Does board-level employee representation impact firms’ value? Evidence from the European countries

Abstract

Purpose

Using, for the first time, a sample of European listed firms from 30 countries with different legal regimes of board-level employee representation (BLER), the purpose of this paper is to examine the impact of BLER on firms’ value of European public companies, where employee representation is voluntary or imposed by law depending on the country of origin.

Design/methodology/approach

Using a difference-in-differences approach and a matching procedure, the authors analyze the impact of BLER on firms' value.

Findings

The results of this paper suggest that BLER adopted voluntarily affects positively firms’ value comparing to a group of firms where employee representation is in some way mandatory. Moreover, the findings of this paper show that firms from countries where BLER is not imposed by law tend to pay higher dividends. Nevertheless, the evidence presented in this paper only holds for low levels of employee representation on the board.

Research limitations/implications

This research not only provides some evidence in favor of the codetermination on corporate governance but also offers new avenues for discussing the conditions necessary for codetermination to be effective, especially the level of employees' participation on board.

Practical implications

This study provides to policymakers new insights for them to gain perspective, analyze and decide if codetermination is a useful tool to improve firms’ performance or at least in what conditions it should be applied.

Social implications

This study incentives the discussion of the proper way to include workers in firms’ boards with expected benefits on firms’ performance, economies and societies.

Originality/value

This paper provides evidence of a positive (but limited) impact on firms’ value derived from voluntary codetermination.

Citation

Silva, S., Silva, A. and Machado, R.B. (2022), "Does board-level employee representation impact firms’ value? Evidence from the European countries", Journal of Financial Regulation and Compliance, Vol. 30 No. 4, pp. 465-488. https://doi.org/10.1108/JFRC-10-2021-0082

1. Introduction

In recent years, there has been a growing debate about how firms should adapt their corporate governance structures to cope with new challenges posed by sustainability, economic growth and social inclusion concerns. The necessary adaptation process grounds on new business dynamics headed for more inclusive labor markets, where employee ownership and employee participation in business decisions play a major role. For start, this could be achieved with a set of company policies guided by a long-term outlook, which includes, among others, the reformulation of shareholder and executive incentives and compensation plans (OECD, 2018), to moderate short-termism and its perils for a “sustainable company” outline (Jackson and Petraki, 2010). Even if both shareholders and managers align their interests on short-term outcomes, precluding any agency conflict will be made at the expense of long-term prospects, harming employees and other stakeholders’ interests (Dallas, 2012).

As Conchon and Waddington (2011) point out, the main issue emerges from the bulk of empirical studies on the effectiveness of corporate governance systems: almost all of them proceed from a shareholder-oriented view. This perspective usually overlooks the importance of employee representation at the corporate decision-making level, as it does not fit the foundations of the reigning principal–agent theory (Allen et al., 2015; Gold and Waddington, 2019). Furthermore, the plethora of legal systems and institutional settings across European Union (EU) countries do not lead to a clearer picture, making it more difficult to capture the role of employees’ representation on firms’ performance.

Even so, the widespread of board-level employee representation (BLER) throughout European companies is already a reality, and not a singularity (Conchon and Waddington, 2011), though there is no clear-cut evidence on how this corporate governance mechanism affects firms’ value.

Therefore, in face of such a lack of conclusive evidence on the role of BLER on firms’ value, we contribute to the literature discussing the effects of BLER on European listed firms. Differently from the majority of prior research (that focuses on single-country studies or on a small group of countries with the same legal regime), we explore different legal BLER regimes. Our general goal is to determine the effect of codetermination on firms’ value and firm’s payout ratio on a sample of European public firms. More specifically, we pretend to test the role of the voluntary codetermination process on firms' performance expecting that voluntary codetermination has advantages over imposed BLER processes. Additionally and having in mind the pros and cons outlined in BLER literature, we also pretend to examine if there is a recommended level of voluntary employees’ participation or at least an interval of recommended level participation that impacts positively firms’ value.

In fact, the assortment of corporate governance systems and company laws in the EU, despite the efforts for harmonization, has contributed to a European landscape that can be portrayed by different incidence levels of BLER. Following Conchon and Waddington (2011), Conchon et al. (2015) or Gold and Waddington (2019), EU Member States, plus Iceland, Norway and Switzerland, can be classified into three groups (until 2015 [1]):

1. no regulation – countries with no legal provisions or another type of institutional arrangements supporting BLER: Belgium, Bulgaria, Estonia, Iceland, Italy, Latvia, Lithuania, Malta, Romania, Switzerland and the UK;

2. limited regulation scope – countries where BLER regulation is restricted to some public sector or privatized enterprises: Greece, Ireland, Poland, Portugal and Spain; and

3. wide regulation scope – countries where BLER is widespread over national laws and other institutional settings [2]: Austria, Croatia, Czech Republic [3], Denmark, Finland, France, Germany, Hungary, Luxembourg, the Netherlands, Norway, Slovakia, Slovenia and Sweden.

Thus, we assign firms from countries included in the group:

• (1) and (2) as the non-codetermination subsample, which is also our group of interest; and

• firms from countries included in (3) as the codetermination subsample, acting as our control group.

Our final sample comprises 15,236 firm-year observations, related to 1,671 listed firms from 30 European countries over the 2006–2015 period.

We test our research hypotheses using different regression methodologies. Our main results provide evidence that BLER adopted voluntarily by firms (from countries where there is no wide BLER regulation scope) impacts positively firms’ value relative to a control group of firms (from countries where BLER is in some way mandatory). Moreover, firms from BLER non-codetermination group seem to pay higher dividends than firms included in the control group.

To rule out the possibility of our evidence is resulting from differences in firm-level characteristics, in particular, size differences, we adopt as a robustness check a procedure based on matching observations in the same industry and year and by the closest Size, Sales Growth, ROA and Leverage variables. This further analysis supports our main results. In addition, we uncover that our previous evidence, on average, holds for low levels of employee representation (on the board of firms from non-codetermination BLER countries).

Overall, our conclusions contribute to address the role of BLER in the corporate governance framework; moreover, we discuss this issue from the perspective of the advantages of voluntary BLER participation.

The remaining of this study is organized as follows. Section 2 provides a review of the related literature to our research hypotheses. Section 3 describes the data, empirical models and variables. Section 4 presents and discusses the empirical results. Section 5 outlines our main conclusions.

2. Literature review and research hypotheses

One can define BLER, also known as codetermination, as the employees’ right to attend and participate in the firm’s board [4]. However, the question of if and how BLER impacts firms’ value is still waiting for robust supporting evidence. At one hand, the mainstream literature on corporate governance, namely, the shareholder-value-oriented approach, says very little about BLER. On the other hand, the few existing studies provide mixed evidence about BLER’s effect at the performance level.

