Mandatory Worker Representation on the Board and Its Effect on Shareholder Wealth

• Published date: 01 March 2018
• Author: Stefan Petry
• DOI: http://doi.org/10.1111/fima.12181
• Date: 01 March 2018

Mandatory Worker Representation

on the Board and Its Effect

on Shareholder Wealth

Stefan Petry∗

Several countries legallymandate representation of workers on boards of directors. The evidence

on the shareholder wealth effects of such a corporate governance design is mixed. I examine

abnormal announcement returns around major milestones leading to the passing of the German

Codetermination Act in 1976. I find that news about the act causes an average decline in the

equity value of firms that are certain to have been affected by the new law of up to 1.5% relative

to the control firms. Firms close to the regulatorythreshold of 2,000 employees remain unaffected

implying an expectation of avoiding compliance.

I’ve never known a corporation in Germany where somebody could survive with

10 votes of the labor side against him.

Ferdinand Piech, Chairman of the Supervisory Board of Volkswagen AG1

A number of countries, including most European countries, as well as China and Egypt, legally

mandate that a certain proportion of members of the board of directors be drawn from the ranks

of a firm’s workers(Botero et al., 2004). It is typically expected that such regulatory intervention

into corporate governance has a negative effect on firm value as it moves firms away from their

optimal board design. However, existing studies find mixed results.

Most of these studies examine the effects of the German 1976 Codetermination Act. The

Codetermination Act requires that firms with more than 2,000 employees in Germany have

boards in which 50% of the directors are worker representatives. The chairman of the board,

a shareholder representative, has a tie-breaking second vote providing the shareholder-elected

directors with a majority on the board. This may attenuate the influence of workers on the

board. However, some board decisions, such as the reappointment of the management team

and the chief executive officer (CEO), require approval from at least two-thirds of the board

members. In firms with 50% of the board made up of workers, a corporate policy that is biased

I thank an anonymous referee and Raghavendra Rau (Editor), as well as Bruce Carlin, Daniel Ferreira, Aleksandra

Gregoric,Han Kim, Rose Liao, Spencer Martin, Alexandra Niessen, John Nowland,Richard Roll, Jared Stanfield; seminar

participants at the Universities of Mannheim, Melbourne, Oxford, and UCLA Anderson; and conference participants

at the Australasian Finance and Banking Conference, the EFM Symposium on Corporate Governance and Control, the

FMA Meeting, and the German Finance Association Annual Meeting for helpful comments. I am particularlygrateful to

Bruce Grundy, Mark Garmaise, and David Yermack for their suggestions and support. I thank Hoppenstedt for giving

me access to their company archive. Financial support from the Cambridge European Trust, the Cambridge Political

Economy Society Trust,FAZIT Stiftung, and the LCCI CommercialEducation Trust is gratefully acknowledged. All errors

are my own responsibility.

∗Stefan Petry is an Assistant Professorof Finance in the Alliance Manchester Business School, University of Manchester

in Manchester,United Kingdom.

1“Volkswagen Board Is Divided Over CEO’s Future at Car Maker,” WallStreet Journal, March 01, 2006.

Financial Management •Spring 2018 •pages 25 – 54

26 Financial Management rSpring 2018

toward worker interests may become more important for the CEO than corporate policies that

may create conflict. In the extreme, this could lead to collusion, as argued in Tirole (1986)

and Pagano and Volpin (2005). The introductory quote by Ferdinand Piech in reference to the

Codetermination Act seems to confirm that the CEO’s job security depends upon the support of the

worker-directors.

Yet, the empirical evidence regarding the effect of the Codetermination Act on firm value is

ambiguous. For example, Fauver and Fuerst (2006) and Benelli, Loderer, and Lys (1987) report

an insignificant effect on firm value. Renaud (2007) f inds positiveproductivity effects in affected

firms 15 years after the Codetermination Act. Gorton and Schmid (2004) note a negative effect

on firm value. However, theyestimate a sur prisingly largerelative stock market discount of 21%

in 1989 and 43% in 1992 for firms whose boards were composed of 50% workers compared to

those whose boards were composed of only one-third workers.

A likely explanation for the mixed findings in the papers cited above is data limitations in the

form of small sample sizes, which may not be representative, and identification issues due to

either analyzing the effect of the Codetermination Act over a very long time period of up to

20 years (Renaud, 2007) or to examining data many years after 1976 (Fauver and Fuerst,

2006). To identify the Codetermination Act as the cause of the observed firm behav-

ior, it is important to compare firm outcomes over a relatively short period after the

2,000 employee threshold of the Codetermination Act became known. Once the mandated em-

ployeethreshold was known, the legal threshold wouldhave become part of the f irms’ optimization

sets.2That is also the reason why a regression discontinuity (RD) design comparing the outcomes

of firms around the 2,000 employee threshold after the 1976 Codetermination Act’s enactment

cannot establish causality.The RD design requires that f irms are randomly distributed around the

threshold. Since this is not the case in Germany, as the threshold is part of the firms’ optimization

sets, RD results still suffer from an endogeneity bias.

This paper is among the first to employ an event study methodology using daily return data

around 11 announcements related to the passing of the Codetermination Act from 1974 to 1979.

Since the 2,000 employee threshold was not established before the first draft of the law was

announced, the results are less likely to be impacted by an endogeneity bias and provide a more

accurate estimate of the wealth effects. The advantage of using stock returns is that they measure

the net effect given that the act could havehad positive and negative effects. The sample consists

of 476 German firms, of which 124 were affected by the law (“affected firms”) and 352 were

unaffected (“control firms”).

I find that fir ms that would certainly have been affected by the act, which I label “certainly

affected” firms, experience a significantly negative reaction to the announcement of the first draft

of the law, of −1.5% relative to the control firms, during the one-day announcement window.The

abnormal returns of the certainly affected firms around the remaining events are economically

smaller or insignificantly different from zero highlighting the importance of the first event in

this analysis. Interestingly, all abnormal event returns for firms that would have been affected by

the law, but close to the 2,000 employee threshold, are insignificantly different from zero. This

implies a market expectation among such firms that they could make adjustments and fall below

the 2,000-employee threshold.3

2For example, if a negativeimpact of 50% workers on the board was expected, firms would compare the cost of having

workers on the board with the cost of avoidingthe regulation (e.g., through outsourcing, f irm restructuring, and/or laying

off employees).

3For evidence of governancelaw avoidance measures undertaken by firms, see Gao, Wu, and Zimmerman (2009).