A comparison of the effectiveness of employee voice arrangements in Australia

Blackwell Publishing LtdOxford, UKIRJIndustrial Relations Journal0019-8692Blackwell Publishing Ltd, 20062006375543559Original ArticleEmployee voice arrangements in AustraliaAmanda Pyman et al.

Industrial Relations Journal 37:5, 543–559 ISSN 0019-8692

A comparison of the effectiveness of employee voice arrangements in Australia Amanda Pyman, Brian Cooper, Julian Teicher and Peter Holland ABSTRACT The incidence and effectiveness of different forms of employee voice were compared across three measures: perceived managerial responsiveness to employee needs, job control and influence over job rewards. Multiple regression analyses revealed that voice was perceived as most effective when an amalgam of different forms was present. The effective coexistence of different forms of voice challenges the notion that nonunion voice acts as a substitute for union representation.

INTRODUCTION At its simplest, voice describes how employees raise concerns, express and advance their interests, solve problems, and contribute to and participate in workplace decision making. Employee voice has traditionally been channelled through union recognition and representation, but this has never been the exclusive means of communication and influence at the workplace. Indeed, as union membership and the coverage of collective bargaining have declined markedly in advanced western nations over the last two decades, there has been increased interest in alternative forms of representation (Bryson, 2004). It has been argued that a significant gap exists between the actual level of employee representation provided and the desired level of employee representation (Bray et al., 2001; Freeman and Rogers, 1999; Towers, 1997). To some extent this gap has been redressed by a growth in non-union and direct forms of voice. Drawing on survey data from 1,000 employees in union and non-union workplaces in Australia, this article compares the effectiveness of different voice mechanisms across three criteria: perceived managerial responsiveness to employee needs, perceived job control and perceived influence over job rewards. The effectiveness of employee voice mechanisms has become particularly pertinent since the 1996 election of a Liberal– National Party coalition government, which is committed to marginalising union influence and encouraging direct and non-union voice. Despite the importance of employee voice, there is still a dearth of empirical research that examines the incidence and effectiveness of different voice mechanisms (Gollan, 2005; Willman et al., 2003). ❒ Amanda Pyman is Lecturer in Industrial Relations and Human Resource Management, Kent Business School, Canterbury. Brian Cooper is Research Fellow, Julian Teicher is Head of Department and Peter Holland is Lecturer, Department of Management, Monash University, Australia. Correspondence should be addressed to Dr Amanda Pyman, Kent Business School, The University, Canterbury, Kent, UK; email: [email protected]

© 2006 The Author(s) Journal compilation © 2006 Blackwell Publishing Ltd, 9600 Garsington Rd, Oxford OX4 2DQ, UK and Main St., Malden, MA 02148, USA.

544

Amanda Pyman et al.

This article is structured in five main sections. In the first section the concept of employee voice is examined. This is followed by a historical analysis of the patterns of employee voice in Australia. Sections 3 and 4 present the method and results. The fifth section discusses the findings in light of the literature and considers the implications of those findings.

EMPLOYEE VOICE Discussions of employee voice in the academic literature have been linked to organisational performance, climate and commitment (Dundon et al., 2004); employee satisfaction and industrial citizenship (Bishop and Levine, 1999); and high involvement workplaces and high performance work systems (Boxall and Purcell, 2003). It is the work of Freeman and Medoff (1984) that first adapted the exit/voice concept to the labour market. They argued that voice, in the sense of direct two-way communication, was important for both employers and employees. Unions were, however, seen as the central, independent vehicle for the collective voicing of employees’ dissatisfactions (Freeman and Medoff, 1984). Consequently, much of the industrial relations literature views the articulation of grievances, whether on an individual or collective basis, as the sole component of voice (Gollan, 2001). An emphasis on grievances and union-only voice reflects a narrow conceptualisation of the alternative regimes available to employees to advance their rights and interests at work, and to contribute and participate in organisational decision making. As recognised in the human resource management literature, the meanings of employee voice can be categorised into multiple strands, including: articulation of individual dissatisfaction; communication/exchange; collective representation and organisation; upward problem solving; engagement and contribution to management decision making; and demonstration of mutuality and cooperative relations (e.g. Dundon et al., 2004; McCabe and Lewin, 1992; Wilkinson et al., 2004). While employee voice has historically been channelled through union representation and collective bargaining, declining unionisation and the growth of human resource management have been accompanied by a burgeoning interest in alternative forms of employee voice, particularly non-union representative voice (joint consultation) and direct voice (e.g. Bryson, 2004; Dundon et al., 2004; Haynes et al., 2005; Kaufman and Taras, 2000; Kleiner, 2001; Millward et al., 2000; Storey, 2001). The incidence of union versus non-union and direct voice mechanisms has been widely debated. For example, Benson (2000) found that Australian employees in unionised workplaces had a greater number of voice mechanisms present than their counterparts in unorganised workplaces. Delery et al. (2000) identified a similar pattern of voice arrangements in the US. The superiority of union versus non-union voice regimes has also received widespread attention in the academic literature, and the theoretical arguments are well known. Direct voice has been argued to be superior to union voice for two reasons: first, barriers between employers and employees can be disintegrated by dealing directly with employees rather than through an intermediary (Bryson, 2004; Storey, 1992); and second, direct voice allows managers to better respond to the heterogeneous interests of workers (Storey, 1992). In contrast, union voice is assumed to be superior on the basis of ‘independence’. In the western tradition, trade unions are independent representatives of employee interests by their very nature. Enterprise or © 2006 The Author(s) Journal compilation © Blackwell Publishing Ltd. 2006