As pointed out by Forcillo (2017), some researchers argue that BLER has negative effects on firms’ performance (or at least may produce negative consequences after the BLER threshold is reached [5]), as the interests of workers on the board could not be aligned towards those of shareholders. In fact, the primary goal of shareholders is to maximize their investments’ return, whereas the workers’ interests lie in their job role and their salary (Hansmann, 1990). Moreover, it is argued that the presence of workers on the board could delay decisions in terms of planning and innovation of processes, looking to maximize their own interests and not shareholders’ interests (Pejovich, 1978). Another negative argument for BLER is that the presence of workers on the board requires specific competencies (e.g. knowledge of corporate strategies) that often they do not have, and therefore, they may be not qualified to cover such role (Huse et al., 2009). This argument is consistent with the idea of workers’ influence over firm governance could impede efficient decision-making and lead to “hold-up” problems that discourage capital formation because potential investors know that the outcome of their investment can be captured by workers (Jensen and Meckling, 1979; Dammann and Eidenmueller, 2020).

In addition, and according to Williamson (1975), it is not possible to reach an effective model of governance through a system of BLER because it is imposed exogenously by the law (in certain countries and under specific conditions). If it was efficient, then the shareholders would voluntarily adopt such a model. This view is consistent with Jensen and Meckling (1979, 474), who claim that if BLER is “[…] beneficial to both stockholders and labor, why do we need laws which force firms to engage it? Surely, they would do so voluntarily”, meaning that in cases where workers are able to add value to the firms’ management, the codetermination could be voluntary and produce positive effects. Gorton and Schmid (2004), addressing the case of the German codetermination system (domestically mandatory), namely, the BLER’s intensity impact wherein workers representatives must comprise one-third (a half) of the supervisory board of companies with 500 (2,000) or more employees, find that BLER could be detrimental to firms’ value (measured by Tobin’s Q). Thereby, firms, where equal representation is applied, underperform by 31% of those with one-third of worker representatives on the board. In a recent study involving German firms, Eulerich et al. (2020) uncover evidence that codetermination reduces, on average, firms’ value. Moreover, the authors’ findings reveal that employees are able to extract benefits from shareholders through increases in salaries and in employees’ number and decrease the value of dividends paid to shareholders; these effects are mitigated by the presence of strong shareholder monitors or high analyst coverage.

On the other side of this discussion, there is a branch of literature that provides evidence of the advantages deriving from the adoption of BLER (Boneberg, 2010). Some of those studies sustain the positive effects of BLER on workers’ motivation, which in turn impacts positively firms’ performance. The findings of Levine and Tyson (1990) support the presence of workers on the board because this gives them more responsibility, more motivation and, therefore, more involvement in business decisions. Freeman and Lazear (1995) suggest that codetermination allows for a better exchange of information between the board and workers. So, during any period of crisis, codetermination allows a reduction of any probability of strikes and, therefore, greater cooperation between shareholders and workers. In fact, studies carried out by Cable and FitzRoy (1980) and FitzRoy and Kraft (1993), analyzing the performance of German companies, also support evidence that parity codetermination had brought positive effects on productivity to the bigger companies. Moreover, Renaud (2007) analyzed the differences in terms of productivity and profitability of German firms; his results indicate that with the transition from a quasi-parity to parity codetermination, there was an increase in productivity and firms’ value. Supporting the Lisbon Strategy of the EU based on active workers’ participation, Kluge and Wilke (2007) claim that countries with participation rights for employees, on average, perform better in the rankings of the World Economic Forum’s Business Competitiveness Index than those without this type of rights. Furthermore, Kleinknecht (2015) emphasizes that a greater diversity of the board can bring better decisions, increasing the quality of monitoring and making better decisions (in the long run) for the company, without thinking merely of short-term profit. Still, with the focus on German codetermination, Freeman and Lazear (1995) and Lin et al. (2018) find that BLER is a good instrument for interest alignment between managers and debtholders, avoiding expropriation and proposing that employee representatives and debtholders risk profiles are very alike, thus inciting managers to address long-term interests of all of the company’s stakeholders and weakening the dominant short-term performance orientation. Furthermore, stronger work participation in decision-making increases firms’ leverage with a lower cost of debt, longer maturities and fewer covenants. Moreover, those firms present more stable cash flows and have lowered their idiosyncratic risk. Further, Fauver and Fuerst (2006), using a sample of German listed firms, assert and show evidence that firms with BLER are more prone to pay dividends, which can indicate that BLER triggers the moderating effect of dividends on opportunistic costs (e.g. cronyism) incurred by insiders and large shareholders (Easterbrook, 1984; Faccio et al., 2001; and, in some way, Jensen, 1986). This argument is well expressed by Fauver and Fuerst (2006, 682):

We therefore anticipate that a labor presence on the board should reduce asset stripping, management perk-taking, and management salaries, and increase the payout of cash flows in the form of shareholder dividends.

Still in the German context, Chyz et al. (2019) find employee representation on audit committees as the most influential codetermination mechanism associated with reduced aggressive financial reporting (such as tax policy aggressiveness and earnings manipulation). GregoriČ and Rapp (2019) found that firms with BLER are associated with a lower sensitivity of employment cuts to firm performance during crises periods; they have shown that this lower sensitivity was in part ensured through downward adjustments in the labor costs per employee, presumably through the implementation of flexible time agreements, reorganizations of bonus systems, temporary layoffs and work-sharing agreements.

In another context, Hollandts et al. (2009) find a significant and positive relationship between employee representation and firm performance (measured by the return on equity) using a French sample.

One of the strongest results on the benefits of BLER came from a study provided by Jäger et al. (2021), which showed that shared governance resulted in positive effects on capital formation as the result of worker involvement in investment decisions; such fact could be derived from worker representatives have longer-term views than shareholders or executives or also because shared governance generally facilitates cooperation between firms and their employees.

Anyway, for the positive effects of BLER to be effective, some conditions must hold, as it is not automatic; in literature, BLER efficacy has been linked to the employers’ attitudes towards workers’ voice (Holland, 2014; Jirjahn and Mohrenweiser, 2016) or to the ability of employees (Huse et al., 2009). More specifically, GregoriČ and Poulsen (2019, 244) state:

The latter will be more likely when employers acknowledge the positive effects of BLER, in terms of higher employee commitment, investments in firm-specific knowledge, productivity and improved employee–employer cooperation.

Complementarily, the benefits of the codetermination can also be associated with better socially responsible behavior of firms, then resulting in competitive advantages for organizations and, in turn, be beneficial for shareholders in terms of corporate financial performance (Van der Laan et al., 2008), where company social responsibility activities act as a moderator between BLER and financial performance (Lopatta et al., 2020); in this context, Scholz and Vitols (2019) show a positive relationship between codetermination and corporate social policies as the adoption of targets for reducing pollution in Germany. Surprisingly, corporate social responsibility (CSR) is only marginally addressed in literature but could be an important argument in favor of codetermination (Nekhili et al., 2020). Besides CSR, Rosenbohm and Haipeter (2019) also find that corporate structures (such as headquarters location or the level at which supervisory boards are established) are powerful resources available to employee representatives.

