Global norm ‐ making processes in contemporary multinationals

This paper contextualises and analyses global norm ‐ making concerning the nature of work within multinational companies. We develop a framework for analysing the relations between formal and informal elements of global norm ‐ making, stressing dynamic interdependencies between formality and informality on four dimensions: the codification of norms; enforcement mechanisms; the use of ‘platforms’ and aggregation with wider contexts. We investigate four cases of transnational norm ‐ making spaces across two multinationals, and analyse how these interdependencies work, both for norms and practices that emanate from the upper levels of corporate hierarchies, and those developed at lower levels among global operational teams.

substantial advantages in operationalizing their ideas, they do not have a monopoly on organizational norm creation, as 'formal authority is one power resource among many; others include control of key knowledge, strategic location within the chain of production, possession of critical technical, personal or managerial skills' (Ferner, 2000, p. 525).
This paper aims to develop analysis of how formality and informality interact in the creation, diffusion and implementation of organizational norms that are intended to have international application. By norms, we refer to 'shared ideas and standards of appropriate behaviour' (Finnemore & Sikkink, 1998, p. 191), which have some degree of governing force concerning the nature of work. To advance our understanding of the puzzle of how formality and informality interact, we identify and analyse key conceptual elements of norm-making processes. This allows us to assess a range of organizational contexts within which global norm-making takes place, accommodating both HQsubsidiary transfer and norms that travel largely outside this structure-notably through integrated business units and work groups. Our research question is: how do formality and informality combine during the creation, diffusion and implementation of global norms to create a range of types of norm-making processes affecting work in MNCs?
We deploy our framework by exploring four cases of global norm-making drawn from two large European business services firms, based on in-depth, qualitative research. This allows us to contextualise how norm-making processes may be affected by corporate contexts-in particular the difference between federal and more hierarchically-driven patterns of control. It also allows us to show how norm-making occurs in different 'transnational spaces' within MNCs, and to analyse the involvement of a wide range of types of actors in processes of norm creation, dissemination and implementation.

| GLOBAL NORM-MAKING IN CONTEXT: AN ANALYTICAL FRAMEWORK
The processes by which ideas about, or potential standards of, contextually appropriate behaviour become norms are strongly shaped by organizational and societal contexts. Turning an idea into a norm involves the exercise of legitimate authority; the organizational studies literature emphasizes that interests are both advanced and challenged by new norms, and compliance with them cannot simply be assumed. The ways in which those driving new norms seek to ensure that they are implemented is therefore likely to vary according to the context of organizational power and authority structures (Ferner & Edwards, 1995). Corporate-level managers in loosely coupled 'federal' MNCs are unlikely to be able to rely on managerial power to diffuse norms in the same way as those in more hierarchically-structured firms, for example.
There is also variation among MNCs in the extent to which a geographical segmentation of work is occurring (Dicken, 2015); where this is strong, work groups are increasingly spread across multiple locations worldwide (Hinds et al., 2014). If norms governing a range of issues-a fair distribution of work, how performance is managed, etc.-can emerge from within such groups, then we must allow for global norm-making to be driven by relatively low power actors well outside the scope of international management cadres, such as the leaders of international operational teams.
Norms also depend on societal contexts; new norms always have to mesh with those in the wider societies in which the operations of MNCs are based, something that is, a starting point for much of the literature concerning 'institutional duality'. Norms must either cohere with an existing web of norms in the organisational and societal contexts in question, or the actors instigating them must be able to manipulate this context to allow norms to be implemented (Muzio & Faulconbridge, 2013).

