Article body

Introduction

The difficulties faced by trade unions in Advanced Market Economies (AMEs) (Towers, 1997; Kohler and Jimenez, 2015; Kretsos and Vogiatzoglou, 2015), including a decline in density and membership, reduced bargaining power and political influence (e.g. Gall, 2005), and the rise of ‘never members’, that is, employees who have never become union members (Bryson and Gomez, 2005; Haynes et al., 2008), have been studied extensively. More generally, unions face a problem of declining instrumentality—that is, a perceived decline in efficiency and effectiveness, manifest in concessions on wages and benefits alongside relentless change in labour markets (Lévesque and Murray, 2006). If trade unions in AMEs are to successfully renew and respond to changed circumstances, what needs to be done? One response is to look to new sets of resources and ‘levers of power’ (Lévesque and Murray, 2006). One lever is ‘potential recruits’ or unmet demand: employees who would like to join a union but are not union members (Pyman et al., 2009). Examining the unorganized as a potential source of renewal is an important line of inquiry following several decades of largely unsuccessful union revitalization programs in countries like Canada, the US, the UK and Australia.

This paper investigates ‘unmet demand’[1] in Australia by focusing on the predictors of employees’ willingness to join a union in non-unionized workplaces if a union were available. Using a national survey of employees, the Australian Worker Representation and Participation Survey (AWRPS), we address two important research questions: 1- To what extent are non-union members willing to join a union in non-union workplaces?; 2- What predicts non-union members’ willingness to join a union in non-union workplaces?

Our paper is organized as follows. Section one provides an overview of trade unions and the institutional framework governing industrial relations in Australia. The second section outlines the theoretical debates relevant to willingness to join a union in unionized and non-unionized workplaces, in order to contextualize the study’s objectives and to develop hypotheses. The third section outlines the survey, variables and measures before turning to a summary of the findings and a discussion in sections four and five respectively.

Trade Unions and the Institutional Framework in Australia

The most recent Australian Bureau of Statistics (ABS, 2016) data estimate that there are 1.6 million union members in Australia (August 2014) with a union density of 15.1% (Pekarek and Gahan, 2016). Both the number of union members and union density have been steadily declining since the 1990s (Teicher et al., 2013). Of particular relevance are the high proportion of never members; 90% of persons aged 15-19 years who have never been a union member and more than half of all other age cohorts who have also never joined a union (Pekarek and Gahan, 2016). These findings underscore the untapped potential for unions to recruit among non-union workplaces and especially younger workers.

Aside from the challenges of declining membership and density, Australian unions have also experienced a fall in the number of working days lost (to disputation), low real wage growth and a steady decline in collective agreements. The latter is indicative of a declining capacity of unions to engage in collective bargaining (Pekarek and Gahan, 2016). This decline is all the more surprising because Australian unions benefit from a favourable institutional environment that enables unions to compel employers to participate in collective bargaining provided that the union can establish coverage rights and demonstrate that it has members in that workplace.

Having provided an overview of the challenges facing Australian trade unions, we now introduce the evolving institutional framework within which they operate. Historically, the Australian system provided for compulsory conciliation and arbitration. However, in effect, for most of the twentieth century, this was a system where collective bargaining operated at the margins and arbitrated industrial awards were dominant. Beginning in the 1980s, political pressures for greater labour market flexibility have seen the scope for legally binding arbitration vastly reduced and a continuing shift toward a system based on workplace level bargaining in which unions are optional bargaining agents. Whilst there have been some constants, at the national level, there has been nearly continuous legislative innovation since 1996.

The first period of change, 1996-2007, came with the election of a conservative (Liberal-National Party) government and the end of a partnership (Prices and Incomes Accord) between unions and the social democratic (Labor Party) government that was first elected in 1983. The Workplace Relations Act 2006 (Cth.), inter alia, weakened unions and industrial tribunals through reforms aimed at creating a more competitive labour market, including: providing for individual contracts that could undercut legislated minimum standards; forcing a further shift to enterprise-level bargaining by restricting the arbitrated award safety net of minimum wages and conditions to 20 ‘allowable matters’; reducing the capacity for unions to organize and represent workers and bargain collectively; and, creating stronger rights for employers to lockout their workforce. Together, it has been argued that these measures sent a message encouraging employers to adopt hostile anti-union strategies (Wright and Lansbury, 2014; Cooper et al., 2009).

