Whincop, Michael J --- "The Role of the Shareholder in Corporate Governance: A Theoretical Approach'" [2001] MelbULawRw 14; (2001) 25(2) Melbourne University Law Review 418

	Home \| Databases \| WorldLII \| Search \| Feedback Melbourne University Law Review

[The attention of policy-makers has recently been focused on the entitlement of shareholders to participate in corporate governance. This issue is analysed from a perspective which combines social choice theory with other economic considerations. The appropriate degree of collective action that should be imposed by constitutional constraints or by the law is analysed over a range of different types of issues. The article argues that the law on the requisitioning of meetings should be contractible or should include a ‘counter-requisition’ procedure; that the law on the proposal of resolutions should be differentiated across issue-type and should embrace various forms of rationing; and that the case for mandating particular rules for the election of directors is not soundly based.]

INTRODUCTION

In the ongoing process of corporate law reform in Australia, the improvement of corporate governance and the enhancement of the accountability of directors and managers to shareholders have been the guiding stars. This is implicitly linked to the 1990s imperative of creating a regulatory environment that would win the confidence — so shaken in the late 1980s — of both domestic and international investors in Australian equities. However, there is a curious asymmetry in the means employed to undertake this task. On the one hand, the obligations of directors have waxed considerably, in the adoption of, for example, the related parties’ provisions, the civil penalty regimes, the intensification of duties of care, and the increased benchmarking of corporate governance principles. On the other, the law has been much more wary of creating individualised or collectivised rights for shareholders. Sometimes, there are specific instances where rights have been extended, as is the case with the new derivative suit provisions of the Corporations Act 2001 (Cth) (‘Corporations Act’).[1] In other cases, as in recent analyses of the participatory entitlements of shareholders in listed public companies, the view is taken that shareholders need to be constrained, both to protect them from themselves, and to protect directors against officious or malicious interference with their function.[2] I study this latter subject in this article.

My purpose is not to be specifically critical of the logic underlying these proposals. I agree with many of them. Nor is it my intention to allege that there must be symmetry between the obligations of directors and the entitlements of shareholders. Symmetry may be very elegant, but it is not a useful organising principle in the law. Instead, I mean to take a broader theoretical approach to some of the questions associated with the relation between directors and shareholders. My aim here is, first, to provide an economic explanation, much influenced by theories of social choice, of why the law has developed along the participative lines that it has. Australian scholars have made very little use of the social choice literature. Second, I want to show how the legal principles in this area have become confused over time. Third, I seek to synthesise this material to shed some light on several of the reform questions that we are currently facing. Let me provide a little more detail on each of these aims in turn.

To take the first point, I proceed from the premise that an aim of corporate law is that the corporation function as much as possible as a self-governing system. That is, the rights and obligations of shareholders and officers are allocated in ways that permit the value of the corporation to be maximised and divided amongst those with claims to it, with as minimal a role for courts in judicial review and enforcement as is necessary. A number of doctrinal principles related to the role of shareholders in corporate governance are consistent with this aim. These include the legal principles concerning the allocation of functions between the general meeting and the board, and other laws that determine the ways in which majoritarianism is embedded in doctrinal principles. I argue that four matters influence the structure of rights and the scope for majoritarianism. These are: the costs of arriving at collective decisions; the magnitude of externalities that can be imposed by the empowered decision-maker(s); the risk of indeterminacy or ‘cycling’ in giving effect to majority preferences; and, finally, the risk of corporate contracts being held up opportunistically. These factors explain why the law sometimes relies on simple majority rule, in other cases requires a supermajority, and, at times, takes away rights from shareholders completely. I show how there is a fine interplay between the individual and collective rights of shareholders in responding to overreaching and fiduciary concerns by directors. I argue that the rule in Foss v Harbottle[3] reflects and reinforces these concerns.

The second point is that the corporate law of the 20th century confused and distorted the principles relating to shareholder participation, not so much in outcomes as in reasoning and justification. We have come to think that, because shareholders have a limited role in the management and governance of the firm, they should be seen as having no role, unless explicitly provided for in the constitution or a legal rule. I show how these principles have emerged in the law, and the problems that the modern law creates.

