University of Birmingham Employee Choice of Voice and Non-Union Worker Representation

Existing research emphasises employer choice in determining the form and content of non ‐ union employee representation (NER) structures. This article puts forward an alternative but complementary thesis: we conceptualise NERs as an ‘ offer ’ that employees can choose to either accept or reject, with several shades of settlement in between. This article argues that employee choice matters insofar as it determines the likely impact and sustainability of the NER offering. Contributing an original typology of employee choices and the associated prospects for the NER, we illustrate our argument via six workplace case studies assessing NER trajectory.


| INTRODUCTION
Interest in non-union forms of collective employee voice tends to focus on employer motives (Donaghey et al., 2012), the institutional arrangements of such bodies and the resulting outcomes (Dobbins & Dundon, 2020;Laroche & Salesina, 2017;McCloskey & McDonnell, 2018). The literature on non-union employee representation (NER) schemes observes a mix of employer motives for such bodies (Gollan & Xu, 2014), typically depicting employers as having a high degree of choice in determining both the structure of NER and the scope of issues over which employees can have a say (Kaufman & Taras, 2016). Ultimately, employer choice is a significant influence in a context where the prospects for NER voice tend to be chequered: a mix of impacts ranging from ineffectual and marginal to genuinely effective consultative outcomes (Ackers et al., 2005;Wilkinson et al., 2021). This article does not dispute these findings. Instead, it reconfigures the dominant lens of analysis away from focusing on employer choice towards a focus on worker agency. We argue that employees' choice of how they respond to NER structures matters: we conceptualise NERs as an 'offer' for voice by employers that employees can accept or reject, with various shades of settlement in between. Crucially, we argue that employee choice matters because it can determine the eventual impact of the NER. Derived from a review of the existing literature, we propose a typology of employee responses developed from two key factors: first, the extent to which the response of employees and their agents is positive or negative to the offer; and secondly, the degree to which those employees have agency to engage with the offer. We deploy a multi-case study design of six NERs, charting the trajectory of such structures in each case over time and across issues. Our article thus contributes an original conceptualisation of NER structures, one that shifts attention from employer dominance to consider the capacity and agency of labour in effecting change to the scope of NER voice. We structure the article as follows: in Section 2, we review the existing literature on NER. Section 3 presents a new typology of NER employee choice responses. The following two sections concern the empirical demonstration of our ideas: the underpinning fieldwork methodology that supported the typology and illustrations in practice. The final section discusses the implications of our analysis.

| LITERATURE REVIEW
With the decline in union representation and bargaining coverage, there is ample space for employers to experiment with new forms of employee voice. Existing research focuses on how employers design and control the working of such structures, their motives and outcomes (Avgar, 2021;Butler, 2009a;Della Torre et al., 2021;Dobbins & Dundon, 2020;Wilkinson et al., 2018). Ramsay (1977), in a seminal contribution, proposed that such regimes resulted from workers' potential to contest or consent to managerial authority. For example, as workers' collective power and mobilisation challenged employer authority, employers created voice structures to accommodate or contain workers' demands. As the 'challenge from below' receded, employers would progressively withdraw from collective voice arrangements. In contrast, an analysis of employee involvement and participation (EIP) schemes put forward by Marchington and colleagues Marchington et al., 1993) argued that the radical thesis was inadequate in explaining newer forms of worker participation: with a mix of employer choice, deregulatory political forces, and labour and product market pressures, both individual and collective EIP schemes existed as uneven waves, rather than cyclical shifts in collective worker power. The choice was not simply motivated by a desire to control labour but wrapped up alongside new managerial techniques such as TQM or Quality Circles .
If these analyses tilted the study of non-union structures in the direction of employer behaviour, later research tends to solidify this approach as standard. Several studies consider the provision of voice structures chosen by employers (Lavelle et al., 2010;Marginson et al., 2010). Bryson et al. (2019) and Willman et al. (2006) propose that employer preference for a particular voice regime is likely to be a prime factor in its emergence, particularly in liberalmarket economies where institutional constraints on employers are lax. Less sensitive to multiple motives, these authors assume employers are simply trying to reduce transaction 2 | DONAGHEY ET AL.