In 2005, the conservative government embarked on a second and more radical wave of labour market re-regulation known as ‘Work Choices’. This legislation decentralized and decollectivized industrial relations further, imposing severe restrictions on union activity and, again, curtailing the powers of industrial tribunals; for example, arbitration was barred even where this was included in a dispute resolution procedure negotiated by the parties and included in their agreement (Stewart, 2006; Wright and Lansbury, 2014).

Following the enactment of ‘Work Choices’ and a high profile union campaign led by the Australian Council of Trade Unions (ACTU) (Muir, 2008), the Australian Labor Party was elected in 2007 and held office until 2013. While the earlier Accord partnership was not renewed, unions exerted considerable influence in and over labour market regulation deliberations, leading to the passage of a new statute - the Fair Work Act 2009 (Cth.). Notable changes implemented in the 2009 Act were: an extended list of National Employment Standards (NES); requirements that enterprise agreements meet a ‘better off overall’ test[2]; and, restoration of employee unfair dismissal rights (Wright and Lansbury, 2014).

The subsequent defeat of the federal Labor government in 2013 produced a resurgence of discussion of the need for further legislation to curb union activity. However, without a majority in both houses of parliament, the conservative government has been unable to achieve substantive legislative change thus far (Forsyth, 2016). However, re-election of the conservative government in July 2016 has seen a re-affirmation of their commitment to further legislative change. In particular, the recent reports of the Productivity Commission Inquiry into the Workplace Relations Framework and the Royal Commission into Trade Union Governance and Corruption are likely to encourage the government in its efforts to enact new laws that will adversely impact on unions[3] and the institutional framework within which they operate (Forsyth, 2016; Wright, 2016).

This section has provided an overview of the position of Australian trade unions and the evolving institutional framework in which they operate in order to contextualise the central issue of this paper: the predictors of union joining in non-union workplaces in an environment where non-union workplaces are an overwhelming majority and there is an increasing proportion of never members (Pekarek and Gahan, 2016). We now turn to the key theoretical debates that underpin an examination of the predictors of employees’ willingness to join in non-union workplaces.

Predictors of Union Joining: Debates

The literature on union membership is substantial. However, for our purposes, three strands are relevant. These strands are: 1- the individual propensity to unionize; 2- the rise and characteristics of non-union workplaces and alternative forms of employee representation (employee voice); and, 3- managerial responsiveness to employees and unions. We consider each of these research strands in turn.

The Individual Propensity to Unionize

There is an extensive literature examining the factors affecting an individual’s propensity to unionize in predominantly unionized workplaces (e.g. Adams, 1974; Deery and De Cieri, 1991; Farber and Saks, 1980; Fiorito and Greer, 1982; Kochan, 1979; Charlwood 2002; Cullinane and Dundon, 2014). Charlwood (2002: 469) summarized this literature into three models encompassing both employee and workplace characteristics: dissonance, utility and political views. As our survey instrument was focused on workplace characteristics and attitudes, we did not include a measure of individuals’ political views. On the one hand, the absence of this individual measure may be regarded as a limitation of our empirical work given evidence that political values do influence confidence in unions and given that political views and the link to union joining have been previously examined (Blanden and Machin, 2003; Frangi and Hennebert, 2015; Riley, 1997). On the other hand, this limitation is in part offset by evidence that immediate work attitudes are stronger predictors of willingness to join a union than political beliefs (Charlwood, 2002), and, more generally, by the theory of planned behaviour (Azjen, 1991). Empirical evidence provides strong support for the theory of planned behaviour: that is, intentions to perform behaviours of different kinds can be predicted with high accuracy from attitudes toward the behaviour (Azjen, 1991). We return to this issue of individual attitudes in the conclusion and discussion of future research.