In brief and drawing on these studies, one can expect that BLER’s efficacy could be higher when employers are more supportive of it, and employees are more suited to play the role of managers. Complementarily, as pointed out by Jackson (2005), the difference in BLER benefits across countries is also explained by unions’ strength, political systems and the degree of concentrated corporate ownership.

Building on this study objectives, on the overall previous evidence and on Jensen and Meckling arguments, we postulate as our first set of hypotheses that BLER will have positive effects on firm’ value, as well as on payout ratio, in countries that voluntarily adopt the employee representation on their boards. Thus, we defend the idea that just when the stakeholders believe in codetermination (and find a proper equilibrium between shareholders and workers’ interests) that it really works and can be a governance mechanism leading to improved firm’s performance.

H1a.

Board representation by directors elected by employees has a positive impact on firms’ value from countries included in the non-codetermination group.

H1b.

Board representation by directors elected by employees has a positive impact on firms’ payout ratio from countries included in the non-codetermination group.

Nevertheless, concerning the impact of BLER on firms’ value and discussing with further detail the previous argument, Gorton and Schmid’s (2004) observed that the prudent use of labor in corporate governance can be value-enhancing, but the excessive influence of labor can create a firm that is a “country club” for workers.

Also, Germain and Lyon-Caen (2016) developed a corporate board theoretical model showing that low levels of employee representation on the board may increase the shareholder’s value, affecting positively the strategic choices made by the firm, which leads to more investment in long-term projects. They also showed that employee representation increases the shareholder value if the conflict of interest is not too severe, and employees can provide valuable enough information. This is consistent with Fauver and Fuerst’s (2006) research, who found that a limited employee representation increases significantly the value of large German firms.

Accordingly, we also intend to test if there is evidence of such alleged harmful effects of BLER (on firms’ value) when labor representation surpasses a certain level. Thus, we will test further this prediction. Therefore, we formulate our second set of hypotheses.

H2a.

Low levels of employee representation on the board have a positive impact on firms’ value from countries included in the non-codetermination group.

H2b.

Low levels of employee representation on the board have a positive impact on firms’ payout from countries included in the non-codetermination group.

3. Data and methodology

3.1 Data

Following prior studies (Knyght et al., 2010; Vitols, 2010; Aubert et al., 2017; Kim and Patel, 2017), we use the information on publicly traded companies from the European Foundation of Employee Share Ownership (EFES). The EFES data set includes all economic and financial information about employee share ownership, employee share plans and participation, as well as about corporate governance and about the ownership structure of the largest 2,747 European public companies. We then matched with Thomson’s Datastream and Worldscope databases to collect market variables and accounting data for the 2006–2015 period. We exclude financial firms because their accounting standards are structurally different from commercial firms and also exclude observations without available information on total assets, market and book value of equity. To reduce the effect of outliers, we winsorize all continuous variables at the 1% in each tail of the distribution. Thus, we assign firms from countries included in Groups (1) and (2) as the non-codetermination subsample, our group of interest, which includes countries where there is no wide regulation for BLER as Belgium, Bulgaria, Estonia, Greece, Iceland, Ireland, Italy, Latvia, Lithuania, Malta, Poland, Portugal, Romania, Spain, Switzerland and the UK and firms from countries included in Group (3) as the codetermination subsample (where there is a wider regulation for BLER) such as Austria, Croatia, Czech Republic, Denmark, Finland, France, Germany, Hungary, Luxembourg, The Netherlands, Norway, Slovakia, Slovenia and Sweden.

Our final sample comprises 15,236 firm-year observations, related to 1,671 listed firms, from 30 European countries across the 2006–2015 period, where the non-codetermination (codetermination) group sums 925 (746) firms.

Panel A of Table 1 describes our sample by country, and Panel B (Panel C) describes it by industry (the type of control).

In Panel A, we observe that most firms in the non-codetermination group are from Italy, Switzerland and the UK; together, they count two-thirds of observations. Regarding the codetermination group of countries, France and Germany sum about 50% of total observations. Panel B shows that most firms belong to the manufacturing sector in both groups. Concerning the type of control, we can see in Panel C that most firms have dispersed ownership, whereas family management has a significant weight mainly in firms from codetermination BLER countries (21% in the codetermination versus 11% in the non-codetermination group).

3.2 Empirical methodology

To test the impact of BLER on firms’ value (and also on payout ratio), we follow the intuition of prior research (Fauver and Fuerst, 2006; Ginglinger et al., 2011) and estimate the following models:

(1.1) Yi,t=αi+β1DEMPi,t+γ1(Firm  Level)+γ2(Governance  level)+μi+γt+εi,c,t

(1.2) Yi,t=αi+β1DEMPi,t+β2DEMPi,t×Non−Codeterminationc+γ1(Firm  Level)+γ2(Governance  level)+μi+γt+εi,c,t

where Y_i,t proxies for firms’ value (Q_i,t), measured as the market value of equity plus book value of assets minus book value of equity, divided by total assets. In regression estimates, we also use the ratio market-to-book value and payout ratio as alternates dependent variables. Our difference-in-differences model, shown in equation (1.2), displays our variable of interest is DEMP_i,t × Non – Codetermination, that captures the difference in firms’ value between the non-codetermination and codetermination groups, where DEMP_i,t is an indicator variable that equals 1 if directors of firm i were elected by employees and 0 otherwise. We do not include the variable Non – Codetermination, which assumes 1 if firm i is included in the group where BLER is voluntary because of the fixed effects framework. Following previous research (Fauver and Fuerst, 2006; Ginglinger et al., 2011), we also include a set of controls that comprise firm- and governance-level variables, as follows: Size is the natural logarithm of total assets; ROA is measured as earnings before interest and taxes scaled by total assets; Capex is capital expenditures scaled by total assets; Leverage is the total debt over total assets; Control stake is the percentage of shares owned by controllers; and Employees’ stake is the percentage of shares owned by employees. All variables are summarized in Appendix. We also include μ_i that represents firm (or industry/country) fixed effects. All regressions include year (γ_t) fixed effects. We clustered standard errors in different combinations (as shown in Tables 3–6). Table 2 reports the descriptive statistics for the main firm- and governance-level variables for the full sample and by non-codetermination and codetermination groups.

In Table 2, we observe that the non-codetermination group displays higher (Tobin’s) Q and Market-to-Book value than the codetermination group, and those differences are statistically significant. On the contrary, firms from countries of the non-codetermination group have a lower number of Executive Board members, their employees hold a lower stake in companies and they present a lower concentration of ownership (measured by the control stake) than firms from countries in the codetermination group.

4. Empirical results

4.1 The impact of board-level employee representation on firms’ value

According to our H1a, our first empirical task is to test if BLER has a positive impact on firms’ value in the set of countries included in the non-codetermination group (Table 1). Consistent with Jensen and Meckling (1979) arguments, we expect that firms will adopt voluntarily BLER when they predict a net positive benefit in doing so. Thus, we test this argument by running different specifications of equations (1.1) and (1.2). Table 3 shows the results.

