Union representation,collective voice and job quality: An analysis of a survey of union members in the UK finance sector

Abstract

This article seeks to identify whether employee perceptions of job quality are better in instances where an onsite union representative is present. It also seeks to identify whether the relationship between onsite representative presence and job quality is explained by employee perceptions of union collective voice. The analysis, based on a survey of union members in the UK finance sector, demonstrates that employee perceptions of several dimensions of job quality are better where an onsite representative is present, and that this can be explained by the higher perceptions of union collective voice that onsite representatives engender.

Keywords

Collective voice finance sector job quality trade unions union representative

Introduction

Research on the activities of shop stewards and onsite union representatives has been central to the study of workplace union organisation for many years (Charlwood and Terry, 2007). Past studies have focused on the benefits that such representatives generate for employees in terms of improvements to terms and conditions of employment, equality of outcomes and disciplinary procedures (Waddington, 2013). There has, however, been a widely recognised decline in recent times in the presence of onsite union representatives in British workplaces (Kersley et al., 2006: 124; Terry, 2003: 259), and this has been accompanied by concerns as to whether unions still have sufficient power to influence employer decision-making (Charlwood and Forth, 2009). Given these concerns, there has been a resurgence of debate among policy-makers, employers’ organisations and trade unions over the contribution onsite representatives make to building effective relationships at work, with codes of practice for statutory time-off for trade union duties and activities having been revised and subjected to ongoing review (ACAS, 2008; BERR, 2007, 2009; Mitchell et al., 2012; TUC, 2012).

The aim of this article is to contribute towards this debate by exploring the impact of onsite union representatives on an outcome that has for many years been accorded significant importance within the employment relations literature – job quality. This is defined in the literature as incorporating both intrinsic and extrinsic outcomes such as high pay, job security, participation in decision-making, task discretion, skill, autonomy, work–life balance and job stress (Clark, 2005; Doellgast et al., 2009: 490; Green, 2009: 9). Although unions have arguably focused traditionally on the economic aspects of the employment relationship, prioritising bargaining over terms and conditions, the goal of improving job quality has also been a matter of longstanding union concern (Bryson et al., 2004: 441; Doellgast et al., 2009: 506; Holman, 2103a: 476), and onsite union representatives may have an important role to play in achieving this goal. The achievement of better job quality has also been accorded significant political importance. For example, the European Union’s employment strategy has focused on the promotion of job quality (as well as job quantity) since its inception in the late 1990s (Green et al., 2013: 753; Holman, 2013a: 476; Smith et al., 2008: 586). The goal of job quality also underpinned the former British Labour government’s education and training policies and its support for the employer–union ‘partnership’ agenda (Brown et al., 2007: 943), and it has featured in discussions on happiness and subjective well-being as an area of public policy and government action (Reeves, 2009), having become central to attempts to measure such outcomes (Green et al., 2013: 753).

Mirroring the importance accorded to job quality in both political and union circles is extensive evidence pointing to its positive impact on a range of both individual and organisational outcomes. In particular, it has been found to be associated with overall individual well-being (Green, 2008; 2009: 8, 18), and has been closely linked to job satisfaction (Pfeffer and Davis-Blake, 1990: 262–263), which in turn has been identified as an important antecedent of higher productivity, increased discretionary effort, fewer quits and lower absenteeism (Clark, 2005: 380; Coats and Lekhi, 2008; Esser and Olsen, 2012; Gazioglu and Tansel, 2006: 1163). It has also been associated with smoother labour market transitions and higher labour market participation rates (Smith et al., 2008: 588), and is viewed as an important antecedent of lower levels of workplace conflict (Green, 2009: 9), better mental health and reduced stress (Wood, 2008). The achievement of higher job quality would thus appear to have the potential to enhance a range of important socio-economic outcomes.

There is, however, growing concern that several dimensions of job quality have deteriorated since the early 1990s (Kalleberg, 2003; Smith et al., 2008: 588), with work intensity having increased (Brown et al., 2007: 942; Clark, 2005: 393; Handel, 2005: 84; Green et al., 2013: 766) and employee discretion and task autonomy having declined (Brown et al., 2007: 942; Clark, 2005: 383; Green, 2009: 18; Olsen et al., 2010: 222). Such changes appear particularly pronounced in the UK (Clark, 2005: 394) having been driven in part by the recent government policy emphasis on labour market flexibility (Brown et al., 2007). Beyond this, higher unemployment and weaker labour markets in the aftermath of the 2008 financial crisis have increased the pressure on governments to prioritise job creation over job quality, while pressure on employers to focus on job quality as a retention strategy has weakened. As Kalleberg (2012: 431) argues, during periods of high unemployment, concerns over job quality tend to diminish given that ‘having any job is generally regarded as better than having no job at all’.

Hence, identifying ways in which job quality can be influenced positively might be deemed particularly important in the current climate. This article seeks to contribute towards this issue by drawing on a survey of union members in the UK finance sector within the Unite trade union (Britain’s largest trade union with 1.4 million members and over 140,000 members in the finance sector) in order to address whether, and in what ways, onsite union representatives are able to influence job quality positively in the workplaces within which they are located.

Theorising the association between onsite union representatives and job quality

From the extant literature, it is possible to discern a number of ways in which onsite union representatives might influence levels of job quality. In exploring this issue from a theoretical perspective, a useful starting point is Freeman and Medoff’s (1984) collective voice/institutional response model. This model suggests a ‘positive voice effects’ hypothesis, whereby if unions voice their members’ job quality concerns via either informal communication channels or collective bargaining processes, this will bring job quality problems to the attention of management, who will then subsequently respond by making changes within the workplace to seek to address these problems (Wood, 2008: 154). Such collective voice effects may well be amplified where onsite representatives are present, given the role they play in seeking out the views of members, listening to their concerns and bringing these concerns to managers’ attention (Bryson and Forth, 2010).

One must keep in mind, however, that positive voice effects of this nature will only emerge where the expression of collective voice is heard by managers (Guest and Conway, 2004) and if managers are willing (or compelled) to respond constructively. This in turn is likely to depend on the union’s scope for involvement in decision-making and on its bargaining strength. Where there is an absence of institutional support (as provided by Germany’s codetermination laws, for example) aimed at enhancing the ability of workers’ representatives to exercise voice in company decisions (Kalleberg, 2012: 436), or where union bargaining power is weak, the union’s ability to pressure employers to improve job quality is likely to be restricted (Brown et al., 2007: 949). Hence, even if onsite union representatives listen to their members’ job quality concerns and convey those concerns to management through collective voice mechanisms, there is no guarantee that this will result in job quality improvements.

There is considerable empirical evidence in support of this argument. For example, Holman (2013a, 2013b) finds in studies exploring data from call centres and from the European Working Conditions Survey that while unions do have the potential to influence job quality positively, this is more likely to happen in countries where they have sufficient power to influence employer decision-making (in Denmark and Sweden, for example). Similarly, Green et al. (2013) attribute higher levels of job quality in coordinated market economies in part to union strength. Doellgast et al.’s (2009) research on front-line service employees across different countries also finds the ability of unions to influence job quality to be dependent on their strength, while Esser and Olsen (2012), using data from the European Social Survey, find aspects of job quality (in particular job autonomy and job security) to be higher in countries where unions have sufficient power to constrain the actions of employers.