| Four dimensions of norm-making
The organizational and societal contexts within which norm-making occurs in MNCs are inherently multiple and complex. It would often be extremely difficult for norm-makers to fully anticipate the variety of formal and informal ALMOND ET AL.
-3 social processes that may occur as norms are created, diffused and implemented, as actors work out what a norm is, what its implications are, and how it relates to other organizational and/or societal norms.
To analyse these processes, we focus on four necessary dimensions of norm-making processes, each of which may, in principle, be more or less formalized in nature. In developing these, we draw on conceptualizations of international norm-making from within the international relations literature (Finnemore & Sikkink, 1998;Therborn, 2002), with adaptations to contexts of norms governing work within contemporary MNCs.
Our first dimension is codification. Ideas and standards about appropriate behaviour are sometimes expressed in the form of codified rules prescribing specific behaviours. However, norms also exist in the less formal nature of statements, guidelines or enunciated ideas which do not have explicit regulatory ambition, but shape behaviour by providing frames of reference for action (Therborn, 2002, p. 864).
In organizational contexts, international safety-critical norms, for example, in the aviation industry, are often identified as instances where norms are highly codified (Andrén et al., 2010). Rules and standard protocols are deliberately created to provide formality in how and what information may be exchanged, and adherence to these rules creates certainty and familiarity among actors who may span regulatory jurisdictions (Kontogiannis & Malakis, 2012). But the international management literature has long reflected the fact that international codification is difficult. Edstrom and Galbraith (1977) argued that the formal codifications of MNCs' bureaucratic control systems often needed to be complemented by social control-a requirement that underpinned the use of expatriate managers. Codification is even more difficult where norms are developed by less powerful actors, especially where these collide with senior managers' interests. For example, Edwards et al. (2005) show the difficulties faced by UK managers in this respect when attempting to diffuse norms internationally within US MNCs.
While MNCs may attempt to use highly codified norms to manage international workforces-for example, by precisely specifying what needs to be done to achieve a performance target-this is not always the case. Some norms may be intentionally vague in their codification, in order to maintain flexibility of interpretation. For example, corporate values statements are clearly intended to normalize particular types of behaviours, but often provide a broad framework within which organizational members seek to illustrate how their behaviours cohere with norms, rather than providing precise prescriptions.
Second, given the contested nature of both global norm-making (Therborn, 2002) and of the employment relationship (Edwards & Bélanger, 2009) a norm typically requires some form of enforcement mechanism. This may be formal-employees will be dismissed for specific unacceptable behaviours-or informal-a workgroup norm is followed to avoid social disapproval from colleagues. Enforcement becomes easier if those subject to a norm become convinced of its value. For Kostova and Roth (2002), processes of ''internalization'' are important in explaining the persistence and stability of attempts at international transfer. However, as they also reflect, it is often difficult to persuade employees to buy into norms developed elsewhere. Failure to persuade subjects to internalize a norm may be a problem if it affects the anticipated outcomes (Björkman et al., 2011). Where norm-makers cannot offer rewards or threaten costs, they can only enforce their norms through persuasion. In other circumstances, however, a failure to fully internalize may be less problematic; subjects may see a norm as inevitable, and comply instrumentally without fully accepting that it is a good idea (Therborn, 2002).
Third, actors promoting a norm 'need some kind of organizational platform from and through which they promote their norms' (Finnemore & Sikkink, 1998, p. 199). This refers to the organisational structures through which a norm is promoted, communicated, and negotiated upon.
The nature of platforms will vary across organizational, societal and actor-specific contexts. Within MNCs, the established formal platforms for norm promotion are organisational structures: established national/regional hierarchies, or international HR functions (Farndale et al., 2010). Formal platforms offer benefits to norm-makers by reducing uncertainty and providing some degree of quasi-constitutional clarity. However, they may lack flexibility, and may not be the fastest way of dealing with emergent issues (Ferner, 2000). Therefore, even senior hierarchical actors sometimes seek to apply norms through less structured platforms: examples include transnational working parties or the appointment of local 'champions' outside standard hierarchical structures Thompson & Driver, 2005). These may be designed as temporary apparatuses for norm-making, and may at least partially bypass intermediate hierarchical levels. For actors without the advantages of hierarchy, platforms are likely to be still less structured. Work groups may well create their own norms, but must rely to a much greater extent on informal persuasion or framing.
Finally, norms must relate to their organisational and societal contexts: a norm's capacity to reduce uncertainty is typically contingent on the ability to situate it within the constellation of norms present in a particular social space, in a process of aggregation which creates 'institutional norm complexes' (Therborn, 2002, p. 871). In MNCs, aggregation includes how a focal norm relates to existing norms and employment practices, for example, how attempts at pursuing norms of workforce diversity interact with embedded norms concerning performance management . Organizational norms generally also need to have some coherence with, or implicit relationship to national, sectoral and occupational institutional logics (Kostova et al., 2016(Kostova et al., , p. 2616. Norm-makers can attempt to work out the relations between different norms formally, but informal processes by which actors on the ground interpret what happens when norms designed for different purposes intersect or collide are probably inevitable in most cases (Ferner, 2000).
Thus, in Table 1, a fully formal norm would be highly codified, with an unambiguous enforcement mechanism, a clear platform for its development and potential modification, and its coherence with other norms would be clearly and explicitly expressed. A fully informal norm would lack clarity in any of these dimensions. 'Full' formality is unlikely in most organizational contexts; even in the case of the safety-critical norms indicated above, deviations in practice have been noted because of the need for individuals to make trade-offs within the work context in relation to efficiency, resource availability, or the complexity of tasks (Kontogiannis & Malakis, 2012). Therefore, we do not propose a binary opposition between formal and informal norms. Instead, we propose that our four dimensions enable us to detect distinct elements of (in) formality throughout the social processes involved in the creation, diffusion and implementation of norms.
This enables us to use our empirical cases to build a granular understanding of a range of types of global norm-making processes.