Dissonance theories propound that dissonance between the expectations and the experience of work trigger unionization (Charlwood, 2002; Farber, 1989; Fenwick and Olson, 1986). Dissonance may arise from: a- the work environment and become manifest in job dissatisfaction, income dissatisfaction and/or a perceived lack of justice, and/or b- the desire for influence, based on the perception that unions as a representative mechanism can deliver more effective influence over organizational decision making (Charlwood, 2002: 469). In a study of non-union employees in the US, Leigh (1986) found that the desire for unionism rises sharply with job dissatisfaction and that satisfaction with pay increases is associated with a lower demand for union representation. More recently, using a sample of 3372 unionized and non-unionized manual employees in the US, Friedman et al., (2006) found that career prospects and the work environment were negatively related to employees’ desire to join a union, whereas employees’ degree of job satisfaction and satisfaction with benefits were positively related to the desire to join a union: that is, those employees who experienced job dissatisfaction or dissatisfaction with benefits were more likely to join a union[4]. More generally, the notion that job dissatisfaction is linked to a desire for representation by a union is well established (e.g. Buttigieg et al., 2007); therefore we advance hypothesis 1:

  • Job satisfaction in non-union workplaces will be negatively related to willingness to join a union (H1).

Utility theories suggest that the decision to join a union results from a rational calculation of the costs and benefits of unionization (Charlwood, 2002: 470). Individuals’ perceptions of union instrumentality are a direct influence on such a calculation (Charlwood, 2002; Cregan, 2005; Gordon and Long, 1981). The concept of union instrumentality is multi-faceted, with strong links between perceived instrumentality and individual behaviour (union joining decision) being identified consistently (e.g. Gahan, 2012; Brown Johnson and Jarley, 2004; Fiorito, 1987; Lévesque and Murray, 2006; Silverblatt and Amman, 1991).

Union instrumentality is defined as the perceived effectiveness of a union in improving wages and working conditions and can be measured inter alia by collective bargaining performance, workplace outcomes and/or union organizing success (Badigannavar and Kelly, 2005; Fiorito and Greer, 1982; Rose and Chaison, 1996). In Britain, Charlwood (2002) found that union instrumentality was the most significant predictor of the willingness to join a trade union. In an Australian study of union joining and leaving, Buttigieg et al., (2007) also found that extrinsic union instrumentality, that is, perceptions of union effectiveness and satisfaction with union services with respect to wages and benefits, was important in explaining union joining and leaving.

While there is empirical support for utility theories and union instrumentality as a driver of the individual joining decision, Bryson (2003) has developed the notion of ‘frustrated demand’ in which the financial and personal costs of joining outweigh demand and perceived effectiveness. Although this concept may help explain employees’ willingness to join in non-union workplaces, for example, if management is perceived to be hostile to union members (personal cost), this construct has not been empirically explored in Australia. However, there is Australian evidence that free riding impedes perceived union instrumentality. Quite simply, instrumental motivations to free ride in a workplace prevail over ideological or personal motivations and organizational characteristics such as organizational size and hours of work in explaining why people do not join unions in Australia (Haynes et al., 2008).

Another potentially important construct relevant to the individual joining decision and linked to union instrumentality is workplace justice. For example, in a rare study of both union and non-union members’ individual attitudes toward union membership, Cregan (2005) found that non-union employees were more likely to join if their awareness of workplace injustice was greater. Her cluster analysis also revealed a larger pattern of non-union members who were categorized as: those with pro-union attitudes (collective values including identification with the union and a sense of solidarity) and an intention to join; those who were neutral (understood the benefits of collectivism but were unconvinced that joining was worthwhile); those who were uninformed; those who showed no interest; and, disillusioned former members who viewed the union as ineffective and undemocratic. Cregan (2005) concluded that few members fell into the category of holding anti-union or unitarist views as a reason for not joining. This finding is akin to D’Art and Turner’s (2008) conclusion that a substantial majority of European employees wanted the protection of strong unions and held positive attitudes towards unions.