By contrast, in countries where union strength is low and institutional support for union voice is lacking, the extant research suggests that unions have had significantly less success in influencing job quality positively. Indeed, Green et al. (2013: 757–758) argue that poorer job quality in the US and the UK should not come as a surprise given the weakness of unions within these countries, while Holman (2013b: 26) suggests that poorer levels of job quality in market regimes such as the UK can be explained by the absence of institutions such as works councils that facilitate the ability of unions to express collective voice. Further to this, Willman et al. (2007) argue that union voice effects may have become muted in the UK in recent years given that declining union density means that union representatives are less able to claim to speak for the majority (or even a sizeable proportion) of the workforce, thus the impact and credibility of union voice has diminished. It is far from certain, therefore, whether one would expect onsite union representatives in the UK to be able to raise levels of collective voice and to translate this successfully into positive outcomes with regard to job quality. It is quite possible that even if union representatives listen to their members’ views and attempt to voice these views to management, they may lack the leverage and credibility to elicit a positive employer response. As such, it is unclear how far one would expect a ‘positive voice effects’ hypothesis to be supported.

Beyond this, however, even if positive voice effects do emerge, there are reasons to doubt that they will hold across all facets of job quality. This is because local level managers may have insufficient discretion to be able to amend employer policy or implement informal rules in response to pressure from onsite representatives. As such, while some intrinsic job characteristics (such as whether the job is interesting/enjoyable, makes good use of individuals’ skills, minimises job stress and is adjusted to help work–life balance) may be subject to local level discretion, issues such as pay and job security are likely to be subject to centralised decisions that lie beyond the scope of local level managers. Hence, while union full-time officers may, via negotiations with senior managers, have the ability to influence these latter elements of job quality at organisational or national level, onsite representatives are likely to be less able to influence them in their dealings with local level managers. Further to this, onsite representatives’ attempts to address job quality at local level may be restrained by their own union if, for example, union full-time officers are pursuing a focus on pay and employment rather than intrinsic elements of job quality, or if challenges to managerial prerogatives over the organisation and pace of work are viewed as having the potential to threaten the future recognition of the union (Carter et al., 2012).

There is, however a further reason, specific to the finance sector (on which the research presented in this article is based), why support for a ‘positive voice effects’ hypothesis might be muted. Onsite representation and activity in the sector was, in the 1990s, viewed as ‘rudimentary’ (Waddington, 2013: 349), and as a result, finance sector unions subsequently developed a general policy of increasing onsite representation and reducing reliance on the activities of seconded representatives¹ and full-time officers. However, assessment of the impact of this policy in four clearing banks between 1999 and 2008 suggested that it had failed, with the coverage of onsite representatives having reduced, the proportion of union members prepared to seek their advice not having increased, and member satisfaction with the quality of onsite representation having declined (Waddington, 2013). This in turn leads to doubt over whether, in the case of the finance sector, onsite union representatives will have been successful in enhancing members’ perceptions of union voice, or if they have been successful in doing so, subsequently using this enhanced voice to improve levels of job quality.

There are, therefore, a number of reasons why the ‘positive voice effects’ hypothesis might not be supported. Beyond this, however, one might argue that the collective voice effects that union representatives engender may impact on perceived levels of job quality in an opposite direction from that hypothesised above, hence it is possible to posit a ‘negative voice effects’ hypothesis. One element of this hypothesis, as argued by Hammer and Avgar (2005) and Meng (1990: 1635), is that collective voice channels may be seen as providing a mechanism by which workers who are dissatisfied with their jobs can express their discontent. Given this, dissatisfied unionised workers will have a lower propensity to quit the organisation than will non-union workers who have less scope to express their grievances. If onsite union representatives strengthen the collective voice effects that encourage workers who are dissatisfied with their job quality to remain with the organisation, this might result in lower aggregate reported perceptions of job quality than in workplaces where onsite representatives are not present.

A further element of the ‘negative voice effects’ hypothesis is that unions may seek to utilise collective voice channels to increase workforce politicisation. One aspect of this might involve raising workers’ consciousness of the negative aspects of their jobs (by highlighting the lack of participation in decision-making, limited training and skill development, low job security, job stress or a poor work–life balance, for example) in order to generate a sense of grievance and thereby strengthen the basis for collective action (Bryson and Freeman, 2103: 4).

Such activity on the part of the union might be expected to have a number of effects. First, it may socialise union members to expect higher standards with regard to job quality (Guest and Conway, 2004), and given this they may adopt stricter criteria when making subjective judgements about their jobs than might non-unionised workers (Bryson and McKay, 1997; Hammer and Avgar, 2005: 244; Pfeffer and Davis-Blake, 1990: 281). Onsite representatives may play an important role in this socialisation process, hence members’ ‘internal standards’ (Hammer and Avgar, 2005: 258) might be particularly high in workplaces where an onsite representative is present. Second, raising workers’ consciousness of the negative aspects of their jobs may result in greater workforce discontent and a more adversarial industrial relations climate. Should managers equate this with a reduction in employee discretionary effort and commitment (Gordon and Denisi, 1995: 223; Guest and Conway, 2004; Hammer and Avgar, 2005: 243; Pfeffer and Davis-Blake, 1990: 260), they may subsequently seek to exercise greater control over work processes via the introduction of more bureaucratised work systems. This in turn is likely to reduce the scope for employee autonomy and discretion (Hammer and Avgar, 2005: 241), thereby negatively influencing job quality. This might be particularly likely to happen in workplaces with an onsite union representative given that they may be a primary vehicle by which grievance identification and workforce politicisation occurs.

This politicisation argument has been criticised, however, given that over time workers may come to blame unions as much as managers for unresolved job quality problems. Hence there are limits to which union representatives will wish to politicise workers by generating a sense of grievance over the negative aspects of their jobs (Gordon and Denisi, 1995: 225; Hammer and Avgar, 2005: 243; Pfeffer and Davis-Blake, 1990: 261). It nevertheless remains a theoretical possibility that the voice effects potentially engendered by onsite union representation could influence job quality negatively. This article seeks to identify whether there is empirical support for this ‘negative voice effects’ hypothesis.

A further reason, however, for questioning whether a positive relationship will exist between onsite union representation and job quality is that it is unlikely that onsite representatives will be randomly distributed among the population of unionised workplaces. It is widely argued that unions are more likely to organise workplaces that have job quality problems (Bryson et al., 2004: 441; Farber and Saks, 1980; Gordon and Denisi, 1995; Hammer and Avgar, 2005: 243; Miller, 1990; Pfeffer and Davis-Blake, 1990; Renaud, 2002). Indeed, a central notion of Kelly’s (2005) mobilisation theory is that where a collective sense of injustice develops (in relation to issues such as poor job quality), workers will seek union representation in order to pursue a remedy. In line with this, Bryson and Freeman (2013: 21) find in their analysis of large-scale datasets in both the UK and the US that poor working conditions are strongly correlated with the desire for union representation. Similarly, it is plausible that onsite representatives are more likely to be found in workplaces within which working conditions are poor, given that unions might perceive a particular need for onsite representation within such workplaces, and individuals with a greater propensity to engage in union activism may be particularly willing to take on a representative role. Hence, if there is a negative association between onsite representative presence and employee perceptions of job quality, this could be explained by the possibility that onsite representatives are located in workplaces that had poor job quality in the first instance.

