The Determinants and Effects of Governance Provisions: Evidence from The Netherlands$

(1)

The Determinants and Effects of Governance Provisions: Evidence from

The Netherlands

I

N. Meijer1, supervised by prof. dr. C.L.M. Hermes2

Abstract

Recent events in The Netherlands have caused the Dutch government to introduce a national governance provision. However, to my knowledge, no empirical evidence has yet been presented of the determinants and effects of Dutch governance provisions, even though the national set-ting of governance studies is very important. In this thesis, I aim to investigate the determinants and effects of Dutch governance provisions. Using fixed effects panel regressions with a unique dataset on Dutch governance provisions, I find a positive effect of Dutch governance provisions on the firm’s focus on environmental stakeholders, but a negative effect on capital expenditures. I also conclude that factors such as leverage and firm size are more important factors explaining financial performance than Dutch governance provisions, and find only a weak negative effect of governance provisions on firm profitability. I do find evidence that the strength of Dutch gover-nance provisions differs using a Dutch entrenchment index: the entrenchment index in most cases shows significant effects while the other governance provisions do not. I conclude that the fea-tures of Dutch corporate governance already emphasize balance of power between management and shareholders, which mitigates the effect of Dutch governance provisions on firm output. Keywords: Corporate Governance, Governance Provisions, Stakeholder Value, Financial Performance, Entrenchment Index

1. Introduction

Corporate governance concerns the mechanisms by which the owners of the firm exercise con-trol over management to ensure their interests are protected. This stems from the separation of ownership and control of the firm; the main advantage of this is that specialists can be hired for managing the firm one the one hand, and taking risk by owning its shares on the other. However, this conversely gives rise to agency costs due to a possible misalignment of incentives between the owners of the firm (shareholders) and its management, for instance if managers are opportunistic and consequently extract private benefits (Jensen and Meckling,1976). Agency costs have a nega-tive effect on the efficient operation of the firm and should thus be managed. To this end, several

I_{Master’s thesis MSc Finance, University of Groningen, January 2018.} 1_{Student number: S2022915.}

2_{Department of Economics, Econometrics and Finance, Faculty of Economics and Business, University of}

(2)

governance mechanisms are available to shareholders that ensure management acts in their inter-ests. One such governance mechanism is the market for corporate control. It essentially enables a replacement of incumbent management by a corporate board that better represents the owner’s interests. It is this threat of being replaced that ensures managerial decision making reflects the interests of the shareholders (Fama and Jensen,1983). While such governance mechanisms which emphasize shareholder control mitigate agency costs by ensuring the shareholder’s interests are represented, some scholars have argued that the interests of no one stakeholder (e.g. solely share-holders) should have ultimate priority over other stakeholder groups. Managerial decision mak-ing should instead balance the interests of all stakeholders simultaneously (e.g. customers, sup-pliers, employees and the community as well as corporate owners)(see e.g. (Cornell and Shapiro,

1987); (Jensen, 2001)). As an example of the rationale behind this stakeholder view, Prahalad

(1994) posits that excessive shareholder control leads to an emphasis on easily quantifiable finan-cial metrics such as stock price fluctuations. Even though cutting out potential inefficiencies in a firm’s business may at first be detrimental to current financial performance, timely structural change is at times necessary to create sustainable value. While the debate between shareholder and stakeholder value advocates inspire a stream of research on the purpose of the firm in its own right, the insights from both perspectives suggest that the way a firm’s control structure (which I refer to as shareholder power versus managerial power) is organized can significantly influence its strategic decision making. Recent events with regard to corporate governance in The Netherlands demonstrate that pressure from (in this case) shareholder control can indeed alter the strategic course of the firm. In early 2017, consumer goods firm Unilever unexpectedly received a takeover bid from U.S. competitor Kraft Heinz Company (Schwartz,2017). Even though the firm had pursued a long-term value creation strategy by focusing on sustainability (Unilever ranked number one on the Dow-Jones Sustainability Index), pressure from shareholders led the firm to pivot to a more shareholder oriented focus. Unilever has sold divisions, bought back shares and increased dividends by 12 percent with the proceeds as a result (Kooiman and Van der Heijden,

2017). Similarly, AEX listed paint and chemicals firm AkzoNobel received multiple consecutive takeover bids from U.S. competitor PPG Industries over the course of 2017. PPG Industries argued that AkzoNobel’s long-term stakeholder value strategy towards employees, the environment and R&D had neglected shareholder value, and therefore promised immediate shareholder wealth by offering a 31 percent premium (Verbraeken,2017). Multiple AkzoNobel shareholders demanded AkzoNobel management to consider PPG’s offer as a result. AkzoNobel management pivoted to shareholder value creation to convince its shareholders to reject PPG’s takeover bid, by promising to sell off its specialty chemicals division and increase dividend by 50 percent with the proceeds, with an extra one billion Euros special dividend by November 2017 (Zevenbergen,2017).

(3)

have the right to buy the bidder’s stock at a steep discount in case of an unwanted takeover. As such, management is to a certain extent protected against being easily replaced under shareholder pressure (Gompers et al.,2003). Returning to the example described above, the Dutch government has announced several governance provisions on a national level that should protect firms from hostile takeovers and thereby provide more power to management. For instance, the Dutch gov-ernment will be able to veto hostile takeovers of firms that are "vital" to the Dutch economy. Also, a provision is introduced that provides management the option to take 250 days to contemplate a drastic strategic change proposed by shareholders (Hoeks and Dobber,2017). The issue with these plans is that the effect of the governance provisions as proposed by the Dutch government on firm output in a Dutch setting specifically, remains unclear. Most empirical evidence of gov-ernance provisions on firm output is U.S. based and yields both positive and negative results. On the one hand, the adoption of governance provisions can result in more corporate attention to non-investor stakeholders (Kacperczyk,2009). On the other hand, governance provisions can also remove managerial incentive to perform.Bebchuk et al.(2009) refer to this as managerial en-trenchment, and find a negative relationship between this entrenchment index (containing the six most protective U.S. governance provisions found inGompers et al. (2003)) and firm value and returns. More integral to the context of this thesis, however, is the fact that the national setting of governance studies is very important, since e.g. legal constructs and financial and economic developments which have a large effect on the optimal governance conditions differ from coun-try to councoun-try (Doidge et al.,2007). For instance, Dutch law requires that the responsibility for the strategy and well being of the firm lies with management, not the shareholders, regardless of the way the firm’s control structure is organized3. So, even though the Dutch government’s national governance provisions will most likely provide more power to management, the ques-tion remains what management does with this (extra) power and how this pronounces itself in this specific Dutch setting. Furthermore, it is unclear whether these national governance provi-sions equally affect all Dutch firms in all industries; e.g. Gillan et al.(2003) argue that since the severity of agency costs differ from firm to firm, the firm’s optimal control structure defined by specific governance provisions differs between firms as well. This leads in the U.S. to varying firm and industry determinants of the presence and strength of governance provisions. To my knowledge, no empirical evidence has yet been presented on the determinants and the effects of governance provisions in a Dutch context. In this thesis, I aim to investigate these issues by asking the following main research question:

What are the determinants and effects of Dutch governance provisions?