On the basis of the above literature, we advance hypothesis 2:

  • Perceived union instrumentality in non-union workplaces will be positively related to willingness to join a union (H2).

Non-union Workplaces and Alternative Forms of Representation

Alternative forms of employee representation may substitute for unionization, but the evidence is mixed. For example, Belfield and Heywood (2004: 279) found that in the UK, the presence of human resource management[5] and advanced employee involvement schemes were associated with a lower desire for unionization. Similarly, in Canada and the US, Campolieti et al., (2013: 378) found that non-union forms of employee representation were negatively related to the presence of unionization at the workplace. Conversely, Charlwood (2002) found that the presence of alternative voice did not predict reduced willingness to join. A relationship between the presence of alternative forms of representation and unmet demand may be explained by the concept of the ‘representation gap’ in non-union workplaces; that is, that there is no alternative model of employee representation that employees will accept as an effective substitute for union representation (e.g. Millward et al., 1992). Notwithstanding this explanation, evidence in the UK and Australia has also shown that alternative voice practices (dual channel employee voice), including non-union and/or direct voice, can also complement traditional unionization (e.g. Pyman et al., 2006; Uhe and Perkins, 2007). On balance, we advance hypothesis 3:

  • The availability of alternative voice in non-union workplaces will be negatively related to willingness to join a union (H3).

Managerial Responsiveness to Employees

A third important consideration of the willingness to join a union and in explaining unmet demand is managerial responsiveness to employees. We know, for example, that employer choice, defined as employers’ decisions as to whether to adopt a ‘voice’ regime, is a critical determinant of the type and impact of voice practices in workplaces (e.g. Charlwood, 2006; Willman et al., 2006; Bryson et al., 2013).

There is little empirical research addressing the influence of managerial responsiveness on employees’ resultant desire for unionization. There is, however, a related body of research that examines the influence of management attitudes to unions on an individual’s joining decision (e.g. Bryson and Gomez, 2005; Campolieti et al., 2013; Friedman et al., 2006; Gall, 2005; Pyman et al., 2010). Importantly, Belfield and Heywood (2004) found that employees in the US and the UK who anticipated or perceived managerial opposition to unions were much less likely to desire or join a union. They also found at the individual level that non-union employees’ desire for unionization was lessened if they perceived a high level of influence over information-exchange; a proxy of the industrial relations climate. We argue that information exchange is a proxy for the industrial relations climate in that the industrial relations climate can be understood as the extent to which relations between employers and employees are positive (Pyman et al., 2010). The more employers share information with employees, the more positive relations between employers and employees are likely to be. Similarly, formal HRM practices at the group level, such as team working and performance related pay have also been found to reduce the desire for unionization, whereas utilization of problem-solving practices raised the desire for unionization (Belfield and Heywood, 2004). Related to this, Cregan and Brown (2010) found that the willingness of non-union employees in Australia to participate in joint consultation varied directly with their expectation that a joint consultative committee would lead to more democratic representation. Drawing on these findings, we advance hypothesis 4:

  • Perceived managerial responsiveness to employees in non-union workplaces will be negatively related to willingness to join a union (H4).

Data and Methods

Sample and Procedure

The data reported in this paper are drawn from responses to the 2010 Australian Worker Representation and Participation Survey (AWRPS), a national survey that investigated employees’ responses and attitudes to workplace participation, representation, and influence. The instrument was based on the 1994-1995 Worker Representation and Participation Survey conducted in the US (Freeman and Rogers, 1999), the 2001 British Worker Representation and Participation Survey (Diamond and Freeman, 2002), the 2003 New Zealand Worker Representation and Participation Survey (Haynes et al., 2003), and a previous 2007 AWRPS (see Freeman et al., 2007; Teicher et al., 2007). Questions from other country surveys were adapted, to conform to the institutional and demographic contexts in Australia.