As the above discussion demonstrates, on the basis of the extant theory and evidence it remains open to question whether one would expect the presence of onsite union representatives (and the collective voice effects they potentially engender) to be associated with more favourable employee perceptions of job quality in the UK context (and in the finance sector in particular). This article seeks to shed light on this issue by addressing the following research aims. The first is to evaluate whether union members’ perceptions of the extent to which the local union provides effective collective voice are more favourable where an onsite representative is present (if not, there will be no grounds to argue for either the positive or negative voice effects hypotheses outlined above). The second is to evaluate whether union members’ reports of job quality are more or less favourable where an onsite representative is present than elsewhere. The third is to consider whether perceptions of the extent to which the local union provides effective collective voice mediates this relationship. If support for mediation is found, this will suggest support for the ‘positive voice effects’ hypothesis (if the relationship between workplace representatives and job quality is positive) or the ‘negative effects’ hypothesis (if the relationship between workplace representatives and job quality is negative). Should the results suggest a negative association between onsite union representatives and job quality, and should this remain once respondents’ perceptions of collective voice have been controlled for, this could suggest that the relationship between representatives and job quality is endogenous, such that onsite union representatives are located in workplaces that had particularly low levels of job quality in the first instance.

Method

Data and context

As mentioned above, the data used in this article are taken from the Unite trade union finance sector members’ survey conducted in spring 2008. Unite is the largest trade union in the sector and represents workers in all grades and occupations across a range of employers including the major English and Scottish banks, insurance companies, building societies, finance houses and business services companies. The survey involved distributing a questionnaire by post to a random sample of 35,600 working members from a population of 142,400. There were 3087 responses received, constituting a response rate of 8.7%. Although this is in line with previous research drawing on surveys of union members in the finance sector (see, for example, Waddington, 2013), it nevertheless means that the results should be treated with caution. That said, the dataset has certain advantages in exploring the issues of interest to this article given that it provides insights in terms of perceptions of collective voice effects and workplace experiences that are not usually available in broader cross-sector studies.

In terms of industry composition, the UK finance sector includes banks and building societies (accounting for more than half of the output), insurance and pension companies and financial intermediation services. This is reflected in the profile of respondents to the Unite finance members’ survey, a large proportion of whom (41%) are from the large banks, with the next largest (13%) being from large insurance companies. As such, just over half of respondents come from 10 employers: Barclays (15% of respondents); Royal Bank of Scotland (14%); HSBC (14%); NatWest (6%); Lloyds TSB (4%); Bank of Scotland (3%); Allianz Cornhill (3%); Friends Provident (3%); Royal Sun Alliance (2%); and Norwich Union (2%). In total, responses were received from employees across 174 companies. Cross-checking the sample against the characteristics of the population suggests it is broadly representative by gender (58% of the sample are female against 53% of the population), but slightly skewed towards younger people (10% in the sample are aged under 30 against 6% of the population).

The sector on which the analysis is based (finance) is arguably particularly appropriate in exploring the research questions outlined above, given the nature of union activity in the sector. The UK Labour Force Survey reports trade union density in ‘financial intermediation’ in the UK to be 20.5% (Achur, 2010). Union presence (defined as the proportion of employees whose workplace has a union present) and collective agreement coverage (defined as the proportion of employees whose pay and conditions are agreed in negotiations between the employer and a trade union) are reported as 43.3% and 28.4% respectively. By comparison, in the economy as a whole, 24% of workplaces in 2004 and 21% of workplaces in 2011 had a recognised trade union (14% of private sector workplaces in 2004 and 11% of private sector workplaces in 2011), having fallen from 64% of workplaces in 1980. As such, the collective organisation of finance workers is deemed to be ‘moderate’ (Gall, 2008: xiii). Given this, one might argue that finance is an appropriate sector in which to seek to identify union effects, and it would not be unreasonable to expect the results to hold in other sectors with similarly moderate (or stronger) levels of unionisation.

The sector is also important in economic terms, having accounted for between 6% and 10% of UK GDP in the last decade. It accounted for 3.6% of employment (1.1 million employees) in 2012 (a figure broadly similar to before the financial crisis in 2008) located in approximately 34,000 workplaces (Bank of England, 2011; House of Commons Library, 2012). More than one-third of employees are in managerial occupations, with fewer than one-third in administrative or secretarial roles, and a quarter in associate professional and technical roles. The average salary of £49,826 per year is heavily skewed by the London job market, with almost a quarter of all financial services employment located in London.

Measures

Onsite union representative

The presence of an onsite union representative is measured using a single item: ‘Is there an office representative at your place of work?’ In total, 48% of respondents are in workplaces in which an onsite union representative is present.

Collective voice

Seven items measured on a five-point scale assess the extent to which respondents believe their local union provides effective collective voice. These are how good the local union is at: keeping everyone up to date about proposed changes; providing everyone with a chance to comment on proposed changes; responding to suggestions from members; dealing with work problems members may have; consulting members on matters that affect them at work; listening to their views; and representing members on issues that matter to them. These issues are considered pertinent to assessing the effectiveness of onsite union representation and collective voice in the finance sector (Waddington, 2013: 347), given the role onsite representatives play in seeking out the views of members, listening to their concerns and bringing these concerns to managers’ attention (Bryson and Forth, 2010). A principle component factor analysis demonstrated that all of these items loaded onto a single factor that explained 72% of the total variance in the items, hence they were combined into a single scale (Cronbach alpha reliability of .93).

Job quality

There are multiple dimensions to job quality comprising both extrinsic and intrinsic factors (Gallie, 2013: 454), and as Smith et al. (2008: 590) argue, it is a complex concept to measure. As demonstrated by Table 1, there are no agreed definitions for it within the literature, with a wide and varied range of dimensions having been used in previous studies. Indeed, Table 1 points to 22 different dimensions of job quality in the extant research. Some have been used frequently (pay, autonomy and job discretion, job security, for example), while others are specific to individual studies.

Table 1.

Dimensions of job quality used in the extant literature.