Employee voice arrangements in Australia

545

in-house unions have been challenged on the grounds that they are driven by managerial objectives specific to the enterprise (Gaffney, 2002). Indeed, one of the major challenges to the effectiveness of non-union representative and direct voice mechanisms is in their capacity to transform the power relations in an organisation. Because they lack sanctions for non-compliance, collective power and access to independent sources of advice or assistance, non-union voice mechanisms are more susceptible to managerial influence and control (Gollan, 2000a; Terry, 1999; Wilkinson et al., 2004). The substitution thesis is also a major theme in the industrial relations and human resource management literature examining the incidence and superiority of different voice regimes. Research in the US and UK often portrays non-union employee representation as a union avoidance strategy, suggesting that such voice regimes are more likely to be found in non-union workplaces because human resource management practices act as a substitute for unionisation and collective bargaining (Fiorito, 2001; Machin and Wood, 2005). That this is a simplistic view of the operation of different voice mechanisms is evident from the ‘social partnership’ literature, which demonstrates that non-union representative and direct voice measures can coexist with union voice (Budd, 2004). The coexistence of different voice mechanisms in practice is clearly illustrated in the European countries through the operation of works council structures (Terry, 1999). Despite a widespread interest in alternative forms of non-union representative and direct voice, previous research has focused on the impact of union activity on the exit/ voice model, including employee turnover, quit rates, retention and job tenure, job satisfaction, and training (e.g. Batt et al., 2002; Delery et al., 2000; Freeman, 1980; Iverson and Buttigieg, 1995; Iverson and Currivan, 2003; Miller and Mulvey, 1991). There remains a paucity of research on the effect, operation and effectiveness of different combinations of voice mechanisms in practice (Dundon et al., 2004), a deficiency that is particularly pronounced in Australia. Further, the available research on the effectiveness of voice regimes has produced inconsistent results. Kim and Kim’s (2004) comparison of union and non-union representation, for example, found that unionists were more satisfied than non-members, particularly in regard to distributive and employee advocacy issues. In contrast, Bryson’s (2004) analysis of the 1998 British Workplace Employee Relations Survey (WERS) data found that non-union voice was more effective than union voice in eliciting managerial responses to employee needs. This article addresses the gap in the Australian literature by comparing the perceived effectiveness of different configurations of employee voice. We use voice measures similar to Bryson’s (2004: 239) to compare the effectiveness of union, non-union and direct voice on three outcomes: perceived managerial responsiveness to employee needs, perceived job control and perceived influence over job rewards. An important aspect of this article and an extension of the work of Bryson (2004) is the inclusion of job control and job reward measures in evaluating the effectiveness of employee voice. Voice is judged to be effective when it is associated with more positive responses from managers to employee needs, greater levels of control over the work process and increased influence over job rewards.

PATTERNS OF EMPLOYEE VOICE IN AUSTRALIA In Australia, union voice has historically been the most prominent means by which workers were organised and represented in the employment relationship. The © 2006 The Author(s) Journal compilation © Blackwell Publishing Ltd. 2006

546

Amanda Pyman et al.

dominance of union voice dates back to the enactment of compulsory conciliation and arbitration in 1904. Notions of collective representation and third-party dispute resolution underpinned this regulatory system, and unions were legally constituted as representative of employee interests, in relation to both members and non-members. The legitimacy afforded to unions under the arbitral system was embedded in extensive institutional guarantees, including exclusive coverage of defined categories of employees and guaranteed recognition and representation rights. Despite the prominence of unions under the arbitral system, it has been argued that employee voice was marginalised by the centralisation of industrial relations processes at the national and industry level (Bray et al., 2001). The gradual decentralisation of the industrial relations system since the 1980s and a series of legislative reforms predicated on neoliberal ideology have contributed to a significant decline in union voice, and the emergence of alternative forms of representation. In 2004, Australian unions represented less than 23 per cent of the workforce (ABS, 2004). Formal regulation of direct and non-union employee representative structures first emerged in the early 1970s in the form of joint consultative committees, with the passing of specific statutes governing occupational health and safety in the Australian states (Bray et al., 2001). Similarly, during the 1970s and 1980s, various anti-discrimination and equal employment opportunity laws were enacted at state and national levels. In creating an environment free from discrimination, employers resorted to representative and direct forms of consultation. More extensive forms of direct and non-union representation awaited the development of decentralised bargaining over wages and conditions, beginning in the late 1980s. It could be argued that the development of a public policy framework for enterprise bargaining was central to the National Wage Case decisions in 1987 and 1991, but enterprise bargaining was not regulated by statute until the passage of the Industrial Relations Reform Act 1993 (Cth.). Two types of representative structures were introduced in 1993: certified agreements with unions and enterprise flexibility agreements (EFAs). The latter were promoted as a non-union form of agreement making. In reality, the legislation mandated employer consultation and negotiation with unions over EFAs at the formulation stage, greatly diminishing the incidence of direct bargaining between employers and employees (Campling and Gollan, 1999). Following the enactment of the Workplace Relations and Other Legislation Amendment Act 1996 (Cth.) (WRA), the newly elected Liberal–National government sought to create a pure non-union agreementmaking option. Under the WRA, there are two streams of direct and non-union voice structures, in addition to the availability of union voice mechanisms. Direct and nonunion representative structures are provided for in the form of (i) non-union certified agreements between an employer and a valid majority of employees, and (ii) Australian Workplace Agreements, which constitute individual agreements between an employer and an employee. Informal initiatives promoting direct employee voice have also been common in Australia since the mid-1970s. This decade witnessed the introduction of a wave of modest participative practices such as job redesign and consultation (Teicher, 1992). Interest in employee participation grew in the 1980s and was paralleled by the growth of the human resource management practices (Sheehan et al., 2002). In the 1980s, following the negotiation of the Accord Agreements between the Labor Party government and the Australian Council of Trade Unions (ACTU), participation assumed a collectivist orientation, characterised by an emphasis on labour–management © 2006 The Author(s) Journal compilation © Blackwell Publishing Ltd. 2006