costs. In their view, the choice boils down to four mutually exclusive options: buy union voice, make non-union voice, hedge (a combination of the former two), or maintain the no voice default. While these authors add that employer preferences can change due to employee dissatisfaction or union campaigns, employers tend to stick to regime choice based on the high switching cost. Hall and Purcell (2012) and Hall et al. (2013) find that employer choices determine the nature and extent of workplace consultative structures. Employers decide whether information and consultation (I&C) is 'active', mostly restricted to 'communication', or prone to becoming 'defunct' over time. Indeed, the contribution of other actors to such arrangements is a byproduct of employer practice. For example, employers' approach to consultation influenced the degree of organisation among workforce representatives. Where employers pursued an 'active' approach to consultation, developing representatives' competence and coordination skills were required. Initial problems experienced by some organisations who opted to be 'active consulters' were overcome by workforce representatives 'growing into their role' as sought and defined by the employer. In those cases where 'communicators' dominated, there was no imperative for coordinated input from workforce representatives and, consequently, little opportunity for representatives to develop the necessary cohesion to press for more effective consultation. Those employers pursuing a communications approach required little workforce interaction or cohesion with workforce representatives feeling less need to meet or network between meetings with the employer. In a minority of instances, Hall and Purcell found that representative capacity to influence NER bodies is confined to where union recognition exists or unions seek to colonise such structures to establish influence. Only in one of the non-union cases, and considered by Hall and Purcell as 'untypical', are representatives assertive in pressing for institutional amendments to the voice regime.
On the one hand, there are reasonable grounds for accepting the various strands of what might be termed the employer dominance thesis to NER. Subject to various motives, employers appear to drive the 'vision' behind the NER and its forms, scope and functions (Butler, 2009b;Kougiannou et al., 2021). On the other hand, however, there is a long-standing tradition of interdependency and reciprocity as core to industrial relations (Valizade et al., 2016). It is plausible that reciprocity dynamics play an important role in addition to, or alongside, the accounts of employer dominance mentioned above. The workforce invariably has significant agency over the extent to which they receive or engage with such bodies. As Kaufman (2015) argues, advancing frameworks that purport to explain voice in organisations but include only one actor as a protagonist are akin to cutting a piece of paper with one-bladed scissors. Similarly, Wilkinson et al. (2010) argue that the 'returns' from NER are proportional to both sides' investments. Rather than seeing employer choice as the sole determinant setting the tone of NER trajectory, a more helpful interpretation is to see NERs as part of a dynamic set of reciprocal 'offers' and 'counter-offers' over the scope of issues for voice on the part of the employer. The workforce may choose to accept or reject in whole or in part elements of the offer and the nature of that response will have consequences for the NER arrangement.
Some studies recognise that workforce responses to NER are likely to be essential and highlight workers as important actors (see Fairris, 1997). One of the earlier studies by Broad (1994) found that employer NER strategies are subject to counter-mobilisation by the workforce when a company employee council was used as an alternative voice channel to suppress unionisation. When workers realised the mechanism protected employer interests over their concerns, workers rejected the body and eventually supported the idea of unionisation (see also Badigannavar & Kelly, 2005). Similarly, Taras and Copping (1998) and Gollan (2001) observed non-union employees rejecting NERs developed by management in their Imperial Oil and Eurotunnel cases, respectively. In contrast, research on NERs by Chaykowski (2015), Gollan and Xu (2014), Dietz et al. (2005) and Kaufman and Taras (2010) found that employees embraced the structure as an effective means of representing their concerns. In between these opposites of accepting and rejecting are halfway houses where employees see NER as marginal or modest (Johnstone et al., 2010;Johnstone & Wilkinson, 2015;Slinn, 2014). In these cases, whether the workforce accepted or rejected an employer's 'offer' was crucial for determining the NER impact.

| ADVANCING THE EXISTING LITERATURE
This section proposes a typology of employee choice that captures a broad spectrum of responses to employers' NER offers and the impact of such responses on the structures' impact. Specifically, we advance that there are likely to be four types of employee response and that, rather than employer choice being the sole decider, worker responses to what is offered can potentially influence the NER (see Figure 1). These four responses are derived from two conditions. The first is the extent to which employees are willing to engage (or not) with the F I G U R E 1 A typology of worker responses 4 | DONAGHEY ET AL. NER, as offered by management. The second condition is whether workers have high or low agency regarding their capacity to influence the employer's offer. From this, we devised a 2 × 2 matrix with our associated worker responses. Conceptually, this is not a static model. Rather, it is intended to serve as more fluid and dynamic set of types that can change over time and space, even within a single organisation or for a particular class or employees in a work unit.