Dimension of job quality	Study
Pay	Brown et al. (2007); Clark (2005); Doellgast et al. (2009); Green (2009); Holman (2013a, 2013b); Leschke and Watt (2008); Olsen et al. (2010); Tilly (1997)
Autonomy and job discretion	Brown et al. (2007); Clark (2005); Doellgast et al. (2009); Esser and Olsen (2012); Gallie (2007, 2013); Green (2009); Green et al. (2013); Holman (2013a, 2013b); Olsen et al. (2010); Smith et al. (2008); Tilly (1997)
Job security	Clark (2005); Doellgast et al. (2009); Esser and Olsen (2012); Gallie (2007); Green (2009); Holman (2013a, 2013b); Leschke and Watt (2008); Tilly (1997)
Promotion prospects/ career advancement	Clark (2005); Leschke and Watt (2008); Olsen et al. (2010); Tilly (1997)
How hard, stressful or dangerous the job is	Brown et al. (2007); Clark (2005); Green et al. (2013); Olsen et al. (2010); Smith et al. (2008)
Skill	Brown et al. (2007); Gallie (2007); Green (2009); Holman (2013a, 2013b)
Effort	Brown et al. (2007); Green (2009); Olsen et al. (2010)
Participation in decision-making	Doellgast et al. (2009); Gallie (2007, 2013); Holman (2013b); Leschke and Watt (2008)
How interesting the work is and the monotony/ complexity of tasks	Brown et al. (2007); Olsen et al. (2010); Smith et al. (2008)
Relations with colleagues	Brown et al. (2007); Clark (2005); Olsen et al. (2010)
Creative content of work	Brown et al. (2007); Smith et al. (2008)
Flexible forms of work and work–life balance	Brown et al. (2007); Holman (2013a, 2013b); Leschke and Watt (2008)
Hours of work	Clark (2005); Green et al. (2013); Tilly (1997)
Representation	Gallie (2007); Holman (2013a); Leschke and Watt (2008)
Non-standard forms of employment	Leschke and Watt (2008)
Position within organisation and class hierarchy	Brown et al. (2007)
Work that helps others and is useful to society	Olsen et al. (2010)
Prestige	Clark (2005)
Absence of detailed monitoring and surveillance	Doellgast et al. (2009)
Fringe benefits	Tilly (1997)
Due process in discipline	Tilly (1997)
The intellectual/ emotional demands of the job	Smith et al. (2008)

Further to this, there is also a division within the extant research between studies that take a subjective approach to the measurement of job quality, focusing on employees’ perceptions of particular aspects of the quality of their job (Holman, 2013b: 23), and studies that take a more objective approach, focusing on assessments of the tangible and observable job characteristics and features (Green et al., 2013). As Bryson and Freeman (2013) suggest, it is arguably more important to assess job quality from the former perspective, as it is in instances where employees perceive their job quality to be poor that negative socio-economic outcomes (higher labour turnover, lower productivity, poorer mental health and higher stress, for example) are likely to emerge.

In line with this view, the Unite survey included nine items that relate to employee perceptions of several of the dimensions of job quality commonly referred to in the literature. Each of these items was measured on a four-point scale ranging from 4 = ‘strongly agree’ to 1 = ‘strongly disagree’. Six of the items referred to intrinsic content of the job, and a principle components factor analysis demonstrated that the items loaded onto a single factor that explained 62% of the total variance in the items (‘my job is interesting and enjoyable’; ‘my job makes me feel that I have accomplished something’; ‘I have a say in how the work that I am responsible for gets done’; ‘my job makes full use of my skills and talents’; ‘my job enables me to learn new skills and develop my abilities’; ‘my job offers opportunities for advancement at work’). The six items were combined into a single scale labelled ‘job content’ (Cronbach alpha reliability of .87). The remaining three items referred to distinct dimensions of job quality and were therefore used as single item measures: job security (‘my job is secure’); job stress (‘my job is stressful’); and work–life balance (‘my job enables me to work the hours that suit my personal requirements’).

Control variables

A range of factors that might otherwise influence the relationship between union representation, collective voice and job quality are controlled for in the analysis. Both individual and establishment characteristics are controlled for, including gender, ethnicity, work status (i.e. whether part-time or full-time), respondent age, salary, the size of the establishment the respondent works in (i.e. number of employees) and the geographical region of the respondent’s workplace. The Appendix contains details of the control variables used.

Analytic procedure

The article’s first aim is to identify whether respondents’ perceptions of the extent to which the local union provides effective collective voice are more favourable where an onsite representative is present than where an onsite representative is not present. This is tested in an ordinary least squares equation that treats the collective voice measure outlined above as the dependent variable and the onsite representative dichotomous variable as the independent variable. The equation controls for the individual and establishment characteristics outlined earlier.

The article’s second aim (to ascertain whether reports of job quality are more favourable where an onsite representative is present) is assessed in four equations within which the dependent variables are the four measures for job quality outlined above, and the independent variable is the onsite union representative dichotomous variable.² These equations also include controls for individual and establishment characteristics.

The article’s third aim is to assess whether respondents’ perceptions of the extent to which their local union provides them with effective collective voice mediates the relationship between the presence of onsite union representatives and job quality. This is tested following the mediation procedures recommended by Baron and Kenny (1986). These procedures state that the independent variable must affect the mediator (as identified by the article’s first aim) and also the dependent variable (as identified by the article’s second aim). Mediation exists when the dependent variable is regressed simultaneously onto the mediator and the independent variable, and the effect of the independent variable on the dependent variable is less when the mediator is present. Full mediation occurs if the independent variable has no significant effect when the mediator is present. The size of the indirect effect of the independent variable on the dependent variable via the mediator is assessed using a Sobel test (Sobel, 1982).³

Regression models were checked for multicollinearity among the independent variables using variance inflation factors (VIFs) and were below 2.1 among the main study variables presented in Table 2 and therefore well below the standard cut-off of 10 (Levin et al., 2006: 1167).⁴

Table 2.

Means, standard deviations, alpha reliability and zero-order correlations between main study variables.

	M	SE	1	2	3	4	5
1 Onsite representative	.48	.50
2 Voice	2.88	.97	.41^***	(.87)
3 Job content	2.69	.63	.08^***	.16^***	(.93)
4 Job security	2.52	.80	−.04†	.06^**	.39^***
5 Job stress	3.13	.78	−.14^***	−.07^**	−.07^***	−.09^***
6 Work life balance	2.63	.88	.18^***	.18^***	.30^***	.17^***	−.28^***

Notes: N = 1926. † p < .10; ** p < .01; *** p < .001. Intercorrelations involving control variables are omitted from the table.

Results

Means, standard deviations, Cronbach alpha reliability coefficients and correlations among the main variables are presented in Table 2.

The first stage of the analysis is to assess whether respondents’ perceptions of the extent to which their local union provides effective collective voice are more favourable where onsite union representatives are present than where they are not present. As argued earlier, if no differences emerge there will be no grounds to argue for either the positive or negative voice effects hypotheses outlined above. However, the results, given in Table 3, provide strong evidence to suggest that respondents’ perceptions of the extent to which their local union provides effective voice are indeed more favourable where onsite representatives are present than where not present (beta = .432, p < .001; the change in R² following the addition of the union representative variable to the model indicating that it explains 10% of the variation in voice). The evidence would appear to demonstrate, therefore, that the presence of onsite representatives is strongly associated with workers’ perceptions of how far the local union provides effective collective voice.

Table 3.

Regression model examining the relationship between union representative presence and perceptions of employee voice.

