Voice at work: The effects of worker representation on worker welfare and firm performance
Workers in liberal market economies like the US and the UK are excluded from participating in the governance of their firms. While workers may be informally consulted by their managers or employers, in the US the National Labor Relations Act puts restrictions on workers’ legal rights to exercise ‘voice’ through cooperative representation institutions in their firms. However, survey evidence shows that a majority of workers in the US and UK would prefer more opportunities to exercise voice in their workplaces (e.g. Bryson and Freeman 2013).
Recent policy proposals in both countries aim to cater to this demand for worker voice by giving workers a right to elect representatives to company boards or to establish shop-floor representation committees (for example, board-level representation laws have been proposed by Theresa May in 2016, the UK Labour Party in 2018, and US Democratic senators in 2018; see Jӓger et al. 2021b for a discussion of recent policy proposals).
Many continental European countries give workers a formal right to voice via board-level or shop-floor elected representation. Credible evidence on the economic impacts of these ‘voice’ arrangements is still scarce, due to a paucity of ‘natural experiments’ in which certain firms are quasi-randomly assigned to greater worker voice. Most introductions or extensions of formal voice arrangements in Europe occurred in the 1960s and 1970s, too early to study with modern administrative datasets.
A priori, giving workers a voice in firm governance could improve worker satisfaction by giving them an avenue to lobby for improvements in working conditions; it could reduce turnover and separations by giving workers an opportunity to shape their firm from within rather than quitting; and it could raise wages by strengthening worker bargaining power (Freeman and Lazear 1995, Freeman et al. 2020).
On the other hand, opponents of voice arrangements warn that giving workers decision-making power will impede efficient decision-making and deter capital formation because prospective investors know that workers will be able to capture the fruits of potential capital investments (Jensen and Meckling 1979).
Two recent studies examine the economic effects of one type of worker voice – the allocation of board seats to worker representatives (Jӓger et al. 2021 and Blandhol et al. 2020). Both studies find limited or zero effects of the arrangement on wage-setting. Moreover, Jӓger et al. (2021) also study effects on firm performance and find limited (if anything, positive) effects on capital formation and productivity.
In a recent paper (Harju et al. 2021), we examine a pair of reforms in Finland in 1991 and 2008 that introduced or expanded workers’ rights to a broader and richer class of voice institutions. These reforms are unique in occurring late enough to study with high-quality modern datasets that enable credible causal identification. We study the economic effects of these reforms using administrative data, and additionally draw on extensive survey evidence to understand how worker voice in Finland functions in practice. Our paper complements a recent survey article that draws on a wide range of quantitative and survey evidence to understand the impacts of worker voice institutions (Jӓger et al. 2021a).
Background and reforms
In 1991, Finland introduced a law establishing a right to worker voice in firms with 150 or more employees. By statutory default, workers have a right to elect 20% of the members of either the firm’s board of directors or the management group, with the employer choosing between the two options. However, the law leaves wide discretion to employers and workers to negotiate alternative voice arrangements.
Survey evidence indicates that workers in about 15% of covered firms take up the statutory default right to board representation, while workers in 40% of firms negotiate alternative forms of voice, such as advisory panels or regular meetings between top management and worker representatives. In the remaining 45% of firms, workers do not take up their right to voice; surveys suggest that this is mostly due to employers opposing the introduction of worker voice and successfully blocking the institution.
Figure 1 Firm governance in Finland before (left) and after (right) the 1991 voice reform
Finnish workers also have a right to shop-floor representation via elected ‘shop stewards’ dating back to a 1978 law. Prior to 2008, this right existed in firms with 30 or more employees; in 2008, a reform extended this right to firms with between 20 and 29 workers (Keskinen 2017).
We study the effects of both reforms, which lets us analyse the effects of rights to worker voice (instantiated through a wide diversity of specific voice institutions) on worker outcomes and firm performance. Our data sources are linked employer-employee data from Statistics Finland, matched onto a job-quality survey that lets us estimate the effects of worker voice on intangible aspects of job quality.
The size-dependent nature of both reforms allows us to estimate the causal effects of rights to worker voice by sorting firms based on their size before the reforms and comparing firms above and below the size thresholds before and after the reforms. This difference-in-differences strategy builds on the assumption that firms above and below the thresholds would have followed similar trajectories in the absence of the reforms. We test the plausibility of this assumption by showing that their trajectories were parallel in the years leading up to the reforms.
Worker outcomes: Turnover and wages
We start by studying the effects of worker voice on voluntary and involuntary separations. The classic ‘exit-voice’ model of Hirschman (1970) predicts that giving workers a right to voice will reduce voluntary quits, and voluntary quit rates are considered a standard revealed-preference measure of overall job quality.
Meanwhile, worker representatives state that one of their highest priorities is preventing involuntary layoffs. Since we cannot distinguish between voluntary and involuntary separations, we use job-to-job separations as a proxy for voluntary separations and employment-to-nonemployment separations as a proxy for involuntary separations.