Employee voice arrangements in Australia

547

cooperation and a role for unions in decision making at the workplace (Bray et al., 2001). Productivity bargaining was also pervasive at this time, emphasising decentralised bargaining and lessened third-party intervention. The ACTU championed industrial democracy during the 1980s, while non-union strategies gained momentum in the private sector as a means for increased competitiveness (Campling and Gollan, 1999). Mutual support for increased employee involvement and direct relations was evidenced in a joint statement between the ACTU and employer associations, supporting voluntary employee participation schemes at the workplace (Lansbury, 1980). However, because these schemes were voluntary, they were rarely introduced and remained management initiated. Wright (1995) has argued that ‘constitutionalism’ was widespread among employers at this time, characterised by a refusal to consult with unions so as to retain managerial prerogative. The reluctance of employers to consult with unions was compounded by the inadequacy of union workplace structures; a historic weakness of Australian unions (Campling and Gollan, 1999; Rimmer and Verevis, 1990). The Termination Change and Redundancy Case in 1984 was, however, an important turning point in reducing managerial discretion, allowing employees greater involvement in a wider range of workplace issues including the introduction of technological change (Ludeke, 1991). The test case decision was progressively incorporated into awards regulating the employment of two-thirds of the Australian workforce, resulting in unions gaining an unprecedented right to be consulted by employers. Since the mid-1990s, employee voice has assumed a more individualistic character, fuelled by an increasingly aggressive focus on managerial prerogative, the decline of unionism and the primacy of regulation of the employment relationship at the enterprise level. There was consensus among employers, unions and the government during this time that the introduction of enterprise bargaining would enhance employee voice in the determination of terms and conditions of employment and workplace decision making, by lowering the locus of employment regulation (Bray et al., 2001). However, the legislative amendments contained in the WRA in 1996 were radical, placing less stringent requirements on employers to negotiate with unions. For example, good-faith bargaining requirements were removed from the legislation, and individual forms of agreement making were elevated over and above collective forms, as the primary instrument of employment regulation. These changes have resulted in a lack of pluralistic decision making and negotiation (Deery and Walsh, 1999). The rise in alternative forms of voice in Australia and the emphasis on direct participative and representative arrangements at the policy level mirrors international developments. In the UK in particular, newly established European directives upholding information and consultation rights for employees have highlighted the importance of non-union employee representation (Gollan, 2005). The debates in the academic literature across Australia, the UK and the US also follow similar lines, focusing on the issue of substitution versus complementarity, and the extent to which management strategies for using direct voice arrangements are aimed at union avoidance, or the building of organisational commitment through legitimate and effective participative decision making and consultation procedures. This article contributes to the debate on employee voice by comparing the effectiveness of different voice arrangements in Australia on three outcomes: managerial responsiveness to employee needs, level of job control and influence over job rewards. © 2006 The Author(s) Journal compilation © Blackwell Publishing Ltd. 2006

548

Amanda Pyman et al.

METHOD Sample and procedure The data reported in this article were drawn from responses to the Australian Worker Representation and Participation Survey (AWRPS, 2004), which investigated workers’ responses and attitudes to workplace participation and representation, unions and union–management relations. The AWRPS instrument was based on the 1994–95 Worker Representation and Participation Survey conducted in the US (Freeman and Rogers, 1999), the British Worker Representation and Participation Survey (2001) (Diamond and Freeman, 2002), and the New Zealand Worker Representation and Participation Survey (2003) (Haynes et al., 2003). The Australian survey questions were closely modelled on the New Zealand survey, but were adapted to conform to the institutional and demographic contexts in Australia. One thousand employees were surveyed nationally using computer-assisted telephone interviewing techniques. Potential respondents were selected randomly from the residential telephone directory and contacted between October 2003 and March 2004. The interviews were conducted in the evenings between 5.30 PM and 8.30 PM. The sample was limited to Australian residents in paid employment of more than 10 hours per week who had left secondary school. Self-employed persons and company owners were excluded from the survey. Mobile, business and commercial telephone numbers were excluded from the sample. The sample was stratified by Australian State/Territory to reflect the geographical distribution of the population as reported in the Australian Bureau of Statistics Census of Population and Housing 2001. Fifty-nine per cent of the respondents were female, and the mean age of the sample was 41.5 years (SD = 11.41). The mean number of hours worked per week was 36.50 (SD = 12.19), with 67 per cent of the respondents working full-time (defined as 35 hours or more per week). The majority of the respondents (80.8 per cent) were non-manual workers and were born in Australia (77.9 per cent). Just under half of the respondents (45.5 per cent) reported that they worked in organisations with 500 or more employees. The mean number of years worked for their current employer was 8.46 (SD = 8.02). Although our sample overestimates women and non-manual workers, particularly professionals (census figures indicate 45 per cent of Australians are female and 70 per cent work in non-manual occupations), it is broadly representative of the Australian population in terms of demographic characteristics. However, there is overrepresentation in some industries, with a large concentration of our sample in the health and education sectors (33.8 per cent compared with 17.2 per cent using census data). Because of the skew across industries, where appropriate, we weighted the sample using the inverse probability of selection by industry (using ABS, 2001 Census data) to compensate for sample non-response bias.

Measures Measures of managerial responsiveness to employee needs Employees were asked to rate the performance of the management at their workplace on a five-point scale ranging from 1 = ‘failure’ to 5 = ‘excellent’ on six dimensions: (i) concern for employees; (ii) giving fair pay increases; (iii) willingness to share power © 2006 The Author(s) Journal compilation © Blackwell Publishing Ltd. 2006

Employee voice arrangements in Australia

549

and authority; (iv) keeping everyone up to date with proposed changes; (v) promoting equal opportunities for all employees; and (vi) 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). Principal components analysis of the six items indicated a single factor, thus providing support for the unidimensionality of the scale. Scores on this measure were internally consistent with a Cronbach’s alpha coefficient of 0.91. The scores on the six items were averaged to form a composite measure (possible range of scores is 1–5), with higher scores indicating greater managerial responsiveness to employee needs. The mean score on our measure of managerial responsiveness to employee needs was 3.55 (SD = 0.83). Perceived job control and influence over rewards Employees were asked to rate their actual level of influence and involvement over selected aspects of their employment on a four-point scale ranging from 0 = ‘none’ to 3 = ‘a lot’ across six items: (i) deciding how to do your job and organise your work; (ii) setting working hours; (iii) deciding how much of a pay rise people in your workgroup/department should get; (iv) the pace of work; (v) deciding how to work with new equipment or software; and (vi) deciding what kinds of perks/bonuses are offered to people. Based on principal components analysis of the six items, two composite measures were constructed: (i) four items measuring job control (items i, ii, iv, v; alpha = 0.70), and (ii) two items measuring influence over rewards (items iii, vi; alpha = 0.68). Although the latter reliability estimate is less than Nunally’s (1978) recommended 0.70 cut-off, internal consistency coefficients greater than 0.60 are considered acceptable for newly developed measures (Hair et al., 1998). Higher scores on both measures indicate greater reported job control and greater influence over rewards (possible range of scores is 0–3). The mean score on our measure of job control was 2.02 (SD = 0.77). The mean score on influence over rewards was 0.44 (SD = 0.73). Voice measures The voice measures used in this study are similar to those outlined by Bryson (2004: 239–240). Union voice was operationalised as collective representation and managerial recognition, referring to the presence of a union of any kind at the workplace. The limitations of this instrument prevent measurement of the extent and strength of union activity, for instance, if there was a union delegate or union committee on-site and the level of union–management interaction. Non-union representative voice is present where a joint consultative committee exists and the representatives are not chosen through union channels. Direct voice is present where at least one of the following mechanisms exists: (i) regular meetings with all staff; (ii) a committee of employees that discusses problem management on a regular basis; or (iii) a formal employee involvement programme such as quality circles. Other measures of direct voice such as the presence of an ‘open-door policy’ were excluded from the measure, as this was identified by an overwhelming majority of the sample as universal practice. Control variables Consistent with the recommendations of Bryson (2004), we controlled for a range of individual, job and workplace characteristics in the regression analyses. The analyses controlled for gender, age, occupation, country of birth, hours worked each week, industry, gross weekly wage, organisational tenure, individual union membership and © 2006 The Author(s) Journal compilation © Blackwell Publishing Ltd. 2006