First, we propose the employee response of acceptance, where workers accept the employer's offer and exercise sufficient agency to respond to the structures offered by employers to resolve their concerns adequately. As such, the NER's structure is satisfactory for workers and there is no effort to alter these fundamentally. In terms of impact, the structure is more likely to be sustained over time and retain influence, which Johnstone and Wilkinson (2015) highlight for WebBank and Dietz et al. (2005) at Sportasia. Sustainability does not imply that the NER is necessarily one of 'mutual gains' or 'positive sum' type outcomes as is sometimes expected of what constitutes a successful NER (Dobbins et al., 2017;Valizade et al., 2016). Rather it implies that the body functions as an acceptable means of securing employee concerns, through the eyes of employees, as Johnstone and Wilkinson (2015) have argued. Employees may experience some gains and some losses, as is the nature of employment relations exchange, but do not seek to reject the existing structure. The second employee response is adjustment. Adjustment is where employees find some limitations or are frustrated with the existing structure offered by the employer but have the agency required to seek to change it rather than rejecting it outright. Here, the outcome is contested and there is a pattern of ebbs and flows, as Taras and Copping (1998) highlight in the movement between non-union and union-based representation. In the short run, this may lead to a mixed bag of positive and negative engagement with the NER as employees attempt to make the NER responsive to their concerns. In the long run, the response pattern may move toward the acceptance category if employees successfully get the employer to address their frustrations with the structure.
Alternatively, the NER's impact will likely recede into either of the following two categories if the employer does not resolve employee frustrations with the structure. Acquiescence is the response pattern where employees view the structure offered by management as inadequate but cannot seek an alternative and 'withdraw into silence' (Donaghey et al., 2011). In these circumstances, the offered structure might endure for some time but will likely have marginal or zero influence and may wither over time. One can observe such a scenario in Broad's (1994) study of Deniko, where substantive actions did not match the employer's commitment to a structure, and thus while the representative structure persisted, it drew little interest from workers. The final response employees may choose to a NER offer is avoidance. Under this scenario, we propose that employees will not only discard the NER structure offered by the employer as inadequate, but crucially, given the continued prevalence of issues of concern, they seek an alternative mode of voice articulation that will act as a competitor to the employer's NER arrangement. Gollan (2002) documents such examples at News International and Markey (2007) at SunCorp. Despite these companies investing significant resources in NERs and giving them a high degree of autonomy, in both cases workers engaged with their preferred unions instead for representation.
Our proposed typology has an advantage over the 'employer dominance' approach as it brings the elements of reciprocity and workers' agency back into the fold. In addition, the typology highlights a greater degree of political dynamism to our understanding of NER, appreciating that the exchange can move from one pattern to another over time. The third advantage of the typology over existing management-dominated approaches for NER in the literature is the recognition that workers can impact the depth and scope of NER voice.
Emphasis on the latter allows for a synthesis of the employer's offer, the employees' counteroffer (should it exist) and the outcome (the depth and scope of NER voice structure). The remainder of the paper illustrates the value of the typology by drawing on several cases of NER. We show how the NER's trajectory depends on the workers' response to the employers' offer.

| CASES, CONTEXT AND METHODS
Cases are drawn from Ireland and the UK, where the regulation of NERs is permissive: Employers are generally permitted to organise such structures without requiring a rigorous mapping of efforts to statutory prescription (Butler et al., 2018;Dundon et al., 2014). Legislatively backed NERs exist on one-off items like transfer of undertakings/redundancy consultations which retain relevance for non-unionised workers and, more generally, under information and consultation regulations in these jurisdictions. The latter stem from a European Union directive and provide for (and still do in the UK context despite the UK leaving the European Union) employee information and consultation rights. This article is part of a wider project investigating the impact of the EU Directive General framework for informing and consulting employees (2002/14/EC) in Ireland and the UK. Insofar as that Directive promised the potential for permanent structures of formally non-union representation, that project provided a highly complementary basis for studying NER, mainly as regulatory oversight in both jurisdictions was permissive and not intrusive of firm-level choices. However, in both jurisdictions, such provisions' coverage and their potency is generally assessed as being very much at the margins (Donaghey et al., 2013) with significant flexibility for local actors to act without strict recourse to the regulations or simply to ignore them.
Six of the 16 firms involved in the wider study are most attuned to this article's conceptual concerns. The remaining 10 were a mix of unionised cases or cases that did nothing in response to the Directive and are thus irrelevant to this article. Engagement with and reflection on the six cases provided the inductive impetus for the ideas contained in this article and subsequent engagement with the literature on that basis. To be sure, impelling the initial cases were different empirical concerns and an analysis of an existing case set is presented herein. However, given the initial research objectives, there were no apparent imbalances in the original case reports. Instead, characterising case reports was a breadth and depth of thematic evidence regarding firm-level voice arrangements, their motivation, evolution and evaluation by multiple organisational stakeholders. Such thematic evidence provided the opportunity to develop and refine initial emergent ideas. We thus inductively generated the typological framework as a heuristic device for advancing understanding in the field through an iterative process of within and across case secondary analysis and reviews of the existing literature.
