Supplementary written evidence from the TUC (CGV0177)

Introduction

1.1       The Trades Union Congress (TUC) is the voice of Britain at work. We represent nearly six million working people in 52 unions across the economy. We campaign for more and better jobs and a better working life for everyone, and we support trade unions to grow and thrive.

1.2       The TUC submitted written evidence to the Committee’s corporate governance inquiry in November 2016. In addition, the TUC has given oral evidence to the Committee at hearings on 15th November and 20th December 2016.

1.3       This note augments the evidence previously submitted by the TUC, in particular in relation to workers on boards, which was discussed in detail at the Committee’s evidence hearing on 20th December 2016. In addition, the note will respond to supplementary evidence CGV0173, which includes some inaccuracies relating to the evidence on workers on boards, as well as inaccurate and misleading references to the TUC’s previous evidence.

2.1                   Quantitative inter-company studies of the impact of worker board representation on company performance

Methodological caveats

2.1       There are a significant number of academic research studies that set out to assess the impact of worker board representation on company performance. There are around 30 German studies dating from 1982 to the present, and a small number that focus on other countries (including France, Norway and Sweden)[1].

2.2       There are some important methodological caveats that should be understood in relation to these studies. The most important is the difficulty of finding an appropriate control group of companies. This reflects the fact that where worker board representation is in place, this stems from a legal requirement that applies to certain categories of companies in that country (for example, the requirements often apply to companies over a minimum size, some countries limit requirements to plcs while in some requirements vary according to sector and so on). This means that all companies that share these characteristics in terms of size, sector and so on, will all either have worker board representation in place or not. This creates a problem in terms of what can be used as a control group of companies with which to compare the outcomes of those with worker board representation, as there will, by definition, be other differences between the two groups, in addition to the difference of worker board representation.

2.3       The second methodological caveat stems in part from this first: there is considerable variation between the studies in terms of what they use as a control group, which will inevitably limit the comparability of the different studies. Some studies examine company performance before and after the introduction of requirements on worker board representation; some use smaller companies, that are not subject to requirements on worker board representation, as a control group; and some of the German studies compare companies where one third of the board are worker directors with those where one half of the board are worker directors. These differences will inevitably have an impact on the findings of the studies. They also affect the quality of the studies, as arguably some methodologies are more robust than others, and the relevance of each study to the debate today.

2.4       In addition, the studies examine a wide range of different impacts, including productivity, profitability, return on equity, shareholder value, share price, price-cost margin, company innovation, employment development, company efficiency, market value, reliable accounting, company survival and return on assets. Again, this means that comparisons between the studies need to be undertaken with care.

Interpreting the quantitative inter-company studies[2]

2.5       Given the number of studies that exist, it is misleading to point to just one study and assert that this demonstrates the impact of worker board representation.

2.6       It is also the case that assessing the studies as a whole is made more difficult by their differences in terms of methodology and indicators. There are different ways of assessing the range of studies, as set out below.

2.7       Focussing on the studies set out in the publication Board-level employee representation rights in Europe Facts and trends, and looking at the different impacts studied, on a simple numerical basis, positive effects of worker board representation are found by 13 studies, negative effects by nine studies and no significant impact by 11 (please note that where studies have found a mixture of positive or negative or neutral effects these impacts have been counted separately, so the total number of studies is less than the number of impacts given here). However, this approach gives equal weighting to different indicators of company performance.

2.8       Assessing each study to derive an overall finding, and ignoring those that give mixed results for different indicators, ten studies find positive effects while four find negative effects and eight find no significant effects. Including those studies that give mixed results and interpreting the results as positive or negative or neutral overall from a narrow perspective of company performance (ie, ignoring factors such as higher wage levels that would denote a positive impact for the workforce but to a conventional economic approach could be seen as a ‘cost’ to the company, regardless of the potentially positive impact on other factors, such as labour turnover) ten studies find positive effects, while seven find negative effects and 11 find no significant overall effects.

2.9       While the results of these studies are clearly mixed, they are overall more positive than negative. In addition, more recent studies tend to be more positive, possibly reflecting improvements in research methodologies over time.