A total of 500 employees sampled randomly from residential telephone directories were surveyed nationally using computer-assisted telephone interviewing (CATI) over a two-month period in 2010. The sample was limited to Australian residents engaged in paid employment for more than 10 hours[6] per week who had left secondary school. Self-employed persons and company owners were excluded from the survey. The sample was stratified by Australian state/territory to reflect the geographical distribution of the population as reported in the ABS Census of Population and Housing.

In this study, we are focusing on the respondents, just over half of the sample (n = 268 or 54 per cent), who reported that there was no union at their workplace that people doing their type of work could join. Of these respondents, 57 per cent were male and the mean age was 40.05 years (SD = 12.00). The mean number of hours worked per week was 38.26 (SD = 14.15). The majority of respondents (74 per cent) were non-manual employees, and 78 per cent were employed in the private sector. Two-thirds of respondents (66 per cent) reported that they worked in workplaces with fewer than 50 employees and just under a quarter (23 per cent) in workplaces with 100 or more employees. The mean number of years that employees had worked for their current employer was 5.57 (median = 3, SD = 6.75).

Measures

Dependent variable

Willingness to join a union: Following Charlwood (2002), respondents in non-union workplaces were asked: “If a union was formed at your workplace, how likely would you be to join?” Responses were rated on a 4-point ordinal scale ranging from 1 = “not at all likely” to 4 = “very likely.”

Independent variables

Perceived union instrumentality: Following Charlwood (2002), employees in non-union workplaces were asked: “Would you be better or worse off with a union that you could join at your workplace, or would it make no difference?” Responses were rated on a 5-point scale ranging from 1 = “a lot worse” to 5 = “a lot better.”

Job satisfaction: Employees were asked to respond to the statement: “Overall, I am satisfied with my job.” Responses were rated on a 5-point scale ranging from 1 = “strongly disagree” to 5 = “strongly agree”. Our measure is an indicator of global job satisfaction. Single-item measures of global job satisfaction are commonly used in large-scale survey research in which an overall assessment of employee attitudes is desired (Spector, 1997). It has also been found that such measures have excellent test-retest reliability and show convergent validity with multi-item measures of job satisfaction (Saari and Judge, 2004).

Perceived managerial responsiveness to employees: Employees were asked to rate the performance of the management at their workplace on a 5-point scale ranging from 1 = ‘failure’ to 5 = ‘excellent’ across six dimensions: 1- concern for employees; 2- giving fair pay increases; 3- willingness to share power and authority; 4- keeping everyone up to date with proposed changes; 5- promoting equal opportunities for all employees; and 6- making work interesting and enjoyable. These items were drawn from the New Zealand Worker Representation and Participation Survey (Haynes et al., 2003) and are similar to those used by Bryson (2004). A principal components analysis of the six items supported a single-factor, with a Cronbach’s alpha coefficient of .88. The scores on the six items were averaged to form a composite measure, with higher scores indicating greater perceived managerial responsiveness.

Presence of alternative voice: Following Willman et al. (2006), alternative voice was operationalized by the presence of the following two-way forms of communication between employees and management: daily walk around the workplace by senior management; an ‘open door’ policy so employees can tell senior management about problems with their supervisors; suggestion schemes; regular hard copies of a workplace newsletter; notice boards; surveys or ballots of employees’ views and opinions; email and workplace intranet (including magazines or staff bulletins); regular staff meetings between senior management and employees; team briefings; non-union employee representatives (such as an agent, lawyer or advisor); an employee involvement program (such as quality circles, semi-autonomous or self-directed work groups); a formal grievance or dispute resolution procedure; or, a committee of managers and employees who meet regularly to consult over a range of workplace issues. We constructed an index averaging how often these two-way communication arrangements were reported to be used in the workplace (on a scale from 1 = not at all to 4 = daily). This index has a potential range from 1 to 4, with higher scores representing greater alternative voice in the workplace.