The third point is a review, amongst other things, of some of the proposals in the CSAC report on shareholder participation.[4] I have no argument with much of the report, but I do discuss the role of requisitioning meetings, proposing resolutions, and shareholder voting. I criticise aspects of the report’s findings in these areas, as well as proposals from other sources. I emphasise the desirability of greater state-contingency in legal principles, and also of contractual resolutions of the law in this area. The article is divided into three substantive parts that correspond to these issues.

II A THEORETICAL ANALYSIS OF SHAREHOLDER PARTICIPATION

A The Calculus of Corporate Consent

1 Introduction

The questions as to which functions should be vested in shareholders in general meeting, and which in the board of directors, and the preconditions to be met before these functions can be exercised, are fundamentally constitutional questions. They have much in common with the issues relating to the processes of consent that are adopted in the constitutions of states.

Before specifically developing a constitutional framework for corporations, let me first take a common observation that is cited to explain the division of powers between the board and the general meeting. The CSAC report says:

Few would doubt this apparently self-evident point. The greater the size of the shareholding body, the more impoverished decision-making is likely to be, given the reduced opportunities for deliberation and the weak or non-existent incentives for shareholders to become informed.[6] The trouble with this argument is twofold. Most of the common law, and some of the statutory principles relating to the rights of shareholders vis-à-vis the board are wholly indifferent to the size and organisation of the shareholding body and the consequent incidence of collective action problems. Indeed, we may go further and point out that the shareholding bodies of English corporations were small and concentrated for many long years, including the period in which most of the crucial common law principles were formulated.[7] Moreover, the argument perhaps proves too much. If the argument were right, why would shareholders have any power at all in some of the transformative decisions in the life of the corporation, such as certain issues related to capital reductions, voluntary windings-up, changing constitutional provisions, and so on?[8] The answer is not obvious on the arguments of CSAC and the Organisation for Economic Co-operation and Development (‘OECD’).

To begin, it is helpful to adopt a premise that corporations are substantially self-governing.[9] That is, the contracts between shareholders and other constituencies and the constitution between shareholders and managers create appropriate incentives for the value of the firm to be maximised and distributed amongst the various claimants. The need for judicial intervention should be limited to exceptional cases (although the definition of ‘exceptional’ may be debated). Much attention has been paid to the role of contracts in the theory of the firm, but the role of the constitution as a regulator of collective action by shareholders has not been analysed systematically.[10] In the section that follows, I initially analyse the role of organisational constitutions in regulating collective action in general terms, having regard to the constitutional economics of James Buchanan and Gordon Tullock.[11] I then apply this model to corporations.

2 Constitutional Economics

In order to understand the influences on the content of constitutions, it is useful to recall that there are three ways by which humans can achieve the ends they desire.[12] The first two are individual action and trade. Whereas trade necessarily involves two or more people, with responsibilities predefined by agreement, individual action involves essentially solitary or unconcerted exertions. The third is collective action. Collective action involves action by or on behalf of some group of persons. It is generally the case that the benefit obtained by collective action is shared by the group, but not necessarily — or indeed usually — in equal shares, or in a manner in which costs and benefits are shared proportionally.[13] Collective action need not be unanimously supported, except where the constitutional process for the group imposes a rule of unanimous consent to the proposed collective action. It follows that collective action can make some members of the collective worse off, depending on how costs and benefits are shared. This is not true of individual action or trade, since a person’s decision to engage in either activity is, presumably, based on potential to increase his or her own welfare.

Buchanan and Tullock argue that a constitution should be treated as a set of constraints on, and procedures regulating, collective action. These restrictions, in order to have constitutional status, must command nearly unanimous support as the basis for future collective action.[14] A principal variable that Buchanan and Tullock examine is how much support particular forms of collective action would have to enjoy before they could be pursued. In other words, what proportion of the population must a constitutional rule require for the approval of a specific (post-constitutional) collective action proposal, in order for that constitutional rule to be supported unanimously?[15] This heuristic has some attributes in common with the Rawlsian analysis of the social contract that would be supported behind a veil of ignorance, since the constitutional rules are completely general, providing only the process for co-operation by collective action.[16] The considerations that Buchanan and Tullock explore as influences on these constitutional rules are (a) the costs of decision-making, and (b) the externalities associated with different rules.