Taken together, and as one can observe in Table 3, our main results provide evidence that BLER adopted voluntarily by firms from countries where such regulation is not compulsory impacts positively firms’ value relative to a control group of firms from countries where BLER is in some way mandatory.

Models (1)–(6) show estimates from regressing equation (1.1) and considering only the non-codetermination group, whereas Models (7)–(12) provide results for the full sample estimating equation (1.2). Taking Model (1) as an example, (Tobin’s) Q is 0.0712 higher for firms that voluntarily adopt BLER compared to firms that did not do that among the non-codetermination group (variable DEMP). Besides, this effect is even more pronounced when compared the differences between the non-codetermination and codetermination groups. Hence, in equation (1.2), the estimates of our variable of interest, DEMP_i,t × Non – Codetermination, which captures the difference in firms’ value for the non-codetermination group relative to the codetermination group, is positive and significant across estimations. As an example, in Model (7), (Tobin’s) Q for the non-codetermination group is 0.2770 higher relative to the codetermination group, ceteris paribus. And this effect does not depend on the proxy used to capture firms’ value, (Tobin’s) Q or the market-to-book value, although the economic magnitude is slightly different (0.2770 vs 0.3375).

Our findings suggest that the decision to voluntarily adopt BLER results in a positive and significant net benefit comparing to firms that do not adopt BLER, in the same pool of non-codetermination employee representation regime countries, and with firms from countries where BLER is largely regulated by law.

Overall, the results support the argument that BLER when adopted on a voluntary basis impacts (more) positively firms’ value.

4.2 The impact of board-level employee representation on firms’ payout

Following the intuition of Fauver and Fuerst (2006), we also expect that firms adopting voluntarily BLER are more prone to pay dividends. So, as postulated in H1b, we predict that BLER would have a higher positive impact on the payout ratio of firms from countries included in the non-codetermination group. Thus, we test our assumption by estimating our set of equations. Table 4 reports the results.

Table 4 shows coefficients’ estimates from regressing equation (1.1), including only the non-codetermination group in Models (1)–(3). As expected, coefficients’ estimates of DEMP_i,t, our binary variable that assumes 1 if a firm included in the non-codetermination group adopted voluntarily employee representation on the board, are positive and significant in all models. Taking Model (1), firms with BLER pay, on average, more 8.66% than firms without employee representation. Estimations resulting from regressing equation (1.2) are presented in Models (4)–(6). Our main variable of interest, DEMP_i,t × Non – Codetermination, is positive and significant, meaning that those firms from countries included in the non-codetermination group seem to pay higher dividends than firms in the codetermination group.

4.3 The impact of board-level employee representation on firms’ value (and payout) robustness checks

Our previous findings suggest that voluntary BLER impacts positively firms’ value. Furthermore, results also suggest that firms from countries included in the non-codetermination group are more prone to pay higher cash dividends. However, and aiming to find further evidence to support our results, we adopt as a robustness check a procedure based on matching observations between non-codetermination and codetermination groups in the same industry and year and by the closest Size, Sales Growth, ROA and Leverage variables (as mentioned before, all variables are as described in Appendix). The coefficients’ estimates were obtained by regressing equation (1.2) on the matched sample. Table 5 provides the results.

Panel A of Table 5 replicates our analysis of Table 3, and Panel B replicates the analysis in Table 4. Our variable of interest is DEMP_i,t × Non – Codetermination that captures the impact of BLER on firms’ value (payout ratio) for the matched sample. Although the coefficients’ estimates of our variable of interest are positive and significant across estimations, the economic magnitude is lower in coefficients displayed in Panel A – impact on firm’s value – and higher in the case of payout ratio (Panel B). This effect was expected because of the matching procedure, which selects identical observations from non-codetermination and codetermination groups based on the criteria described above, which in turn decreases the variance across observations. However, the BLER effect still holds for firms that voluntarily adjusted their boards to include employee representation compared with firms from codetermination regimes but with very similar firm-specific attributes.

Finally, we address potential endogeneity associated with the adoption of BLER in the non-mandatory group of countries, using a two-stage least squares estimation procedure. In the first stage, we estimate BLER using the five-industry classification groups, based on SIC codes (see Panel B of Table 1); then we use the fitted values in the second-stage regressions. The results, not reported by brevity [6], are consistent and qualitatively similar to the ones shown previously.

4.4 The impact of board-level employee representation levels on firms’ value (payout)

So far, our evidence supports the argument that voluntary BLER impacts positively firms’ value and also that those firms are more prone to pay higher dividends. However, if such positive impact is only observed in firms from non-codetermination countries, then we hypothesize that there must be a difference in employee representation levels between firms from non-codetermination and from codetermination BLER countries. Following the insights of, for example, Gorton and Schmid (2004), Fauver and Fuerst (2006) and Balsmeier et al. (2013), whose studies provide evidence that different levels of BLER impact firms’ value in different ways, we re-estimate equation (1.1), replacing the DEMP_i,t by three indicators of employee representation levels:

1. BLER 0–33% (DEMP 0–33%) that equals 1 if employee representation is strictly higher than 0 but is strictly less than 33% and 0 otherwise;

2. BLER 33–50% (DEMP 33–50%) that assumes 1 if employee representation is equal or strictly higher than 33% but is strictly less than 50% and 0 otherwise; and

3. BLER > 50% (DEMP > 50%) that assumes 1 if employee representation equals or exceeds 50% and 0 otherwise.

Panel A shows the results using (Tobin’s) Q as dependent variable (and market-to book value as an alternative variable), while Panel B displays results using as dependent variable the payout ratio. The estimates provided in both panels for DEMP 0–33%, the lowest level of employee representation, are very similar in sign and magnitude to the ones found in our main results, namely, in Tables 3 (Models (1)–(6)) and 4 (Models (1)–(3)). Regarding moderate and the highest level of employee representation, DEMP 33–50% and DEMP > 50%, respectively, results are on average not statistically different from 0. This evidence is consistent with the idea that only low levels of employee representation impact positively firms’ value and also the payout ratio.

5. Main conclusions

As Hayden and Bodie (2021) pointed out, there is a continuous disintegration of the intellectual foundations of the modern corporation, which raises questions about the core principles. That is the case of the shareholder primacy norm and the exclusive shareholder franchise; in fact, nowadays, it is clear that such arguments are sustained by flawed assumptions about the nature of shareholders or even stay in contradiction with the fundamental precepts of standard economics that are supposed to constitute their foundations.

As firms’ corporate governance moves away from the existing corporate order, we all need further and further to discuss alternatives to a new effective corporation framework. That is why is so important to study the shortcomings and the strengths of the BLER as a model of shared corporate governance. Moreover, this research is also important, as codetermination not only is about workers participation in the corporations’ decisions but may also benefit other corporate counterparts, as their interests line up with those of employees; that is the case of creditors because both groups are mainly interested in the long-term firm survival.

Using, for the first time, a sample of European listed firms from 30 countries with different legal regimes of BLER that we assign as non-codetermination and codetermination groups, we investigate the impact of BLER on firms’ value and on cash dividends payout ratio. For the period 2006–2015, we test our hypotheses using different regression techniques and alternative proxies for firms’ value.