550

Amanda Pyman et al.

organisational size.1 Following Bryson’s (2004) argument, individual union membership and organisational tenure may be endogenous with respect to managerial responsiveness, job control and influence over job rewards. To address this issue, we compared models including and excluding individual union membership and tenure. The results were almost identical. Our rationale for including these two variables is to obtain estimates of voice effects, after the effects of union membership and job tenure have been controlled for. Method of analysis As the three dependent measures were continuous in nature, multiple linear regression analysis was used. The relationships were tested using multiple regressions to control for variables simultaneously, and to take into account the interrelationships among the independent variables. We conducted three separate multiple linear regression analyses using the measures of managerial responsiveness to employee needs, job control and influence over job rewards as the dependent variables. For ease of interpretation, these analyses utilise unweighted data (the parameter estimates from weighted analyses were almost identical). The eight combinations of voice mechanisms were dummy coded (1 = present, 0 = not present) and compared with the reference category of voicelessness (no voice). In addition to the eight measures of voice, the control variables identified earlier were included. Following Bryson (2004), we included managerial employees in our analyses, as they are also likely to be managed by other senior people in an organisation. Collinearity diagnostics using the tolerance test indicated that the regression estimates were not affected by multicollinearity. Because there were missing data, the regression models were estimated using the full-information maximum likelihood (FIML) procedure in LISREL Version 8.7 (Jöreskog and Sörbom, 2002). FIML generally outperforms more common methods of handling missing data, yielding parameter estimates with less bias in large samples (Enders, 2001; Schafer and Graham, 2002). There were no consequential departures from normality, linearity and homoscedasticity for the measures of managerial responsiveness to employee needs and job control. Scores on the measure of influence over job rewards were positively skewed. A logarithmic transformation was performed, but did not improve the results over the untransformed data reported here. As the voice arrangements were all measured on the same (dummy coded) scale, unstandardised partial regression coefficients are reported as measures of effect size. Finally, we report 95 per cent normal theory confidence intervals (CIs) to assess the precision of our estimates. Note that if a 95 per cent CI excludes zero, then the estimate is statistically significant at the 0.05 level.

RESULTS Table 1 summarises the different voice arrangements within our sample. We report both unweighted and weighted estimates. As Table 1 shows, the majority of employees were located in workplaces where a plurality of voice mechanisms was present (29.6 per cent unweighted, 23 per cent weighted). Direct voice arrangements were also common, as a single channel and in combination with other forms. Interestingly, only 1

The organisational size variable was inverse-transformed due to severe positive skewness.

© 2006 The Author(s) Journal compilation © Blackwell Publishing Ltd. 2006

Employee voice arrangements in Australia

551

Table 1: Summary of employee voice arrangements Voice arrangements Union-only Union and direct Union and non-union representative Union, direct and non-union representative Direct-only Direct and non-union representative Non-union representative-only No voice Not known

n

Unweighted (%)

51 138 25 296 128 100 15 90 157

5.1 13.8 2.5 29.6 12.8 10.0 1.5 9.0 15.7

Weighted (%) 4.8 11.9 2.6 23.0 16.1 12.1 1.9 12.4 15.3

5.1 per cent of respondents (4.8 per cent weighted) reported their workplace to have union-only voice. Voicelessness (no voice) was present in 9 per cent of cases (12.4 per cent weighted), and about half of these employees were found in organisations employing fewer than 20 employees (46.1 per cent unweighted, 50.3 per cent weighted). It should also be noted that in 15.7 per cent of cases, no accurate voice mechanism could be identified because of missing data on at least one or more of the measures of employee voice. Table 2 presents the results of the multiple regression analyses2 against the reference group of no voice. The regression analyses showed that the strongest predictor of managerial responsiveness to employee needs was the combination of direct and nonunion representative voice arrangements (Β = 0.94, 95 per cent CI: 0.72–1.16). The combination of union, direct and non-union representative voice arrangements was also positively related to managerial responsiveness to employee needs (Β = 0.80, 95 per cent CI: 0.58–1.02), as was the combination of union and direct voice (Β = 0.62, 95 per cent CI: 0.39–0.86), and direct voice on its own (Β = 0.58, 95 per cent CI: 0.36– 0.80). Further analysis focusing solely on direct voice arrangements found that regular staff meetings (Β = 0.44, 95 per cent CI: 0.32–0.56), problem-solving groups (Β = 0.28, 95 per cent CI: 0.18–0.38) and a formal employee involvement programme (Β = 0.21, 95 per cent CI: 0.11–0.31) were associated with greater managerial responsiveness to employee needs. Union-only voice was not a statistically significant predictor of managerial responsiveness (B = 0.21, 95 per cent CI: −0.06–0.48), and the CI for non-union representative-only voice (although it did not contain zero) was wide (Β = 0.49, 95 per cent CI: 0.08–0.90), meaning that we cannot place any great weight on these observed effects. It is interesting to note that of the control variables, those respondents working in smaller organisations (Β = 1.87, 95 per cent CI: 1.24–2.50) were more likely to report greater managerial responsiveness to employee needs, whereas union members were more likely than non-union members to report lesser managerial responsiveness to employee needs (Β = −0.15, 95 per cent CI: −0.29 to −0.01). The strongest predictor of perceived job control was the combination of direct and non-union representative voice (Β = 0.59, 95 per cent CI: 0.39–0.79). Direct-only voice 2