	Voice
	B	SE	β	p
Sex (female)	.054	.049	.028
Ethnicity	.033	.088	.009
Part-time	−.042	.069	−.018
Age (< 30)	.077	.125	.018
Age (30–39)	−.056	.100	−.022
Age (40–49)	−.065	.094	−.032
Age (50–59)	−.057	.094	−.028
Workplace size (<10)	.179	.079	.067	*
Workplace size (10–24)	−.046	.075	−.019
Workplace size (25–49)	−.057	.084	−.018
Workplace size (50–99)	−.337	.103	−.075	**
Workplace size (100–199)	−.259	.099	−.059	**
Workplace size (200–499)	−.210	.076	−.066	**
Workplace size (500–999)	−.087	.075	−.028
Pay (£12–19.9K)	.062	.080	.029
Pay (£20–29.8K)	.112	.087	.049
Pay (£29.9–42K)	.090	.094	.039
Pay (>£42K)	−.064	.107	−.020
Region (NE)	−.112	.070	−.036
Region (SW)	−.037	.080	−.010
Region (NW)	−.073	.068	−.024
Region (Wales)	−.025	.125	−.004
Region (N & R Ireland)	.100	.139	.017
Region (Midlands E)	−.114	.116	−.021
Region (SE)	−.109	.076	−.032
Region (Midlands W)	−.095	.087	−.024
Region (Scotland)	−.044	.083	−.012
Union rep	.843	.055	.432	***
F	16.125***
R²	192

Notes: * p < .05; ** p < .01; *** p < .001. Omitted reference categories: age (> 59); workplace size (> 999); pay (< £12K); region (London). N = 1926. OLS analysis.

The second aim of the article is to ascertain whether respondents’ reports of job quality (with regard to job content, job security, job stress and work–life balance) are more favourable where an onsite representative is present than where not present. The results, presented in Table 4, demonstrate that respondents in workplaces with an onsite representative report higher job quality than do those in workplaces without an onsite representative where two of the four measures are concerned: job content (beta = .071, p < .05) and work–life balance (beta = .123, p < .001). There is also weak evidence (at the 10% significance level) that they report lower levels of job stress (beta = –.056, p < .1). There is, however, no evidence to suggest that respondents in workplaces with an onsite representative perceive job security to be greater than do those in workplaces without a representative.

Table 4.

Regression models examining relationships between union representatives, employee voice and job quality.

	Job content				Job content				Job security
	B	SE	β	p	B	SE	β	p	B	SE	β	p
Sex (female)	.071	.034	.057	*	.066	.034	.053	*	−.031	.044	−.020
Ethnicity	−.080	.061	−.033		−.083	.060	−.035		.019	.079	.006
Part-time	−.009	.047	−.006		−.005	.047	−.003		.061	.062	.032
Age (< 30)	−.044	.086	−.016		−.051	.085	−.018		.016	.113	.004
Age (30-39)	−.145	.069	−.090	*	−.140	.068	−.087	*	−.III	.090	−.054
Age (40-49)	−.168	.064	−.131	**	−.162	.064	−.126	*	−.191	.084	−.1 16	*
Age (50-59)	−.195	.065	−.146	**	−.190	.064	−.142	**	−.176	.085	−.103	*
Workplace size (< 10)	.082	.054	.048		.064	.054	.037		.161	.071	.073	*
Workplace size (10-24)	.067	.052	.042		.071	.051	.045		.212	.068	.105	**
Workplace size (25-49)	.082	.057	.040		.088	.057	.042		.278	.075	.105
Workplace size (50-99)	−.012	.071	−.004		.020	.070	.007		.009	.092	.003
Workplace size (100-199)	−.047	.068	−.017		−.022	.067	−.008		.166	.089	.046	†
Workplace size (200-499)	−.015	.052	−.007		.006	.052	.003		.028	.068	.01 1
Workplace size (500-999)	−.070	.052	−.035		−.062	.051	−.031		.044	.068	.017
Pay (£I2-I9.9K)	.022	.055	.016		.016	.054	.011		.006	.072	.003
Pay (£20-29.8K)	.249	.060	.170	***	.238	.060	.163	***	−.007	.079	−.004
Pay (£29.9-42K)	.330	.065	.220	***	.321	.064	.215	***	.094	.085	.049
Pay (>£42K)	.508	.073	.244	***	.514	.073	.246	***	.051	.096	.019
Region (NE)	−.023	.048	−.012		−.012	.048	−.006		−.102	.063	−.040
Region (SW)	.01 1	.055	.005		.014	.054	.006		−.076	.072	−.026
Region (NW)	.039	.047	.020		.046	.046	.024		−.019	.061	−.008
Region (Wales)	−.129	.086	−.034		−.126	.085	−.034		−.074	.113	−.015
Region (N & R Ireland)	.091	.096	.024		.081	.095	.021		.425	.125	.088	***
Region (Midlands E)	.157	.080	.045	*	.168	.079	.049	*	.056	.105	.013
Region (SE)	.007	.052	.003		.018	.051	.008		−.079	.068	−.028
Region (Midlands W)	.039	.060	.015		.048	.059	.019		−.063	.078	−.020
Region (Scotland)	−.047	.057	−.020		−.043	.056	−.018		−.018	.074	−.006
Union rep	.088	.038	.071	*	.007	.040	.005		.010	.049	.006
Voice					.097	.016	.151	***
F	5.362***				6.397***				2.334**
R ²	.073				.092				.033
Change F					38.235***
Change R²					.018

Notes: † p < .10; * p < .05; ** p < .01; *** p < .001. Omitted reference categories: age (> 59); workplace size (> 999); pay (< £12K); region (London).N = 1926. OLS analysis.

	Job stress				Job stress				Work–life balance				Work–life balance
	B	SE	β	p	B	SE	β	p	B	SE	β	p	B	SE	β	p
Sex (female)	.127	.042	.081	**	.128	.042	.082	**	−.117	.047	−.066	*	−.124	.047	−.070	**
Ethnicity	.007	.076	.002		.008	.076	.003		−.163	.084	−.048	†	−.167	.083	−.049	*
Part-time	−.120	.059	−.064	*	−.121	.059	−.065	*	.470	.065	.221		.476	.065	.224	***
Age (< 30)	.160	.108	.046		.162	.108	.047		−.090	.119	−.023		−.100	.118	−.025
Age (30-39)	.007	.086	.003		.005	.086	.003		−.056	.095	−.025		−.049	.095	−.022
Age (40-49)	.114	.080	.072		.113	.080	.071		−.100	.089	−.055		−.091	.088	−.050
Age (50-59)	.092	.081	.055		.091	.081	.054		−.196	.090	−.104	*	−.189	.089	−.100	*
Workplace size (< 10)	.204	.068	.096	**	.209	.068	.098	**	−.162	.075	−.067	*	−.185	.075	−.076	*
Workplace size (10-24)	.229	.065	.1 16	***	.227	.065	.115		−.240	.072	−.107		−.234	.071	−.105	***
Workplace size (25-49)	.159	.072	.062	*	.157	.072	.061	*	−.110	.080	−.038		−.103	.079	−.035
Workplace size (50-99)	−.121	.088	−.034		−.130	.089	−.036		−.140	.098	−.034		−.096	.097	−.024
Workplace size (100-199)	−.046	.085	−.013		−.053	.085	−.015		−.224	.094	−.056	*	−.191	.093	−.048	*
Workplace size (200-499)	−.104	.065	−.041		.110	.065	.043	†	.033	.072	.01 1		.060	.072	.021
Workplace size (500-999)	−.089	.065	−.036		−.092	.065	−.037		.098	.072	.035		.109	.071	.039
Pay (£I2-I9.9K)	.167	.069	.098	*	.169	.069	.099	*	−.109	.076	−.056		−.1 17	.075	−.060
Pay (£20-29.8K)	.178	.075	.098	*	.181	.075	.100	*	−.117	.083	−.057		−.132	.083	−.064
Pay (£29.9-42K)	.257	.081	.138	**	.260	.081	.139		−.048	.090	−.023		−.059	.089	−.028
Pay (>£42K)	.286	.092	.1 10	**	.285	.092	.110	**	−.056	.102	−.019		−.047	.101	−.016
Region (NE)	.083	.060	.034		.080	.060	.032		−.166	.067	−.059	*	−.152	.066	−.054	*
Region (SW)	−.005	.068	−.002		−.006	.068	−.002		−.046	.076	−.014		−.041	.075	−.013
Region (NW)	.091	.058	.038		.089	.059	.037		−.064	.065	−.024		−.055	.064	−.020
Region (Wales)	.104	.108	.022		.103	.108	.022		−.316	.120	−.060	**	−.313	.118	−.059	**
Region (N & R Ireland)	−.188	.1 19	−.040		−.186	.1 19	−.039		.193	.132	.036		.180	.131	.034
Region (Midlands E)	−.099	.100	−.023		−.102	.100	−.024		.091	.11 1	.019		.106	.110	.022
Region (SE)	.046	.065	.017		.043	.065	.016		−.053	.072	−.017		−.039	.071	−.013
Region (Midlands W)	.091	.074	.029		.089	.074	.028		−.035	.083	−.010		−.023	.082	−.006
Region (Scotland)	−.062	.071	−.021		−.063	.071	−.021		.033	.079	.010		.038	.078	.01 1
Union rep	−.087	.047	−.056	†	−.066	.050	−.042		.217	.052	123	***	.108	.055	.061	*
Voice					−.026	.020	−.032						.129	.022	142	***
F	4.644***				4.544***				8.248***				9.326***
R ²	.064				.065				.109				.125
Change F					1.697								35.323
Change R²					.001								.016