The case design is illustrative (Epler, 2017, p. 21). Under such designs, the intention is to depict a particular set of circumstances to explain a situation. Thus, the cases are not designed to test or falsify particular conceptual concerns but to demonstrate the typology's utility as a heuristic framework. Cases are not necessarily discrete episodes in that one case exclusively typified one type of worker response to the employer's NER offer. Instead, most cases exhibited change through time, meaning responses could evolve given the capacity of the NER structure to accommodate such dynamics. Our cases are not matched to a common set of empirical features other than that they were characterised by the employer adoption of NER structures. Instead, our selected cases display a high degree of cross-case dissimilarity in ownership structure, size and type of operations, product and labour markets, management styles, numbers employed and conditions of employment. While there are disadvantages associated with such dissimilarity, most notably in making comparisons, it has one strength in evidencing how the qualitative dynamics outlined in the typology hold across very different cases, strengthening its analytical value (see Mills et al., 2010).
Evidence from each case was acquired through company and site-level fieldwork across five years, between 2008 and 2012. Companies were selected and targeted due to prior desk research, which evidenced the existence of employer-provided NER structures. After successfully negotiating access to targeted sites or securing access to key respondents, evidence collection involved multiple location visits and interview rounds across five years by a team of researchers undertaking interviews with employers and workforce representatives. As shown in Table 1, we conducted 127 interviews across the six cases. Complementing interview evidence was documentary material amassed during the fieldwork. That information, provided by the employer and worker representatives, offered additional documented accounts of NER structures and practices, including minutes from NER meetings held between management and union and non-union staff representatives, provided data on the issues channelled through the voice system that existed at particular cases and sites. Examples included employee grievances like redundancy concerns, pay and bonus issues, overtime working, grading and performance appraisals (among others).
The 127 interviews focused on the extent to which employees could share in decisionmaking outcomes and whether NER structures allowed for consultation or negotiation or were confined to communication or downward information flows from the employer. We also asked about the range of issues or matters typically addressed by the NER structure. In turn, we asked for evaluations of NER depth and scope in terms of employee satisfaction with such arrangements. Where arrangements were evaluated negatively, we sought to determine what efforts employees made to address dissatisfaction. We analysed our evidence via protocols appropriate to the case study design, as suggested by Eisenhardt (1989) and Yin (2018). Conceptual categories included 'depth', 'scope', employee evaluations (positive/neutral/ negative) and employee counter-responses where negative assessments prevailed. The coding was a refining process that underwent several rounds of review among multiple researchers who checked and cross-referenced the coding of both interview reports and documentary results to ensure validity, accuracy and consistency of interpretation and analysis. It is to the results of that exercise that we now turn. volunteered for the role rather than subject to election. However, grievances with a collective dimension for the workforce were channelled to the structure by representatives on issues like rosters on bank holidays or seasonal periods. Most of these from the period of fieldwork were minor operational issues. As the NER routinely met at specified monthly intervals, staff raised the issues more due to routine requests from representatives to bring something to the meetings. Representatives saw the NER as a query raising rather than an issue resolution structure, and any issues representatives raised were not particularly pressing or a source of contention. Instead, employees at the site were more pressed by individual grievances, which were channelled to the first-line supervisors. In this context, the resolution of individual grievances was discreet and timely. The informal conduit for resolving individual grievances obviated the need for grievances to accumulate collectively and require raising at the NER. As a result, we found a twofold approach to raising grievances amongst employees: the individual level of first-line supervisors and the intermittent level of the NER. There was an acceptance amongst worker representatives interviewed that the latter was fit for purpose as a communication mechanism on queries and the airing of minor collective concerns. The NER prevailed throughout the fieldwork, was used intermittently and was not subject to contention. At BritCo., the NER offered emanated from union substitution and occurred in the circumstances of a union organising drive. Although a prior NER existed, it had fallen into disuse due to inactivity and disinterest on the employer's part. The union threat revamped the NER as the HR Director of BritCo. observed: The information and consultation forum was in place before we received advances from the union, but when we rebranded to the NER, there was certainly a little bit more noise in the system. So we enhanced the relationship because if we got that right, we felt employees wouldn't see the need for a third party.
Union activists made up a third of the elected representatives to the revamped NER. Union activists were 'neither encouraged to engage with the forums, but neither were they told to disengage' (Union Organiser). The union performed a background role in assisting union activist-cum-NER representatives with advice and guidance on dealing with workplace grievances as they arose. The other exclusively non-union representatives volunteered and were elected, motivated by an interest in the NER and a desire to work with the employer.