The constitutional rule that has the lowest decision-making costs is one that allows an individual — any individual, not a single dictator — to mandate collective action. Under such a rule, I might be able to require that the state build a new railway line between the suburb in which I live and some other point.[17] The rule does not require me to consult or negotiate with any other person. By contrast, the rule that imposes maximum decision-making costs is one that requires unanimity as a condition of collective action. If every citizen must agree, the costs of procuring unanimity will be very high. That requirement affords every individual with the capacity to veto collective action. This can thwart even the most beneficial forms of collective action, because each individual has an incentive to ‘hold out’ from giving consent in order to procure a larger share of the gains from collective action, for instance, by requiring a side payment in order to give consent. Thus, the task becomes one of dividing the gains from trade (the process being substantially the same as relying on consensual trade to obtain social ends), which may prove to be very difficult with as few as three citizens.[18] Hence, decision-making costs, D, rise as the proportion of citizens required to consent rises, as Figure 1 shows:

The costs associated with the externalities imposed, with respect to the required consent, behave in the opposite way. The fewer the people who must consent, the higher the net social costs of collective action. People will mandate action that benefits them, regardless of whether it maximises social welfare. As the number of individuals required for collective action rises, it becomes harder to impose an externality on a party who has not consented, and it is impossible to impose an externality if unanimity is required. Figure 2 suggests a shape for this curve, E.

Thus, if the total costs of collective action, C, are conceived as the sum of the costs functions D and E, it becomes apparent that rational citizens would support constitutional rules which require that collective action be supported by that proportion of citizens at which the joint cost function lay at a minimum, C*, which we will call c/n, as Figure 3 suggests:

Although c/n lies between 1/n and n/n, the fact that c/n appears in Figure 3 to be around a bare majority of 50% is simply a coincidence of the shape of the costs functions. Buchanan and Tullock point out that there is nothing special about the 50%+1 rule.[19] There is no reason to believe that costs of collective action are consistently minimised at this point. For example, in a smaller community, the cost curve for decision-making may be much flatter because of the lower costs associated with decision-making (including both negotiation costs and the costs of beneficial collective action). Thus, c/n would be much closer to n/n.

3 Application to Corporations

The economics of constitutions may seem remote from corporations, but there are very clear parallels. Corporations combine features of exchange and collective action. In the first instance, they are constituted by contracts — most obviously so when parties choose to invest in a securities offering.[20] However, because the corporation is a species of hierarchy, it embraces various forms of fiat so that perpetual recontracting is unnecessary. This point, originating with Coase,[21] is central to most modern theories of the firm.[22] We may therefore expect the constitution of a corporation to pursue an objective analogous to the organisation of states, namely, minimising the costs of future collective action. This should have three effects on the corporate governance of corporations:

4 Cycling and Indeterminacy

Factors besides decision-making costs and externalities also influence the shape of these constitutional provisions. The first of these is the tendency for cycling in the determination of the forms of collective action.[23] What is cycling? Processes of collective action frequently produce outcomes over time that are at variance with those reached in the past. This tendency causes cycles of outcomes rather than a larger deterministic set of outcomes. This property of collective action arises from the impossibility of determining a meaningful set of collective preferences from the preferences of individual members. The work done by social choice theorists, taking their cue from the original development of Kenneth Arrow’s ‘impossibility theorem’,[24] shows how collective action processes dependent on voting mechanisms can produce results that are inconsistent with the will of a majority. This seeming anomaly is a result of the way in which preferences are expressed and aggregated.[25] The simplest illustration of this is Condorcet’s paradox, where three persons, A, B and C, must vote their preferences for three different options, F, G and H. Their preferences are described in Table 1:

Given this distribution of preferences, the outcome depends entirely on the sequence of voting. If voters must choose between F and G, F will be chosen because A and C prefer it. Likewise, A and B will choose G over H. If F is paired against H, however, B and C will vote for H, even though a majority prefers G to H, and a different majority prefers F to G. Technically, majoritarian preferences are said to be no longer ‘transitive’ — a basic condition of individual rationality, in which an ordering of preferences where A is preferred to B, and B to C, requires that A be preferred to C. This means that it is difficult to claim that a majoritarian choice deserves special deference, and also that outcomes can be influenced by ordering the sequence of voting choices.