| METHOD
We utilized a comparative case study approach (Eisenhardt, 1989) to examine how formality and informality combined across our four dimensions of norm-making. We conducted case analyses of four norm-making processes, two from Company A, a publicly quoted MNC with a largely hierarchical structure, and two T A B L E 1 Formality-informality in global norm-making -5 from Company B, a federal MNC which had a partnership ownership structure and was a relatively loose collection of national organizations. The two organizations were selected from a wider group of MNCs as cases in which the depth and breadth of access allowed a thorough examination of the formal and informal aspects of global norm-making and the interaction between the two. They contrasted markedly different forms of international management, which we envisaged affecting the process and substance of norm-making.
Adopting Eisenhardt and Graebner's (2007) theoretical sampling, we selected four cases of norm-making within specific transnational spaces, which were driven by different types of actor. These four cases comprised moves to change the norms that informed the operation of a pre-existing, largely globally consistent set of performance management practices in the hierarchical organisation; moves by team leaders to establish norms within global virtual teams in the same firm; an attempt within the federal MNC to formalize the norms that underpinned a company-wide framework or set of practices for competence development; and attempts within a specific business unit in the same firm to build a global norm promoting and standardizing 'proactivity' in interactions with clients. We used this diversity of transnational spaces to explore the many ways in which there is an interaction between formality and informality in global normmaking.
We conducted in-depth interviews in both companies, using semi-structured interviews that generated 'both retrospective and real-time accounts by those people experiencing the phenomenon of theoretical interest' (Gioia et al., 2012, p. 19). While we prepared a set of questions to guide the interview, we used the flexibility of this approach to pursue instances of norm-making which informants led us towards. Each interview was conducted by two researchers in English.
Interviewees were chosen using a maximal variation purposive method (Creswell, 2013, p. 204) to achieve diversity of respondents (see Table A1). In Company A, we interviewed 34 staff ranging from directors at corporate HQ through to customer-facing staff who provided data services to clients, with fieldwork taking us across 10 countries. In Company B, we interviewed 32 staff ranging from International Partners to junior trainees in functions such as consultancy and tax advice, covering 17 countries, as we 'followed the stories' of the norm-making processes researched. Interviewing such diverse respondents allowed us to triangulate accounts of particular norm-making spaces by interviewing multiple individuals with roles in the creation, dissemination and/or implementation of norms. In Company A, this included a number of leaders of small virtual global teams.
Interviews were transcribed and coded using NVivo software. We took an abductive approach to data analysis, involving constant movement between theory and empirical material (Van Maanen et al., 2007). We iteratively combined predefined categories based on the project's underpinning conceptual framework with categories based on themes emerging from the data. We developed a codebook consisting of five principal codes and 170 sub-codes together with labels and definitions of each in a way that was 'specific yet encompassing of the constructs we were trying to capture' (DeCuir-Gunby et al., 2011, p. 143) and which established and maintained a common interpretative framework among three coders. Inter-rater reliability was ensured through the transcripts being coded and then checked by a second coder, with degrees of similarity checked and any diverging interpretations leading to a revisiting of the codes and the establishment of 'decision rules' (Gioia et al., 2012, p. 22).
The content analysis allowed us to select cases of norm-making by discerning stories and exemplars that offered theoretical insights. By adopting Eisenhardt and Graebner's (2007) 'theory-building from cases', we developed constructs within the cases and attempted to recognise cross-case patterns while developing emergent theory. This process involved iteration between the literature, data and emergent theory to yield 'logical explanations', 'the whys' for the underlying logics of the emergent relationships among constructs (Gehman et al., 2018, p. 288). In this process, the content analysis helped us to organise case evidence around focal constructs and discover the dynamics of the relations between formality and informality across our four norm-making spaces.