2. Theory and hypotheses

2.1. Determinants of governance provisions

To investigate the determinants of governance provisions for Dutch firms, it is firstly impor-tant to further understand the reason why firms have governance provisions. I will then inves-tigate whether empirical evidence points to specific firm determinants of governance provisions,

(4)

which may be extended to my own analysis. As briefly introduced in section1, the underlying reason for the divergence of ownership and control considers the unique characteristics of the capital markets versus the labor market. Investors are able to devote time on capital markets, thereby holding well diversified portfolios and are thus best suited to bear the firm’s risk by own-ing its shares. Corporate owners consequently hire specialists with management experience and expertise from the labor markets to operate the firm. This is referred to as the agency relation-ship, where the principal (shareholder) hires an agent (manager), the agent exerts effort on the job required by the principal, and the principal measures the agent’s performance and awards him accordingly (Fama and Jensen,1983). The issue with the agency relationship is a potential misalignment of incentives. To illustrate this,Jensen and Meckling(1976) describe the notion of perquisites: whenever an entrepreneur needs additional capital and thus attracts outside financ-ing, the agent is no longer fully accountable (i.e. does not own 100% of the firm). Consequently, the agent’s incentives change as he is only liable for part of the firm’s expenditures, and may thus engage in extracting private benefits at the expense of the corporate owners (shareholders). This inefficiency is referred to as agency costs. These agency costs cannot be resolved by using contracts alone. It is for this reason that various control structures (which I refer to as managerial power versus shareholder power) arise to mitigate agency costs and thus determine the power shareholders have versus managers. These control structures can consist of elements internal to the firm such as boards of directors and managerial pay structures (e.g. stock options), as well as external mechanisms such as legal requirements and the market for corporate control (Fama and Jensen(1983);John and Senbet(1998)). While such structures increase shareholder control, some scholars have argued that power should not be reserved to shareholders only (see the example from Prahalad (1994) in section 1), since this may lead to too much focus on current financial performance metrics, neglecting sustainable value to various stakeholders as a result (e.g. Cor-nell and Shapiro(1987);Jensen(2001)). Empirical evidence shows that management indeed tends to sacrifice long-term investment such as R&D to improve short-term earnings when short-term shareholder ownership (e.g. hedge funds) is large. The effect of this phenomenon on long-term firm value is negative (Aspara et al.(2014);Brunzell et al.(2015);Bushee(1998)).

(5)

pro-visions, which implies more power to shareholders). They then investigate the impact of several industry-level and firm-level factors on the three indices in explaining governance structure varia-tion. Based on evidence from existing literature, they hypothesize that four industry factors affect the firm’s governance structure: investment opportunities, degree of competition, monitoring en-vironments and information enen-vironments. There may be two opposite effects of competitive pressure on governance provisions: either the marginal cost of bad managerial performance is increased by highly competitive environments, thus increasing the need for monitoring mech-anisms such as governance provisions, or highly competitive environments already discipline managers to perform in a way that governance provisions are less necessary (i.e. there are no op-portunities in a highly competitive environment for managers to shirk). I will investigate which is true for Dutch firms in section4.1. Furthermore, monitoring environments as a result of leverage may introduce self-governance characteristics which reduce the need for alternative governance mechanisms (Jiraporn and Gleason,2007).Opler and Titman(1994) also show that this is the case for highly leveraged firms due to the danger of financial distress cost.

On an industry-level, Gillan et al.(2003) find that an industry’s investment opportunities is positively related to the strength of board monitoring (that represent shareholders). Thus, firms operating in industries with high investment opportunities generally exhibit higher shareholder power. On the other hand, product uniqueness, degree of competition and average leverage are all negatively related to the governance indices, and thus a higher level of any of these industry factors lead, on average, to firm governance rules that provide more power to management.

On a firm level, holding industry constant, the authors find that higher growth opportunity firms tend to have weaker boards, and thus have higher management power. Firm size is also neg-atively related to the governance indices, implying that management generally has more power in larger firms. Incentive-based compensation is positively related with the board index, but nega-tively related with the charter index, thus suggesting that firms with stronger boards rely more on incentive-based pay, whereas firms with more governance provisions compensate their managers less on an incentive basis. Finally, they find that firms operating in a more uncertain information environment (as measured by return volatility) have fewer governance provisions.

These findings indicate that industry and firm variation can indeed explain the use of gover-nance provisions in a U.S. context. I hypothesize that the same holds in Dutch context, thus: Hypothesis 1 Firm specific factors can explain the variation in the use of governance provisions for Dutch firms

2.2. Effects of governance provisions 2.2.1. Stakeholder value

According to the previous Dutch Minister of Economic Affairs who introduced the national governance provision introduced in section1, the main goal of issuing the national governance provision is to remove shareholder pressure, so that firms can focus more on non-investor stake-holders, thereby creating sustainable value4. The issue with this argumentation is that

sharehold-4_{See the Minister’s note to the Dutch parliament at https://www.rijksoverheid.nl/documenten/kamerstukken/}

(6)

ers need not per se have a short-term orientation. Opposite to the potential negative effects of short-term shareholder ownership described in section 2.1, shareholders may have a long-term view as well. An example of such a shareholder is a pension fund, which exhibits very low portfolio turnover (some pension funds may even hold one particular stock for over a decade) (Gilson et al.,1991). Pension fund managers also are usually not compensated for their portfolio performance, but rather receive a fixed salary. Therefore, a pension fund will have a long-term investment view, in fact positively influencing the firm’s focus on sustainable value (Johnson and Greening,1999). Removing shareholder control may thus not necessarily lead to more corporate attention to non-investor stakeholders. Furthermore, as outlined in section2.1, without some form of shareholder control to mitigate agency costs, management may engage in empire building and extracting private benefits. Therefore, to ultimately be able to determine the effect of governance provisions on firm output, evidence from existing literature has to be assessed.

Some empirical evidence points to a positive relationship between governance provisions and corporate attention to non-investor stakeholders. For instance,Kacperczyk(2009) investigates the effect of one specific type of governance provisions (anti takeover provisions), on five KLD data variables that measure corporate attention to stakeholders: community, minorities, employees, the natural environment and customers. She finds that an increase in anti takeover provisions increases corporate attention to non-investor stakeholders in U.S. firms. These findings would thus support the goals of the Dutch Minister of Economic Affairs of introducing the national governance provision in The Netherlands.

However, it is important to keep in mind the conclusions from Bebchuk et al. (2009), who argue that governance provisions may have negative effects when they protect management too well. They scrutinize the 24 governance provisions in Gompers et al. (2003), and find six gov-ernance provisions that they refer to as “entrenchment” provisions, which are provisions that protect incumbent management from removal (or its consequences). The authors posit that out of those 24 governance provisions, these six entrenchment provisions will have the strongest effect of shifting power to management, away from shareholders. They conclude that entrenchment has negative effects: as the disciplinary threat of being replaced is removed, managers are more inclined to engage in shirking and empire-building, and extracting private benefits. Furthermore, unsuccessful managers within firms that already exhibit low value, may have more incentive to entrench themselves through these mechanisms to prevent them from being replaced.

(7)

other managers instead). However, in The Netherlands, it is required by law that management at all times bears the responsibility for the well being and strategy of the firm5. A true shareholder power control structure may thus not exist in The Netherlands. This is reinforced further when looking at the board structure of U.S. firms versus those in The Netherlands. U.S. firms primarily function with one tier boards that simultaneously assume executive and supervisory responsi-bilities in one board of directors. Conversely, Dutch firms primarily have two tier boards where the executive board is advised and monitored by a separate supervisory board. In an integrative comparison of U.S. and German board structures and responsibilities,Block and Gerstner(2016) find that U.S. one tier boards represent shareholder value, whereas German two tier boards are driven by stakeholder primacy.Douma(1997) shows that German and Dutch two tier boards are comparable. It thus seems that Dutch corporate governance is shaped by legal requirements and board structures that emphasize balance of power between managers and shareholders more than the U.S. context (with balance of power between shareholders and management, I mean that no one group has ultimate authority over the other; power is more equalized between shareholders and management).