Notes: † p < .10; * p < .05; ** p < .01; *** p < .001. Omitted reference categories: age (> 59); workplace size (> 999); pay (< £12K); region (London).N = 1926. OLS analysis.

The third aim of the article is to assess whether respondents’ perceptions of the extent to which their local union provides them with effective voice mediate the relationship between the presence of onsite union representatives and job quality. Mediation is tested in instances where the presence of an onsite representative is significantly positively related to job quality.

The results reported in Table 4 suggest significant support for mediation. When the presence of an onsite representative and respondents’ perceptions of the extent to which their local union provides effective collective voice are entered into the equations simultaneously, collective voice is significantly related to job content (beta = .151, p < .001) and to work–life balance (beta = .142, p < .001), the previously significant relationship between onsite representatives and job content reduces to non-significance (beta falls from .071 to .005, p n.s.), and the previously significant relationship between onsite representatives and work–life balance reduces in statistical significance (beta falls from .123, p < .001 to .061, p < .05). A Sobel test indicates that the indirect effect of union representatives on job content and work–life balance are significant (respectively, Sobel = 5.64, p < .001, and 3.40, p < .001). In other words, respondents’ perceptions of the extent to which the local union provides effective collective voice fully mediates the relationship between onsite representatives and job content, and partially mediates the relationship between onsite representatives and work–life balance. Where job stress is concerned, the weakly positive relationship between the presence of an onsite representative and job stress becomes non-significant when respondents’ perceptions of the extent to which their local union provides effective collective voice is entered into the equation (beta falls from –.056, p < .1 to –.042, p n.s.). Hence, the results overall suggest that the positive relationship between onsite representatives and job quality is explained at least in part by the collective voice effects they engender.

Discussion and conclusions

This article drew on a survey of union members within the Unite trade union in the UK finance sector in order to evaluate the relationship between the presence of onsite union representatives and employee perceptions of job quality, and to evaluate whether the existence of such a relationship can be explained by the collective voice effects onsite representatives engender.

In the event, the analysis revealed several notable outcomes. First, a strong positive relationship emerged between the presence of an onsite representative and respondents’ perceptions of the extent to which their local union provides them with effective collective voice. Second, respondents’ perceptions of three of the four aspects of job quality under observation (job content, work–life balance and job stress, but not job security) were more favourable where an onsite representative was present (although the association was weak for job stress). Third, the association between onsite union representation and respondents’ perceptions of job content and work–life balance was mediated by their perceptions of the extent to which their local union provides effective collective voice, while the results for job stress, though non-significant, showed a similar pattern.

On balance, therefore, the results appear to support the ‘positive voice effects’ hypothesis posited by Freeman and Medoff (1984) that perceptions of job quality will be higher where an onsite representative is present, and that this is explained by the higher levels of collective voice that such representatives engender. As such, the results provide support for the conclusions reached within previous empirical analysis concerning the potential for unions to influence job quality favourably (Doellgast et al., 2009; Esser and Olsen, 2012; Green et al., 2013; Holman, 2013a, 2013b). They also support the conclusions in the extant literature concerning the need for effective institutional support to facilitate the participation of unions in decision-making if a relationship between unions and job quality is to emerge (Holman, 2013b; Kalleberg, 2012), the results here suggesting that onsite union representation is one such form of institutional support. That said, one might argue that the analysis in this article also suggests that the magnitude of the overall union effect on job quality in the UK context is likely to be limited, given the decline in recent times in the presence of onsite union representatives in UK workplaces noted earlier (Kersley et al., 2006: 124; Terry, 2003: 259).

The findings have several implications for unions and public policy. Where unions are concerned, as discussed above, previous research on the UK’s four clearing banks suggested that attempts by finance sector unions to develop onsite representation and reduce reliance on the activities of seconded representatives and full-time officers had failed (Waddington, 2013). One might argue on the basis of the results from the broader finance sector presented here, however, that this policy may have been more successful than previously claimed, with members in workplaces with an onsite representative reporting better job quality than those in workplaces without a representative. This suggests that the Unite trade union at least may well have improved onsite representation and support to its members in recent years.

Where public policy is concerned, the findings might be deemed particularly timely given recent revisions to codes of practice relating to statutory time-off for trade union duties and activities, and recent concern over the costs associated with it (Cabinet Office, 2012). The results presented here suggest, however, that assessments of the costs associated with union representative time-off need also to take into account its positive effects (Mitchell et al., 2012; TUC, 2012). In particular, one might argue that by boosting job quality via voice effects, onsite union representatives are contributing indirectly towards a range of socio-economic outcomes that, as discussed earlier, have been shown to stem from higher job quality, including: higher overall individual well-being (Green, 2008, 2009); higher job satisfaction (Pfeffer and Davis-Blake, 1990: 262–263); higher productivity; fewer quits; lower absenteeism (Clark, 2005: 380; Coats and Lekhi, 2008; Gazioglu and Tansel, 2006: 1163); smoother labour market transitions; and higher labour market participation rates (Smith et al., 2008: 588). As such, any future moves to weaken rights to time-off (as has happened already in the British civil service, for example) that reduce the ability of onsite union representatives to boost job quality via the enhancement of collective voice could in turn reduce their indirect positive influence on the socio-economic outcomes outlined above.