2.10   An updated version of this assessment has been set out in a 2014 PhD thesis[3] by Aline Conchon, co-author with Professor Jeremy Waddington of the University of Manchester of a recent book Board-Level Employee Representation in Europe[4]. The findings of the additional studies included in the updated version are again mixed, but clearly more positive than negative. In terms of the different impacts studied, four are positive, three neutral and one negative. Looking at the studies as a whole, one is positive, two neutral, and one mixed, finding a positive impact on three indicators (R&D investment, payment of dividends and reducing stock volatility, no significant impact on one indicator (return on sales) and a negative impact on one indicator (Tobin’s Q firm value). In addition, the TUC is aware of several studies that have been published since 2014, which are generally positive in their findings[5][6].

2.11   Although the overall findings of these studies are more positive than negative, the TUC does not generally choose to refer to these studies in our work on worker board representation. This is because we believe that the methodological caveats set out above significantly reduce the robustness of the research on company performance; without a control group of companies that is genuinely equivalent to those with worker board representation, it is difficult for reliable comparisons to be made. In addition, the overall findings are mixed and explaining them correctly, including their limitations and variations, is not straightforward, as this brief analysis demonstrates.

2.12   Given the range of studies that exist, it is clearly not robust to simply quote one study as evidence for the effectiveness or otherwise of worker board representation. For example, evidence CGV0006’s response to the question Should there be worker representation on boards and/or remuneration committees? is

“No. Gorton and Schmid (2004) show that greater representation reduces long-run firm value.”

2.13   Evidence CGV0173 (by the same author), refers to the same study:

“Gorton and Schmid (2004)…show that German firms with greater worker representation trade at a 31% valuation discount….”

2.14   To claim that one study shows that worker board representation either works or does not work, with no mention of the other studies and the variation in their findings, is at best misleading.

2.15   In addition, this particular study (Gorton and Schmid 2004) should not be used to assess the impact of worker board representation in comparison with no worker representation, which is clearly the most relevant issue in the context of the UK debate. This is because Gorton and Schmid 2004 compares German boards with equal worker and shareholder representation (parity representation) with those with one third worker representation, and finds that the boards with parity worker representation perform worse than those with one third worker representation. It does not include a comparison with companies with no worker board representation, and therefore should not be used to assess the impact of introducing worker board representation in a situation where worker board representation does not currently exist. (It is worth noting that there are other studies (for example, FitzRoy and Kraft 2005[7] and Renaud 2007[8]) that find that boards with parity representation perform better than those with one third worker representation, so it should  not be assumed that the conclusion that companies with parity representation perform less well than those with one third representation is uncontested).

2.16   The findings of Gorton and Schmid have been discussed in later studies. A 2006 study by Fauver and Fuerst[9] published in the Journal of Financial Economics compares companies with parity representation, one third worker representation and no worker representation. This widely-cited study concludes:

“We show that prudent levels of employee representation on corporate boards can increase firm efficiency and market value. This result contrasts with that of Gorton and Schmid (2004). Although the optimal representation is likely below 50%, the level often mandated by law for large German corporations, it is certainly positive. We interpret our results as implying that there is an inverse U-shaped relation between firm value and employee representation on German corporate boards. We propose that employee representation provides a credible channel for the flow of information to the highest levels of the firm. Consequently, this superior information improves decision-making by the board. Moreover, we find that industries that require more intense coordination, integration of activities, and information sharing, that is, industries such as trade, transportation, computers, pharmaceuticals, and other manufacturing, benefit more from employee board representation” (emphasis added).

2.17   The Fauver and Fuerst study echoes the finding of Gorton and Schmid that company performance in companies with one third worker representation is better than those with parity representation, but finds that companies with parity representation and one third representation both outperform those with no worker board representation.

2.18   Evidence CGV0173 also cites a 2006 study by Faleye, Mehrotra, and Morck, which it refers to as a study of “US firms where employees own a large share of the firm and are thus involved in governance”.

2.19   This study examines a sample of 226 US companies in which the workforce owns 5% or more of the companies’ shares. In 101 of the 226 companies studied the workforce owned less than 10% of the company and the workforce owned more than 25% of the companies’ shares in just 15 of the 226 companies.

2.20   Worker board representation is not part of the corporate governance system in the US. There is no indication that any of the companies examined in the research had company workers sitting on their boards.

