Employee representatives’ influence on continuing vocational training: The impact of institutional context

Abstract

What factors determine the involvement of employee representatives in continuing vocational training? This article uses data from the European Company Survey 2009 to examine firm-level and country-level characteristics in 27 European Union (EU) countries. Multilevel logistic regressions indicate that employee representatives increase the probability of companies providing training needs assessments and time off for training. This association is stronger for disadvantaged groups of employees (low-skilled or temporary employees) if the representatives receive support from encompassing collective agreements, strong trade unions or regular training.

Keywords

Continuing vocational training employee representatives Europe trade unions

Introduction

It is increasingly argued that with lifelong learning, employees are more prepared for unstable labour markets and changing job requirements (European Commission, 2010; OECD, 2012). This article examines the relationship between employee representatives and firm-level continuing vocational training (CVT). Studies have shown that employees are more likely to receive CVT in firms with works councils or a trade union presence (Böheim and Booth, 2004; Booth et al., 2003; Cantner et al., 2014; Green et al., 1999; Heyes and Stuart, 1998; Leisink and Greenwood, 2007; Stegmaier, 2012; Waddoups, 2014). Hoque and Bacon (2008) found greater equality of training provisions for different groups of employees in unionized firms that had learning representatives. The majority of these studies are restricted to single-country analyses, whereas I test the association of employee representatives and CVT for 27 EU countries and for different groups of employees (skilled, low-skilled, temporary and older employees).

Faced with complex duties and scarce resources, many employee representatives find it very difficult to influence the implementation of CVT. This influence can be expected to be higher if they receive institutional supports from trade unions, multi-employer collective agreements and regular training for their activities (Iller et al., 2016). Rees (2013) argues that national differences in adult learning systems are shaped by economic and social regimes and power relations. In many countries, employee representatives have codetermination rights, but they also require external resources so that they can support training-related activities at firm level (Rainbird et al., 2003). Employee representatives need proper skills to identify training needs, and unions can provide these.

There has been little comparative research analysing the relationship between training and the characteristics of firms and employee representatives and external supports. This article therefore analyses the association between employee representatives and training needs assessment for different employees as well as time off provided for training, accounts for different characteristics of employee representatives and considers the moderating effect of country-level institutions and external support.

The next section discusses the role of employee representatives in CVT and the moderating role of external support, followed by the description of the data, variables and empirical strategy. I then present the empirical findings for the association of employee representatives with training needs assessment and time given off for training (step 1), for the association of employee representatives’ specifications with training needs assessments and time given off for training (step 2) and for factors that are related to the involvement of employee representatives in training decisions (step 3).

CVT, employee representatives and trade unions

Employee representatives and CVT

Participation in CVT may be low, particularly among less skilled employees, because of lack of knowledge of their own learning needs and rights. Support and guidance can be crucial: trade unions and workplace employee representatives can perform a coordinating role, increasing participation and reducing inequalities among different groups of employees. By ‘employee representatives’, I mean any type of formal employee representation within a company, including works councillors, trade union representatives and shop stewards.

Employees are more likely to participate in CVT if employee representatives and trade unions support and negotiate over training (Heyes and Stuart, 1998). Representatives can monitor who receives and has access to training, help employees to overcome information asymmetries and promote mutual trust between employer and employees (Green et al., 1999; Stegmaier, 2012). They are especially important for the assessment of training needs because they have close contact with employees. When asserting their rights, employees’ voices are more effective as a collective than as individuals (Stegmaier, 2012), and bodies independent of the employer, such as trade unions and employee representatives, are best able to provide effective voice:

H1. In companies with formal employee representatives, it is more likely that training needs are assessed and time off is provided for training.

Temporary, low-skilled and older employees often face training gaps: companies are more likely to train permanent staff because the returns are higher when tenured employees apply their training and skills within the same company (Canduela et al., 2012; Finegold et al., 2005; Frazis et al., 2000). The return for companies is also higher for employees with high educational attainment because they are better able to learn new skills (Becker, 1964; Wexley and Latham, 2001). Temporary employees could improve their labour market situation by participating in CVT, and risk obtaining only insecure and low-quality jobs if they do not have access to training, which reinforces their disadvantages.

Further training inequalities result from age. Older employees are less likely to have access to training (Canduela et al., 2012), often because of employer decisions (Taylor and Urwin, 2001). Age-related negative stereotypes among employers, such as the perception of change-resistant, inflexible and past-oriented older employees, as well as the shorter time remaining for investment returns, hinder employer investments in training (Lazazzara et al., 2013; Raemdonck et al., 2015). At the individual level, the lower motivation and capability of older employees to learn (Raemdonck et al., 2015) are important obstacles for CVT (Setti et al., 2015). However, if employees must extend their working life for demographic reasons, older employees require training to update their skills and knowledge. They are less likely to stay in the labour market and to find new jobs without access to training.

Therefore, CVT can reduce inequalities among employees. Employee representatives can in principle monitor equal training opportunities for all employees, ensuring that training needs are assessed equally among different employees and motivate groups of employees with low CVT participation.