For reasons of brevity, we do not display all the control variables. Full versions of the models, including the control variables and input covariance matrices, are available from the authors upon request. © 2006 The Author(s) Journal compilation © Blackwell Publishing Ltd. 2006

© 2006 The Author(s) Journal compilation © Blackwell Publishing Ltd. 2006

0.21 0.62 −0.05 0.80 0.58 0.94 0.49 0.45 0.20

0.14 0.12 0.18 0.11 0.11 0.11 0.21

SE (−0.06,0.48) (0.39,0.86) (−0.40,0.30) (0.58,1.02) (0.36,0.80) (0.72,1.16) (0.08,0.90)

CI 0.11 0.33 −0.32 0.41 0.42 0.59 0.44 0.47 0.22

Β 0.13 0.11 0.16 0.10 0.10 0.10 0.20

SE

CI (−0.15,0.37) (0.11,0.55) (−0.63,−0.01) (0.21,0.61) (0.22,0.62) (0.39,0.79) (0.05,0.83)

Job control

0.17 0.30 0.14 0.38 0.17 0.53 0.22 0.35 0.12

Β

0.13 0.11 0.17 0.10 0.10 0.10 0.20

SE

(−0.09,0.43) (0.08,0.52) (−0.19,0.47) (0.18,0.58) (−0.03,0.37) (0.33,0.73) (−0.17,0.61)

CI

Influence over rewards

Β, unstandardised regression coefficient; CI, lower and upper limits of the 95 per cent confidence interval for regression coefficient; SE, standard error. N = 1,000.

Union-only Union and direct Union and non-union representative Union, direct and non-union representative Direct-only Direct and non-union representative Non-union representative-only R R2

Β

Managerial responsiveness

Table 2: Multiple regression analyses of independent variables: Managerial responsiveness, job control and influence over rewards

552 Amanda Pyman et al.

Employee voice arrangements in Australia

553

(Β = 0.42, 95 per cent CI: 0.22–0.62) and the combination of union, direct and nonunion representative voice (Β = 0.41, 95 per cent CI: 0.21–0.61) also had positive effects on job control. Further analysis focusing solely on direct voice arrangements revealed that regular staff meetings (Β = 0.27, 95 per cent CI: 0.17–0.37) and formal employee involvement programmes (Β = 0.16, 95 per cent CI: 0.06–0.26) were associated with greater job control. Union-only voice was not a statistically significant predictor of job control (Β = 0.11, 95 per cent CI: −0.15–0.37). Although non-union representative voice and the combination of union and non-union representative voice were statistically significant predictors, the latter being negatively related to job control, the respective CIs are sufficiently broad that we cannot place great weight on the observed effects. It is noteworthy, yet not surprising, that managerial workers (Β = 0.23, 95 per cent CI: 0.05–0.41) were more likely to report greater job control, whereas those working in routine clerical (Β = −0.34, 95 per cent CI: −0.56 to −0.12) and intermediate production/transport occupations (Β = −0.47, 95 per cent CI: −0.76 to −0.18) were less likely to perceive control over their jobs. Three voice measures were positively related to influence over job rewards: the combinations of direct and non-union representative voice (Β = 0.53, 95 per cent CI: 0.33–0.73), direct, union and non-union representative voice (Β = 0.38, 95 per cent CI: 0.18–0.58), and union and direct voice (Β = 0.30, 95 per cent CI: 0.08–0.52). In contrast to the previous two models, direct-only voice did not have a statistically significant relationship with perceived influence over job rewards. However, further analysis of direct voice arrangements showed that regular staff meetings (Β = 0.20, 95 per cent CI: 0.10–0.30) and problem-solving groups (Β = 0.11, 95 per cent CI: 0.01– 0.21) were associated with greater perceived employee influence over rewards. Controlling for a range of organisational, job and demographic characteristics, union-only voice and the combination of union and non-union representative voice were not statistically significant predictors of influence over job rewards. It is notable that managerial workers (Β = 0.26, 95 per cent CI: 0.08–0.44) and employees in smaller organisations (Β = 0.83, 95 per cent CI: 0.26–1.40) were more likely to report greater influence over job rewards.

DISCUSSION Our results indicate that employees perceived a plurality of voice arrangements as most effective in eliciting managerial responsiveness, having job control and an influence over job rewards. Of the eight different voice arrangements examined, the results showed that the combinations of direct and non-union representative voice, and union, direct and non-union representative voice were most predictive of the three outcomes. Nevertheless, the combination of direct and non-union representative voice was not necessarily superior to the combination of union, direct and non-union representative voice, as evidenced by the overlapping CIs of the estimates. The effectiveness of direct and non-union representative voice in predicting outcomes is consistent with Bryson’s (2004) findings that this combination outperformed all other configurations. Although direct-only voice did affect managerial responsiveness, and job control and influence over job rewards in isolation, stronger effects were generally observed when this form of voice acted in concert with other voice mechanisms. Nonetheless, direct employee voice is the common channel that underpins the two most effective configurations of voice, suggesting direct employee involvement and © 2006 The Author(s) Journal compilation © Blackwell Publishing Ltd. 2006

554

Amanda Pyman et al.