As the top two panels of Figure 2 show, we do not find that a right to worker voice through the 1991 reform reduces the frequency of voluntary separations. The same holds for the 2008 reform. Thus, there is no evidence that a right to worker voice reduces turnover or improves job quality as proxied by a revealed-preference measure.
Figure 2 Turnover effects of worker voice
a) Voluntary (employment-to-employment) separations
b) Involuntary (employment-to-nonemployment) separations
However, the bottom two panels do show evidence for a slight reduction in involuntary separations as a result of worker voice through the 1991 reform. Jӓger et al. (2021b) discuss a few mechanisms through which worker voice could prevent involuntary layoffs, including by letting firms adjust wages and hours more flexibly in the face of crises.
We also examine whether worker voice raises wages by boosting workers’ bargaining power. As the top two panels of Figure 3 show, we find evidence for a very small increase (about 1.6%) in composition-adjusted wages as a result of the 1991 reform, though this estimate is not statistically distinguishable from zero, and the 2008 reform does not appear to affect wages, consistent with the results of Keskinen (2017). As the bottom panel of Figure 3 shows, there is evidence that this slight wage increase operates through a boost to wages of the lowest-earning (bottom deciles) workers.
Figure 3 Wage effects of worker voice
a) Effects on composition-adjusted wages
b) Effects by wage decile
Finally, we study the effects of both reforms on subjective job quality, drawing on the job-quality survey linked to our administrative data. We find evidence that both reforms mildly increased subjective job quality.
To test the hypothesis that worker voice worsens firm performance or deters capital formation (e.g. Jensen and Meckling 1979), we study the effects of the 1991 and 2008 reforms on firm survival, productivity, and capital intensity. We find no evidence for negative effects on any of these outcomes and, if anything, find slight positive effects on firm survival and productivity.
However, we note that the fact that employers in about 45% of firms appear to successfully take efforts to block the introduction of worker voice provides some suggestive revealed-preference evidence that the institution negatively impacts firm performance. Alternatively, this may reflect workers not valuing the institution enough to take it up.
Interpretation and survey evidence
To understand the apparently limited effects of worker voice, we draw on surveys of Finnish worker representatives. Respondents emphasise that worker voice institutions are primarily a venue for information sharing between workers and employers, not for co-decision-making.
As the two panels of Figure 4 show, worker representatives do not believe their position affords much decision-making power, and believe that managers often ignore their advice or are unwilling to take worker voice into account. This finding complements evidence surveyed by Jӓger et al. (2021b) suggesting that most existing worker voice institutions fail to convey much power to workers.
Figure 4 Survey evidence from worker representatives
Overall, we find non-existent or small positive effects of two Finnish reforms introducing rights to worker voice on worker and firm outcomes. The limited effects of the reforms may plausibly reflect the limited power conveyed by the resulting worker voice institutions.
We conclude the paper by noting that Finnish worker voice institutions appear to operate through largely benign channels of information sharing and cooperation that do not substantially improve worker outcomes but also do not harm firm performance.
There remain many open questions about the effects of worker voice institutions. As Jӓger et al. (2021b) describe, we lack evidence on the interaction of worker voice with other labour-market institutions such as unions or collective bargaining; there exists little evidence on the general-equilibrium economy-level effects of worker voice laws; and there may be important interactions between worker voice and changes to the future of work stemming from automation or the rise of non-standard work arrangements.
Blandhol, C, M Mogstad, P Nilsson and O Vestad (2020), “Do employees benefit from worker representation on corporate boards?”, NBER Working Paper w28269.
Blandhol, C, M Mogstad, P Nilsson and O Vestad (2021), “Benefits to employees of worker representation on corporate boards”, VoxEU.org, 13 February.
Bryson, A, and R Freeman (2013), “Employee perceptions of working conditions and the desire for worker representation in Britain and the US”, Journal of Labor Research 34: 1–29.
Freeman, R, and E Lazear (1995), “An economic analysis of works councils”, in J Rogers and W Streeck (eds.), Works Councils: Consultation, Representation, Cooperation in Industrial Relations, NBER Comparative Labor Markets Series.
Freeman, R, D Blanchflower and A Bryson (2020), “Unions raise worker wellbeing”, VoxEU.org, 11 November.
Harju, J, S Jäger and B Schoefer (2021), “Voice at work”, CEPR Discussion Paper 15874.
Hirschman, A (1970), Exit, voice, and loyalty, Cambridge, MA: Harvard University Press.
Jäger, S, B Schoefer and J Heining (2020), “Labour in the boardroom: The effects of codetermination on firm performance and wages”, VoxEU.org, 8 April.
Jäger, S, B Schoefer and J Heining (2021a), “Labor in the boardroom”, Quarterly Journal of Economics 136 (2): 669–725.
Jäger, S, S Noy and B Schoefer (2021b), “What does codetermination do?”, NBER Working Paper w28921.
Jensen, M, and W Meckling (1979), “Rights and production functions: An application to labor-managed firms and codetermination”, The Journal of Business 52 (4): 469–506.
Keskinen, M (2017), “Workplace cooperation and firm performance: Evidence from Finland”, Master’s thesis, Aalto University.