‘Insiders’ and ‘outsiders’

Some scholars argue that trade unions mainly protect the jobs of ‘insiders’ (permanent and skilled employees) as their core members (Lindbeck and Snower, 2001); others stress that trade unions defend the interests of the entire workforce, including ‘outsiders’ (Benassi and Vlandas, 2016). Insider–outsider theory distinguishes between permanent and temporary employees (Lindbeck and Snower, 2001: 166), but some versions also regard unskilled or low-skilled employees as outsiders. With regard to the different groups of employees, temporary employees are outsiders and permanent high-skilled employees are insiders. In practice, lower skilled employees with permanent contracts, as well as older employees, are difficult to classify as either insiders or outsiders. A risk-based conceptualization focuses on the chances of exclusion, which reflect an individual’s social and occupational group, rather than on the current (short-term) labour market situation (Schwander and Häusermann, 2013). Lower skilled employees face a higher risk of unemployment or disadvantageous employment over their life course, compared to higher skilled employees. We therefore classify lower skilled employees with permanent contracts as at least partially outsiders.

Older employees belong to the group of outsiders only if they lose their job and do not find another; otherwise, their jobs are better protected than those of younger workers and they have higher seniority. Their risk of experiencing either atypical employment or unemployment is much lower than for younger employees (Schwander and Häusermann, 2013). We therefore classify older employees as insiders.

Do employee representatives actually foster training activities only for ‘insiders’, or do they act (also) on behalf of ‘outsiders’?

H2. If employee representatives give priority to the training needs of skilled employees with permanent contracts or of older employees, they focus on core workers (insiders).

H3. If employee representatives give priority to the training needs of temporary employees or of low-skilled employees, they focus on disadvantaged workers (outsiders).

Country differences and external support for employee representatives

As the ‘varieties of capitalism’ approach (Hall and Soskice, 2001) suggests, training decisions at company level and the effect of employee representatives are shaped by the national institutional context (Desjardins and Rubenson, 2013). While Hall and Soskice focus on initial vocational training, this is also the case with CVT (Dieckhoff et al., 2007).

Both employees and employers have an interest in vocational training. Streeck (1992) argues that trade unions and employers can cooperate on skill formation issues, but trade unions must maintain their strength and independence to be able to impose employee-friendly rules. Strong unions are more likely to draw employers’ attention to training issues. Unions may negotiate collective agreements at the company, sectoral or national level, with the aim of promoting lifelong learning and covering a wide range of employees (Winterton, 2007). Wage compression in collective agreements encourages employers to increase CVT (especially for low-skilled employees), to keep productivity levels in line with wage floors (Dieckhoff et al., 2007; Stegmaier, 2012). More specifically, CVT is covered by collective agreements in almost all European countries, with a positive impact on participation (Giaccone, 2009). Many such agreements prescribe regular training needs assessments, the participation of employee representatives and providing time off for learning (Munro and Rainbird, 2004). Therefore, differences in trade union strength and collective bargaining coverage may explain CVT variation across countries.

The involvement of workplace representatives is necessary for effective implementation at firm level of sectoral and nation-wide collective agreements (Munro and Rainbird, 2004). Employee representatives are very important because they can ensure that agreements are applied within their company. If employees realize that they have access to training thanks to the initiative of trade unions and their representatives, this can have a positive effect on union recruitment. In addition, union activities with regard to training issues can extend their influence and power into other policy fields (Munro and Rainbird, 2004).

Even in the absence of formal rules and rights, trade unions may encourage employee representatives to initiate or supervise CVT (Iller et al., 2016). They may disseminate good practices and information on how to perform training needs assessments. In the private sector in the United Kingdom, for example, trade union initiatives have resulted in more systematic training, especially for low-skilled employees (Rainbird et al., 2003):

H4. The relationship between employee representatives and training needs assessment, as well as time off for training, is stronger in countries with external support through encompassing collective agreements or strong trade unions.

Statutory rules also support the work of employee representatives. These include rights to information, consultation and codetermination with regard to different topics such as working conditions, working time, remuneration and workplace training. In Germany, for instance, works councils can initiate training needs assessments (Stegmaier, 2012):

H5. The relationship between employee representatives and training needs assessment, as well as time off for training, is stronger in countries in which employee representatives have encompassing social and economic rights.

Data, methods and variables

Data

This article uses data from the 2009 Eurofound European Company Survey (ECS) (European Foundation and TNS Infratest Sozialforschung, 2010) and includes all EU member states except Croatia, for which country-specific information is missing. The ECS contains questions about company characteristics and company policies and covers a representative sample of private companies and public-sector organizations with 10 or more employees (TNS Infratest, 2009). The highest ranking manager responsible for human resources and, where available, the chair of the formal employee representative body were interviewed independently by telephone. The great advantage of the ECS is that it provides firm-level information from the perspective of both the employer and the employee sides. The two different questionnaires, and data for the managers and the employee representatives, were linked and merged into one file. Macro-level variables with country-specific information were added to the dataset using the ICTWSS database (Visser, 2013). The final sample consists of interviews with 24,640 managers and 6214 employee representatives.

Dependent variables

The ECS data contain six dichotomous variables with information about training needs assessment and time off for further training. While the data lack information about course details and who receives training, they allow us to analyse training needs assessments for different groups of employees.