Thus, again, Dutch corporate governance emphasizes balance of power between management and shareholders more than in the United States. This result, together with the entrenchment conclusions fromBebchuk et al.(2009), would suggest for Dutch firms that shifting power from shareholders to managers even more through the use of governance provisions could increase the inclination of managers to shirk and extract private benefits. Or, financially weaker firms that already lack the financial resources needed to address sustainable value creation adopt more governance provisions to protect themselves. This would suggest the relationship between gov-ernance provisions and corporate attention to non-investor stakeholders in The Netherlands is negative, leading to the following hypothesis:

Hypothesis 2 The number of governance provisions is negatively related with non-investor stakeholder value of Dutch firms

However, the same Dutch corporate governance features could also mitigate the entrenchment effect of the use of governance provisions. As Dutch corporate governance again emphasizes bal-ance of power between managers and shareholders already, the effect of governbal-ance provisions to shift power from shareholders to managers may also be less pronounced. Moreover, empirical evidence byKacperczyk(2009) as described earlier shows that one type of governance provisions (anti takeover provisions) has a positive effect on corporate attention to non-investor stakehold-ers. Thus, it can also be hypothesized that:

Hypothesis 3 The number of governance provisions is positively related with non-investor stakeholder value of Dutch firms

Thus, governance provisions provide management more power to focus on sustainable value. The question is, will management of Dutch firms actually do so? I will investigate empirically which hypothesis is more supported in section4.2.

(8)

2.2.2. Financial performance and differences in provision strength

The conclusions from the literature in the previous section on the effects of governance pro-visions on firm output leave the question what this means for the firm’s financial performance.

Gompers et al.(2003) find a negative wealth impact for U.S. firms. The authors created a Gov-ernance Index out of 24 govGov-ernance provisions for a sample of 1500 U.S. firms during the 1990s. They then compared the sample of firms with the strongest shareholder rights (i.e. lowest num-ber of governance provisions) to those with the weakest shareholder rights (i.e. highest numnum-ber of governance provisions), and found negative performance results. The firms with weakest share-holder rights exhibited e.g. lower firm value, lower profits and lower sales growth compared to the strongest shareholder rights firms. Furthermore, buying long the strongest shareholder rights firms and selling short the weakest shareholder rights firms would have resulted in an abnormal return of 8.5 percent per year. Thus, the more governance provisions a particular firm has, the weaker its financial performance.

An important consideration is that the study by Gompers et al.(2003) assumed that all gov-ernance provisions have the same effect. That is, for every govgov-ernance provision reported in the data for a certain firm, one point is added to the Governance Index, regardless of the specific type of governance provision. To investigate whether this assumption holds,Bebchuk et al.(2009) com-pare the effect of their entrenchment index (containing six governance provisions as described in section2.2.1) with the other eighteen provisions of the Governance Index. They find that a higher entrenchment index (i.e. more power to management) leads, on average, to lower firm value and lower abnormal returns. Importantly, however, the other eighteen provisions were not correlated with firm value or negative abnormal returns. Other studies by Brown and Caylor (2009) and

Larcker et al.(2007) using other definitions of governance provisions thanGompers et al.(2003)’s Governance Index, similarly show that most governance provisions are not linked to firm perfor-mance. This suggests that some provisions are more important than others. That is, the strength of the effect of shifting power from shareholders to management seems to depend on the specific type of governance provision.

Placing the previous literature in a Dutch context is subject to the same considerations as section2.2.1; the features of Dutch corporate governance emphasize more equal power between managers and shareholders. Thus, the effect of governance provisions on firm output may either be exacerbated or mitigated by the governance features of this Dutch context (see hypothesis

2and3). However, conversely to the empirical results of the effect of governance provisions on stakeholder value in the previous section, the literature described in this section strongly suggests the impact of governance provisions on financial performance is negative. I thus hypothesize for Dutch firms:

Hypothesis 4 The number of governance provisions is negatively related with financial performance of Dutch firms

(9)

Hypothesis 5 The relative strength between the types of Dutch governance provisions differs in the extent to which power is shifted from shareholders to management

3. Data and sample selection

3.1. Dutch governance provisions

To investigate the determinants of Dutch governance provisions and its effects on firm out-put, I will analyze governance provisions data on Dutch public firms listed on the Amsterdam stock exchange that has been provided by Eumedion6_{. The data ranges from 2006 to 2017 and}

includes 14 unique governance provision types. A description of the 14 governance provision types included in this analysis is provided in table1.

Table 1: Description of the 14 corporate governance provisions examined in this analysis

(1) Binding nomination (absolute)is a provision that provides the board of directors, or a predefined

third party, the right to nominate a candidate-director or candidate-executive, which the Annual General Meeting (AGM) of shareholders is bound to accept.

(2) Binding nominationis a provision that provides the board of directors, or a predefined third party, the

right to nominate a candidate-director or candidate-executive. The AGM can still vote this candidate away when a two thirds majority that represents at least fifty percent of shareholder’s capital votes against.

(3) Certification (non Tabaksblat)is a provision that indicates that the firm has issued certificates of the

firm’s shares on the stock exchange. The shares underlying these certificates are held by an

"administration office foundation", that under normal conditions passes the voting rights belonging to the shares on to the owners of the certificates. However, in contrast with the Dutch corporate governance code ("Tabaksblat"), this specific provision provides the administration office the possibility to withdraw voting rights of the holders of the certificates in case of "hostile conditions". Voting rights are then exercised only by the administration office.

6_{Eumedion is an independent Dutch foundation that represents institutional investor’s interests in the}

(10)

Table 1: (continued)

(4) Certification (Tabaksblat)indicates that the firm has issued certificates of the firm’s shares on the stock

exchange. In contrast with the previous provision, owners of these certificates can at all times and under all conditions claim their voting rights belonging to their certificates from the administration office. The administration office can thus only vote for those shares on which the holders of the respective certificates have not claimed voting rights.

(5) Combined meetingsprovides special rights for meetings between the executive board and the board

of directors. For example, this can include binding nominations for candidate-directors and candidate-executives, or statutory amendments.

(6) Dual classis a provision that is similar to the priority stock provision in the sense that it allows the

firm to issue stocks with differing levels of voting power. For instance, owning stock A grants one vote per stock, whereas owning stock B grants 10 votes per stock.

(7) Financial preferred stock (certificated)is a provision that indicates that the firm has issued preferred

stock to a select group of shareholders to finance the firm.

(8) Financial preferred stockis identical to the previous provision, but is not certificated (see provision 3).

(9) Golden shareis a provision that gives the owner of this particular type of stock special veto rights.

The owner of the golden stock is not required to own a majority of the firm’s shares to be able exercise its veto rights in the AGM. In The Netherlands, this shareholder is the Dutch government. This allows the Dutch government to e.g. block a takeover of a firm if that serves the national interest.

(10) Preferred stock (availability)is a provision that allows a firm to issue special stock that a special

protection foundation can take on (via a call option), in case the independence and continuity of the firm is under threat (for instance as a result of a hostile takeover bid). Subsequently, the foundation owns a substantial part of the voting rights in the AGM, often a majority.

(11) Preferred stockindicates that the firm has issued preferred stock, that can be taken on by the special

protection foundation in case the independence and continuity of the firm is under threat (for instance as a result of a hostile takeover bid).