Again, the structure appeared to work well in the first year. The offered agenda by the employer focused on two key issues of contention stimulating the union organising drive. One grievance was the redundancy terms at BritCo, as company restructuring had cultivated fears about job losses. The grievance was that the prescribed payout in the company handbook compared unfavourably with industry standards. The employer conceded that the human resource department should rewrite the redundancy programme in consultations with worker representatives on the NER. The employer offered representatives a month to review the revised terms by holding meetings with their constituencies on any pertinent issues. The outcome was improved terms and the introduction of a redundancy pool, providing employees at risk of losing their jobs with 8 weeks to secure a new position or project in the company. Worker representatives interviewed expressed satisfaction with how the NER addressed the grievance.
Nevertheless, the NER was less satisfactory when worker representatives sought a solution to the employer practice of not disclosing salary scales, the second pressing issue behind the union organising. At NER sessions, the employer remained unwilling to shift position on the practice. Although the employer attempted to compromise by offering to annually post a confidential sheet to individual employees-outlining relevant payroll codes, job families and benchmarked pay range-this proved unsatisfactory to representatives in its lack of transparency. In the main, employees nonetheless accepted the outcome as a compromise arrangement: the structure seemed to at least offer a vehicle for the resolution-if only partial -of grievances. Further evidence of the structure's acceptance was the stalled momentum of the union organising campaign in the same period attributed by the organiser to worker representatives' attention turning to the NER, at least in the initial 12-month period after the employer revamped the structure.

| The adjustment pattern
We observed the adjustment pattern across three of the six case studies comprising this study. Furthermore, we found the pattern to act as a precursor to other patterns emerging. In this pattern, representatives accepted the general principle behind the NER in offering an opportunity to resolve grievances-indeed, in the case of BritCo., the historical origin of the NER was a pattern of acceptance in the first instance. However, in these cases, we tended to find employee representatives articulate sustained criticisms of the NER's depth and scope, which introduced a degree of contention and instability into the structure. Throughout the fieldwork, employees could splinter off from the structure and move closer to the acquiescence or avoidance category. Such outcomes were not, however, universal, and we often observed a degree of fracturing amongst representatives on the continued utility of the NER.
At Retail Co., workforce representatives used the NER to complain about classic 'tea and toilet roll issues'. Utilising the structure for such grievances was to the annoyance of store managers who tried to promote the NER as a quality circle. The NER may well have engendered a pattern of acceptance before the fieldwork, but representatives were generally highly critical of the structure during the fieldwork period spanning 5 years. There was widespread cynicism amongst representatives about the effectiveness of the NER. As one employee representative observed: You have an opportunity to voice an opinion, but whether any heed will be taken of that is another thing altogether.
Employee representatives voiced criticism of the employer for the superficial response to issues raised at the store level and the bureaucratic process of passing the grievances through various upper-level, non-store NER structures. Representatives assessed the elaborate nature of the NER structure as generating unnecessary delays to store-level problems. Several representatives perceived it as a tactic to 'take the heat' out of grievances. The issue of shopfloor ventilation is one example. Ventilation was reportedly a common grievance across all stores, with complaints that the facilities were too cold in the winter and too warm in the summer. At two stores, however, we observed representatives repeatedly raising the grievance at the NER regarding temperature levels inside the facility. In both cases, representatives complained first to the Store NER who claimed it was a matter appropriate to the regional NER; yet upon bringing their grievance to the regional structure, representatives were informed that EMPLOYEE CHOICE OF VOICE it was relevant for hearing at the divisional level NER. Despite these repeated attempts conducted across 24 months, the employer offered no resolution of the issue.
Despite cynicism and a belief that the NER structure needed reform to be more responsive to store-level issues, employee representatives continued to engage with the structure. Over the lifetime of the fieldwork, representatives routinely utilised the NER for information and consultation and were asked to by their constituents. Despite the structure's reported failings, most representatives saw the utility of the NER for small requests like new uniforms and jackets, canteen changes and staff training on pallet trucks. On the other hand, as discussed in the avoidance pattern, representatives were dropping out and others were relying on alternative sources for grievance resolution.