The manager questionnaire asked, ‘Is the need for further training periodically checked in a systematic way?’ If answered in the affirmative, the follow-up question asked ‘whether or not [their] needs for further training are systematically checked at regular intervals’ for ‘permanent employees in skilled or high-skilled positions’, ‘permanent employees in low-skilled or unskilled positions’, ‘employees with fixed-term contracts’ and ‘older employees’. Additionally, the questionnaire asked all managers, ‘Have any of your employees been given time off from their normal duties in the past 12 months in order to undergo further training?’ In addition to these questions for managers, employee representatives were asked whether they had been involved in establishing rules and procedures for access to training, either by way of consultation or negotiation. This allows direct examination of the impact of employee representatives on CVT. Although access to training is only the first step in the process – others include, for example, training intensity and course content – rules and procedures about access to training are very important for the subsequent steps.

Independent and moderating variables

The variables of interest for the explanation of the first set of analyses, based on management interviews, are the existence of an employee representative at the firm level (independent variable) and collective agreement coverage (moderating variable). A dichotomous employee representative variable measures whether there is a formal employee representative at the establishment. It covers statutory works councils, trade union representatives and shop stewards, thus encompassing different systems of employee representation across countries. In terms of collective agreement coverage, a dummy variable indicates whether more than 60 percent of a company’s employees are covered.

For the second set of analyses, based only on firms where both a manager and an employee representative were interviewed, the characteristics of employee representatives are included, thus identifying differences in employee representation bodies. For the proportion of employee representatives belonging to a trade union, a variable was constructed with three categories (⩽19%, 20%–79% and ⩾80%). We assume that a high number of trade union employee representatives indicate close contacts between the representatives and the trade union and therefore measure external support. Training for employee representatives is an additional form of external support for their CVT activities; this was measured by the question whether representatives receive regular training on issues specific to their role. A dichotomous variable shows whether employee representatives have been involved in establishing rules and procedures with regard to access to training, which measures their direct influence. We constructed a variable for years of experience as an employee representative in the company (⩽3, 4–8, 9–15, ⩾16 years). A dummy variable, indicating whether more than 60 percent of employees of the company are members of a trade union, serves as a proxy for union power within a company.

All variables in the third set of analyses, based on employee representatives’ answers on their involvement in training decisions, are identical to the second set, but excluding their involvement in training decisions (which is the dependent variable).

All three sets of analyses include other differences across firms that may influence CVT. In line with previous research (Böheim and Booth, 2004; Canduela et al., 2012; Finegold et al., 2005; Stegmaier, 2012), we control for firm size (4-category variable with 1 = 10–19, 2 = 20–49, 3 = 50–249 and 4 = ⩾250employees); employees in high-skilled jobs (3-category variable with 1 = ⩽19%, 2 = 20%–39% and 3 = ⩾40%), part-time employees (⩾40%); economic situation of the firm (3-category variable with 1 = not good, 2 = good and 3 = no answer/do not know – included to avoid losing too many cases); multi-site establishment (one of a number of different establishments belonging to the same company); female employees (3-category variable with 1 = ⩽39%, 2 = 40%–59% and 3 = ⩾60%); and economic sector (NACE rev. 1, collapsed to 7 categories for the second and third sets of analyses to increase the number of observations for each category).

For the moderating macro-level variables, we rely on ICTWSS data (Visser, 2013). High collective bargaining coverage can support employee representatives if agreements include CVT issues or participatory rights of employee representatives. Trade unions may support employee representatives via different channels. Their strength is measured by density (net membership as percent of employees). The rights of employee representatives (1 = information and consultation, 2 = economic and social, 3 = economic and social rights, including codetermination) account for different legal possibilities that can influence training decisions. Information for the country variables is mainly for the year 2009, but in the case of missing information, data from the years 2007–2010 were used. A list of the moderating country variables and their values (Table A.1) is not included for reasons of space but is available from the author. All linear macro-level variables were centred at the mean and standardized using z-scores to assess the effects of country-level variables, controlling for firm-level characteristics (Enders and Tofighi, 2007). This allows comparison of the relative effects of variables with different scales, eases interpretation of effects and causes fewer convergence problems (Hox, 2010).

Methods

We estimate multilevel logistic regression models because the data have a hierarchical structure with two levels: companies nested in countries. The effect of employee representatives on CVT must be assumed to vary according to country, since their existence, their rights and industrial relations regimes differ cross-nationally. All models show significant variation (at 0.001 level) between the levels (firm and country level). Thus, models allowing for random variation (multilevel model) explain more than models that do not allow for random variation. If we do not explicitly account for this multilevel structure, we would get biased estimates. Multilevel models allow combination of predictors at different levels.

For all three steps of the analyses, we applied two different sets of regressions. Models 1a, 2a and 3a included 27 countries as random intercepts and firm-level variables to explain differences across companies. The macro-level variables were then added in Models 1b, 2b and 3b to explain cross-national differences. The multilevel model allows cross-level interactions among country-specific features and the existence and characteristics of employee representatives at firm level. These interactions show whether the effect of employee representatives (and their characteristics) varies systematically across countries. Each regression only included one macro-level variable. This means that for the interactions in Table 1, for example, we ran 18 multilevel logistic regressions (3 macro-level variables × 6 dependent variables). The italicized coefficients show the general effect of the country-level variables without interactions. All models were estimated using the R package lme4 (Bates et al., 2015). We tested the conditional effects for different levels of collective bargaining coverage and union density for all significant interactions (results available from the author, Figure A.1). The interaction results are shown in parentheses in Tables 1 and 2 if the conditional effect is not significant for any country in our sample.