It must be kept in mind, however, that onsite union representation, on the basis of the results presented here, appears to be positively associated with some elements of job quality but not others. In particular, there is no evidence of an association between onsite representative presence and job security. This might be seen as unsurprising given that employer concessions to union demands for greater job security may lead to significant additional labour costs in the event of a downturn, hence employers may be particularly resistant to attempts to influence this. It might also be unsurprising given that, as argued earlier, job security policies and staffing reductions are unlikely to be subject to a significant degree of managerial discretion at local level. As such, decisions on these matters may well be beyond the sphere of onsite union representatives’ influence.

The finding that onsite representatives appear unable to influence all the facets of job quality under observation here may, however, be seen as having implications for the development of a coherent body of research on job quality. As discussed above, there is no consensus in terms of how job quality should be defined within the extant literature, even in terms of whether it should be measured using objective job characteristics or employees’ perceptions of their job quality (Green et al., 2013; Holman, 2013b: 23). If it is the case, therefore, that certain antecedents (in this case, onsite union representation) influence some elements of job quality but not others, this in turn suggests that different studies may achieve different results depending on the definition of job quality that is adopted. Given this, one might argue that a coherent body of studies on job quality is unlikely to emerge unless agreement can first be reached in the literature in terms of how job quality should be defined, and unless acknowledgement is made of the possibility that different elements of job quality might be influenced by different antecedents in different ways.

In interpreting the findings presented here, a number of caveats need to be kept in mind. One caveat is that the analysis is based on subjective measures of employees’ perceptions of particular aspects of their job quality (Holman, 2013b: 23). Future studies may also usefully assess whether the findings hold in analyses based on more objective measures of tangible and observable job characteristics and features (Green et al., 2013). An additional caveat is that the analysis is based on a single (albeit the predominant) union in the UK finance sector, hence the extent to which the findings can be generalised more broadly remains to be seen. Future research might therefore assess the impact of other unions in the finance sector on job quality including those which organise in single banks (Affinity and Accord in Lloyds Banking Group, and Advance in Santander Group, for example, which together represented 90,000 employees in 2009). A final caveat is that the survey on which the analysis in this article is based was conducted in spring 2008 before the worst effects of the ‘credit crunch’ and subsequent crisis in the finance sector took hold, leading to the British government taking stakes in some banks. It is open to question, therefore, whether union representatives will have continued to be able to mobilise collective voice to improve job quality outcomes during this period. That said, it is possible that in the period following the credit crunch, onsite unions may have had an important role to play in preventing redundancies. As such, while the analysis presented here did not find (as discussed above) a relationship between onsite union representation and perceived job security in the period leading up to the credit crunch, it is quite possible that such a relationship will have emerged subsequent to this.

The results presented here suggest a number of other avenues for future research further to those outlined above. In particular, a more nuanced analysis of the antecedents of onsite union representative influence on job quality is arguably needed that considers the activities representatives engage in, distinguishing, for example, between the impact of active and inactive representatives. Onsite representative activity levels are likely to reflect the provision of facility time and access to management, suggesting that the impact of these factors requires further exploration. More research is also needed on the mechanics of how onsite union representative voice effects emerge. For example, it could be that while unions voice members’ concerns to onsite managers, these managers may not have sufficient influence to change employment practice themselves to address these concerns. They may, however, in turn voice the union’s concerns to more senior managers in the organisation, who do have sufficient influence to make the requisite changes.

Further research is needed, therefore, to help address the caveats outlined above and to explore the precise mechanisms by which a union representative effect emerges. However, should this research reach similar conclusions to those reached here, this will reinforce the argument that onsite union representatives, as a result of the positive voice effects they engender, are capable of having a significant positive impact on job quality and, by extension, on a broad range of socio-economic outcomes.

Footnotes

Appendix

Appendix:

Means for control variables

	Mean	SD
Sex (female)	.55	.50
Ethnicity	.07	.26
Part-time	.22	.42
Age (< 30)	.05	.23
Age (30–39)	.19	.39
Age (40–49)	.39	.49
Age (50–59)	.33	.47
Age (60+)	.05	.21
Workplace size (<10)	.17	.37
Workplace size (10–24)	.20	.40
Workplace size (25–49)	.11	.31
Workplace size (50–99)	.05	.22
Workplace size (100–199)	.06	.23
Workplace size (200–499)	.11	.31
Workplace size (500–999)	.12	.32
Workplace size (1000+)	.19	.39
Pay (< £12K)	.13	.32
Pay (£12–19.9K)	.30	.46
Pay (£20–29.8K)	.24	.43
Pay (£29.9–42K)	.23	.42
Pay (>£42K)	.10	.30
Region (NE)	.15	.36
Region (London)	.13	.34
Region (SW)	.11	.32
Region (NW)	.16	.37
Region (Wales)	.04	.19
Region (N & R Ireland)	.04	.19
Region (Midlands E)	.05	.21
Region (SE)	.13	.33
Region (Midlands W)	.09	.29
Region (Scotland)	.10	.30
N = 1926

Funding

This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

Notes

Author biographies

Kim Hoque is Professor of Human Resource Management at Warwick Business School. His research explores trade unions, training, workplace equality, small and medium sized firms and agency working. He has published widely on these topics in journals such as British Journal of Industrial Relations, Human Relations, Public Administration and Work, Employment and Society.

John Earls is Head of Research at Unite the Union. He is a member of the Advisory Committee of the Centre on Skills, Knowledge and Organisational Performance (SKOPE).

Neil Conway is a Professor in Organisational Behaviour in the Department of Management, Royal Holloway. His current research interests include the psychological contract, forms of employment contract, work motivation and everyday workplace affect and behaviour.

Nick Bacon is Professor of Human Resource Management at Cass Business School. His recent publications have explored aspects of trade unions, equality, partnership arrangements, SME performance and private equity ownership. He co-edited the Sage Handbook of Industrial Relations (2008) and the Sage Handbook of Human Resource Management (2009).

References

ACAS (2008) Employee representatives: Challenges and changes in the workplace. ACAS Policy Discussion Paper, London.

Achur

(2010) Trade Union Membership in 2009. London: Department for Business, Enterprise and Regulatory Reform.

Bank of England (2011) Measuring financial sector output and its contribution to UK GDP. Quarterly Bulletin Q3: 234–246.

Baron

Kenny

(1986) The moderator–mediator distinction in social psychological research: Conceptual, strategic and statistical considerations. Journal of Personality and Social Psychology 51(6): 1173–1182.

BERR (2007) Workplace Representatives: A Review of their Facilities and Facility Time. Government Response to Public Consultation. London: Department for Business Enterprise and Regulatory Reform.

BERR (2009) Reps in Action: How Workplaces Can Gain from Modern Union Representation. London: Department for Business Enterprise and Regulatory Reform.

Brown

Charlwood

Forde

Spencer

(2007) Job quality and the economics of New Labour: A critical appraisal using subjective survey data. Cambridge Journal of Economics 31(6): 941–971.

Bryson

Forth

(2010) Union Organisation and the Quality of Employment Relations. London: National Institute of Economic and Social Research.

Bryson

Freeman

(2013) Employee perceptions of working conditions and the desire for worker representation in Britain and the US. Journal of Labor Research 34(1): 1–29.

10.

Bryson

McKay

(1997) What about the workers? In: Jowell

Curtice

Park

. (eds) British Social Attitudes: 14th Report. Aldershot: Ashgate, pp. 23–48.