Control variables: Based on the prior literature and a selection of relevant studies (e.g. Charlwood, 2002; D’Art and Turner, 2008; Toubol and Jensen, 2014), we included several control variables in the regression analyses. First, we included gender, age, hours worked per week, occupation and organizational tenure, since these individual employee characteristics have been shown to affect the individual propensity to unionize (Toubol and Jensen, 2014). Gender was coded (1 = male, 0= female) and occupation was coded (1 = non-manual, 0 = manual). Age and organizational tenure were measured in years. We also included organizational characteristics that have been shown to affect individuals’ propensity to unionize: workplace size and sector. Workplace size (total number of employees) was coded using a five-point ordinal scale (1 = < 20 employees, 2 = 20-49 employees, 3= 50-99 employees, 4 = 100-499 employees and 5= 500 or more employees). Sector was coded (1 = private, 0 = public/not-for-profit).

Methods of Analysis

Because our dependent variable (willingness to join a union) was ordinal in nature, we used ordered probit regression (Tabachnick and Fidell, 2007). To facilitate the comparison of effect sizes from independent variables measured with different scales, we standardized (z-scored) the predictors prior to analysis. Following Tabachnick and Fidell (2007), we found no evidence of multicollinearity among the predictors (see Table 2, all correlations among predictors were under the commonly used .70 threshold). Due to positive skewness, we log transformed organizational tenure; however, this did not change the pattern of results so we retained the non-transformed variable.

Results

Among respondents employed in non-union workplaces, 34 per cent said they were fairly likely or very likely to join if a union were available at their workplace (see Table 1); a finding consistent with previous surveys in Australia (see Teicher et al., 2007). However, only 11 per cent of our sample expressed a strong desire to join a union if given the opportunity. In this regard, we noted that about three-quarters of workers believed that a union would make no difference to them personally, suggesting a high degree of indifference.

Table 1

Willingness to Join a Union in Non-Unionized Workplaces (%)

Willingness to Join a Union in Non-Unionized Workplaces (%)

N = 268

-> See the list of tables

Table 2 presents the means, standard deviations and correlations of the study variables. There were statistically significant correlations between willingness to join a union and perceived union instrumentality, perceived managerial responsiveness to employees, job satisfaction, age, organizational tenure, occupation (manual vs. non-manual), and presence of alternative voice in the workplace. The signs of these coefficients were in the expected direction, and the strongest correlate with willingness to join a union was perceived union instrumentality, followed by perceived managerial responsiveness to employees.

Table 2

Means, Standard Deviations, and Correlations among the Study Variables

Means, Standard Deviations, and Correlations among the Study Variables

p < .05.

-> See the list of tables

The results of the ordered probit regression analyses predicting the willingness to join a union are shown in Table 3. Following Bryson (2004), we included managerial employees in the analyses because they are likely to be managed by other senior people in an organization and have the right to unionize in Australia. The effect on the results is negligible if managerial employees are excluded from the analysis.

Table 3

Results of Ordered Probit Regression Predicting Willingness to Join a Union

Results of Ordered Probit Regression Predicting Willingness to Join a Union

Notes: B = partially standardized regression coefficient. Standard errors in parentheses. N = 268.

* p < .05.

-> See the list of tables

Only two predictors of the willingness to join a union were statistically significant at the 95 per cent level of confidence: perceived union instrumentality and perceived managerial responsiveness to employees. Thus, hypotheses two and four were supported and hypotheses one and three were not supported. The positive association between perceived union instrumentality and a willingness to join a union was particularly strong (B = .69, p < .05), consistent with Charlwood’s (2002) research in Britain and the work of Buttigieg et al., (2007) in Australia on union joining and leaving. In terms of managerial responsiveness to employees, as hypothesized, employees’ perceptions of the responsiveness of managers were negatively related to willingness to join a union (B = -.27, p < .05). This finding shows that if employees rated their manager as being poor in responding to employees, they were more likely to join a union if given the opportunity.