Table 1.

Multilevel logistic regression models (odds ratios) for employee representatives and training activities.

	Training needs assessment					Time off
	Systematic	Skilled	Low-skilled	Temporary	Older
Models 1a: Only firm-level variables
(Intercept)	0.93	4.00***	4.48***	0.42***	1.57	0.72
Employee representative	1.57***	1.31***	1.28***	1.12***	1.29***	1.33***
⩾60% CBC	1.27***	1.27***	1.05	1.11**	1.10*	1.14***
Firm size (ref. 10–19 employees)
20–49	1.37***	1.17**	1.16***	1.20***	1.13**	1.32***
50–249	2.01***	1.54***	1.56***	1.74***	1.70***	1.91***
⩾250	3.19***	2.30***	2.04***	2.59***	2.51***	3.46***
Employees in high-skilled jobs (ref. <20%)
20%–39%	1.51***	1.78***	1.21***	1.21***	1.35***	1.50***
⩾40%	1.74***	2.93***	0.87***	1.44***	1.58***	1.75***
⩾40% part-time employees	0.98	0.82*	1.16**	1.16**	0.92	0.85***
Economic situation (ref. not good)
Good	1.43***	1.11*	1.05	1.07*	1.02	1.11***
No answer/DK	0.96	0.87	1.03	1.05	1.05	0.82**
Multi-site	1.53***	1.24***	1.20***	1.21***	1.29***	1.26***
Female employees (ref. <40%)
40%–59%	0.96	1.12	1.04	1.01	1.01	0.97
⩾60%	0.92*	1.09	0.94	1.10**	1.00	0.90***
13 industry dummies	yes	yes	Yes	yes	yes	yes
N (firms)	23,275	17,525	17,310	16,497	17,213	23,196
Models 1b: With cross-level interactions
Bargaining coverage (CBC)	0.94	1.05	1.45***	1.16	0.99	1.14
ER × CBC	0.99	0.97	1.12***	0.97	0.94	1.00
Union density (UD)	0.90	1.25***	1.34**	0.99	0.92	1.06
ER × UD	0.95	0.94	1.08*	1.08*	1.01	1.01
ER rights^a	0.82/0.82	1.23/1.37	1.43/1.77	0.89/1.27	0.62/1.45	1.06/1.31
ER × ER rights2	1.10	0.91	1.28**	0.98	0.94	1.32***
ER × ER rights3	1.05	1.28	1.26**	0.98	0.85	1.25***

CBC: collective bargaining coverage; DK: don’t know; ER: employee representative.

Models 1b contain the control variables as in Models 1a.

ER rights ranges from 1 (few) to 3 (many) with 1 as the reference category.

p < 0.1; **p < 0.05; ***p < 0.01.

Table 2.

Multilevel logistic regression models (odds ratios) for employee representative details and training activities.

	Training needs assessment					Time off
	Systematic	Skilled	Low-skilled	Temporary	Older
Models 2a: Only firm-level variables
(Intercept)	1.30	7.79***	2.14***	0.41***	1.21	1.04
% of employee representatives belonging to trade union (ref. < 20%)
20%–79%	0.83	0.88	1.16	1.05	0.92	0.89
⩾80%	0.98	1.11	1.23	1.21*	1.06	1.12
Regular training	1.04	0.98	0.95	1.16*	1.19*	1.05
Involved in training decisions	1.21**	1.45***	1.44***	1.07	1.18*	1.18**
Experience as employee representative (ref. ⩽3 years)
4–8 years	0.93	1.06	1.03	0.97	1.11	0.83**
9–15 years	0.99	1.15	0.90	1.03	1.06	0.79**
⩾16 years	1.02	1.11	0.93	0.92	1.11	0.88
Union membership of employees (ref. ⩽20%)
20%–79%	1.15	1.03	0.80*	0.91	0.88	1.05
⩾80%	0.98	1.17	0.80	0.83	0.70**	0.88
N (firms)	5583	4695	4663	4528	4645	5553
Models 2b: With cross-level interactions
Bargaining coverage (CBC)	0.96	1.10	1.58***	1.09	0.98	1.21
Trade union ER2 × CBC	0.99	0.85	1.20	1.04	1.23*	1.10
Trade union ER3 × CBC	(0.81*)	0.91	1.19*	1.00	1.12	1.11
Regular training × CBC	1.16	1.16	1.15	0.97	0.93	1.15*
Union density (UD)	0.91	1.28**	1.39**	1.03	1.02	1.17
Trade union ER2 × UD	1.06	1.34	1.15	1.14	1.12	1.01
Trade union ER3 × UD	0.94	1.28	1.30**	1.11	1.04	0.98
Regular training × UD	1.10	(1.26*)	1.20**	0.92	0.94	1.02
ER rights^a	0.98/0.92	1.10/1.59	1.47/2.02	0.90/1.30	0.61/1.56	1.20/1.47
Trade union ER2 × ER rights2	1.09	1.08	1.21	1.93***	1.55*	1.68**
Trade union ER3 × ER rights2	0.50***	1.01	0.92	0.86	1.11	1.64**
Trade union ER2 × ER rights3	0.63*	1.08	1.21	1.71**	0.89	1.04
Trade union ER3 × ER rights3	0.60*	1.55	0.91	1.17	0.97	1.06
Regular training × ER rights2	1.21	1.39	0.98	0.83	0.78	1.36*
Regular training × ER rights3	1.13	1.41	1.09	1.19	1.03	1.31

CBC: collective bargaining coverage; ER: employee representative.