Our main results suggest that firms that adopt voluntarily BLER (i.e. firms from countries where BLER is not imposed by law) experience higher valuations relative to a control group of firms from countries where BLER is in some way mandatory. Additionally, this BLER effect also impacts positively the payout ratio of our non-codetermination group. Further, we match observations from both groups, non-codetermination and codetermination groups, in the same industry and year and by the closest Size, Sales Growth, ROA and Leverage variables, to control for firm-specific characteristics that might impact our main results derived from non-observable features not captured by our fixed effects framework. This analysis gives additional support to our main results.

Moreover, we examine if this BLER effect holds across different levels of employee representation; our evidence points out that this effect only holds for lower levels of BLER. Overall, these results are consistent with Jensen and Meckling (1979) view; firms from non-codetermination regimes will voluntarily adopt BLER when they believe that will improve shareholders’ wealth.

In short, our study shows evidence of a positive (but limited) impact on firm value of voluntary codetermination, which contributes to reinforce the argument (already stated by other authors as Berglund and Holmén, 2016) that for well-run firms is recommended to take employee interests properly into account, independently of whether they are represented on the board.

However, it is noteworthy to mention that our study is also subject to some limitations, in which the main one is the coverage of our primary source of data. The EFES database is only available since 2006 and only includes listed firms with more than €200m in stock market capitalization, which turns reasonable to assume that a part of listed companies is left aside. Despite that, our study makes a contribution to BLER literature and can be used by decision-makers to incentive the discussion of the proper way to include workers in firms’ boards with expected benefits on firms’ performance, economies and societies.

Table 1.

Sample description

Subsample:	Non-codetermination		Codetermination
	No. Firms	Obs.	No. Firms	Obs.
Panel A – sample description by country
Austria			27	263
Belgium	50	464
Bulgaria	6	38
Croatia			11	102
Czech Republic			8	73
Denmark			42	375
Estonia	6	58
Finland			55	523
France			204	1927
Germany			172	1590
Greece	36	350
Hungary			5	50
Iceland	3	19
Ireland	18	161
Italy	117	1,075
Latvia	3	29
Lithuania	4	31
Luxembourg			6	46
Malta	5	48
Netherlands			51	460
Norway			77	697
Poland	81	734
Portugal	23	218
Romania	4	31
Slovakia			1	10
Slovenia			9	86
Spain	63	586
Sweden			78	733
Switzerland	110	1,007
The UK	396	3,452
All countries	925	8,301	746	6,935
Panel B – sample description by industry
Industry
Industry 1	146	1,314	112	1,065
Agriculture, Forestry, Fishing (1–9); Mining (10–14);
Construction (15–17)
Industry 2	259	2,353	250	2,341
Manufacturing (20–39)
Industry 3	221	2,000	170	1,590
Transportation (40–49)
Industry 4	52	447	31	305
Wholesale (50–51); Retail (52–59)
Industry 5	247	2,187	183	1,634
Services (70–89)
All	925	8,301	746	6,935
Panel C – sample description by type of control
No controlling owner	491	4,557	236	2,363
State ownership	37	354	58	505
Private	138	1,094	83	694
Corporate	86	744	114	1,028
Executive	41	363	58	530
Founders	18	136	14	128
Family	105	973	160	1,486
Foundation	8	71	20	171
Employees	1	9	3	30
All	925	8,301	746	6,935

Notes:

Panel A describes the number of firms (“No. Firms”) and the number of observations (“Obs.”) by the non-codetermination and codetermination groups over 2006–2015. Non-Codetermination represents a group of countries where there is no specific legislation on BLER or where legislation is restricted in scope. Codetermination represents a group of countries where BLER legislation is wide in scope. Panel B (Panel C) describes the sample by five-industry classification groups, based on SIC codes (type of control)

Table 2.

Descriptive statistics

Subsample:	Full sample			Non-codetermination			Codetermination
	n	Mean	Median	n	Mean	Median	n	Mean	Median
Size	15,236	13.969	13.743	8,301	13.867	13.581	6,935	14.091***	13.952***
ROA	14,964	0.0338	0.0865	8,134	0.0394	0.0926	6,830	0.027	0.0808***
(Tobin’s) Q	15,236	1.2657	0.9364	8,301	1.298	0.9569	6,935	1.2271***	0.9134***
Market book (MBV)	15,236	1.2946	0.9017	8,301	1.3446	0.9319	6,935	1.2349***	0.871***
Payout ratio	15,234	0.3999	0.2804	8,298	0.3924	0.2842	6,935	0.4088	0.2769
Leverage	15,234	0.2323	0.2147	8,300	0.2303	0.2146	6,934	0.2346	0.2147**
Capex	13,481	0.118	0.037	7,338	0.1125	0.0364	6,143	0.1244**	0.0378***
Sales growth	13,560	0.0838	0.0497	7,373	0.0862	0.051	6,187	0.0809	0.0482
Cash flow	13,532	0.0787	0.0769	7,361	0.0809	0.077	6,171	0.076***	0.0769
Top executives	15,024	3.5625	3	8,164	3.4366	3	6,860	3.7124***	3***
Employees’ stake (%)	15,236	8.4993	1.0617	8,301	7.7265	1.1688	6,935	9.4242***	0.9182***
Control stake (%)	15,019	29.8127	27.045	8,124	24.0066	0	6,895	36.6537***	38.62***

Notes:

Table 2 reports descriptive statistics for the full sample and by subsamples of non-codetermination and codetermination, between 2006 and 2015. Non-Codetermination represents a group of countries where there is no specific legislation on BLER or where legislation is restricted in scope. Codetermination represents a group of countries where BLER legislation is wide in scope. Size is the logarithm of total assets. ROA is measured as earnings before interest and taxes scaled by total assets. (Tobin’s) Q is measured as the market value of equity plus book value of assets minus book value of equity scaled by the book value of assets. Market book is the market value of equity divided by the book value of equity. Payout ratio is measured as cash dividends divided by net income. Leverage is the total debt over total assets. Capex is capital expenditures scaled by total assets. Sales growth is the percentage change in sales over a year. Cash flow is net income plus depreciation and amortization costs divided by total assets. Top executives is the number of board members. Employees’ stake is the percentage of shares owned by employees. Control stake is the percentage of shares owned by controllers. Variables are defined in Appendix. For each variable, the mean, median and number of observations are displayed. Differences in means are tested using t-statistic test and differences in medians are tested using Wilcoxon rank sum test.

***,

** and

* mean statistical significance at the 1%, 5% and 10% levels, respectively

Table 3.