participation is a positive predictor of outcomes. Of the different types of direct voice arrangements, the use of regular staff meetings was found to be most predictive of managerial responsiveness, job control and influence over job rewards. The effectiveness of multiple channels of voice rather than a single channel in Australia is consistent with other empirical studies in Europe, the UK and New Zealand (e.g. Bryson, 2004; Haynes et al., 2005; Machin and Wood, 2005; Sako, 1998; Sisson, 2000; Wilkinson et al., 2004). Sako’s (1998) research on the European car industry, for instance, revealed that a combination of direct and indirect forms of voice was linked to better performance and a greater willingness among employees to participate and contribute to organisational decision making. Similarly, Sisson (2000) found that in cases where a greater number of different participative forms were used in the UK, managers were more likely to report benefits. Also in the UK, Wilkinson et al. (2004) found that union and non-union channels coalesced in a systematic alliance so as to complement one another. This is supported by Machin and Wood’s (2005) analysis of the British WERS, which found that there is evidence of a complementarity between union and human resource management practices. The evidence of a complementarity between different voice arrangements challenges the widely held view that direct forms of voice act as a substitute for union voice (e.g. Dundon, 2002; Gollan, 2005). Collectively, the results reported above also reaffirm the value of building and developing a complementary presence of alternative forms of employee representation and involvement in the workplace. Nevertheless, it is important to reiterate that a simple model of choice of consultative and representative forms may not be straightforward, for either employers or employees, because there are many external factors that will impinge on this decision (Marchington et al., 2001). The relative ineffectiveness of union-only voice as demonstrated by the finding of a negligible observed effect on the three outcome measures is noteworthy. However, we cannot be confident of the direction of the association. In part, the absence of a statistically significant relationship between union voice and the outcome measures may be due to the relatively small number of employees (less than 6 per cent) who identified the presence of union-only voice arrangements (see Table 1). Nevertheless, from a union perspective, the ineffectiveness of union voice in isolation challenges the very logic of collective organisation. Our results indicate that union members had poorer perceptions of managerial responsiveness to employee needs than did nonunionists. This effect may of course be attributable to ‘voice induced complaining’ (Bryson, 2004; Freeman and Medoff, 1984), that is, the notion that union members hold more critical attitudes of management because they have a greater awareness of labour–management interaction and a greater willingness to speak out. Non-unionists may simply be unaware of what they are missing out on in the employment relationship. Unfortunately, the present sample was not large enough to conduct separate analyses of the effectiveness of employee voice arrangements for union members and non-unionists. The limitations of the measure of union voice must also be reiterated, because it precludes an in-depth analysis of the strength and impact of union voice regimes such as union density and the presence of union delegates on the shopfloor. The findings have numerous implications. First, they reveal that contrary to the substitution argument, union, non-union representative and direct voice are not mutually exclusive. This finding resonates with the ‘copartnership’ model which advocates a shared agenda, emphasising the compatibility of high involvement human resource management practices and union voice (Boxall and Purcell, 2003). Indeed, a preference for cooperative union–management relations was also identified in the © 2006 The Author(s) Journal compilation © Blackwell Publishing Ltd. 2006

Employee voice arrangements in Australia

555

AWRPS among a majority of both union and non-union members (Teicher et al., 2005). In the Australian context, the operation and interaction of multiple voice channels transcends the unitarism/pluralism dichotomy and underscores the rationale for a plurality of interests and cooperative relations in the employment relationship. For unions and employers, the findings suggest the desirability of an incorporated, mutually reinforcing approach to employee voice which is predicated on multiple and inclusive channels, collaboration and cooperation. This model resonates with integrative approaches to human resource sustainability and a ‘soft’ model of human resource management that is underpinned by workforce development and cooperation with unions (Fiorito, 2001; Gollan, 2000b). As Guest (1995: 136) has suggested, unions should be seen as being compatible with principled advocacy of human resource management. From a theoretical stance, and consistent with the arguments of Benson (2000), the duality of voice arrangements in Australia suggests a rethinking of simple dichotomies such as human resource management versus industrial relations, and individualism versus collectivism. The critical issue appears to be a configuration of multiple channels of voice rather than a single channel, and it is important to analyse how and why the different voice channels complement each other, under what conditions multiple arrangements flourish and the structural underpinnings of the operation of multiple voice channels. The research findings are also consistent with the studies of dual commitment, whereby employees have been found to exhibit commitment to both the organisation and collective representation by a union (Deery et al., 1999; Guest, 1995). The results raise several avenues for further research in light of the limitations of this study. First, the data are self-reports, and we acknowledge that the results may have been influenced by common method variance (Podsakoff et al., 2003). The inclusion and analysis of objective data, such as enterprise agreements, minutes of staff meetings and employee committees and human resource management policies, in addition to the use of a variety of data collection methods, is warranted in a more extensive analysis of the efficacy of different voice arrangements. Certainly, qualitative data analysis drawing on case data at the workplace level would be necessary to investigate the argument of Dundon et al. (2004) that it is the extent to which voice is embedded that is fundamental. It is imperative to analyse both the structural aspects of voice (e.g. the frequency of joint consultative committee and staff meetings) and the cultural practices that underpin the operation of these arrangements in the workplace. The latter issue is indeed critical to furthering our understanding of the relationship between different types of employee voice arrangements and organisational outcomes. The second limitation is that inferences concerning the relationship between employee voice arrangements and organisational outcomes must be made with caution, particularly because the data are cross-sectional and the sample sizes for some voice measures were small, thus limiting statistical power and generalisability. The use of a longitudinal design can help to strengthen causal inferences, while larger sample sizes within each of the voice categories would have also strengthened the study. Third, further research on the efficacy and interaction of employee voice arrangements must take greater account of contextual factors, including industry, technology, workforce demographics and geographical location. Employee self-reports obtained from our large-scale survey do not permit a rich and detailed analysis of such factors; yet, these factors may have an appreciable impact on the effectiveness of employee voice arrangements. A related issue that warrants further qualitative research on employee voice in © 2006 The Author(s) Journal compilation © Blackwell Publishing Ltd. 2006

556

Amanda Pyman et al.

Australia is the extent to which the strategies and motivations of the primary actors shape the various channels of voice and how they are embedded in organisations. For instance, managerial attitudes and practices toward unionisation are likely to have a direct influence on the form and effectiveness of union voice at the workplace level, and are important considerations in subsequent analyses.