Discussion

The initial impetus for this research was a paucity of knowledge on union joining in Australia, and especially on employees’ willingness to join a union in non-union workplaces. Gaining this knowledge is important, in view of the Australian evidence of latent unmet demand and an increasing proportion of ‘never members’ and non-union workplaces (Pekarek and Gahan, 2016). Using a national survey, we addressed two important research questions: 1- To what extent are non-union members willing to join a union in non-union workplaces?; 2- What predicts non-union members’ willingness to join a union in non-union workplaces?

Our analyses found that about a third of non-union members were willing to join a union if one were available. After controlling for a range of personal, job and workplace characteristics, only two statistically significant predictors of non-union members’ willingness to join a union in a non-union workplace were identified: perceived union instrumentality and perceived managerial responsiveness, with managerial responsiveness working in the opposite direction. That is, a lack of managerial responsiveness to employees was a predictor of non-union members’ willingness to join.

Surprisingly, in terms of individual characteristics, holding constant other variables in the model, the level of job satisfaction was not a statistically significant predictor of willingness to join a union, thus failing to provide support for Charlwood’s (2002) dissonance model. However, we found that union instrumentality had the strongest influence on willingness to join a union, therefore providing support for Charlwood’s (2002) utility model. Interestingly, union instrumentality has also been found to be a significant influence on the desire to join a union in unionized workplaces (e.g. Cregan, 2005; Gall, 2005; Belfield and Heywood, 2004).

The importance of instrumentality as a predictor of non-union members’ willingness to join a union highlights a lack of emphasis in the existing research on union renewal regarding the need for unions to convince non-union members that unions can make a material difference to their substantive terms and conditions of employment and to justice and fairness in the workplace. In countries such as Australia, the institutional environment, currently existing under the Fair Work Act 2009 (Cth.), provides some scope for unions to demonstrate instrumentality through periodic reviews and test cases before the national industrial tribunal (Fair Work Commission), which can change labour standards on matters such as flexible working hours and penalty payments for weekend work. More generally, in an era of declining union influence, unions will find it difficult to demonstrate effectiveness unless they can redefine the considerations that constitute effectiveness in the eyes of the potential recruits (non-unionists).

Interestingly, in terms of organizational (workplace) characteristics, we found that the presence of alternative forms of voice were not predictive of non-union members’ willingness to join a union, but the level of perceived managerial responsiveness to employees was. That is, it was found that if employees perceived management to be responsive to their needs, the presence of alternative voice did not add further predictive power of the willingness to join in non-union workplaces. One possible explanation for managerial responsiveness predicting an unwillingness to join a union, but not the presence of alternative voice, may be that voice is mediated by the degree of managerial responsiveness. In other words, the effect of alternative voice may be indirectly transmitted, if responsiveness is an outcome of voice.

Our finding that perceived managerial responsiveness to employees was associated with decreased willingness to join a union is an important contribution to the literature, since few studies have explored this factor, particularly in non-union workplaces. This finding provides confirmatory evidence that management sincerity and authenticity are crucial to employees’ positive experiences of work and that these are preconditions for perceived effectiveness of employee voice (e.g. Boxall and Purcell, 2016). Our findings also lend support to the notion that there is a difference between structural voice and the materiality of voice (Couldry, 2010); that is, the forms or channels of voice vis-à-vis the substance of voice, the latter relating to how managers actually listen, act and respond to employees’ concerns or issues. Similarly, Holland et al., (2012) argued that it is not the specific voice practices that are important (akin to structural voice), but rather employees’ direct interactions with management and subsequent perceptions of managerial responsiveness (substance of voice) that are most influential in producing positive employee outcomes, including trust in management. Our results regarding managerial responsiveness may also be linked to the sample characteristics: a greater proportion of non-manual workers in small private sector workplaces in which workers have closer relationships with management and a higher level of influence in the workplace.