| Company A
Company A is a large provider of telecommunications and IT services. A high proportion of the workforce is based in the country of origin, but it has expanded globally within two product-based divisions over the last 2 decades in order to provide global clients with IT services. Virtual working across borders has increased sharply through the formation of hundreds of work groups that are geographically dispersed across the world.

| Performance management
The first transnational space we examined related to changes in Company A's performance management system. Overall, this system was in principle relatively formal: it was highly codified with formal enforcement mechanisms and diffused through formal hierarchy. Yet there was substantial informal negotiation both of how the system applied at individual levels. Difficulties in aggregating between the performance management system and a range of norms and practices, such as those relating to employment security, coaching and development, led to mistrust and uncertainty, which fed into informal interpretations of the system.
Because of the industrial relations context in the country of origin, the firm has long had a formal policy of avoiding compulsory redundancies. An unintended consequence of this was that the performance management system became the main means of managing people out of the organisation. Individuals with poor performance ratings would be obliged to retrain, apply for new positions within the firm, or take voluntary severance. This, alongside the use of a forced distribution, led to the system falling into some disrepute. For one manager involved in implementing the system: the performance management system is basically a systematic way of punishing your people…every quarter there was an expectation from the company that 20% of your people would be underperforming and the second expectation was that if they underperformed for three or more quarters, you would sack them because they were clearly useless. Of course, we had no easy tools to sack them, all we had was a very painful process (Interview 8).
A forced normalized distribution in the system was 'expected' (Interviews 6, 11), but neither codified nor explicitly discussed. Despite some partial denials from senior managers involved in disseminating the system (e.g., 'there is the myth of the bell-curve', Interview 4), managers of small teams repeatedly pointed to having to put precise numbers of appraisees into the bottom-performing category (e.g., 'we have got to make it fit the bell-curve, so you have got to choose at least one person to go in it' [Interview 13]). Different managers reacted to such pressures in different ways, with some 'shielding people from the insanity that seems to come down from the top', while others 'just like to cascade the pain and excitement downwards' (Interview 7).
Therefore, the implementation of the performance management system depended on informal negotiations between individuals and appraisers. Lack of official clarity on how the practice worked did not prevent it from being taken very seriously by appraisers and appraisees; indeed, it was taken so seriously that those involved in implementation saw it as counter-productive, with employees focusing on managing their performance reviews at the expense of productive activity: if you were perceived as underperforming you would spend days prepping for your quarterly review process, whereas, you know, that time should be spent serving customers or getting on with a project (Interview 14).
ALMOND ET AL. -7 Those implementing the policy would behave in ways that were not officially sanctioned, for example, by informally rotating employees in the bottom category in order to ensure that no single employee spent repeated quarters in it (e.g., 'I was told it was my turn to take one for the team', Interview 7).
In response to a question regarding the possibility of modifying the practice, one manager brought up other formal norms which aggregated with performance management: the problem is we've linked everything together. We've linked the performance management… (with) coaching, so the need for future performance is tied into the same system, as is pay and remuneration' (Interview 8).
The performance management system became overloaded, in a context of formal and informal normative requirements of varying degrees of codification. These included, notably: a perceived need for senior management to send a message to shareholders that performance was being taken seriously; a long-standing collective agreement in the country of origin setting a norm of no compulsory redundancies; and norms of fairness and equity within work groups concerning how their performance was assessed and what happened with the decision in the annual review. This left those subject to the system in highly ambiguous situations, requiring continual informal interpretation/negotiation. What on the surface was a highly formal practice became substantially informal in process and in substance.
Recently, following a merger with a smaller multinational, the company made the unusual decision of adopting the performance grading structure of the acquired company, replacing a five-tier system with a simpler three-tier system. The decision to replace the relatively aggressive, if informal, forced distribution of the previous system, was in some senses a recognition of the problems that had arisen. This decision, according to interview respondents, had 'the personal endorsement' of the CEO and an established hierarchical platform was employed for introducing the new system. This formal platform worked effectively to 'brief downwards' the new system in that employees felt they had 'got to just get on with it' (Interview 7). Although the new system was seen more positively than its predecessor, trust remained limited. For one senior manager involved in disseminating the system, 'I have just got this awful feeling that our three-point scale might start to be corrupted with people saying, 'Well, if you are a two maybe you are a rising two or maybe you are a sinking two' (Interview 4). Strong path dependencies were at work; while the new system did not use a forced distribution, given denials that this was the case previously it seemed difficult to persuade employees of this.
The new system was codified around a set of the firm's 'values'-a series of single words emphasizing creativity and excellence, used as framework type norms which were widely diffused and meant to be 'dovetailed in with performance management system' (Interview 26). However, these had uncertain meanings in practice. Employees pointed to having to 'squeeze' (Interview 1) their achievements into each of the values; for one senior employee, this meant 'I think probably people's minds are made up as to what grade you've got anyway and they just pick and choose the evidence that they want to support it' (Interview 6).

