the management control and governance mechanisms of the HNets. This understanding could be significant for management accounting and organisational researchers as the clan form of organisation has been considered more of a metaphor than a practice. However, as organisations become more intertwined and independent at the same time; understanding how this clan-based structure is governed, managed and controlled could provide a unique perspective on alliance management. The results should be useful for various practice oriented stakeholders including, regulators, law enforcement agencies, formal value transfer systems, financial institutions as well as the HNets members.
Since 2001 Qatari government through their regulatory and law enforcement agencies have been actively trying to curb the use of HNets with slow and limited success (The Peninsula, 2008). The continued growth of these informal remittance systems in the GCC region (including Qatar) indicated the lack of impact the enhanced regulations have had on Hawala transaction volume.
Existing research provides only limited information concerning the governance, control structures and mechanisms of HNets. Hence, approaches to regulation up till now can be best described as "shots in the dark" and unlikely to be successful. Understanding the current governance and control systems of these networks may well aid in understanding the nature and endurance of HNets. This research might also be useful in identifying alternate or additional mechanisms to curb the use of such networks should this be deemed a relevant public policy goal.
Developed research model could be tested in different environments (i.e. in a different set of countries or a region). For instance, the study could be rolled out to Western European and North American regions to assess and identify differences and similarities that exist between different cultural, social, ethnic and regulatory settings. These results should allow for a better understanding of this global network and examination of regulatory strategies, monitoring structures and mechanisms that may be developed and tested. It would be imprudent to declare HNets illegal. Instead, mechanisms and policies need to be developed that would incorporate the HNets self-monitoring features and ensure the system's safe and judicious use. Also, there is a growing need to estimate the global size of HNets transactions, the number of transactions, average size of each transfer and the aggregate flow of currency. Furthermore, the changing world economic environment has impacted the HNets as well. These need to be understood to make future monitoring policies effective.