All models also control for firm size; high-skilled, part-time and female employees; economic situation of the firm; multi-site establishment; and industry (7 dummies). Models 2b contain the control variables as in Models 2a.

See notes to Table 1. Trade union ER takes the values 1 (<20%), 2 (20%–79%) and 3 (⩾80%) with 1 as the reference category.

p < 0.1; **p < 0.05; ***p < 0.01.

Given the cross-sectional nature of the data, we seek to describe patterns of association of employee representatives and external support with CVT and cannot identify causal effects.

Results

The association of employee representatives with training needs assessment and time off for training (step 1)

To assess the effect of employee representatives on training needs assessment and time off for training, Models 1a display the results with only firm-level variables and Models 1b with firm- and country-level variables.

Training needs are more likely assessed periodically and systematically in firms with employee representatives (Table 1), confirming Hypothesis 1 and previous studies. The average gap for training needs assessment among the 27 countries for the ECS 2009 data is 20 percent between low-skilled and skilled employees, 22 percent between older and skilled employees and as high as 45 percent between temporary and skilled employees, signalling considerable training inequalities. As Table 1 indicates, the presence of employee representatives also shows a significant positive relation with training needs assessments for all groups of employees. In line with Hypotheses 2 and 3, they apparently act on behalf of both labour market insiders and outsiders. In addition to training needs assessment, firms with employee representatives also show higher provision of time off for further training.

Similarly, if more than 60 percent of a firm’s employees are covered by a collective agreement, the probability of time off for training and training needs assessment is higher for all except low-skilled employees. Although these agreements do not necessarily address training issues (we lack information about their content), they are positively related to CVT.

The control variables affect the dependent variables as expected: employees have higher chances of training needs assessment and time off for training in firms which are larger, have many high-skilled employees, enjoy favourable economic situations, and have multi-site establishments.

Models 1b in Table 1 investigate whether the effect of employee representatives varies systematically across countries with different macro-variables. High collective bargaining coverage intensifies the positive effect of employee representatives on training needs assessment only for low-skilled employees. Strong unions, measured by density, increase the probability of a positive effect of employee representatives on the training needs assessment for low-skilled and temporary employees. Apparently, collective agreements and trade unions increase representatives’ awareness of inequalities and disadvantages, resulting in more equally distributed training activities. The conditional effect of employee representatives on low-skilled employees is significantly positive for countries with collective bargaining coverage of more than 33 percent or with union density of more than 16 percent and on temporary employees for all levels of union density.

If employee representatives exist at firm level and are equipped with more than just information and consultation rights, then they increase the probability of training needs assessment for low-skilled employees and for time off provision for training. Step 1 concludes as follows: employee representatives alone increase time off and training needs assessments for all employees; their effect is stronger for employees with a high outsider risk if they receive support from collective agreements and strong unions and if they have economic and social rights.

The association between differences in employee representation, training needs assessment and time off for training (step 2)

For the second step, we use additional information from the employee representative questionnaire to examine factors through which differences in employee representation are related to training needs assessment and time off for training. Highly unionized employee representative bodies have closer and more regular contacts with their trade unions compared with less unionized bodies. Similarly, those who receive regular training on issues specific to their role as employee representatives receive external support and can be expected to have a higher awareness of employee training.

If more than 80 percent of the employee representation body belongs to a trade union, temporary employees are more likely to receive training needs assessments (Table 2). There is no such relation for any other group of employees. Furthermore, regular training of employee representatives increases the likelihood of training needs assessments for temporary and older employees. It seems that regularly trained employee representatives consider the interests of both insiders (older employees) and outsiders (temporary employees).

As expected, time off for training is higher – and all except temporary employees experience higher levels of training needs assessment – if representatives are involved in establishing rules and procedures, which confirms the association in Models 1a and 1b. Experience as an employee representative is not related to training needs assessments and is negatively associated with time off. We also controlled for the unionization of employees, as this could influence the association between representatives and training activities. Our results indicate that training needs are less likely to be assessed for low-skilled and older employees in firms with high union density.

In addition to the effect of employee representative characteristics at the firm level, we are interested in their interaction with country-level variables. High collective bargaining coverage combined with unionized employee representatives improves the training needs assessment of low-skilled and older employees (Table 2, columns 3 and 5, rows 2 and 3 of Models 2b), partly supporting Hypothesis 4. The combination of encompassing collective agreements and regularly trained employee representatives increases the probability of time off for training, maybe because employee representatives participating in training courses are more familiar with the content of agreements, including training issues. Strong trade unions positively affect the association of highly unionized as well as regularly trained employee representatives with training needs assessments for low-skilled employees (Table 2, column 3, rows 7 and 8 of Models 2b), signalling more an outsider than an insider focus of trade unions. Although the interaction term is significant for bargaining coverage and unionized employee representatives on training assessments (Table 2, column 1, row 3 of Models 2b) and for union density and trained employee representatives on training needs assessments for high-skilled employees (Table 2, column 2, row 8 of Models 2b), the conditional effects are not significant for any country in our sample. Moreover, only in countries with high levels of bargaining coverage and union density are unionized employees positively related to the assessment of training needs for low-skilled employees.