| Work organization within client-facing teams
The second transnational norm-making space we investigated in Company A consisted of a developing work environment within international client-facing teams. The firm had developed a division providing globally stan- and enforcement relatively informal. The creation, dissemination and implementation of norms was largely reliant on the informal leadership of the project manager. The potentially competing demands of corporate, team and client norms aggregated very informally.
We investigated a specific project serving a large European-owned multinational client with a key facility in the United States. The team comprised 10 people, all of whom interacted with each other across borders daily. A USbased professional managed a team of one other US-based worker, three based at the South American, and five based at the Indian, service centres. The manager visited the client firm's regional headquarters weekly, but otherwise worked from home. She had never visited South America or India, yet organized work across the team in these locations. While her line manager was US-based, on the project she reported to the contract director, who was located near the client firm's global HQ in Europe. She was line manager of the other American employee; the international workers had local line managers, but she, as the international project manager, provided input into their performance reviews.
The pattern of contracting was idiosyncratic to each client; each contract carried its own norms. For example, in contract negotiations, representatives for the client company: wanted us to be very careful about how we go about contacting their client…they wanted us to contact them once in terms of what work we were going to do…the scope of work we are going to do, and make sure that was done very carefully (Interview 29).
While Indian IT operatives essentially instructed local field engineers across the Americas, they were not permitted to contact the local client. Interviewees reported continual (re)negotiation, involving fairly junior organizational members, of complex contracts, affecting basic work organization, such as who was permitted to move items within the client's buildings. In this process, global company-wide norms, such as those on product and task standardisation, and company-wide practices, such as performance management, were continually supplemented or even replaced within the project team.
Coordinating this work required management across borders on a continual basis. Yet, there were no real codified procedures for this 'micro-global' management, and surprisingly little instruction as to how to operate in this environment: they don't really believe in giving us anything that helps us or guides us other than just I think they try to help us to try to say we all work in a global environment and keep that at the forefront of when you are discussing and working with folks in different cultures (Interview 29).
There is no formal training saying bear in mind that you have differences between this culture and this culture, no, but there is a formal training…in the learning platform that Company A has where you will learn about the ethics, the values (the firm) has, how the organization works, but less specific about the culture itself (Interview 23).
In other words, instruction as to the general norms and values of the organization, provided largely through an IT platform, was the limit of training to the more specific challenges of managing, in real time, routinely global operations. This led to substantial ambiguity; workgroup norms lacked both codification and clarity of enforcement mechanisms within the project team. As one regional disseminator and implementer of these norms put it: Company A puts up some tools, some emails…if you read it or not, well it's up to you (Interview 23).
A whole series of norms-task-related norms, those concerned with managing people, and those around dealing with client demands-had to be (re)aggregated repeatedly, without the benefits arising from face-to-ALMOND ET AL.
-9 face interaction, or specific knowledge of the local working environment of colleagues. The role of project manager, not recognized as an international management role in the corporate hierarchy, implied substantial experimentation in recreating and adapting norms and practices at the international team level. At the same time, the group had to remain compliant with the twin constellations of norms at corporate level (norms on performance, corporate values, ethical statements etc.), and those arising from specific contracts, where Company A and client company actors continually negotiated with each other in enacting the formal norms of the business contract. For example, the operation of performance management, while formally driven by the global Company A system, was inflected by client demands, such as the desire of the client to work with (or not work with) specific individuals in the international work group, and informal processes by which client managers reported back on the performance of specific individuals to the global work group leader.