2.21   Employee share ownership is very different from workers sitting on company boards and one cannot be taken as a proxy for the other. The paper doesn’t discuss the voting strategies or practicalities of the employee votes, so it is not clear whether there is any coordination of the workforce votes, and in nearly half (45%) of the companies studied the employees owned less than 10% of the shares. Overall, the findings of this study cannot be used to reach conclusions about workers’ voice in corporate governance and the relevance of this paper to the debate about worker board representation is marginal at best.

2.22   It is also important to note that a number of other, more recent UK studies show a positive relationship between employee ownership and company performance. A government-commissioned review of employee ownership carried out by Graham Nuttall in 2012 reviewed the evidence on this and concluded that on balance existing academic studies associate employee ownership with higher productivity levels. It also quoted research showing that employee owned companies had demonstrated greater economic resilience over the economic downturn following the financial crisis[10]. The results included variations, including by company size, that are discussed in the report. Interestingly, the review states that “The benefits of employee ownership are most likely to be realised when ownership co-exists with wider drives for employee participation in decision-making.” This is an interesting conclusion in the context of the debate on worker representation on company boards.

3.1                   Inter-country comparisons

3.1       An alternative to trying to assess the impact of worker board representation on a company basis is to compare the economic and social performance of countries that have worker board representation in place with those without. The TUC’s original evidence to the inquiry included findings from the European Participation Index, which compares countries with high standards of worker participation (i.e. widespread rights and practices for board representation, workplace representation and collective bargaining) with countries with comparatively low standards in terms of their performance on the Europe 2020[11] headline indicators. Updated information on this is now available, and is set out below.

Comparative performance of countries with stronger rights versus weaker worker participation rights on five Europe 2020 headline indicators (2009-2014)

3.2       As noted in our original evidence, correlation does not prove causation. Nonetheless, it is striking that countries with strong workers’ participation rights out-perform countries with weak workers’ participation rights on all of the Europe    headline indicators. Of particular relevance to debate about the UK’s weak investment record and our approach to industrial strategy and is the finding that gross domestic expenditure on R&D is twice as high in countries with strong rights for workers’ participation as in countries with weak participation rights. Again, correlation is not the same as causation; but nonetheless these findings should be of interest to all those with an interest in the future success of the UK economy.

4.1                   Additional comments on research quality

4.1       Evidence CGV0173 explicitly criticises the TUC’s evidence to the inquiry. It notes that “The TUC’s written evidence states that employee representatives on boards are perceived favourably by other board members”, but goes on to claim that “almost all of the papers cited are unpublished and have not passed peer review.”

4.2       This assertion is simply wrong. All the papers quoted by the TUC on the attitudes of other board members to worker board participation are published, and the references are clearly given in the footnotes.

4.3       The main study discussed is a Swedish study that compares the results of two different surveys of Swedish chairpersons and company representatives that took place in 1999 and 2009. This is published in a Swedish publication Arbetsmarknad & arbetsliv, or in English ‘Labour and working’. This is a peer-reviewed publication, as is made clear on its website[12].

4.4       An additional study quoted is published in the Industrial Relations Journal. Its website describes it as “a cutting edge, research based, peer reviewed publication focusing on the changing nature, forms and regulation of the employment relationship”[13].

4.5       A third study quoted is by the Think Tank for Social Change (TASC) in Ireland. This is likely not peer-reviewed as TASC is not an academic organisation. However, few policy makers would reject the finding of every think tank and other organisation that produced research that is of a different nature from that of a research journal.

4.6       The TUC welcomes comments on its work and is happy to debate any of the points contained within it on the basis of the available evidence. However, it is important that that debate is transparent, and accurate when representing alternative viewpoints. It is a serious matter to denigrate the integrity or quality of another organisation’s work, especially when done without having approached the organisation directly to discuss concerns. We regret that our evidence to the BEIS Committee corporate governance inquiry has been the subject of inaccurate criticism.

4.7       Evidence CGV0173 goes on to set out a number of conceptual concerns about the respective roles of workers and shareholders and others. This submission will not go through these in detail, but would like to comment on one particular assertion:

“Workers’ risk aversion reduces their incentives to improve long-run value since they may sacrifice it to reduce risk.  Investment in intangible assets such as R&D, brand, customer loyalty, environmental protection, and indeed creation of new jobs, bear significant risk. Shareholders are willing to make these investments if they increase long-run return, but workers will be less willing to do so, since they will be more concerned about preserving their existing jobs.”