(12) Priority stockare shares that have similar veto rights as the golden share provision. In this case, the

firm can issue these shares to any investor and may be constructed to represent specific areas of business strategy. These shares are not sold when the firm is taken over. Therefore, this may deter hostile takeovers in the sense that any firm may take over all available common shares, but in the AGM they can still be overruled in those areas the priority stock has veto rights.

(13) Qualified majorityis a provision that specifies a minimum share of shareholder capital to be

represented at the AGM to reject the nomination or accept the layoff of a candidate-executive or candidate-director.

(14) Shareholder agreementallows the firm to have shareholders enter an agreement on, for instance, the

way they exercise their voting rights. This reduces e.g. the probability of internal shareholder conflicts.

Otherprovisions include the use of pyramid constructions, and arrangements with family owners.

(11)

It is important to consider the fact that Eumedion essentially created two sub-samples. The first Dutch governance provisions sub-sample ranges from 2006 to 2011, and contains 43 to 49 firms listed on the large cap AEX and mid cap AMX index. An overview of the prevalence and summary statistics on the governance provisions types in this sub-sample is presented in table2.

Table 2: Dutch governance provisions sample 2006 - 2011

Percentage of firms with provisions in

Governance provision 2006 2007 2008 2009 2010 2011

Binding nomination (absolute) 2.0 2.0 2.3 2.3 2.3 2.3

Binding nomination 0.0 0.0 0.0 0.0 0.0 0.0

Certification (non Tabaksblat) 4.1 4.1 2.3 2.3 2.3 2.3

Certification (Tabaksblat) 10.2 6.1 9.1 11.4 11.6 11.6

Combined meetings 0.0 0.0 0.0 0.0 0.0 0.0

Dual class 0.0 0.0 0.0 0.0 0.0 0.0

Financial preferred stock (certificated) 20.4 20.4 15.9 15.9 11.6 14.0

Financial preferred stock 0.0 0.0 0.0 0.0 0.0 0.0

Golden share 2.0 0.0 0.0 0.0 0.0 0.0

Preferred stock (availability) 55.1 53.1 56.8 61.4 62.8 62.8

Preferred stock 2.0 4.1 2.3 2.3 2.3 2.3 Priority stock 16.3 12.2 22.7 20.5 20.9 18.6 Qualified majority 16.3 16.3 13.6 11.4 11.6 11.6 Shareholder agreement 0.0 0.0 0.0 0.0 0.0 0.0 Other 4.1 4.1 2.3 2.3 2.3 2.3 None 16.3 18.4 13.6 13.6 11.6 11.6 Summary statistics

Mean governance provisions 1.30 1.22 1.27 1.30 1.28 1.28

Standard deviation 0.83 0.85 0.79 0.79 0.77 0.77

Max governance provisions 4 4 3 3 3 3

Total governance provisions 65 60 56 57 55 55

Total firms 49 49 44 44 43 43

This table shows the prevalence of each Dutch governance provision type described in table1for the sample 2006-2011. The table ends with summary statistics.

In the second sub-sample, Eumedion expanded the data to additionally include the Amster-dam stock exchange small cap indices between 2012 and 2017. This increased the number of firms that are considered in the data significantly: 70 to 91 firms are included in this sub-sample. Sum-mary statistics and the prevalence of governance provisions types in this sub-sample are reported in table3.

(12)

Table 3: Dutch governance provisions sample 2012 - 2017

Binding nomination (absolute) 0.9 1.0 1.0 1.0 1.0 1.0

Binding nomination 29.2 38.4 32.3 30.3 31.0 30.4

Certification (non Tabaksblat) 4.7 5.1 5.1 5.1 5.0 4.9

Certification (Tabaksblat) 8.5 7.1 9.1 9.1 5.0 5.9

Combined meetings 0.0 1.0 1.0 1.0 1.0 1.0

Dual class 0.0 0.0 0.0 0.0 1.0 1.0

Financial preferred stock (certificated) 5.7 6.1 6.1 5.1 3.0 2.9

Financial preferred stock 0.0 0.0 0.0 0.0 5.0 2.9

Golden share 0.0 0.0 0.0 0.0 0.0 0.0

Preferred stock (availability) 43.4 46.5 44.4 43.4 39.0 38.2

Preferred stock 0.9 1.0 1.0 1.0 1.0 1.0 Priority stock 14.2 14.1 12.1 12.1 15.0 14.7 Qualified majority 20.8 18.2 16.2 15.2 22.0 22.5 Shareholder agreement 2.8 10.1 12.1 14.1 18.0 14.7 Other 2.8 3.0 4.0 4.0 4.0 4.9 None 17.1 15.3 15.2 16.0 16.5 16.0 Summary statistics

Mean governance provisions 1.57 1.68 1.61 1.57 1.56 1.56

Standard deviation 1.08 1.17 1.11 1.12 1.09 1.07

Max governance provisions 4 4 4 4 4 4

Total governance provisions 110 121 127 127 142 139

Total firms 70 72 79 81 91 89

This table shows the prevalence of each Dutch governance provision type described in table1for the sample 2012-2017. The table ends with summary statistics.

provisions would be considered the extreme "dictatorship" portfolio in Dutch terms, whereas in the U.S. this would require a total of fourteen or more provisions. One reason for this observation is perhaps rooted in the features of the Dutch governance context outlined in section2.2.1, which suggests that Dutch governance rules already seem to favor balance of power between manage-ment and shareholders as opposed to a U.S. context. Dutch firms seeking balance of power in their control structure therefore may have less need to incorporate governance provisions compared to a U.S. context.

Another reason is perhaps that the strength of governance provisions differs significantly be-tween specific types of Dutch governance provisions. As described in section 2.2.2, empirical evidence finds that most governance provisions have no effect on firm output, suggesting that some provisions are stronger than others. Thus, perhaps some Dutch governance provisions are sufficiently strong already, meaning no more than a few are needed. An interesting question in this regard is whether there is a Dutch equivalent of the entrenchment index byBebchuk et al.

(13)

Instead, I will rely on an expert view to determine the Dutch entrenchment index. To this end, Eumedion’s chief executive Rients Abma considered three Dutch governance provisions to be the "strongest" provisions. Firstly, preferred stock and/or preferred stock (availability), which allows a firm to effectively neutralize the threat of a hostile takeover by issuing shares that are taken on by a protection foundation. Secondly, priority stock, which does not shield the firm from hos-tile takeovers as strongly as the previous provision, but can limit the influence of an "unwanted" shareholder by assigning important decisions at the AGM to the holders of priority stock. Lastly, binding nomination (absolute)/binding nomination, which can be designed in a way that it "re-news" the right of the board of directors to make a binding nomination if an initial nomination gets rejected by the AGM (more detailed explanations are outlined in table 1). Thus, I create a Dutch entrenchment index that adds one point when a firm has adopted: 1. binding nomina-tion (absolute) or binding nominanomina-tion (since firms use either the first, or the second provision), 2. priority stock, 3. preferred stock (availability) or preferred stock (since firms use either the first, or the second provision) and thus takes a value of 0 to 3. I table4and5, I show the preva-lence of Dutch firms with at least one entrenchment provision (i.e. have an entrenchment index bigger than zero) compared to the prevalence of all other provisions. Because firms sometimes use both entrenchment as well as non-entrenchment provisions, I also include the prevalence of Dutch firms adopting at least one entrenchment provision, but none of the non-entrenchment provisions (Entrenchment index only). I observe that Dutch firms indeed overwhelmingly use en-trenchment provisions. On average, 74.6 percent of Dutch firms adopt at least one enen-trenchment provision across the sample, compared to just 43.9 percent that choose for at least one of the non-entrenchment provisions. Furthermore, 30 percent of Dutch firms choose, on average, to adopt at least one entrenchment provision, but none of the other provisions. I will investigate whether this clustering around entrenchment provisions can indeed be explained by differences in provi-sion strength in section4.3. The question that remains is which Dutch firms choose which type of provision. I will investigate this in more detail in section4.1.