As in Models 1b in Table 1, the varying rights of employee representatives across countries do not alone explain any training activities (Table 2, row 9 of Models 2b). Interacted with union membership of employee representatives, they are negatively associated with systematic training needs assessments. Apparently, this does not apply to some outsider and insider groups of employees because unionized employee representatives in countries where they possess economic and social rights have a positive impact on the probability of training needs assessment of temporary (only for medium-unionized employee representatives) and older employees (only for medium-unionized employee representatives with economic and social rights without codetermination) as well as time given off for training (only in countries with economic and social rights without codetermination). Economic and social rights also increase the positive effect of regularly trained employee representatives on time off for training, probably because they can more effectively apply their knowledge gained in training courses.

The association between differences in employee representation and involvement of employee representatives in training decisions (step 3)

We can conclude from the previous analyses that employee representatives have a positive association with training and that this relation is stronger in 10 cases for employees with a high risk of being or becoming outsiders and only in two cases for insider employees (elderly) if they receive external support (see Models 1b in Table 1 and Models 2b in Table 2). As the third and last step, we examine factors that increase the involvement of employee representatives in training decisions as one mechanism for the general positive effect of employee representatives on training. For reasons of space, we do not include Table 3, but this is available from the author.

Although unionized employee representatives have a positive effect on outcomes, they are negatively associated – compared with less unionized employee representatives – with involvement in the training decision-making process. One explanation could be that employers are often hostile to union-organized employee representatives and exclude them from decision-making (Bacon and Storey, 2000; Munro and Rainbird, 2004). High bargaining coverage at the country level plays a moderating role, positively influencing the effect of unionized employee representatives on involvement in training decisions. Therefore, the involvement of employee representatives in firm-level training decisions is more likely if they receive external support, but not in countries with very high bargaining coverage. Collective agreements are central, probably because they contain passages regarding the inclusion or consideration of employee representatives in firm-level training. Similarly, employee representatives attending regular training courses are more likely to be involved in training decisions.

Differences in the rights of employee representatives across countries cannot explain the patterns of involvement. One interpretation could be that there are no differences between unionized and non-unionized employee representatives because all employee representatives have the same rights regardless of trade union membership.

Robustness checks were conducted with different codings for the variables ‘collective agreement coverage at firm-level’, ‘proportion of employee representatives belonging to a trade union’ and ‘experience as employee representatives’. The use of continuous variables would result in a considerably lower number of cases (several hundred) because a further question regarding categories was asked in the case of non-response or ‘do not know’ for the first (continuously scaled) question. The results for bargaining coverage at firm level, coded as continuous or with the cut-off point of 40 percent or with seven categories, are qualitatively the same as the 60 percent cut-off point used. With an 80 percent cut-off, the result for older employees is not significant, which does not change the overall interpretation. The significance of the results for the proportion of employee representatives belonging to a trade union varies for less informative binary variables with cut-off points of 40, 60 and 80 percent. However, we use a 3-category variable with more or less equal numbers of cases because our results indicate that the association is not linear, which would be masked by a continuous or binary variable. Therefore, our variable with three categories is more informative. The results for different codings of experience as employee representatives (continuous variable and binary variable with the median as cut-off point) are qualitatively the same except for time off for training. Our more informative variable with four categories considers a possible non-linear relationship. No alternatively coded variable changes the results for the other variables of the models.

Conclusion

This article had three aims. The first was to explore the association between employee representatives, training needs assessment for different groups of employees and time off for training. Second, we accounted for different characteristics of employee representatives, and third, we considered country-level institutions and external support for employee representatives. A major finding is that, controlling for a range of firm-level characteristics, the existence of employee representatives is positively related to regular and systematic training needs assessments not only for insiders (skilled and older employees) but also for those with a high risk of being or becoming outsiders (low-skilled and temporary employees) and with time off for training. This confirms Hypothesis 1 and previous single-country studies, but in our case for 27 EU countries, and is in line with Hypotheses 2 and 3.

More specifically, the first set of models confirmed Hypotheses 1, 2 and 3, with employee representatives focusing on both insiders and outsiders. Employee representatives increase the likelihood of time off and training needs assessments for all employees, and their effect is stronger and more prevalent for employees with high risks of being or becoming outsiders (low-skilled or temporary employees) if they receive support from collective agreements and strong unions (partially in line with Hypothesis 4). As the second step, we further investigated whether there were differences among employee representative. Highly unionized representation bodies are significantly associated with training needs assessments only for temporary employees, and regular training of employee representatives increases the likelihood of training needs assessments for temporary and older employees. At the country level, collective bargaining coverage, combined with medium and highly unionized employee representatives, improves the training activities of disadvantaged employees (low-skilled) and also insiders (older employees). High levels of bargaining, in combination with regularly trained employee representatives, also increase the probability of time off for training. Trade union density positively affects the association of highly unionized as well as regularly trained employee representatives with training needs assessments for low-skilled employees, confirming the positive relation of employee representatives with training needs assessment for outsiders, especially in countries with strong trade unions.