CONCLUSIONS Employee voice has attracted increasing attention in the industrial relations and human resource management literature, yet the incidence and effectiveness of different voice types have not been well analysed, particularly in Australia. The effectiveness of different voice arrangements is particularly critical in Australia, given the revolutionary changes recently enacted by the Howard Liberal–National government’s Workplace Relations Amendment (Work Choices) Act 2005 (Cth.), which aims to completely deregulate the Australian system, at least in terms of the forms of employee representation, involvement and participation. The efficacy of employee voice arrangements does, however, also have international significance, as increasing numbers of western nations adopt neo-liberal policies emphasising direct employee involvement, participation and representation. The clear message from this research is that an amalgam of employee voice regimes is most effective in predicting organisational outcomes, namely, managerial responsiveness to employee needs, perceived job control and influence over job rewards. While direct-only voice is the common underpinning of the efficacy of the different voice combinations, stronger effects were observed when this voice form interacted with union and non-union representative channels. Further analysis of what drives the observed efficacy of direct-only voice revealed that regular staff meetings were most strongly linked to the prediction of organisational outcomes. These results confirm that it is the interaction and coexistence of multiple channels of voice that are most effective and legitimate, in employees’ minds, in achieving organisational outcomes. In turn, the findings challenge common assumptions that non-union representative voice and direct employee involvement act as substitutes for union voice and that non-union employee representation structures are ineffective in providing a true voice for employees because they lack independence, seek to institutionalise worker cooperation and therefore avoid union representation and action (e.g. Gollan, 2005; Kelly, 1996). In fact, our results resonate with Hyman’s (1997) argument that direct structures can give greater legitimacy and efficacy to the decision-making process, ensure greater organisational commitment and act as a complement to existing union structures. It could be argued that the results require a rethinking of the substitution thesis and the models that underpin employer and employee choice of voice arrangements, because the presence of non-union representative or direct voice per se does not automatically translate to the replacement of union voice channels. In a practical sense, the effectiveness of employee voice in Australian workplaces is dependent on a plurality of arrangements, that is, multiple, mutually reinforcing channels. While achieving this is likely to present challenges for employers, employees and unions, the findings do highlight the significance of labour–management interaction and thus the value of a union–employer partnership approach predicated on mutual benefits. Such an approach is seriously threatened by the recent changes embodied in the Work © 2006 The Author(s) Journal compilation © Blackwell Publishing Ltd. 2006

Employee voice arrangements in Australia

557

Choices Act 2005. This legislation unequivocally privileges direct or non-union representative voice mechanisms, while undermining union voice and multi-channel voice arrangements. References ABS (2001), Census of Population and Housing (Canberra, Australian Government Publishing Service). ABS (2004), Employee Earnings, Benefits and Trade Union Membership. August 2003, Cat. No. 6310.0 (Canberra, Australian Government Publishing Service). Batt, R., A. J. S. Colvin and J. Keefe (2002), ‘Employee Voice, Human Resource Practices and Quit Rates: Evidence from the Telecommunications Industry’, Industrial and Labor Relations Review, 55, 4, 573–594. Benson, J. (2000), ‘Employee Voice in Union and Non-union Australian Workplaces’, British Journal of Industrial Relations, 38, 3, 453–459. Bishop, L. and D. I. Levine (1999), ‘Computer Mediated Communication as Employee Voice: A Case Study’, Industrial and Labor Relations Review, 52, 2, 213–233. Boxall, P. and J. Purcell (2003), Strategy and Human Resource Management (Basingstoke, Palgrave Macmillan). Bray, M., P. Waring, D. Macdonald and S. Le Queux (2001), ‘The “Representation Gap” in Australia’, Labour and Industry, 12, 2, 1–31. Bryson, A. (2004), ‘Managerial Responsiveness to Union and Nonunion Worker Voice in Britain’, Industrial Relations, 43, 1, 213–241. Budd, J. (2004), Employment with a Human Face: Balancing Efficiency, Equity and Voice (Ithaca, NY, Cornell University Press). Campling, J. and P. Gollan (1999), Bargained Out: Negotiating without Unions in Australia (Sydney, The Federation Press). Deery, S. and J. Walsh (1999), ‘The Character of Individualised Employment Relations in Australia: A Model of “Hard” HRM’, in S. Deery and R. Mitchell (eds), Employment Relations: Individualisation and Union Exclusion (Sydney, The Federation Press) pp. 115–128. Deery, S., P. Erwin and R. Iverson (1999), ‘Industrial Relations Climate, Attendance Behaviour and the Role of Trade Unions’, British Journal of Industrial Relations, 37, 4, 533–558. Delery, J. E., N. Gupta, J. D. Shaw, G. Jenkins and M. L. Ganster (2000), ‘Unionization, Compensation and Voice Effects on Quits and Retention’, Industrial Relations, 39, 4, 625–645. Diamond, W. and R. Freeman (2002), What Workers Want from Workplace Organizations: A Report to TUC’s Promoting Trade Unionism Task Group (London, Trades Union Congress). Dundon, T. (2002), ‘Employer Opposition and Union Avoidance in the UK’, Industrial Relations Journal, 33, 3, 234–245. Dundon, T., A. Wilkinson, M. Marchington and P. Ackers (2004), ‘The Meanings and Purpose of Employee Voice’, International Journal of Human Resource Management, 15, 6, 1149–1170. Enders, C. K. (2001), ‘The Performance of the Full Information Maximum Likelihood Estimator in Multiple Regression Models with Missing Data’, Educational and Psychological Measurement, 61, 5, 713–740. Fiorito, J. (2001), ‘Human Resource Management Practices and Worker Desires for Union Representation’, Journal of Labor Research, 22, 2, 335–354. Freeman, R. (1980), ‘The Exit-voice Tradeoff in the Labor Market: Unionism, Job Tenure, Quits and Separations’, Quarterly Journal of Economics, 94, 4, 643–673. Freeman, R. and J. Medoff (1984), What Do Unions Do? (New York, Basic Books). Freeman, R. and J. Rogers (1999), What Workers Want (Ithaca, NY, ILR Press). Gaffney, F. (2002), ‘Reconstituting the Collective? Non-union Employee Representative Structures in the United Kingdom and Australia’, in P. J. Gollan, R. Markey and I. Ross (eds), Works Councils in Australia: Future Prospects and Possibilities (Sydney, The Federation Press) pp. 149–178. © 2006 The Author(s) Journal compilation © Blackwell Publishing Ltd. 2006

558

Amanda Pyman et al.