4.8       The inter-country research presented earlier contradicts this assertion, as it shows that countries with strong rights for workers’ participation invest twice as much in R&D as countries with weak participation rights.

4.9       The assertion that shareholders are willing to back investments in intangible assets is a controversial claim with which many experts would disagree and should be examined carefully on the basis of the evidence. There is considerable evidence that contradicts this claim, including the analysis of Professor John Kay’s Review of UK Equity Markets and Long-Term Decision Making[14].

4.10   A recent survey of over 1,200 firms published by the Bank of England found that the most important reason for underinvestment was a constraint on using profits for investment purposes, with three quarters of firms rating distribution to shareholders (including dividends and share buybacks) and purchase of financial assets (including mergers and acquisitions) ahead of investment as the most important use of internally generated funds. Strikingly, 80 percent of publicly owned firms agreed that financial market pressures for short-term returns to shareholders had been an obstacle to investment[15].

4.11   It has been well-documented that holding periods for shares have reduced considerably in recent years. Generally, workers' time horizons are longer-term than most investors, giving them a greater interest in the long-term success and survival of the firm than shareholders, many of whom are gone after three or five years (putting to one side high frequency trading, which brings average holdings periods down dramatically).

4.12   In addition, a counter, though also normative, argument to the one set out in evidence CGV0173, is that workers are more likely to invest in firm-specific capital if their investments are protected through voice in decision-making.

4.13   More broadly, our experience of industrial relations is that in reality workers and their representatives are often prepared to accept and take risks, precisely in order to secure the long-term future of their enterprise. Workers’ ‘existing jobs’ can never be taken for granted and in reality innovation and change are necessary in order to secure not only existing jobs, but crucially future jobs. Workers and trade unions are aware of this and this is reflected in the deals that they do with company management.

4.14   This submission is not aiming to reach a definitive assessment of the evidence on the respective influence of shareholders and workers on firm-level investment. However, we do aim to demonstrate that the assertion that shareholders are more likely than workers to support long-term investment in intangible assets is highly contestable at best.

4.15   Finally, in its first two pages, evidence CGV0173 makes a direct criticism of another section of the TUC’s original evidence to the BEIS Committee corporate governance inquiry. It says:

“The TUC’s paragraph 3.11 wished to support the claim that high pay ratios are demotivating, and quoted an unpublished draft of a paper as evidence – even though there has been a published version available for three years. This is an example of confirmation bias – any partisan group, with a pre-conceived viewpoint, can almost always find evidence to support this viewpoint.”

4.16   The TUC quoted three papers to support this point, only one of which was unpublished (which was clearly noted in the footnotes). The unpublished paper cited in our evidence was sourced from Will Hutton’s interim report into Fair Pay in the Public Sector[16], in which it was quoted. Without a subscription to the Journal of Banking & Finance, in which the paper was finally published, we were unaware that an updated version had been published.

4.17   This demonstrates not confirmation bias, but simply the fact that the TUC was unaware of the updated version of that particular paper. Had this been raised with us directly, we would have responded immediately by withdrawing the footnote from our evidence.

4.18   The TUC remains of the view that there is convincing evidence in support of the point that high pay differentials are associated with lower firm performance. This area was examined in some depth by Will Hutton in his Fair Pay review. It is worth reproducing some paragraphs from his report to demonstrate that this relationship is not something that the TUC has simply asserted because it suits our position, but is an argument put forward by a range of respected academics and authors on the basis of evidence. To quote from Will Hutton:

“A wide range of academic studies, covering large and small businesses across different sectors, and even sports teams, in North America and Europe, suggest there is a strong correlation between narrower pay dispersion within an organisation and improved organisation performance…Overall, the evidence on pay compression within organisations leans towards it having more benefits than costs.”[17]

4.19   There are more recent academic studies that support this conclusion. For example, a 2016 study based on the MSCI found that “Labor productivity, measured by sales per employee, was lower for companies with higher intracorporate pay gaps on average during the study period, a finding we noted in nine of ten GICS sectors”[18].

4.20   The TUC aims to make the best use of the resources and research available to us in good faith in our work. The TUC welcomes all comments on its work and is happy to debate any of the points contained within it on the basis of the available evidence.

5.1                    

6.1                   March 2017