What follows from the indeterminacies of processes of social or collective choice is that particular subjects for collective action may be removed from the competency of the ‘citizens’ — in corporations, the shareholders[26] — in order to avoid the dissipation and deadweight costs that may be associated with cycling in production and investment decisions. An example would be a decision to commence a project, and then to abandon it in favour of some other proposal. Corporate constitutions and corporate laws are correspondingly likely to do two things to address this problem. One is to lock out the shareholders from being able to take collective action in certain areas. The other is to use supermajorities to mitigate the risk of cycling. A supermajority favours the status quo against any proposal that does not have wide and substantial support (technically, proposals that are at least ‘Condorcet winners’[27]).

5 Hold-Up Problems

The final factor influencing the design of constitutions is the influence that collective choice, occurring ex post, may have on ex ante incentives. This is an aspect of a more general contracting problem associated with the making of idiosyncratic investments. Economists have studied this problem in detail.[28] The gains associated with contracting often depend upon the making by parties of certain investments — these generate quasi-rents where the returns are dependent on the continuance of the exchange and the presence of other assets with which the investments are used. This is sometimes described as an investment ‘in match’.[29] An example is where a manager invests his or her human capital in a firm-specific specialisation which has minimal value to other employers. The problem is that one or both of the parties to the contract may have incentives to try to capture a larger share of quasi-rents by renegotiating the exchange ex post. The specialised assets increase the likelihood of this post-contractual opportunism, given their low opportunity cost. For example, one of the parties threatens not to perform unless the gains from the exchange are redistributed. Such ‘quasi-rent-seeking’ (often described as ‘hold-up’ behaviour) can discourage the making of specialised investments, even though both parties may be better off when the investments are made. Much work in economics has focused on the governance processes and contracts which are used to constrain these processes of quasi-rent-seeking.[30]

Most economic analyses have considered the processes which limit hold-up behaviour by individual parties. However, it is also possible for hold-up to occur by way of collective action.[31] For example, two or more shareholders could engage in quasi-rent-seeking by attempting to hold up the performance of corporate contracts if they were capable of withdrawing their investment from the firm.[32] Whether this is a credible threat depends on the firm’s liquidity constraint (in particular, whether the firm would have to sell assets to which other quasi-rent-generating investments are specific), and on the size of the shareholders’ interests. If the collective withdrawal of assets by a group of shareholders is a credible hold-up threat, it could be directed vis-à-vis other shareholders, and against other constituencies making firm-specific investments (including managers, employees and suppliers). These problems reinforce the earlier argument that corporate contracts and laws are likely to remove certain of these forms of collective action from the competency of shareholders, in view of cycling problems. Further, they influence the pay-offs in a way that minimises the credibility of the threat. Chapter 2J of the Corporations Act substantially circumscribes the entitlements of shareholders to withdraw the capital of the firm, either directly or indirectly, and Part 5.5 requires that, if shareholders wish to enter a voluntary winding-up, they must, amongst other things, demonstrate the support of a supermajority and submit to the statutory set of priorities for dispersing the proceeds of the liquidated firm.[33]

Having identified decision-making costs, externalities, cycling and quasi-rent-seeking as influencing the scope of, and requirements for, collective action in a corporation, I now go on to discuss relevant aspects of the law and constitutions. Here, I explain the role of shareholders, and collective action by them, in corporate governance, and the reasons for the considerable variation that we find therein.