The impact of board-level employee representation on firms’ value: main results

Subsample:	Non-codetermination						Full sample
Dependent variable:	Q	MBV	Q	MBV	Q	MBV	Q	MBV	Q	MBV	Q	MBV
Model:	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)	(9)	(10)	(11)	(12)
DEMP	0.0712*** (2.77)	0.0826*** (2.70)	0.0712 (1.62)	0.0826* (1.72)	0.0712** (3.25)	0.0826** (2.76)	−0.2130* (−1.82)	−0.2494*** (−3.15)	−0.2130 (−1.50)	−0.2494** (−2.40)	−0.2130 (−1.37)	−0.2494* (−2.09)
DEMP*Non-Codetermination							0.2770** (2.35)	0.3375*** (4.13)	0.2770* (1.92)	0.3375*** (3.23)	0.2770* (1.74)	0.3375** (2.66)
Size	−0.3059*** (−5.55)	−0.3408*** (−5.69)	−0.3059*** (−5.28)	−0.3408*** (−7.99)	−0.3059*** (−7.68)	−0.3408*** (−7.44)	−0.2202*** (−5.42)	−0.2619*** (−6.33)	−0.2202*** (−3.47)	−0.2619*** (−4.71)	−0.2202*** (−6.25)	−0.2619*** (−6.34)
ROA	−0.0183 (−0.46)	0.0426 (0.98)	−0.0183 (−1.04)	0.0426** (2.45)	−0.0183 (−0.36)	0.0426 (0.84)	−0.0014 (−0.04)	0.0418 (1.14)	−0.0014 (−0.05)	0.0418 (1.36)	−0.0014 (−0.03)	0.0418 (1.04)
Capex	0.1553** (2.55)	0.2253*** (2.86)	0.1553** (2.95)	0.2253** (2.39)	0.1553** (3.57)	0.2253** (3.93)	0.1212*** (3.41)	0.1914*** (4.12)	0.1212*** (3.07)	0.1914*** (3.18)	0.1212* (2.69)	0.1914** (3.65)
Leverage	−0.2216 (−1.02)	−0.0859 (−0.32)	−0.2216* (−1.77)	−0.0859 (−0.40)	−0.2216 (−0.75)	−0.0859 (−0.23)	−0.1802 (−1.26)	−0.0220 (−0.13)	−0.1802* (−1.80)	−0.0220 (−0.14)	−0.1802 (−1.11)	−0.0220 (−0.07)
Control	0.0076 (0.55)	0.0067 (0.30)	0.0076 (0.59)	0.0067 (0.28)	0.0076 (0.67)	0.0067 (0.35)	0.0041 (0.25)	0.0084 (0.38)	0.0041 (0.32)	0.0084 (0.51)	0.0041 (0.21)	0.0084 (0.36)
Employees’ stake	−0.0005 (−0.34)	−0.0008 (−0.49)	−0.0005 (−0.59)	−0.0008 (−0.67)	−0.0005 (−0.60)	−0.0008 (−0.61)	0.0011 (1.09)	0.0013 (1.04)	0.0011 (1.26)	0.0013 (1.06)	0.0011 (1.85)	0.0013 (1.33)
Constant	5.6834*** (7.53)	6.1967*** (7.54)	5.7896*** (7.57)	6.3498*** (10.74)	5.7896*** (11.28)	6.3498*** (11.59)	4.5163*** (8.04)	5.1043*** (8.89)	4.4530*** (4.85)	5.0253*** (6.05)	4.4530*** (8.88)	5.0253*** (9.01)
SE cluster	Firm	Firm	Country	Country	Industry	Industry	Firm	Firm	Country	Country	Industry	Industry
Firm FE	Yes	Yes	No	No	No	No	Yes	Yes	No	No	No	No
Year FE	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Industry FE	No	No	No	No	Yes	Yes	No	No	No	No	Yes	Yes
Country FE	No	No	Yes	Yes	No	No	No	No	Yes	Yes	No	No
Observations	7,053	7,053	7,053	7,053	7,053	7,053	13,069	13,069	13,069	13,069	13,069	13,069
R2	0.126	0.120	0.126	0.120	0.126	0.120	0.111	0.116	0.111	0.116	0.111	0.116
Number of firms	892	892	892	892	892	892	1,630	1,630	1,630	1,630	1,630	1,630

Notes:

Table 3 presents regression estimates of equation (1.1) in Models (1)–(6) and equation (1.2) in Models (7)–(12) to test H1a. The dependent variable is Q_i,t (MBV) meaning the market value of equity plus book value of assets minus book value of equity divided by total assets (ratio of the market value of equity over book value of equity). DEMP_i,t is an indicator variable that equals 1 if directors were elected by employees and 0 otherwise. Non-Codetermination is a dummy variable that assumes 1 if firm i is included in the group of countries where there is no specific legislation on BLER or where legislation is restricted in scope and 0 otherwise. All variables are as described in Table 2 and also in Appendix. Fixed effects by firm, year, country and industry are used in different schemes to control for any unobservable or omitted factors. Robust t-statistics standard errors in parentheses (…) clustered at firm, industry and country level as indicated in the table.

***,

** and

* mean statistical significance at the 1%, 5% and 10% levels, respectively

Table 4.

The impact of board-level employee representation on firms’ payout: main results

Subsample:	Non-codetermination				Full Sample
Dependent variable:			Payout ratio
Model:	(1)	(2)	(3)	(4)	(5)	(6)
DEMP	0.0866***	0.0866***	0.0866***	−0.1454*	−0.1454*	−0.1454
	(6.10)	(4.72)	(11.88)	(−1.67)	(−1.68)	(−1.60)
DEMP * Non-Codetermination				0.2519***	0.2519***	0.2519**
				(2.89)	(3.05)	(2.69)
Size	0.0235	0.0235	0.0235*	0.0350**	0.0350*	0.0350*
	(1.28)	(1.64)	(1.68)	(2.41)	(2.01)	(2.13)
ROA	0.1453***	0.1453***	0.1453***	0.1393***	0.1393***	0.1393***
	(6.69)	(9.71)	(9.42)	(8.32)	(8.91)	(14.10)
Capex	0.0092	0.0092	0.0092	0.0399**	0.0399*	0.0400
	(0.36)	(0.50)	(0.51)	(2.14)	(1.72)	(1.55)
Leverage	−0.5286***	−0.5286***	−0.5286**	−0.4867***	−0.4867***	−0.4866***
	(−7.05)	(−10.88)	(−4.38)	(−8.46)	(−9.00)	(−7.04)
Control	0.0042	0.0042	0.0042	−0.0061	−0.0061	−0.0061
	(0.19)	(0.17)	(0.28)	(−0.41)	(−0.36)	(−0.47)
Employees’ stake	0.0007	0.0007	0.0007	0.0003	0.0003	0.0003
	(0.99)	(0.88)	(1.35)	(0.51)	(0.48)	(0.56)
Constant	0.4625*	0.5609***	0.5609**	0.3435	0.3435	0.3403
	(1.76)	(3.16)	(2.86)	(1.64)	(1.47)	(1.33)
SE cluster	Firm	Country	Industry	Firm	Country	Industry
Firm FE	Yes	No	No	Yes	No	No
Year FE	Yes	Yes	Yes	Yes	Yes	Yes
Industry FE	No	No	Yes	No	No	Yes
Country FE	No	Yes	No	No	Yes	No
Observations	7,053	7,053	7,053	13,069	13,069	13,069
R2	0.070	0.070	0.070	0.056	0.056	0.056
Number of firms	892	892	892	1,630	1,630	1,630