11.

Bryson

Cappellari

Lucifora

(2004) Does union membership really produce job satisfaction? British Journal of Industrial Relations 42(3): 439–459.

12.

Cabinet Office (2012) Consultation on Reform to Trade Union Facility Time and Facilities in the Civil Service – Government Response. London: Cabinet Office.

13.

Carter

Danford

Howcroft

. (2012) ‘Nothing gets done and no one knows why’: PCS and workplace control of lean in HM Customs and Revenue. Industrial Relations Journal 43(5): 416–432.

14.

Charlwood

Forth

(2009) Employee representation. In: Brown

Bryson

Forth

Whitfield

(eds) The Evolution of the Modern Workplace. Cambridge: Cambridge University Press, pp. 74–96.

15.

Charlwood

Terry

(2007) 21st-century models of employee representation: Structures, processes and outcomes. Industrial Relations Journal 38(4): 320–337.

16.

Clark

(2005) Your money or your life: Changing job quality in OECD countries. British Journal of Industrial Relations 43(3): 377–400.

17.

Coats

Lekhi

(2008) Good Work: Job Quality in a Changing Economy. London: The Work Foundation.

18.

Doellgast

Holtgrewe

Deery

(2009) The effects of national institutions and collective bargaining arrangements on job quality in front-line service workplaces. Industrial and Labor Relations Review 62(4): 489–509.

19.

Esser

Olsen

(2012) Perceived job quality: Autonomy and job security within a multi-level framework. European Sociological Review 28(4): 443–454.

20.

Farber

Saks

(1980) Why workers want unions: The role of relative wages and job characteristics. Journal of Political Economy 88(2): 349–369.

21.

Freeman

Medoff

(1984) What Do Unions Do? New York: Basic Books.

22.

Gall

(2008) Labour Unionism in the Financial Services Sector: Fighting for Rights and Representation. Farnham: Ashgate.

23.

Gallie

(2007) Production regimes and the quality of employment in Europe. Annual Review of Sociology 33: 85–104.

24.

Gallie

(2013) Direct participation and the quality of work. Human Relations 66(4): 453–473.

25.

Gazioglu

Tansel

(2006) Job satisfaction in Britain: Individual and job related factors. Applied Economics 38(10): 1163–1171.

26.

Gordon

Denisi

(1995) A re-examination of the relationship between union membership and job satisfaction. Industrial and Labor Relations Review 48(2): 222–236.

27.

Green

(2008) Work effort and worker well-being in the age of affluence. In: Cooper

Burke

(eds) The Long Work Hours Culture: Causes, Consequences and Choices. Bingley: Emerald Group Publications.

28.

Green

(2009) Job quality in Britain. UK Commission for Education and Skills, November.

29.

Green

Mostafa

Parent-Thirion

. (2013) Is job quality becoming more unequal? Industrial and Labor Relations Review 66(1): 753–784.

30.

Guest

Conway

(2004) Exploring the paradox of unionised worker dissatisfaction. Industrial Relations Journal 35(2): 102–121.

31.

Hammer

Avgar

(2005) The impact of unions on job satisfaction, organizational commitment, and turnover. Journal of Labor Research 26(2): 241–266.

32.

Handel

(2005) Trends in perceived job quality, 1989 to 1998. Work and Occupations 32(1): 66–94.

33.

Holman

(2013a) Job types and job quality in Europe. Human Relations 66(4): 475–502.

34.

Holman

(2013b) An explanation of cross-national variation in call centre job quality using institutional theory. Work, Employment and Society 27(1): 21–38.

35.

House of Commons Library (2012) Financial services: Contribution to the UK economy. Standard Note SN/EP/06193.

36.

Kalleberg

(2003) Flexible firms and labour market segmentation: Effects of workplace restructuring on jobs and workers. Work and Occupations 30(2): 154–175.

37.

Kalleberg

(2012) Job quality and precarious work: Clarifications, controversies, and challenges. Work and Occupations 39(4): 427–448.

38.

Kelly

(2005) Social movement theory and union revitalisation in Britain. In: Fernie

Metcalf

(eds) Trade Unions: Resurgence or Demise? London: Routledge, pp. 62–83.

39.

Kersley

Alpin

Forth

. (2006) Inside the Workplace: Findings from the 2004 Workplace Employment Relations Survey. London: Routledge.

40.

Leschke

Watt

(2008) Job Quality in Europe. WP 2008.07. Brussels: European Trade Union Institute for Research, Education and Health and Safety.

41.

Levin

Whitener

Cross

(2006) Perceived trustworthiness of knowledge sources: The moderating impact of relationship length. Journal of Applied Psychology 91(5): 1163–1171.

42.

Meng

(1990) The relationship between unions and job satisfaction. Applied Economics 22(12): 1635–1648.

43.

Miller

(1990) Trade unions and job satisfaction. Australian Economic Papers 29(55): 226–248.

44.

Mitchell

Coutinho

Morrell

(2012) The Value of Trade Union Facility Time. London: NatCen Social Research.

45.

Olsen

Kalleberg

Nesheim

(2010) Perceived job quality in the United States, Great Britain, Norway and West Germany, 1989–2005. European Journal of Industrial Relations 16(3): 221–240.

46.

Pfeffer

Davis-Blake

(1990) Unions and job satisfaction: An alternative view. Work and Occupations 17(3): 259–283.

47.

Preacher

Hayes

(2004) SPSS and SAS procedures for estimating indirect effects in simple mediation models. Behavior Research Methods, Instruments, and Computers 36(4): 717–731.

48.

Reeves

(2009) Introduction: The new utilitarianism. In: Griffiths

Reeves

(eds) Well-being: How to Lead the Good Life and What Government Should Do to Help. London: The Social Market Foundation, pp. 10–25.

49.

Renaud

(2002) Rethinking the union membership/job satisfaction relationship: Some empirical evidence in Canada. International Journal of Manpower 23(2): 137–150.

50.

Smith

Burchell

Fagan

O’Brien

(2008) Job quality in Europe. Industrial Relations Journal 39(6): 586–603.

51.

Sobel

(1982) Asymptotic confidence intervals for indirect effects in structural equation models. In: Leinhardt

(ed.) Sociological Methodology 1982. Washington, DC: American Sociological Association, pp. 290–312.

52.

Terry

(2003) Employee representation: Stewards and the new legal framework. In: Edwards

(ed.) Industrial Relations: Theory and Practice, 2nd edn. Oxford: Blackwell, pp. 257–284.

53.

Tilly

(1997) Arresting the decline of good jobs in the USA? Journal of Industrial Relations 28(4): 269–274.

54.

TUC (2012) Facility Time for Union Reps: Separating Fact from Fiction. London: Trades Union Congress.

55.

Waddington

(2013) The views of members towards workplace union organization in banking between 1999 and 2008. British Journal of Industrial Relations 51(2): 333–354.

56.

Willman

Bryson

Gomez

(2007) The long goodbye: New establishments and the fall of union voice in Britain. International Journal of Human Resource Management 18(7): 1318–1334.

57.

Wood

(2008) Job characteristics, employee voice and well-being in Britain. Industrial Relations Journal 39(2): 153–168.