One final question arising from the findings is whether managerial responsiveness to employees is a proxy for workplace justice in the workplace. Our survey measure of managerial responsiveness has items that approximate to workplace justice and, as seen in section two, workplace justice has been linked to the individual joining decision (Cregan, 2005). In terms of the scale of perceived managerial responsiveness and the link to justice, and referring directly to the items in the scale (see Measures in section 3), items 1, 4 and 5 equate to distributive, informational and procedural justice; item 2 to distributive and procedural justice; item 3 to employee influence; and, item 6 to intrinsic motivation and job satisfaction. Consequently, it could be argued that our measure of perceived managerial responsiveness to employees brings together influence and justice. This relationship between influence and justice gives rise to the argument that unions can demonstrate instrumentality through a focus on the work itself, and, employees’ influence upon organizational decision making. The importance of influence and justice in predicting willingness to join has implications for union servicing and organizing strategies, suggesting a need to monitor and evaluate, or at least gauge on a regular basis, the climate (state of relations) between employer and employees as measured through the extent of managerial responsiveness to employees.

Conclusion

The reasons employees join unions in unionized workplaces has been the subject of longstanding inquiry, but the predictors of joining in non-union workplaces have rarely been studied. As far as we can determine, this is the first study, using Australian data, which examines the predictors of employees’ willingness to join a union in non-union workplaces. Previous research has implicitly assumed that the decision-making processes and behaviour associated with union joining are the same across union and non-union workplaces. We challenge this assumption and find that perceived union instrumentality and perceived lack of managerial responsiveness to employees are significant predictors of willingness to join a union in non-union workplaces. We also suggest that in order to enhance their instrumentality in the workplace, unions could be more effective in recruiting if they targeted never members, as experience of unionism is closely linked to perceived effectiveness and thus joining (Bryson and Gomez, 2005).

This study of unmet demand and the predictors of non-union members’ willingness to join a union makes important theoretical and practical contributions to the literature. Our contribution to the literature on union joining and ‘frustrated demand’ is twofold. First, we develop a deeper understanding of unmet demand in Australia[7] by examining the predictors of non-unionized employees’ willingness to join a union. Second, we complement research on ‘frustrated demand’ (Bryson, 2003)[8] and union joining in unionized workplaces, by analysing Australian data on non-unionized workplaces for the first time, and based on the premise that willingness to join a union may not be the same across union and non-union workplaces.

Despite the insights this paper provides into the individual decision to join in non-union workplaces and the subsequent implications for union strategy, it has several limitations that require consideration in future research. First, given the cross-sectional design of the study, causality between an expressed willingness to join and the influence of the selected predictors cannot be conclusively determined. Second, the data gathered were based on intended behaviour (self-report) and could be subject to response biases. Related to this potential bias is the possibility for common method variance, a frequently identified problem with self-report data (Podsakoff et al., 2012). Accordingly, we recommend that a longitudinal study be undertaken in order to strengthen causal inferences and to explore individuals’ actual behaviour with regard to union joining, and, with regard to the relevance of workplace factors such as organizational culture, compensation, extent of training provided and trust in management (Kim, Kim and Ali, 2015). Third, while we believe our single-item measures of willingness to join a union and union instrumentality have face validity, we have no data on their reliability. In part, our rationale for using single-item measures was to minimize respondent load and maximize the response rate. Future research could use multi-item measures to address this limitation of a single item measure of willingness to join.

Finally, a related limitation was the exclusion of two variables relevant to this study’s findings and veracity: past membership of a union and employees’ political or ideological beliefs, which have been found to be a determinant of individual attitudes toward unions (Deery and De Cieri, 1991; Toubol and Jensen, 2014). The earlier 2004 AWRPS (see Teicher et al., 2007) did include past union membership, but after controlling for instrumentality, it was not found to be a statistically significant predictor (see also Riley, 1997 for a review of similar findings). With regard to political or ideological beliefs, we instead focused on union instrumentality, because, as we noted earlier, prior research shows that proximal attitudinal variables are stronger predictors of willingness to join a union than more distal factors, such as political and ideological beliefs acquired through socialization (Charlwood, 2002). The findings in this paper therefore provide support for Riley’s (1997: 283) argument that the “hypothesised mechanisms behind significant correlations between respondent, industry and company specific factors are likely to be psychological in essence, and hence may function through attitudinal factors”.