Despite the initial pattern of acceptance at BritCo, we observed that many employee representatives shifted their position over time. The basis for the shift was representative claims that once the momentum of the union organising campaign dissipated, the calibre of issues addressed changed. Representatives claimed that attempts at raising other substantive grievances on terms and conditions were kept off the agenda by the management chair or insufficiently engaged at NER meetings. Instead, the NER appeared to be led by facilities-type issues such as office cleaning and furnishing changes. Over nine months, employee representatives attempted to discuss growing concerns over a recently introduced Performance Improvement Plan (PIP). Representatives sought to address constituent concerns over the lack of transparency in the criteria and claims that first-line supervisors used the PIP process punitively rather than developmentally. However, the employer countered that the practice was not up for discussion as it was a core element of the current corporate strategy. The employer consistently ruled out repeated attempts to raise the matter as an agenda item by representatives. While worker representatives recognised that the provision of information by management had improved on pre-union organising experiences, growing frustrations in the second year emerged with the NER as attempts to raise issues were stonewalled.
At Medical Co., the NER originated from a formal employee opt-in request for negotiations to introduce an I&C arrangement under the EU Information and Consultation regulations. The triggering of the opt-in had involved state bodies, which encouraged the employer to evidence a pre-existing NER structure to circumvent the request. However, employees involved in the optin request challenged the legitimacy of this structure as not meeting legislative requirements. To placate their concerns, the employer appointed the leading organisers behind the opt-in campaign to sit as representatives on the NER, alongside several employer appointees. The employer also offered representatives the opportunity of a project to review how the NER could be improved. However, divisions arose within the representatives between the employer appointees, the majority, and the minority involved in the opt-in campaign. The majority argued that the existing NER was fit for purpose and met legislative requirements, while the minority pressed for the Standard Rules of the legislation to be imposed with the assistance of the state Labour Court.
Furthermore, the majority who accepted the NER also proved content with the proposed constitutional remit to 'inform and consult', whilst the minority sought to change this to 'inform and consult with a view to reaching an agreement' as the basis for the new structure. Divisions over the function of the structure among representatives prompted the employer to put the dispute views on the NER to referendum amongst the wider workforce. A contentious referendum campaign subsequently arose, with the employer warning the workforce that changes to the NER would be looked upon unfavourably by headquarters, whilst those seeking adjustment towards the standard rules were prohibited from circulating written campaign 12 | DONAGHEY ET AL. material. With close to 80% turnout across the plant, the referendum gave support of 62% in favour of the existing NER. However, the accepted NER was relatively remote from day-to-day influence in employment relations, meeting only twice a year. Its influence appeared marginal, confined to a downward communication forum on company developments with minutes of meetings posted on the staff intranet. Despite its relatively marginal role, it appeared to have shifted into the acceptance pattern.

| The acquiesce pattern
The pattern of acquiescence is where employees view the structure offered by the employer as inadequate but do not or cannot actively seek an alternative and withdraw into silence. In these circumstances, the offered NER might endure for some time but likely to have marginal or zero influence. Three cases fell into this pattern: Concrete Co. and Manufacture Co., whose pattern we found to be consistently in this acquiescent pattern and BritCo., which retreated into this pattern over the course of the fieldwork.
At Concrete Co., some work sites had collective 'works committees' tailored to specific site issues concerning production, health and safety and work schedules. Not all sites had such committees, being present only where there was a demand for such bodies by either management or the site workforce. However, there was no coherent election or nomination scheme at most sites. Instead, the employer selected representatives from the workforce, with such choices influenced by personal relationships or convenience. Employer representatives reported it difficult to recruit employees to get involved in the committees or attend. These committees met on average every 2-3 months, but the calling of sessions largely depended on whether individual matters had arisen in the company that management wanted to review. Overall, the committees were a marginal influence in shaping employment relations and were strictly controlled by management to ensure that only production or health and safety issues were featured. According to the HR Director: We know about all the kinds of structures you could have in place but from an employer's point of view, what is the benefit? I know about the cost that would arise from them. The structures that we have in place are negligible. It's all about a desire to work and behave in a safe manner.
The most common response from those employees who participated in the fieldwork was that their views about changes to their work are rarely valued or acted on by the employer. In terms of topics, employees said they are mainly asked for views on health and safety but are never asked views on pay, changes to work practices, staffing issues or plans for the company. As one employee representative commented: Decisions made are not negotiated, and suggestions from employees do not appear to have any effect. Employees are often the last to hear.