| Company B
Company B was a large provider of a range of business services such as audit and consultancy. A federal structure comprised a string of national partnerships and an International Board, regulating issues such as the brand and inter-country agreements concerning resource flows. The influence of client expectations was evident. As one senior respondent put it, the firm needed 'a common set of goals, objectives, platforms and even a standard purpose across the whole network' (Interview 40) in order to deal effectively with clients who wanted a standard service internationally. As another put it: A few years ago we won some very big global audits, one in particular. That global client expected us to be able to tell them what our headcount was, where our experts were located and in what number, and we couldn't do that easily…if we can't do that we're not going to be relevant in the world of global business in five years' time (Interview 66).

| Competency framework
Contextual pressures to have a more coherent approach to global customers created a transnational space, which managers sought to use to develop a global competency framework. Overall, this competence development framework, was quite highly codified, albeit with some leeway for different local interpretations. The platform for its development was reliant both on formal authority-through the leader for global HR-and informal persuasion, in that the lead needed to 'sell' the idea to a designated person in each country. Likewise, in enforcement, respondents were aware that the framework corresponding to express wishes of very senior staff, rather than relying on formal sanctions. The move towards a revised global competency framework aggregated as part of a wider 'paradigm shift' towards much closer global integration, involving parallel moves towards integration in resourcing, finance and sales.
This competency framework consisted of 10 global core competencies, some of which were restated and reinterpreted, while others were new, such as 'global acumen'. These were built into a formal framework used in assessing promotion cases. However, these criteria were only partially codified, and permitted different in- The platform for diffusion was characterized as being reliant on the ability of leaders to frame the case for change. 'I don't have any normal means of execution, I just have the influencing skills', as one senior disseminator (Interview 65) put it, and went on: 'I don't have anybody from the territories who are reporting directly to me and so all of my work has to be done through influence'. She emphasized that her personal credibility and ability to build trust with local implementers were crucial factors in the successful roll-out internationally.
While these informal processes of persuasion and consensus-building were certainly important, it is evident that there is a complementary element of hierarchical influence, even in an apparently federal firm. The realization that the firm had a 'client delivery problem' (Interview 66) in its inability to draw on globally integrated systems led to a defining agreement between the most senior Partners in the firm. This stated clearly that the firm needed global management systems to avoid 'walking into a future that we cannot service' and the message that was unambiguously conveyed was: 'We are going to have global processes' (Interview 66).
Those leading the development of particular global norms and practices may not have had authority to instruct, and needed to use social skills in effectively framing the case for change, but were doing so in an organization in which managers had been sent a signal from the top that they should be working towards, not obstructing, global norm creation. In the case of the competency framework, the senior partner in the United States was a vocal supporter of the initiative. In this context, those championing the initiative found relatively little reluctance among senior country staff to participate. Moreover, once it was formally incorporated into the company's HR systems concerning performance appraisal and development needs, it was difficult for any manager or employee to avoid using it. 'And when you join…you have to learn these because (as) part of your interview process you have to state how you met those five areas' (Interview 40).