Notes:

Table 4 reports regression estimates of equation (1.1) in Models (1)–(3) for the non-codetermination subgroup and of equation (1.2) in Models (4)–(6) for the all sample to test H1b. The dependent variable is payout ratio that is the cash dividends divided by net income. DEMP_i,t is an indicator variable that equals 1 if directors were elected by employees and 0 otherwise. Non-Codetermination is a dummy variable that assumes 1 if firm i is included in the group of countries where there is no specific legislation on BLER or where legislation is restricted in scope and 0 otherwise. All variables are as described before and also in Appendix. Fixed effects by firm, year, country and industry are used in different schemes to control for any unobservable or omitted factors. Robust t-statistics standard errors in parentheses (…) clustered at firm, industry and country level as indicated in the table.

***,

** and

* mean statistical significance at the 1%, 5% and 10% levels, respectively

Table 5.

The impact of board-level employee representation on firms’ value and on payout ratio: robustness tests

Subsample:	Matched sample
Dependent variable:	Q	MBV	Q	MBV	Q	MBV
Model:	(1)	(2)	(3)	(4)	(5)	(6)
Panel A – The impact of board-level employee representation on firms’ value
DEMP	−0.0397	−0.0720**	−0.0397	−0.0720	−0.0397	−0.0720*
	(−1.55)	(−2.19)	(−1.02)	(−1.28)	(−1.32)	(−1.87)
DEMP*Non-Codetermination	0.1220***	0.1617***	0.1220***	0.1617***	0.1220***	0.1617***
	(7.46)	(7.18)	(10.02)	(8.11)	(6.92)	(5.85)
Size	−0.1671***	−0.2127***	−0.1671**	−0.2127**	−0.1671**	−0.2127*
	(−2.95)	(−2.89)	(−2.61)	(−2.21)	(−2.80)	(−2.50)
ROA	0.2787***	0.4721***	0.2787***	0.4721***	0.2787**	0.4721**
	(3.12)	(3.69)	(4.14)	(3.89)	(2.60)	(3.40)
Capex	0.0408	0.0990	0.0408	0.0990*	0.0408	0.0990*
	(0.99)	(1.35)	(1.24)	(1.81)	(1.37)	(1.82)
Leverage	−0.3816**	−0.0598	−0.3816*	−0.0598	−0.3816	−0.0598
	(−2.11)	(−0.21)	(−1.73)	(−0.29)	(−1.45)	(−0.14)
Control	0.0045	0.0201	0.0045	0.0201	0.0045	0.0201
	(0.30)	(0.62)	(0.28)	(0.63)	(0.30)	(0.51)
Employees’ stake	0.0027**	0.0030*	0.0027**	0.0030	0.0027***	0.0030***
	(2.57)	(1.88)	(2.38)	(1.58)	(7.80)	(4.78)
Constant	3.6186***	4.1788***	3.6186***	4.1788***	3.6186**	4.1788**
	(4.33)	(3.91)	(3.77)	(2.99)	(3.73)	(2.98)
SE cluster	Firm	Firm	Country	Country	Industry	Industry
Firm	Yes	Yes	No	No	No	No
Year	Yes	Yes	Yes	Yes	Yes	Yes
Industry	No	No	No	No	Yes	Yes
Country	No	No	Yes	Yes	No	No
Observations	3,463	3,463	3,463	3,463	3,463	3,463
R2	0.156	0.162	0.156	0.162	0.156	0.162
Number of firms	1,268	1,268	1,268	1,268	1,268	1,268
Matching quality	0.665

Panel B – The impact of board-level employee representation on payout ratio
Subsample:	Matched sample
Dependent variable:	Payout ratio
Model:	(1)	(2)	(3)
DEMP	−0.2765***	−0.2765***	−0.2765***
	(−12.47)	(−10.83)	(−8.43)
DEMP*Non-Codetermination	0.3751***	0.3751***	0.3751***
	(28.37)	(37.11)	(44.60)
Size	0.0467	0.0467	0.0467
	(1.21)	(1.38)	(1.09)
ROA	0.8298***	0.8298***	0.8298***
	(5.75)	(6.55)	(5.88)
Capex	−0.0190	−0.0190	−0.0190
	(−0.39)	(−0.29)	(−0.28)
Leverage	−0.4198***	−0.4198***	−0.4198***
	(−3.30)	(−3.30)	(−6.10)
Control	−0.0546**	−0.0546	−0.0546**
	(−2.33)	(−1.60)	(−3.97)
Employees’ stake	0.0009	0.0009	0.0009
	(0.83)	(0.77)	(0.97)
Constant	0.1707	0.1707	0.1707
	(0.30)	(0.34)	(0.30)
SE cluster	Firm	Country	Industry
Firm FE	Yes	No	No
Year FE	Yes	Yes	Yes
Industry FE	No	No	Yes
Country FE	No	Yes	No
Observations	3,463	3,463	3,463
R2	0.138	0.138	0.138
Number of firms	1,268	1,268	1,268

Notes:

Panel A of Table 5 presents regression estimates of equation (1.2) for a matched sample. The dependent variable is Q_i,t (MBV) is the market value of equity plus book value of assets minus book value of equity divided by total assets (ratio of the market value of equity over book value of equity). DEMP_i,t is an indicator variable that equals 1 if directors were elected by employees and 0 otherwise. Panel B reports regression estimates of equation (1.2) using as dependent variable is Payout ratio that is the cash dividends divided by net income. All variables are as described before and also in Appendix. Fixed effects by firm, year, country and industry are used in different schemes to control for any unobservable or omitted factors. Robust t-statistics standard errors in parentheses (…) clustered at firm, industry and country level as indicated in the table.

***,

** and

* mean statistical significance at the 1%, 5% and 10% levels, respectively. The quality of the matching is reported by the p-value of Likelihood-ratio

Table 6.