The NER at Manufacture Co. has never played a prominent role in the company and has existed in a context of general disinterest and low-level commitment from representatives and employees alike. Despite providing for highly detailed electoral procedures, ballots had not regularly occurred, given the limited interest of employees in applying for the role. The employer admitted that getting people to come forward could be challenging. We found only one elected employee representative during the fieldwork; supervisors had appointed the remainder. The employer also doubted how effectively representatives communicated back to their 'constituents.' Indeed, many representatives acknowledged that they did not report back to their constituents and often returned to work, which they were typically under pressure to complete. Line managers typically encouraged a return to work and were noted by employee representatives to be disparaging towards the NER as disruptive of shopfloor operations. Representatives also reported that they had trouble securing points from their constituents. That outcome partly stems from employees not having issues they wished to bring to the attention of representatives and partly due to a preference to address grievances via alternative routes of informally engaging line managers. Despite an initial flurry of activity in the very early stages of the structure's life, where there was discussion of issues over bonus payments and hiring, employee representatives perceived the process and outcome as downward communication from the employer. Alternatively, grievances initially raised in the early phases of the forum around water filters, shopfloor lighting and toilets had shown little progress leading representatives to conclude that there was no point in bringing these issues forward in subsequent meetings. Indeed, representatives felt that the employer should 'compromise more' on specific issues if only to portray an image to the shopfloor that the NER and the representative function could deliver on workplace issues. As one employee representative claimed, "most people have no interest in the Forum. Most things people ask for get turned down." Such outcomes led to widespread frustration.
The pattern at BritCo. was one of initial acceptance to adjustment. When attempts at adjustment failed to accommodate employee issues of concern, the turn of events was towards the acquiescence pattern. Three representatives dropped out of the forum after a year and a half under the perception that it was 'toothless'. Driving that decision was a failure to address grievances over performance management and emerging divisions among the representatives. The non-union representatives outlined in interviews that they saw the position as an opportunity to demonstrate their abilities to management and allow them to advance up the corporate ladder. Indeed, some representatives interviewed outlined that they ran for the position on the advice of line managers, who claimed that it would assist future promotion efforts. Such representatives acknowledged that they were less inclined to raise issues that might fracture their working relationship with management and often declined the opportunity to raise workplace grievances. Postures of that sort were a source of annoyance for the employee representatives who were active union members and felt that it debilitated efforts to raise issues at the NER. A combination of representatives dropping out or being reluctant to raise conflictual issues resulted in fewer issues going before the NER falling off by the end of the fieldwork.

| The avoidance pattern
Under the avoidance scenario, we propose that employees not only discard the voice structure offered by the employer as inadequate but, crucially, given the continued prevalence of issues of concern to them, seek to use an alternative mode of voice articulation. Three cases exhibited this pattern's features, although the alternative did not seem to threaten the NER structure's existence necessarily. However, at Manufacture Co., Retail Co., and BritCo, continued employee dissatisfaction tended to be expressed outside of the NER structure and at an individual level.
One of the reasons given by representatives at Manufacture Co. for employees refraining from providing grievances was that they opted to bring issues individually to their line managers. Employee representatives felt that it was unlikely that some employees would want to share individual grievances with a fellow employee, even if that employee was a formal representative. Instead, the natural port of call appeared to the front-line manager. Supervisors were perceived to be speedier and potentially more effective in resolving individual employee concerns.
Similarly, at Retail Co., there were instances of employees opting to avoid the NER favouring individual engagements with line management instead. As with Manufacture Co., front-line supervisors, not the NER, were assessed as a more effective and timely means for resolving grievances. Indeed, there were also instances of frustrated employees taking a grievance to outside statutory bodies when the NER process had been exhausted. In the grievance mentioned above on shopfloor temperatures, employees turned to the statutory Health & Safety Authority: it investigated the complaint, found it valid and served the company with an enforcement notice.
At BritCo., frustration with the NER manifested itself in avoidance, exemplified by the differences in subsequent caseload between representatives (or former representatives) who were union members and those who were not. Towards the end of the fieldwork, the latter representatives reported a decline in caseload, which they attributed to three factors. One was a lack of training on their part over grievance handling meant that representatives lacked confidence in addressing significant individual grievances from their colleagues, in some cases advising the colleague to seek advice elsewhere or simply abdicating responsibility if they felt the issue was too serious. In some cases, this may have resulted in acquiescence, but tempering this conclusion is that representatives or former representatives who were known union members reported a growing number of inquiries directed towards them. Such grievances were encouraged by these representatives to be brought through the formal union structure to allow for individual representation. By the conclusion of the fieldwork, the union had reported increased attendance at individual disciplinary and grievance hearings, which the company insisted on holding off-site.