The third step examined the direct involvement of employee representatives in training decisions as one mechanism for the general positive effect of employee representatives on employees’ training activities. Unionized employee representatives affect the outcome in a positive way, but they do not result in higher involvement in training decision-making processes. As noted above, this might be the result of hostile employers who fear union influence on management decisions. Nevertheless, higher collective bargaining coverage at the country level reduces this negative relationship. Regularly trained employee representatives not only improve training needs assessments, they also increase the probability of employee representative involvement in decision-making processes at the firm level.

In sum, (unionized) employee representatives require external support such as collective agreements and regular training to affect the CVT decision process at firm level. They have, alone, a positive indirect association with outcomes (training needs assessment and time off for training), which is even stronger for some groups of employees in interactions with high levels of collective bargaining coverage, strong unions and in countries in which employee representatives have more than just information and consultation rights. Non-market coordination mechanisms enable employee representatives, faced with complex duties and scarce resources, to make use of their influence and to inform and motivate employees.

With regard to the debate about labour market insiders and outsiders, our results do not confirm that employee representatives and trade unions protect insiders. Indeed, the probability that they will defend the interests of disadvantaged groups, although these are not the core membership of trade unions, and reduce training inequalities increases with union density and collective bargaining coverage.

Regarding the main limitations of the study, we cannot claim causality, but only associations, given the cross-sectional design. We cannot completely reject reverse causality, in the sense that firms supporting CVT may also favour formal employee representatives, as suggested by the concept of high-involvement work systems (HIWS) (Boxall et al., 2015). Nevertheless, this concept stresses the importance of individual voice and involvement rather than collective voice. Formal employee representatives do not exist only in firms that favour collective employee involvement (and CVT) or where employees are highly involved in work processes. Our set of controls should mitigate this potential bias: formal employee representatives exist not only in firms with HIWS but also in traditional Fordist systems (all our models account for different industries, hence reflecting different work systems). According to the HIWS concept, high employee involvement is more likely in firms relying on high-skilled employees. Our models account for this possible bias by controlling for the share of high-skilled employees. Previous studies solved this endogeneity problem with instrumental variable approaches (Booth et al., 2003; Stegmaier, 2012), confirming our results and interpretation. Longitudinal data are necessary to examine further the causal mechanisms between employee representatives, external support and CVT.

Furthermore, the data do not allow examination of training incidence and intensity for different employees, or of the quality and content of further training. Although we could examine training needs assessment, time off for training and the involvement of employee representatives in training decisions, national data and single-country studies could complement our findings with respect to training details. Further research, in particular case studies, may also investigate details regarding the role of the employer and employee representatives in the decision-making process of CVT.

Policies and initiatives promoting lifelong work-related learning must address barriers and inequalities for different groups of employees. Employee representatives seem to be a promising solution, especially if they receive external support from trade unions through collective agreements and regular training.

Footnotes

Acknowledgements

This article uses the European Company Survey 2009. I am grateful to the anonymous reviewers for valuable comments on previous versions of this article and Richard Hyman for his help during the editorial process.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: Research for this article was funded partly by the Hans-Böckler-Stiftung (HBS, Grant 2013-637-5) and partly by the Austrian Fonds zur Förderung der wissenschaftlichen Forschung (FWF, Grant M 1565-G16).

Author biography

Tobias Wiß is a researcher at the Department of Politics and Social Policy, Johannes Kepler University, Austria.

References

Bacon

Storey

(2000) New employee relations strategies in Britain: Towards individualism or partnership? British Journal of Industrial Relations 38(3): 407–427.

Bates

Maechler

Bolker

Walker

(2015) lme4: Linear Mixed-effects Models Using ‘Eigen’ and S4 (R package version 1.1–9). Available at: https://CRAN.R-project.org/package=lme4

Becker

(1964) Human Capital. Chicago, IL: University of Chicago Press.

Benassi

Vlandas

(2016) Union inclusiveness and temporary agency workers: The role of power resources and union ideology. European Journal of Industrial Relations 22(1): 5–22.

Böheim

Booth

(2004) Trade union presence and employer-provided training in Great Britain. Industrial Relations 43(3): 520–545.

Booth

Francesconi

Zoega

(2003) Unions, work-related training, and wages: Evidence for British men. Industrial & Labor Relations Review 57(1): 68–91.

Boxall

Hutchison

Wassenaar

(2015) How do high involvement work processes influence employee outcomes? An examination of the mediating roles of skill utilisation and intrinsic motivation. International Journal of Human Resource Management 26(13): 1737–1752.

Canduela

Dutton

Johnson

Lindsay

McQuaid

Raeside

(2012) Ageing, skills and participation in work-related training in Britain: Assessing the position of older workers. Work, Employment and Society 26(1): 42–60.

Cantner

Gerstlberger

Roy

(2014) Works councils, training activities and innovation: A study of German firms. Jena Economic Research Papers 2014-006. Jena: Friedrich-Schiller-University and Max Planck Institute of Economics.

10.

Desjardins

Rubenson

(2013) Participation patterns in adult education: The role of institutions and public policy frameworks in resolving coordination problems. European Journal of Education 48(2): 262–280.

11.

Dieckhoff

Jungblut

O’Connell

(2007) Job-related training in Europe: Do institutions matter? In: Gallie

(ed.) Employment Regimes and the Quality of Work. Oxford: Oxford University Press, pp. 77–103.