Gollan, P. (2000a), ‘Non-union Forms of Employee Representation in the United Kingdom and Australia’, in B. Kaufman and D. Taras (eds), Nonunion Employee Representation: History, Contemporary Practice and Policy (New York, M. E. Sharpe) pp. 410–449. Gollan, P. (2000b), ‘Human Resources, Capabilities and Sustainability’, in D. Dunphy, J. Benveniste, A. Griffiths and P. Sutton (eds), Sustainability—The Corporate Challenge of the 21st Century (St Leonards, NSW, Allen & Unwin) pp. 55–77. Gollan, P. (2001), ‘Be Aware of the Voices’, People Management, 7, 6, 52–54. Gollan, P. (2005), ‘Silent Voices: Representation at the Eurotunnel Call Centre’, Personnel Review, 34, 4, 423–450. Guest, D. (1995), ‘Human Resource Management, Trade Unions and Industrial Relations’, in J. Storey (ed.), Human Resource Management: A Critical Text (London, Routledge) pp. 110– 141. Hair, J. F., R. E. Anderson, R. L. Tatham and W. C. Black (1998), Multivariate Data Analysis, 5th edn (Upper Saddle River, NJ, Prentice Hall). Haynes, P., P. Boxall and K. Macky (2003), New Zealanders’ Influence at Work—Report to the Minister of Labour on the New Zealand Worker Representation and Participation Survey, University of Auckland Business School, Auckland (May). Haynes, P., P. Boxall and K. Macky (2005), ‘Non-union Voice and the Effectiveness of Joint Consultation in New Zealand’, Economic and Industrial Democracy, 26, 2, 229–256. Hyman, R. (1997), ‘The Future of Employee Representation’, British Journal of Industrial Relations, 35, 3, 309–336. Iverson, R. D. and D. M. Buttigieg (1995), ‘Union Commitment: Implications for Union Participation and Exit-voice’, Labour and Industry, 6, 3, 95–111. Iverson, R. D. and D. B. Currivan (2003), ‘Union Participation, Job Satisfaction and Employee Turnover: An Event-history Analysis of the Exit-voice Hypothesis’, Industrial Relations, 42, 1, 101–105. Jöreskog, K. and D. Sörbom (2002), LISREL (Version 8.7) [Computer Software] (Chicago, IL, Scientific Software International). Kaufman, B. and D. Taras (2000), Nonunion Employee Representation: History, Contemporary Practice and Policy (New York, M. E. Sharpe). Kelly, J. (1996), ‘Works Councils: Union Advance or Marginalisation?’, in A. McColgan (ed.), The Future of Labour Law (London, Mansell) pp. 44–62. Kim, D. and H. Kim (2004), ‘A Comparison of the Effectiveness of Unions and Non-union Works Councils in Korea: Can Non-union Employee Representation Substitute for Trade Unionism?’, International Journal of Human Resource Management, 15, 6, 1069–1093. Kleiner, M. M. (2001), ‘Intensity of Management Resistance: Understanding the Decline of Unionization in the Private Sector’, Journal of Labor Research, 22, 3, 519–540. Lansbury, R. (1980), Democracy in the Workplace (Melbourne, Longman Cheshire). Ludeke, T. (1991), ‘What Ever Happened to the Prerogatives of Management?’, Journal of Industrial Relations, 33, 3, 395–411. McCabe, D. and D. Lewin (1992), ‘Employee Voice: A Human Resource Management Perspective’, California Management Review, 34, 3, 95–111. Machin, S. and S. Wood (2005), ‘Human Resource Management as a Substitute for Trade Unions in British Workplaces’, Industrial and Labor Relations Review, 58, 2, 201–218. Marchington, M., A. Wilkinson, P. Ackers and T. Dundon (2001), Management Choice and Employee Voice, Chartered Institute of Personnel and Development Research Report (London, Chartered Institute of Personnel and Development). Miller, P. and C. Mulvey (1991), ‘Australian Evidence on the Exit/Voice Model of the Labor Market’, Industrial and Labor Relations Review, 45, 10, 44–57. Millward, N., A. Bryson and J. Forth (2000), All Change at Work: British Employment Relations 1980–98 as Portrayed by the Workplace Industrial Relations Survey Series (London, Routledge). Nunally, J. (1978), Psychometric Theory, 2nd edn (New York, McGraw-Hill). © 2006 The Author(s) Journal compilation © Blackwell Publishing Ltd. 2006

Employee voice arrangements in Australia

559

Podsakoff, P., S. MacKenzie, J. Y. Lee and N. Podsakoff (2003), ‘Common Method Biases in Behavioural Research: A Critical Review of the Literature and Recommended Remedies’, Journal of Applied Psychology, 88, 5, 879–903. Rimmer, M. and C. Verevis (1990), Award Restructuring: Progress at the Workplace, IRRC Monograph No. 28 (Sydney, University of New South Wales). Sako, M. (1998), ‘The Nature and Impact of Employee “Voice” in the European Car Components Industry’, Human Resource Management Journal, 8, 2, 5–13. Schafer, J. L. and J. W. Graham (2002), ‘Missing Data: Our View of the State of the Art’, Psychological Methods, 7, 2, 147–177. Sheehan, C., P. Holland and P. J. Dowling (2002), ‘Human Resource Management and Employee Relations’, in J. Teicher, P. Holland and R. Gough (eds), Employment Relations Management: Australia in a Global Context (Forest, NSW, Pearson Education) pp. 88–109. Sisson, K. (2000), Direct Participation and the Modernisation of Work Organisation (Dublin, European Foundation for the Improvement of Living and Working Conditions). Storey, J. (1992), Developments in the Management of Human Resources (Oxford, Blackwell). Storey, J. (2001), Human Resource Management: A Critical Text, 2nd edn (London, Thomson Learning). Teicher, J. (1992), ‘Theories of Employee Participation and Industrial Democracy: Towards an Analytical Framework’, in B. Dabscheck, G. Griffin and J. Teicher (eds), Contemporary Australian Industrial Relations Readings (Melbourne, Longman Cheshire) pp. 476–494. Teicher, J., P. Holland, A. Pyman and B. Cooper (2005), ‘Australian Workers: Finding Their Voice?’, in P. Boxall, R. Freeman and P. Haynes (eds), Voice and Voicelessness in the AngloAmerican World: Continuity and Transformation in Employee Representation (Ithaca, NY, Cornell University Press). Terry, M. (1999), ‘Systems of Collective Employee Representation in Non-union Firms in the UK’, Industrial Relations Journal, 30, 1, 16–30. Towers, B. (1997), The Representation Gap: Change and Reform in the British and American Workplace (Oxford, Oxford University Press). Wilkinson, A., T. Dundon, M. Marchington and P. Ackers (2004), ‘Changing Patterns of Employee Voice: Case Studies from the UK and Republic of Ireland’, Journal of Industrial Relations, 46, 3, 298–322. Willman, P., A. Bryson and R. Gomez (2003), ‘Why Do Voice Regimes Differ?’, Centre for Economic Performance, Discussion Paper no. 591, London School of Economics (November). Wright, C. (1995), The Management of Labour: A History of Australian Employers (Oxford, Oxford University Press).

© 2006 The Author(s) Journal compilation © Blackwell Publishing Ltd. 2006