| 'Proactivity'
The final transnational norm-making space we cover represents an attempt to bring together a business unit in the tax function to be more globally integrated through a very particular norm-'proactivity'-being constructed. This norm was disseminated through the platform of a network of 'champions' trained in the idea and charged with disseminating it in their territory. It was weakly codified, and, as with the competence framework above, enforcement was dependent on awareness of the wishes of powerful individuals rather than specific sanctions.
'Proactivity' was in keeping with wider developments in the global firm, but its local impact was variable, as aggregation involved meshing with more country-specific business norms.
This norm emerged as senior actors within Company B perceived a need to 'interact with clients in ways that are more globally consistent' (Interview 43) to meet their expectations. The relevant International Partner identified the need for 'a common set of goals, objectives, platforms, and even a standard purpose across the whole network' (Interview 40) but went on to emphasize that what really matters is 'what you do underneath that.' What emerged from the analysis of his team of five senior staff was the aim of generating a higher degree of consistency across borders in the 'proactivity' of staff in relating to clients.
'Proactivity' concerned how staff approached clients and provided standardized services to them. It involved raising the awareness of staff in one country of developments in another so that they could engage in 'cross-flow selling'. Attempts to diffuse this norm relied on training and ongoing conversations within a network. It was recognized that successful implementation of a loose concept like 'proactivity' is difficult to measure. As one respondent involved in dissemination and implementation put it, 'we're given a little bit of allowance' (Interview 47). It was not entirely uncodified, however, as the team that led the initiative created shared written materials on a shared drive and developed training materials that were used through face-to-face delivery and lodged in an online library.
The challenges of disseminating the notion of proactivity and turning it into a norm were considerable. This was both because of the challenges in codifying it and because of the federal structure. As one norm creator put it, 'the ALMOND ET AL.
-11 challenge is how do we influence people (when) you're nobody's boss' (Interview 40). The mechanism used was the creation of a network of 'champions' who were 'a conduit for information to flow from the country back into the global leadership and vice versa' (Interview 47), and attempting to ensure that staff bought into the common standards through a 'bottom-up' route.
'Champions' opted into the coordination of 'proactivity' voluntarily, with incentives of strengthening their claims of good corporate citizenship in performance reviews and extending their personal networks. Once approved by local leaders, they undertook training that brought together small groups on a regional basis. They were encouraged to share experiences through conference calls and an online forum, and expected to present to the International Partner on the 'personal territory business development plan' (Interview 48). They tended to be in mid-career roles, all below Partner, and hence were relatively junior transnational norm disseminators and implementers who interacted with other staff and clients in their territory. Senior respondents were keen to emphasize that it was 'not meant to be us running some sort of a programme, it's meant to be a network where it's all connected.' (Interview 47). However, it would be a mistake to see this as a process governed only by informal interactions leading to coordinated action based on shared interests. The context of the wider patterns of global integration favoured by those in senior positions created a position in which the formal management hierarchy conditioned perceptions of how to react to these global norming processes. While the sanctions for non-compliance might have been weak in practice, there was nevertheless the sense that credit would be earned by enthusiastically engaging. Moreover, the resources at the control of the International Partner allowed him to carry out training to legitimize his initiative.
While 'proactivity' was taken seriously enough by actors across a range of countries to be considered a global norm, its impact appeared to vary markedly from one country to another. Interviews with 10 of the 'champions' demonstrated significant variation in the extent to which a wider group of staff engaged with the initiative. Some reported wide buy-in from staff locally, while others doubted the extent of change. Some within this second group suggested that it was unrealistic to expect behaviour to change given that the way in which people relate to clients is conditioned by longstanding local norms. As one respondent put it, the 'one firm' strategy was something that existed more at a strategic level than at the level of practice: across the firm there is very much an emphasis on people, one firm, our big thing is one firm working collaboratively…When it actually gets down to it, I think quite often different territories do almost what is best for them…they are on the ground listening to the clients themselves (Interview 36).
In summary, we examined four rather different types of norm. Two-the performance management system and competency framework-were largely codified and incorporated into the firms' HR architecture; the other twowork organization within international teams and 'proactivity' in client-facing roles-were less tangible and required ongoing clarification and discussion for them to shape behaviour. Yet all had both formal and informal elements to their substance in practice. Patterns of (in)formality in codification, enforcement, platforms and aggregation for our four cases are summarised in Table 2, reflecting the dynamic interplay of formality and informality across these elements of norm-making in all four cases.

| CONCLUSION
This paper contributes a framework for the analysis of norm-making across the multiple organizational and societal contexts in which MNCs typically operate. It provides a means of performing granular analysis of the construction of global norms and their solidification into employment practices, by identifying elements of formality and informality in the processes of codification, enforcement, the construction and deployment of platforms, and aggregation between norms. Our analytical focus on transnational norm-making spaces extends the remit of context-sensitive international HRM research, by enabling analysis of a wider range of mechanisms for the international coordination and control of work. The IHRM literature has developed sophisticated tools to analyse policy transfer between corporate HQs and national subsidiaries (Kostova et al., 2016), and how this is shaped by organizational and societal contexts.