12.

Enders

Tofighi

(2007) Centering predictor variables in cross-sectional multilevel models: A new look at an old issue. Psychological Methods 12(2): 121–138.

13.

European Commission (2010) Europe 2020: A Strategy for Smart, Sustainable and Inclusive Growth. COM(2010) 2020. Brussels: European Commission.

14.

European Foundation and TNS Infratest Sozialforschung (2010) European Company Survey, 2009. Available at: https://dx-doi-org.web.bisu.edu.cn/10.5255/UKDA-SN-6568-1

15.

Finegold

Levenson

Van Buren

(2005) Access to training and its impact on temporary workers. Human Resource Management Journal 15(2): 66–85.

16.

Frazis

Gittleman

Joyce

(2000) Correlates of training: An analysis using both employer and employee characteristics. Industrial & Labor Relations Review 53(3): 443–462.

17.

Giaccone

(2009) Contribution of collective bargaining to continuing vocational training. Available at: http://www.eurofound.europa.eu/publications/report/2009/industrial-relations/contribution-of-collective-bargaining-to-continuing-vocational-training

18.

Green

Machin

Wilkinson

(1999) Trade unions and training practices in British workplaces. Industrial & Labor Relations Review 52(2): 179–195.

19.

Hall

Soskice

(eds) (2001) Varieties of Capitalism: The Institutional Foundations of Comparative Advantage. Oxford: Oxford University Press.

20.

Heyes

Stuart

(1998) Bargaining for skills: Trade unions and training at the workplace. British Journal of Industrial Relations 36(3): 459–467.

21.

Hoque

Bacon

(2008) Trade unions, union learning representatives and employer-provided training in Britain. British Journal of Industrial Relations 46(4): 702–731.

22.

Hox

(2010) Multilevel Analysis: Techniques and Applications. 2nd edn. New York: Routledge.

23.

Iller

Berger

George

Hauser-Ditz

Wiß

(2016) Unterstützung der Interessenvertretung in der betrieblichen Weiterbildung in Deutschland und Österreich (Study der Hans-Böckler-Stiftung 322). Düsseldorf: Hans-Böckler-Stiftung.

24.

Lazazzara

Karpinska

Henkens

(2013) What factors influence training opportunities for older workers? Three factorial surveys exploring the attitudes of HR professionals. International Journal of Human Resource Management 24(11): 2154–2172.

25.

Leisink

Greenwood

(2007) Company-level strategies for raising basic skills: A comparison of Corus Netherlands and UK. European Journal of Industrial Relations 13(3): 341–360.

26.

Lindbeck

Snower

(2001) Insiders versus outsiders. Journal of Economic Perspectives 15(1): 165–188.

27.

Munro

Rainbird

(2004) Opening doors as well as banging on tables: An assessment of UNISON/employer partnerships on learning in the UK public sector. Industrial Relations Journal 35(5): 419–433.

28.

OECD (2012) Better Skills, Better Jobs, Better Lives: A Strategic Approach to Skills Policies. Paris: OECD.

29.

Raemdonck

Beausaert

Fröhlich

Kochoian

Meurant

(2015) Aging workers’ learning and employability. In: Bal

Kooj

Rousseau

(eds) Aging Workers and the Employee-Employer Relationship. Cham: Springer, pp. 163–184.

30.

Rainbird

Sutherland

Edwards

Holly

Munro

(2003) Employee Voice and Training at Work: Analysis of WERS98 and Case Studies. London: Department of Trade and Industry.

31.

Rees

(2013) Comparing adult learning systems: An emerging political economy. European Journal of Education 48(2): 200–212.

32.

Schwander

Häusermann

(2013) Who is in and who is out? A risk-based conceptualization of insiders and outsiders. Journal of European Social Policy 23(3): 248–269.

33.

Setti

Dordoni

Piccoli

Bellotto

Argentero

(2015) Proactive personality and training motivation among older workers. European Journal of Training and Development 39(8): 681–699.

34.

Stegmaier

(2012) Effects of works councils on firm-provided further training in Germany. British Journal of Industrial Relations 50(4): 667–689.

35.

Streeck

(1992) Training and the new industrial relations: A strategic role for unions?. In: Regini

(ed.) The Future of Labour Movements. London: SAGE, pp. 250–269.

36.

Taylor

Urwin

(2001) Age and participation in vocational education and training. Work, Employment and Society 15(4): 763–779.

37.

TNS Infratest (ed.) (2009) European Company Survey (ECS 2009) – Technical Report: Methodology, Questionnaire Development and Fieldwork. Munich: TNS Infratest.

38.

Visser

(2013) ICTWSS: Database on Institutional Characteristics of Trade Unions, Wage Setting, State Intervention and Social Pacts in 51 Countries between 1960 and 2014. Available at: http://www.uva-aias.net/208

39.

Waddoups

(2014) Union membership and job-related training: Incidence, transferability, and efficacy. British Journal of Industrial Relations 52(4): 753–778.

40.

Wexley

Latham

(2001) Developing and Training Human Resources in Organizations, 3rd edn. Upper Saddle River, NJ: Prentice Hall.

41.

Winterton

(2007) Building social dialogue over training and learning: European and national developments. European Journal of Industrial Relations 13(3): 281–300.