| DISCUSSION AND IMPLICATIONS
Before drawing broader conclusions, a summary of the above patterns is helpful. Categorising cases and outcomes for the NER requires further qualification and comment before developing wider argumentation. In some cases, worker choices evince stability, while in other examples, there was drift across typological characteristics simultaneously and over time and space. Concrete Co. and Hotel Co. exemplify consistency. In the former, the NER was consistently marginal and workers did not or could not seek alternatives. In the latter, workers perceived the NER to work or expressed no grievances with it. Front line supervisors played a complementary role in containing grievances from escalating to the NER structure that might have been unable to resolve them.
In contrast, patterns of acquiescence and avoidance emerged at Manufacture Co. Many spurned the NER in favour of individualised relationships with front-line supervisors who acted as competitors rather than complements to the NER. At Medical Co., the circumstances EMPLOYEE CHOICE OF VOICE | 15 of the NER offer were a source of contention. A pattern of adjustment held sway until the employer triumphed and the worker activists side-lined. Subsequently, the employer orchestrated a NER consistent with its minimalist preferences for information and consultation. With no necessary sign of acquiescence, avoidance or adjustment among the wider employee body, the employer's offer appears accepted by default. Characterising Retail Co's NER was persistent efforts from representatives to bring forward issues the employer preferred not to address: a dynamic meriting the categorisation of adjustment. Limited success in adjustment encouraged avoidance with a preference for line management engagement, although the structure features a simultaneity of some acceptance, some adjustment and some avoidance. The case of BritCo. resided in a mixture of acquiescence and avoidance. Notably, the union and not line management emerged as the competitor.
We can now identify three key interrelated contributions from our cases. The first is that the evidence demonstrates the need for greater attention to worker agency in the context of NER. To date, as we have shown, the literature has generally focussed on the employer's approach. The cases demonstrate that NERs depend on how representatives and employees view the structures; that is, the 'offer' from the employer. Significantly, it is not just relevant to ask whether the employer intends the structure to be 'consultative' or only for 'communications' (Hall & Purcell, 2012), but equally how those patterns of employer choices for voice are interpreted and responded to by workers. Our cases show that the extent to which employees regard the structure as an effective mechanism for issue resolution is vital to the efficacy of NER voice. Worker choices to accept, adjust, acquiesce or avoid can shape the nature of the structure. Undoubtedly, the employer is the actor with the most power. However, having less power is not the same as having none. The implications of these findings are set out in Table 2.
The second point of significance emanating from the cases is that in contrast to existing studies, which see unions as the main competitor to NER structures, we find that front-line supervisors may act as potential competitors or complements (Townsend & Hutchinson, 2017). McCloskey and McDonnell (2018), for example, noted in their study that the availability of multiple voice channels can have counter-productive effects whereby they start to compete with rather than complement each other. Thus, the competitive source to the NER is often T A B L E 2 Implications of employee choice and prospects For NER internal rather than external, although much depends on the quality of relations with frontline supervisors. For example, in Manufacture Co, front-line management effectively attempted to displace the NER. Of course, construing the front-line as a competitor might not be entirely accurate in formal terms. In handling individual grievances, they address issues that the employer may not want in the NER structure. In resolving individual grievances, the front line may be complementary rather than competitive, as in the case of Hotel Co. However, whichever tendency dominates among front-line management is likely to be an empirical question that merits greater attention than hitherto has been the case (Teague & Roche, 2012). Third, and building on this latter point, we propose that caution is necessary if focusing simply on the formal voice structures. For example, Willman et al. (2006) and Lavelle et al. (2010) approach voice regimes based on the presence of structures rather than the depth of the arrangements. We do not advocate abandoning examining the presence or absence of voice structures. However, our cases demonstrate the need for greater cognisance of workers' issues for voice and their capacity to influence employer structures. The extent to which workers engage meaningfully with voice structures depends on the issues that are the focus of their interests at a particular time (Heery, 2010). While Ramsay's (1977) argument stressed that the employer offering was related to issues of labour control, that approach fails to acknowledge worker agency in response to such offerings at the workplace level. In our cases, workers utilised voice structures that they believed most appropriate to resolving an issue at a specific time and in a particular context. We suggest that the appropriateness of voice mechanisms is not derivable from the presence or absence of a structure but from whether they are useful for resolving workers' particular sets of issues.
Finally, in terms of future development, the typology presented here is readily amenable to further falsification to challenge or disconfirm. Case design would thus move from illustrative to testing purposes. For example, one could propose alternative 'prospects' for the NER in light of particular worker choices. Here scholars could demonstrate particular choices occurring under different levels of agency (high/low) or resulting in different types of NER prospects than the ones postulated here. Such efforts would offer a valuable way to challenge the initial insights provided in this study and advance further study in the field around the role of worker agency in shaping voice regimes.