T A B L E 2 Four transnational norm-making spaces
However, our evidence demonstrates that a fuller contextualization of IHRM needs to recognize that the global coordination of human resources in MNCs is not solely, or even perhaps mainly, about flows of HR policies between corporate HQs and national subsidiaries. Focusing on the identification and analysis of transnational norm-making spaces allows us to integrate the analysis of different types of norm-making, including formal HR policies, but also corporate initiatives and the issues involved in international project group coordination.
Second, we build upon the micro-political tradition in MNC research (Ferner, 2000;Geppert & Dörrenbächer, 2011), by constructing and deploying a framework for the systematic investigation of both formal and informal elements of norm-making. Micro-political research has focused on informality in the form of subsidiarylevel resistance to global policies, and/or as the work performed by subsidiary-level managers in working through official policies in diverse national settings. We extend work in this tradition by demonstrating that informality in norm-making is not confined to such local-level resistance or accommodation. For instance, it made sense for senior international managers to allow substantial local interpretation of the 'values' in Company A's performance management system; informality in interpretation is here a constitutive part of the codification of the formal norm-making process. Moreover, informality in enforcement mechanisms can be very effective from senior managers' viewpoints; a notion that compliance will be 'seen well', or that 'something serious' might happen if norms are not followed, can be more effective than more precise enforcement mechanisms-too much precision may be difficult to consistently follow through in practice and, consequently, lack credibility. The platforms for norm-making also combined more and less formal elements, as was particularly evident in the cases of work organization in international teams in Company A, and the proactivity initiative in Company B. Finally, the processes of aggregation of norms and practices often create ambiguities, illustrated most clearly in Company A's performance management system, in which not only did appraisers adapt the system informally, but it also led to other alterations in their behaviour which were kept 'under the radar' and were not intended consequences of the norm in question.
A key practical implication for those engaged in global norm-making is that formality and informality need one another. It often suits senior managers to leave scope for local interpretation concerning how norms are translated into practice by those in operational roles. Equally, the possibility of formal sanctions being exercised keeps informal processes of interpretation and adaptation within certain limits, while the largely informal 'networking' we observed took place within a context in which the wishes of those in formally more senior positions were clearly expressed in a way that structured the incentives to engage in networking. Attempts to change norms, particularly across national borders, are likely to be more successful where formal and informal elements of norm-making are considered together. A key skill for managers, therefore, is to be able to read how the contexts in which the multinational operates will shape the interaction between formality and informality. As our research shows, a large number of employees may be involved in global norm-making, particularly where work flows cross borders.
Training provision in managing across international contexts should be extended to, and adapted for, actors such as the leaders of international operational teams.
With regard to potential future research directions, our research was focussed on developing understanding rather than on prediction, and therefore does not attempt to formally propose determinants of (in)formality across our four conceptual elements of norm-making processes. That said, in deliberately choosing to analyse quite different norm-making contexts, it is possible to consider how corporate and actor characteristics shape the relations between formal and informal elements of codification, enforcement, platforms, and aggregation, and how future research might generalise on this.
Perhaps most obviously, the different corporate characteristics in our research meant that senior managers at the HQs were endowed with varying degrees of formal authority, leading to differences in the extent to which 14 -ALMOND ET AL.
'enforcement' processes and the platforms used were formal in nature. Future research could usefully take the issue of organisational structural characteristics further by investigating comparable norm-making spaces across a wider range of organisations in terms of hierarchical structure across countries. Given known national effects on the extent of hierarchical centralisation and formalisation in MNCs (Belizón et al., 2013), it would be interesting to examine the extent to which the context of the country of origin conditions the nature of emerging transnational characteristics and relations between actors.
Differences in structure also affected the prevalence of types of actor. For instance, those specialising in disseminating norms through informal channels were more prevalent in global norm-making in Company B owing to the difficulty of relying on formal authority alone in a federal structure. A further element of the corporate structures of great relevance in shaping the presence of particular types of actor is the way in which new transnational spaces are opening up in the form of global virtual teams. Particularly evident in Company A, this has led to the emergence of a swathe of team leaders who oversee a largely informal process of global norm creation affecting individuals across borders within these teams.
Future research should consider more fully the roles of the increasingly wide range of actors involved in international norm-making processes; in our initial analysis of data, we identified as categories: corporate level managers who play an important role in creating new norms across the whole of, or parts of, the multinational; actors who engage in 'disseminating' norms across a region or part of the business; those who are expected to implement or enact new norms in the operations for which they have responsibility; and leaders of global virtual teams. While these different roles often overlap, identifying roles relative to norm-making seems more promising methodologically than a sole reliance on job titles, particularly in a context where the processes of managing people internationally are not confined to those labelled as international managers. It would be desirable to develop more fine-grained categorisations of such actors, their agency, and how they utilise formal and informal elements in the task of global norm-making, across a wider range of organisational contexts.

ACKNOWLEDGEMENT
This research was funded by the UK Economic and Social Research Council (ESRC), Award Number ES/N007883/1.

CONFLICT OF INTEREST
The author declares that there is no conflict of interest.

DATA AVAILABILITY STATEMENT
The data that support the findings of this study are in the process of being made openly available by UK Data Services (UKDS). Data will also be made available on request to the first author.