The impact of board-level employee representation levels on firms’ value and on payout ratio

Subsample:	Non-codetermination
Dependent variable:	Q	MBV	Q	MBV	Q	MBV
Model:	(1)	(2)	(3)	(4)	(5)	(6)
Panel A – The impact of board-level employee representation levels on firms’ value
DEMP 0–33%	0.0712***	0.0826***	0.0712	0.0826*	0.0712**	0.0826*
	(2.77)	(2.70)	(1.62)	(1.72)	(3.22)	(2.76)
DEMP 33–50%	0.0227	−0.0014	0.0227	−0.0014	0.0227	−0.0014
	(0.42)	(−0.02)	(0.35)	(−0.02)	(0.66)	(−0.02)
DEMP >50%	0.1473	0.0892	0.1473	0.0892	0.1473**	0.0892**
	(1.33)	(1.32)	(1.29)	(0.66)	(4.45)	(2.60)
Size	−0.3063***	−0.3415***	−0.3063***	−0.3415***	−0.3063***	−0.3415***
	(−5.55)	(−5.70)	(−5.31)	(−8.07)	(−7.66)	(−7.41)
ROA	−0.0184	0.0425	−0.0184	0.0425**	−0.0184	0.0425
	(−0.46)	(0.98)	(−1.04)	(2.44)	(−0.36)	(0.84)
Capex	0.1556**	0.2257***	0.1556***	0.2257**	0.1556**	0.2257**
	(2.55)	(2.86)	(2.96)	(2.40)	(3.57)	(3.93)
Leverage	−0.2217	−0.0852	−0.2217*	−0.0852	−0.2217	−0.0852
	(−1.02)	(−0.32)	(−1.76)	(−0.39)	(−0.75)	(−0.23)
Control	0.0075	0.0064	0.0075	0.0064	0.0075	0.0064
	(0.54)	(0.29)	(0.58)	(0.27)	(0.66)	(0.34)
Employees’ stake	−0.0005	−0.0008	−0.0005	−0.0008	−0.0005	−0.0008
	(−0.34)	(−0.49)	(−0.59)	(−0.67)	(−0.60)	(−0.61)
Constant	5.6918***	6.2091***	5.7978***	6.3619***	5.7978***	6.3619***
	(7.53)	(7.54)	(7.61)	(10.83)	(11.23)	(11.49)
SE cluster	Firm	Firm	Country	Country	Industry	Industry
Firm FE	Yes	Yes	No	No	No	No
Year FE	Yes	Yes	Yes	Yes	Yes	Yes
Industry FE	No	No	No	No	Yes	Yes
Country FE	No	No	Yes	Yes	No	No
Observations	7,053	7,053	7,053	7,053	7,053	7,053
R2	0.126	0.120	0.126	0.120	0.126	0.120
Number of firms	892	892	892	892	892	892

Panel B – The impact of board-level employee representation levels on payout ratio
Subsample:	Non-codetermination
Dependent variable:	Payout ratio
Model:	(1)	(2)	(3)
DEMP 0–33%	0.0866***	0.0866***	0.0866***
	(6.11)	(4.74)	(12.04)
DEMP 33–50%	0.0217	0.0217	0.0217
	(0.50)	(0.96)	(1.33)
DEMP >50%	−0.1826	−0.1826	−0.1826
	(−1.15)	(−1.17)	(−0.98)
Size	0.0233	0.0233	0.0233*
	(1.27)	(1.63)	(1.66)
ROA	0.1452***	0.1452***	0.1452***
	(6.69)	(9.73)	(9.38)
Capex	0.0095	0.0095	0.0095
	(0.38)	(0.52)	(0.53)
Leverage	−0.5261***	−0.5261***	−0.5261**
	(−7.03)	(−10.93)	(−4.31)
Control	0.0041	0.0041	0.0041
	(0.19)	(0.17)	(0.27)
Employees’ stake	0.0007	0.0007	0.0007
	(0.99)	(0.88)	(1.35)
Constant	0.4673*	0.5656***	0.5656**
	(1.78)	(3.19)	(2.86)
SE cluster	Firm	Country	Industry
Firm FE	Yes	No	No
Year FE	Yes	Yes	Yes
Industry FE	No	No	Yes
Country FE	No	Yes	No
Observations	7,053	7,053	7,053
R2	0.071	0.071	0.071
Number of firms	892	892	892

Notes:

Panel A of Table 6 reports regression estimates of equation (1.1) to test H2a, including employee representation by levels as in Fauver and Fuerst (2006): 1) DEMP 0–33% equals 1 if employee representation exceeds 0 but is strictly lower than 33% and 0 otherwise, 2) DEMP 33–50% equals 1 if employee representation equals or exceeds 33% but is strictly lower than 50% and 0 otherwise and 3) DEMP >50% equals 1 if employee representation equals or is higher than 50% and 0 otherwise. The dependent variable is Q_i,t (MBV) is the market value of equity plus book value of assets minus book value of equity divided by total assets (ratio of market value of equity over book value of equity). Panel B reports regression estimates using as dependent variable is Payout ratio that is the cash dividends divided by net income, as postulated in H2b. All variables are as described before and also in Appendix. Fixed effects by firm, year, country and industry are used in different schemes to control for any unobservable or omitted factors. Robust t-statistics standard errors in parentheses (…) clustered at firm, industry and country level as indicated in the table.

***,

** and

* mean statistical significance at the 1%, 5% and 10% levels, respectively

Table A1.

Definitions and sources of the variables

Variable	Definition	Source
Firm level
(Tobin’s) Q	Market value of equity plus book value of assets minus book value of equity, divided by total assets	Worldscope
Capex	Capital expenditures (CAPEX) divided by total sales	Worldscope
Cash flow	Net income plus depreciation and amortization expenses divided by lagged total assets	Worldscope
Leverage	Short-term plus long-term debt divided by total assets	Worldscope
Market book	The market value of equity divided by the book value of equity	Worldscope
Payout ratio	Ratio of cash dividends to net income	Worldscope
ROA	Earnings before interest and taxes (EBIT) divided by total assets	Worldscope
Sales growth	Sales growth is measured as the percentage change in sales over year t − 1 to t	Worldscope
Size	Logarithm of total assets	Worldscope
Governance level
DEMP	Indicator variable that equals 1 if directors were elected by employees and 0 otherwise	EFES
Employees’ stake	Percentage of shares owned by employees	EFES
Executive board members	Number of executive board members	EFES
Type of control	Stake types: i) No controlling owner, ii) State ownership, iii) Private, iv) Corporate, v) Executive director, vi) Founders, vii) Family, viii) Foundation and ix) Employees	EFES
Control stake	Percentage of shares owned by controllers	EFES
Industry level
Sic code	Four-digit standard industrial classification (SIC) code	Datastream
Industry	Five-industry classification groups, based on SIC codes: Industry 1: Agriculture, Forestry, Fishing (01–09), Mining (10–14), Construction (15–17); Industry 2: Manufacturing (20–39): Industry 3: Transportation (40–49); Industry 4: Wholesale (50–51), Retail (52–59); and Industry 5: Services (70–89)

Notes

1. In the time span between 2006 and 2015, there are two distinct classification moments: 2011 and 2015 and there is just one difference related to the Czech Republic.

2. Although such determination may have (in some countries) a threshold of employees’ number before being mandatory,

3. BLER is not mandatory since 2014 for private sector companies.

4. More precisely, it means “the right to elect or appoint some of the members of the company's supervisory or administrative organ; or the right to recommend and/or oppose the appointment of some or all of the members of the company's supervisory or administrative organ”(Article 2 Council Directive 2001/86/EC).

5. Balsmeier, Bermig and Dilger (2013) find a concave relationship between board-level representation and firm’s value; the authors show that moderate employee participation in corporate board decision-making can enhance firm’s value.

6. Results are available upon request.