Table 4: Prevalence of Dutch entrenchment index sample 2006 - 2011

Entrenchment index 64.9 67.6 77.8 82.9 86.1 86.1

All other provisions 43.2 43.2 44.4 42.9 38.9 38.9

Entrenchment index only 21.4 20.8 19.8 21.0 22.0 22.5

This table compares the prevalence of Dutch entrenchment provisions versus all other provisions between 2006-2011.

(14)

Table 5: Prevalence of Dutch entrenchment index sample 2012 - 2017

Entrenchment index 74.3 76.4 72.2 70.4 68.1 68.5

All other provisions 40.0 43.1 45.6 44.4 50.5 51.7

Entrenchment index only 42.9 41.7 38.3 39.5 33.0 32.6

This table compares the prevalence of Dutch entrenchment provisions versus all other provisions between 2012-2017.

2.2.1). Instead, the findings suggest that there must be an effect of Dutch governance provisions in shifting power from shareholders to management. This is investigated in more detail in section

4.2.

(15)

(16)

(17)

Table 7: Dutch governance provisions summary statistics per industry

Industry Mean SD Min Max Firms

Mining 2.06 0.80 1 3 3

Construction 1.77 0.59 1 3 5

Manufacturing 1.71 1.12 0 4 32

Transportation and Public Utilities 1.58 0.83 1 3 5

Wholesale Trade 0.89 0.76 0 2 4

Retail Trade 1.00 0.76 0 2 5

Finance, Insurance, Real Estate 1.76 1.09 0 4 20

Services 1.39 1.24 0 4 22

3.2. Industry and firm determinants

Some evidence that the adoption of governance provisions indeed differs across Dutch indus-tries is presented in table7, which shows summary statistics on governance provisions averaged across eight Dutch industries using SIC division classifications. The industries contain 12 firms on average. Large differences in the mean governance provisions, as well as the minimum and max-imum provisions are observed. Thus, in line withGillan et al.(2003)’s conclusions, I argue that the industry’s investment opportunities, degree of competition, monitoring environments and in-formation environments are important factors that explain the Dutch governance structures. As a proxy for investment opportunities, I recognize that firms with a high market value compared to their book value are often regarded by investors as high growth potential firms. Thus, I include a market to book value variable to represent investment opportunities. To proxy for a particular firm’s or industry’s competitive environment, I include a profit margin factor, defined as the in-dustry’s average or firm’s net income to revenue ratio. I argue that the profit margin measure is an expression of the firm’s competitive environment since in those industries that have high barriers to entry, firms are not bound by competitor pressure in their price setting. These firms therefore tend to have higher profit margins (Semmler,2016). To represent the firm’s or industry’s mon-itoring environment, I include a leverage proxy that is the firm’s total debt to total assets ratio. Lastly, I include the firm’s or industry’s standard deviation of monthly returns as a proxy for the information environment, as this represents the uncertainty and noise that investors face in the financial markets. All variables are derived from the Worldscope database and are averaged on an industry level for the industry analysis, and considers firm level data (holding industry constant) for the firm analysis. This is described in more detail in section4.1.

3.3. Non-investor stakeholder value

(18)

I will rely on these examples from the literature to determine a proper stakeholder value proxy. Firstly, to represent the firm’s long-term focus on a broad range of stakeholders, Corporate Social Responsibility (CSR) factors are included. Dahlsrud(2006) shows that CSR indeed refers to serv-ing many stakeholders in the long-term beyond what is merely legally required. Several CSR variables derived from the ESG ASSET4 database are used as a dependent variable. Firstly, an en-vironmental score is included which reflects how well a firm’s strategic decision making avoids environmental risk and capitalizes on environmental opportunities to generate long-term stake-holder value. Secondly, the economic score is a reflection of the firm’s overall financial health, and its capacity to generate sustainable long-term growth. Thirdly, a social score is included that reflects the firm’s reputation: it measures its capacity to generate trust amongst its employees, customers and society. Lastly, the emission reduction score reflects the firm’s capacity to reduce air pollution, but also waste, spills and other impacts on biodiversity, in order to reduce the envi-ronmental impact of the firm in the community. All variables are percentage scores.

Secondly, in line withBrochet et al.(2015), I posit that if governance structures promote a more long-term decision making horizon through the use of governance provisions, this has a positive effect on expenditures that are likely to only yield a payoff in the future such as R&D (see hypoth-esis3. Furthermore,Gompers et al.(2003) find that U.S. firms with stronger shareholder power exhibit, among others, lower capital expenditures. I thus include two expenditures measures as dependent variables: R&D expenditures and capital expenditures. Both variables are derived from the Worldscope database and are adjusted for firm size by dividing them by total assets.

3.4. Financial performance indicators

In hypothesis4, I argue that the number of governance provisions is negatively related with financial performance for Dutch firms. I therefore include financial performance measures de-rived from the Worldscope database. Firstly, as a proxy for the firm’s profitability controlled for size, I include the firm’s profitability ratio (net income to total assets). Furthermore, I derive the average monthly returns for each firms across the sample’s period, to investigate whether returns of investing in high managerial power firms are indeed lower than investing in high shareholder power firms (in line with conclusions for U.S. firms byGompers et al.(2003)). I lastly include the log of market capitalization as a proxy for the firm’s value.

3.5. Control measures

To complete the stakeholder value data from section 3.3 for empirical analysis, I control for expected determinants of the firm’s focus on stakeholder value due to forces other than the level of protection by governance provisions. These are used as control variables that are in line with

Brochet et al.(2015)’s methodology. Firstly, financial distress is an important determinant of man-agement’s prioritization of current earnings management over stakeholder value. Therefore, I include a leverage variable as a financial distress control, defined as total debt to total assets (sim-ilar to the leverage variable in section3.2). To account for the firm’s operational risk, I include an operating cycle variable that is calculated as:

(19)

The reason for including operating cycle is that firms with very short operating cycles will exhibit volatile cash flows, and thus emphasize decisions that are short-term in nature. Firms with short operating cycles will thus likely score lower on the variables outlined in section3.3. I also control for firm specific growth opportunities. This is because firms that have high growth opportunities will expand their view to reflect strategic decisions that yield a payoff in the future. The control variable to proxy for this is the firm’s market to book value, which is calculated as the firm’s market capitalization divided by its book value of equity (similar to the market to book variable in section3.2). Lastly, I control for firm performance and size using the return on assets (ROA) profitability factor (similar to the ROA variable in section3.4), and a firm size variable7. Firm size is defined as the logarithm of total assets.

Summary statistics of all aforementioned variables are shown in table8.

Table 8: Summary statistics governance provisions sample 2012-2017

Variable Mean SD Min Max N

Governance provisions 1.48 1.06 0.00 4.00 605 Provision determinants Market to book* 1.95 1.20 -0.96 5.52 431 Profit margin 0.03 0.44 -6.67 1.73 372 Leverage* 0.22 0.18 0.00 1.15 382 Return volatility 0.09 0.06 -0.05 0.50 463 Stakeholder value Environment score 0.74 0.25 0.10 0.95 158 Economic score 0.75 0.24 0.09 0.98 158

Emission reduction score 0.70 0.26 0.11 0.96 158

Social score 0.76 0.23 0.06 0.97 158 R&D expenditures 0.05 0.12 -0.09 1.15 106 Capital expenditures 0.07 0.04 0.00 0.36 405 Financial performance Profitability ratio* 0.01 0.31 -2.82 3.65 383 Average return 0.01 0.03 -0.14 0.22 463 Firm value 5.78 1.08 2.81 8.06 382 Control measures Operating cycle 4.83 1.10 1.79 13.14 329 Firm size 6.00 1.19 1.91 9.07 383

This table shows summary statistics of all variables mentioned in section 3 of the 2012-2017 sample. Variables marked with an asterisk (*) are used both as dependent variables in the determinants and financial performance analyses, and as control variables for the stakeholder value analysis (market to book and leverage are also used as controls in the financial performance analysis).

7_{Note that the R&D expenditures and Capital expenditures variables are already controlled for size (see section}_3.3_),

(20)

4. Methodology

4.1. Determinants of Dutch governance provisions

Firstly, the model for explaining the Dutch firm determinants of governance provisions is a fixed effects panel regression model8_{, specified as follows.}

Provisionsit=αi+β1MarketToBookit+β2Pro f itMarginit+β3Leverageit +β4ReturnVolatilityit+uit

(2)

Where the dependent variable Provisions is an absolute governance provisions value between 0 and 4 for Dutch firm i in year t, explained by the industry factors outlined in section3.2 for each Dutch firm i in year t between 2012 and 2017. α is unobserved heterogeneity and u is the error term. Whether entity fixed effects, time fixed effects or both are used is determined via a redundant fixed effects test, and is displayed for each regression in this thesis.

To test the determinants of specific types of Dutch governance provisions, a fixed effects panel regression model is used, specified as follows.

Provisiontypeit=αi+β1MarketToBookit+β2Pro f itMarginit+β3Leverageit +β4ReturnVolatilityit+uit

(3)

Where the Provisiontype represents the most commonly observed types of governance provi-sions in table3. I classify a provision as "most common" when it is observed for at least 10 percent of Dutch firms in the past 5 years. The most common Dutch governance provisions in table 3

are thus: binding nomination, preferred stock (availability), priority stock, qualified majority and shareholder agreement. The Provisiontype variable thus takes a value of 1 when that particular provision is observed for Dutch firm i in year t, and 0 otherwise. The other elements of the model are specified similarly as model2, where year t is in the range 2012-2017.

Lastly, I already introduced the Dutch entrenchment index in section3.1. To investigate which firm determinants explain the use of the entrenchment provisions, a fixed effects panel regression is used, specified as follows.

Entrenchmentit =αi+β1MarketToBookit+β2Pro f itMarginit+β3Leverageit +β4ReturnVolatilityit+uit

(4)

This model is identical to model 3, with the exception of the Entrenchment variable specified earlier. Regression results of all determinants models are shown and discussed in section5.1.

8_{All regressions in this thesis are fixed effects panel regressions instead of a random effects model, as suggested by}

(21)

4.2. Effects of Dutch governance provisions

Firstly, the effects of Dutch governance provisions on stakeholder value are tested by a fixed effects regression model and is specified as follows.

StakeholderValueit =αi+β1Provisionsit+β2MarketToBookit+β3Leverageit +β4Pro f itabilityRatioit+β5OperatingCycleit+β6FirmSizeit+uit

(5)

Where StakeholderValue represents one of the six stakeholder value variables explained in sec-tion3.3 for Dutch firm i in year t between 2012 and 2017. The Provisions variable is constructed similar to the G-Index byGompers et al.(2003), i.e. one point is awarded for each provision for Dutch firm i in year t, regardless of the specific type. While the strength of the types of provi-sions is likely to differ (see the Dutch entrenchment index in section4.1), for this analysis, I only consider the overall effect of shifting power to management. For instance, consider the preferred stock provision, where a special protection foundation can hold the firm’s stock in case the firm is under threat. Management can use this power judiciously, increasing overall shareholder and stakeholder wealth, or they can use this extra protection to maintain private benefits of control. Either way, it is clear that the provision increases managerial power, and decreases shareholder power. This is what matters when constructing this Provisions variable. Since no more than four governance provisions are observed across the entire sample (see section3.1), this variable takes a value between 0 and 4. The other variables are control variables specified in section3.5. FirmSize will not be used as control variable for the R&D expenditures and capital expenditures regres-sions, since these dependent variables are already controlled for firm size.

To test the effect of Dutch governance provisions on financial performance, I use a fixed effects regression model specified as follows.

FinancialPer f ormanceit=αi+β1Provisionsit+β2MarketToBookit+β3Leverageit +β4OperatingCycleit+β5FirmSizeit+uit

(6)

Where FinancialPerformance represents one of the three financial performance variables spec-ified in section3.4. All other variables are specified identically to model 5. Note that Profitabil-ityRatio is one of the three financial dependent variables, and is therefore not used as a control variable in this model.

4.3. Effects of Dutch governance provisions types

(22)

3.1). To this end, I specify a fixed panel regression model as follows.

StakeholderValueit= αi+β1Entrenchmentit+β2OtherProvisionsit+β3MarketToBookit+ β4Leverageit+β5Pro f itabilityRatioit+β6OperatingCycleit+β7FirmSizeit+uit

(7)

Where the Entrenchment variable is specified identical to the dependent variable in model4. I include an OtherProvisions variable to capture the effect of those provisions not in the Dutch entrenchment index and compare it to the effect of the Entrenchment variable. All other variables are specified identically to the variables in model5.

To test the effect of the entrenchment index on financial performance, I use a fixed effects regression model specified as follows.

FinancialPer f ormanceit =αi+β1Entrenchmentit+β2OtherProvisionsit +β3MarketToBookit+β4Leverageit+β5OperatingCycleit+β6FirmSizeit+uit

(8)

Where FinancialPerformance represents one of the three financial performance variables spec-ified in section3.4. All other variables are specified identically to model 7. Note that Profitabil-ityRatio is one of the three financial dependent variables, and is therefore not used as a control variable in this model.

Regression results of all models in this chapter are shown and discussed in section5.

5. Results and discussion

5.1. Determinants of Dutch governance provisions

Table 9: Firm determinants regression results

Variable All provisions Entrenchment Provision 2 Provision 10 Provision 12 Provision 13 Provision 14

(1) (2) (3) (4) (5) (6) (7) Intercept 1.612*** 1.069*** 0.335*** 0.497*** 0.197*** 0.121*** 0.081** (0.095) (0.070) (0.058) (0.030) (0.039) (0.032) (0.036) Market to book 0.000 0.000 0.000 0.000 0.000 0.001 0.000 (0.002) (0.001) (0.001) (0.001) (0.001) (0.001) (0.001) Leverage -0.570* -0.561** -0.326 -0.079 -0.146 0.124 0.017 (0.334) (0.246) (0.204) (0.105) (0.136) (0.113) (0.128) Return volatility -0.055 -0.323 -0.269 -0.251* 0.299 0.052 0.049 (0.471) (0.347) (0.287) (0.147) (0.192) (0.158) (0.180) Profitmargin -0.054 -0.010 -0.004 -0.004 -0.006 0.015 -0.054*** (0.051) (0.038) (0.031) (0.016) (0.021) (0.017) (0.020) R2 _0.907 _0.894 _0.790 _0.958 _0.888 _0.909 _0.805

Entity fixed Yes Yes Yes Yes Yes Yes Yes

Time fixed Yes Yes Yes No No No No

Firms 84 84 84 84 84 84 84

Observations 358 358 358 358 358 358 358

(23)

Regression results of model 2, 3 and4 are shown in table9. I find a significant and nega-tive effect of the leverage variable on all provisions (regression (1)) and the entrenchment index (regression (2)). In other words, more leveraged firms use less governance provisions. This sug-gests that in The Netherlands, the self-governance characteristics of leverage reduce the need for mechanisms such as governance provisions to define the firm’s control structure (see section2.1). Interestingly, I find no significant effect of leverage on all most commonly observed governance provisions individually. This could be explained by the fact that these regressions ((3), (4), (5), (6) and (7)) only consider one governance provision as its dependent variable. Perhaps the self-governance characteristics of leverage in The Netherlands are not strong enough to have an effect on a range of just one provision. I further find a significant and negative effect of the return volatility variable on the preferred stock (availability) provision in regression (4). Dutch firms in less uncertain environments are thus more likely to have this particular provision. This may be explained by the specific protective features of the preferred stock (availability) provision. Since it allows the firm to issue special stock that can be taken on by a special protection foundation in case the firm is under threat, events that cause large volatility in the firm’s stock price (such as a hostile takeover bid) are less likely to occur. I lastly find a significant and slightly negative effect of profit margin on the shareholder agreement provision (regression (7)). This means that Dutch firms in less competitive environments are more likely to have this particular provision. I do not find evidence for any of the other variables explaining the governance provisions. I thus find only limited support for hypothesis1. Future research should identify whether there are alternative factors that may explain the use of governance provisions in The Netherlands.

5.2. Effects of Dutch governance provisions

Table 10: Effects of Dutch governance provisions on stakeholder value

Variable

Environ-ment Economic Emission Social R&D Capex

(1) (2) (3) (4) (5) (6) Intercept 0.564*** 0.733*** 0.673*** 0.693*** 0.000 0.064*** (0.050) (0.051) (0.054) (0.038) (0.040) (0.006) Provisions 0.079** 0.010 -0.009 0.034 0.031 -0.017*** (0.031) (0.032) (0.034) (0.023) (0.026) (0.004) R2 0.715 0.639 0.679 0.817 0.308 0.736

Entity fixed Yes Yes Yes Yes Yes Yes

Time fixed No No No Yes No No

Firms 37 37 37 37 29 93

N 168 168 168 168 122 400

*** p < 0.01, ** p < 0.05, * p < 0.10. Standard errors in parentheses. The top row shows the dependent variable used in the regression.

To find an indication of the effects of governance provisions on stakeholder value, I firstly show the regression results of model5without control variables in table109_{. This yields two}

(24)

nificant results. Firstly, the number of governance provisions is positively related with the firm’s strategic focus on environmental stakeholders. This would offer some preliminary support for hypothesis3. However, I find a slightly negative but significant effect of governance provisions on capital expenditures, which would preliminarily support hypothesis2instead. Again, these results are based on regressions with no control variables, so I will investigate later in this section in more detail which hypothesis is more supported.

Table 11: Effects of Dutch governance provisions on financial performance

Variable Profitability ratio Average monthly

return Firm value

(1) (2) (3) Intercept 0.085 0.002 5.932*** (0.070) (0.009) (0.048) Provisions -0.044 0.004 -0.086*** (0.048) (0.006) (0.032) R2 _0.097 _0.049 _0.967

Entity fixed Yes Yes Yes

Time fixed No No Yes

Firms 95 96 95

N 413 437 412

To find an indication of the effects of governance provisions on financial performance, I firstly show the regression results of model6 without control variables in table11. This indicates that governance provisions are negatively related with firm value, offering some preliminary support for hypothesis4. The other regressions (1) and (2) do not show significant results and have low ex-planatory power (R squared 0.097 and 0.049, respectively). However, again, these regressions do not include control variables. I thus investigate in more detail whether hypothesis4is supported later in this section.

(25)

(26)

The regression results of the full stakeholder value model (model 5) are shown in table 12. Note that the operating cycle control variable is very restrictive to the number of observations, which is why I run each stakeholder value dependent variable regression with and without this control variable. Also, R&D expenditures and capital expenditures are already controlled for firm size by dividing the absolute expenditures data by total assets, thus the firm size control variable is not included in these regressions. Compared to table10, I find that adding the control variables generally has a positive effect on the explanatory power (in terms of R squared10) of all regres-sions. I find that firm size is positively related with all dependent variables, indicating that larger Dutch firms focus on stakeholder value more than smaller Dutch firms. Furthermore, positive and significant effects of governance provisions on the firm’s focus on environmental stakehold-ers and economic well-being, suggesting hypothesis3is supported. Interestingly, the profitability ratio control variable is negatively and significantly related to environmental stakeholder focus, but positively related to economic well-being. Perhaps this is because firms assign any excess financial resources to address sustainable value creation towards environmental stakeholders. This may also explain the negative and significant effect of profitability on the social, emission re-duction and R&D expenditures dependent variables. Conversely, firms promoting the long-term economic well-being of the firm (regression (3) and (4)) may actually hold on to excess financial resources, explaining the positive effect of profitability on the economic stakeholder variable. On the other hand, a slightly negative but significant relationship is observed between the amount of governance provisions and R&D and capital expenditures. These findings support hypothesis2. I find no significant effects of governance provisions on any of the other dependent variables. I thus find inconclusive evidence as to which hypothesis from section2.2.1is more supported. It may be that the effect of Dutch governance provisions is neutral: one the one hand, despite the features of Dutch corporate governance in already emphasizing balance of power between management and shareholders (see section2.2.1), adopting stronger protection through the use of governance provisions does not always exacerbate entrenchment in Dutch firms, thereby contradicting hy-pothesis2. On the other hand, Dutch management does not always focus more on stakeholder value with this extra power, thereby contradicting hypothesis3. Thus, the effect of Dutch gover-nance provisions on stakeholder value is neither always negative, nor always positive. It might also be that Dutch governance provisions significantly differ in strength, thus analyzing the ef-fect of all provisions simultaneously yields no conclusive positive or negative result. A finding supporting this argument is the fact that an average of 30 percent of Dutch firms have at least one entrenchment provision, but none of the other provisions (see section3.1). If these entrench-ment provisions indeed are significantly stronger than the other provisions, then considering all provisions together may explain the neutral result found here. I will investigate this in more de-tail by comparing the effect of the Dutch entrenchment index on stakeholder value and financial performance in section5.3.

Lastly, the regression results of model6are shown in table 13. Adding the control variables has a positive effect on explanatory power in terms of adjusted R squared. However, the adjusted

10_{All reported R squared metrics are adjusted R squared to take into the account the positive effect of adding}

(27)

Table 13: Effects of Dutch governance provisions and control variables on financial performance

(1) (2) (3) Intercept -0.322 0.060 1.241** (0.197) (0.131) (0.549) Provisions -0.038*** 0.003 -0.046 (0.010) (0.007) (0.029) Market to book -0.004 0.003 -0.022 (0.007) (0.004) (0.018) Leverage -0.310*** -0.064 -0.778*** (0.061) (0.039) (0.170) Operating cycle -0.007 0.001 0.036 (0.010) (0.006) (0.028) Firm size 0.090*** -0.008 0.785*** (0.034) (0.023) (0.094) R2 0.428 0.100 0.976

Time fixed Yes No No

Firms 74 70 74

N 323 293 323

(28)

entrenchment index on stakeholder value and financial performance in section5.3.

5.3. Effects of Dutch governance provisions types

In the previous sections I found contradicting and small effects of Dutch governance provi-sions on stakeholder value and financial effects, respectively, which I argued might come from a difference in strength between the Dutch entrenchment provisions and all other provisions. I will investigate in this section in more detail whether this is true. In table14, regression results of model7are shown. Note that these regressions exclude the control variables, in order to find preliminary results of the effect of the entrenchment index and other provisions on stakeholder value. Similar to table10, I only find significant results of the effect of governance provisions on stakeholder value. However, I find that only the entrenchment index is statistically significant (at the 1 percent level). It has a positive relationship with the firm’s focus on environmental stake-holders (-0.153), and a slightly negative relationship with capital expenditures (-0.023). While this preliminary result does not help explain the contradictory results from the previous section, it does indicate that the entrenchment index indeed has a stronger effect on stakeholder value than the other provisions, preliminarily supporting hypothesis5. Again, since these regressions do not include control variables, I will investigate in more detail the effect of the entrenchment index and other provisions on stakeholder value later in this chapter.

In table 15, regression results from model 8 are shown, again without control variables to find preliminary results of the effect of the Dutch entrenchment index and other provisions on financial performance. As these regressions do not include control variables, the R squared is again low for regression (1) and (2). Compared to table 11, I similarly find negative effects of governance provisions on firm value. It seems that both the entrenchment and non-entrenchment governance provisions have a negative effect on firm value (-0.079 and -0.086 at the 10 percent level). Interestingly, I do find a statistically significant (at the 1 percent level) negative effect of the entrenchment index on firm profitability, but not for other provisions. Again, this indicates that the effect of Dutch entrenchment provisions on financial performance is indeed stronger than the effect of non-entrenchment provisions, preliminarily supporting hypothesis5. I will investigate in more detail whether this is true later in this section.

(29)

Table 14: Effects of Dutch governance provisions types and entrenchment index on stakeholder value

Variable

(1) (2) (3) (4) (5) (6) Intercept 0.541*** 0.729*** 0.655*** 0.689*** 0.001 0.065*** (0.051) (0.052) (0.056) (0.039) (0.041) (0.006) Entrenchment 0.153*** 0.004 0.038 0.048 0.035 -0.023*** (0.050) (0.051) (0.054) (0.038) (0.034) (0.006) Other provisions -0.002 0.028 -0.056 0.018 0.022 -0.008 (0.052) (0.053) (0.057) (0.038) (0.067) (0.006) R2 0.719 0.636 0.679 0.816 0.301 0.738

Time fixed No No No Yes No No

Firms 37 37 37 37 29 93

N 168 168 168 168 122 400

Table 15: Effects of Dutch governance provisions types and entrenchment index on financial performance

(1) (2) (3) Intercept 0.099 0.001 5.925*** (0.070) (0.009) (0.048) Entrenchment index -0.116* 0.012 -0.079* (0.066) (0.008) (0.045) Other provisions 0.047 -0.006 -0.086* (0.073) (0.010) (0.049) R2 0.102 0.052 0.967

Time fixed No No Yes

Firms 95 96 95

N 413 437 412

(30)

Table 16: Effects of Dutch governance provisions types and entrenchment index on stakeholder value (with control variables)

Variable

(1) (2) (3) (4) (5) (6) Intercept -1.586* -2.195** -1.999* -1.406** 0.112*** 0.065*** (0.910) (0.891) (1.027) (0.666) (0.025) (0.008) Entrenchment 0.105** 0.076 0.014 0.031 0.007 -0.022*** (0.050) (0.049) (0.057) (0.037) (0.013) (0.006) Other provisions -0.001 0.037 -0.052 0.036 -0.032 -0.010 (0.052) (0.051) (0.059) (0.038) (0.026) (0.007) Market to book 0.000 -0.001 0.000 0.000 0.001 0.000 (0.001) (0.001) (0.001) (0.001) (0.004) (0.000) Leverage -0.211 -0.267 -0.126 -0.027 -0.167*** 0.002 (0.241) (0.236) (0.272) (0.176) (0.057) (0.020) Profitability ratio -0.890*** 1.148*** -0.548* -0.581*** -0.636*** 0.008 (0.251) (0.245) (0.282) (0.183) (0.037) (0.006) Firm size 0.327** 0.409*** 0.393*** 0.307*** (0.132) (0.129) (0.149) (0.097) R2 0.756 0.711 0.702 0.837 0.908 0.738

Time fixed No No No No Yes No

Firms 36 36 36 36 29 92

Observations 164 164 164 164 122 396

table12and16). Thus, both hypothesis2, and hypothesis3are weakly supported. I do find more evidence for hypothesis5, however, as the entrenchment index predominantly shows significant results compared to the other provisions in table14,15and16. Before making final conclusions on this hypothesis, I will investigate the effect of the entrenchment index and other provisions on financial performance including control variables in table17.

Table17shows regression results from model8including control variables. By including the control variables, explanatory power in terms of adjusted R squared increased for all regressions. Similar to the conclusions from table13, I only find a statistically significant negative effect of Dutch governance provisions on firm profitability. Even though preliminary results in table11

and15show that Dutch governance provisions are negatively related with financial performance, adding control variables in table13 and17 captures away most (in case of firm profitability), if not all (in case of firm value) of the negative effect. I can therefore conclude that hypothesis4

(31)

Table 17: Effects of Dutch governance provisions types and entrenchment index on financial performance (with control variables)

(1) (2) (3) Intercept -0.324 0.049 1.239** (0.198) (0.131) (0.550) Entrenchment index -0.038** 0.012 -0.050 (0.015) (0.009) (0.042) Other provisions -0.035** -0.009 -0.035 (0.016) (0.011) (0.046) Market to book -0.004 0.002 -0.022 (0.007) (0.004) (0.018) Leverage -0.311*** -0.057 -0.783*** (0.062) (0.039) (0.173) Operating cycle -0.007 0.001 0.036 (0.010) (0.006) (0.028) Firm size 0.090*** -0.007 0.785*** (0.034) (0.023) (0.095) R2 0.424 0.104 0.976

Time fixed Yes No No

Firms 74 70 74

N 323 293 323

6. Conclusion and limitations

To conclude, I firstly identified the 14 types of Dutch governance provisions and trends in the use of Dutch governance provisions between 2012 and 2017. To answer the research question "What are the determinants and effects of Dutch governance provisions?", I then investigated whether there are specific determinants that explain the use of Dutch governance provisions, the effect of Dutch governance provisions on stakeholder value and financial performance, and whether there are differences in these effects between Dutch governance provision types (i.e. whether the strength of Dutch governance provisions varies). I find that more leveraged firms use less gov-ernance provisions, perhaps because the self-govgov-ernance characteristics of leverage reduce the need for mechanisms such as governance provisions to define the firm’s control structure. Fur-thermore, Dutch firms that have less volatility in average monthly stock returns, are more likely to have the preferred stock (availability) provision, because this particular provision has protec-tive features that restrict shareholder rights in cases that cause significant stock price volatility (e.g. hostile takeover bids). Overall, I do not identify determinants for most provision types. Future research should focus on finding alternative factors that may determine the use of Dutch governance provisions.