B The Scope for Collective Action

Shareholders are undoubtedly entitled to take collective action in relation to two principal subjects — namely board representation, and changes to the articles of association. Even this bare statement of shareholders’ rights contains a germinal idea of importance: neither of these areas constitutes substantive collective action, by which I mean the sort of action which directly commits collective resources, thus indirectly taxing the other shareholders. In this way, the law and constitutions of corporations limit the scope by which shareholder collective action can impose externalities on other shareholders. At the same time, limiting the entitlement of shareholders directly to affect the management of corporate resources limits their capacity to hold up other constituencies for quasi-rent-seeking purposes. In light of this introductory premise, I now seek to examine and explain the nature of permissible collective action in these and cognate areas. I focus in most detail on collective action with respect to self-dealing and fiduciary breaches as these are neglected subjects which play a crucial role in shareholder constitutionalism.

1 Board Representation

Shareholders are entitled to elect, as well as to remove, directors by ordinary resolution (in the case of public companies).[34] However, the specific electoral procedures are largely unregulated, and there are substantial discretions over the specific procedures to be applied. As a general principle, the standard for collective choice in this area is simple majority rule. The appeal of such a rule is obvious — a very low representational requirement (self-appointment being the extreme case) could result in excessively large boards. There is no evidence that board effectiveness increases monotonically with board size.[35] A very high representational requirement, requiring a very high degree of support, has the opposite problems. Where the tenure of a director is limited to a certain number of years before re-election, a high representational requirement could result in a board without directors, since there may be no candidate meeting such an onerous standard. Simple majority rule prevents either outcome.

To put these points differently, a very low representational requirement is likely to impose a higher level of externalities because of the greater incidence of inappropriate or self-interested board appointments. A very high representational requirement imposes high decision-making costs, because of the need for side-payments and negotiation between different factions.

The absence of a mandated voting system is a straightforward reflection of the fact that there is no voting system which satisfies all of the indicia of an optimal system.[36] There are, in theory, risks of cycling. However, experience with large corporations is inconsistent with cycling phenomena. Boards are far more likely to be static and apparently self-perpetuating. This suggests, then, that members of the board (especially the Chief Executive Officer) have and use procedural advantages, such as the holding of proxies and influence over nominations.[37] This result is also significantly reinforced by the presence of business norms such as the famous — to some, infamous — Wall Street Rule, which counsels voting with managers, or selling one’s shares, in the absence of serious violations of other norms which management is expected to observe.

2 Self-Dealing and Moral Hazard

To the extent that the processes and norms associated with determining board representation engender self-perpetuation, it is clearly important for other mechanisms to exist to mitigate the potential for complacency and subgoal pursuit.[38] There are many such mechanisms. Some, such as the takeover, are market-based. Others involve contracts, such as executive remuneration packages. There is also a role for law in limiting certain forms of undesirable conduct, such as fraud and overreaching. Fiduciary duties and certain related statutory constraints are important in this regard.

The law created the scope for collective action in the shadow of the rules it applied to self-dealing.[39] Courts traditionally applied a strict, inflexible standard,[40] but permitted the shareholders to resolve to affirm or ratify fiduciary breaches in order to release directors from the consequences that these rules cause, such as the forfeiture of interests by means of rescission.[41] One cannot have a strict (and thus potentially over-inclusive) rule without having some basis for the release of legal liability outside a court. This enhances the sense that corporations can function in a self-governing manner, since the determination of whether legal sanctions are to be imposed occurs endogenously through the collective action of shareholders. What constitutional constraints should be imposed on this form of collective action by shareholders?

(a) Bare Majority Requirements

What degree of consent must there be, and who should have the right to initiate collective action? The general principle is that shareholders are capable of ratifying an ordinary resolution by simple majority.[42] However, that ratification is dependent on the shareholders receiving full information and on the absence of coercion.[43] On occasions, it has been suggested that unanimity is required, but this view has not prevailed.[44]

On one hand, it might be thought that a bare majority requirement is undesirable because of the greater incidence of externalities that may be borne by shareholders as a result of the lower consent requirement. The fiduciary protection is limited by not demanding unanimity and by making overreaching easier to ratify. However, other factors cut against this point. Equitable doctrine implies that the court is able to review the procedural circumstances of the transaction.[45] This rule, and the possible application of the doctrine of fraud on the minority,[46] are likely to limit the absolute value of the externality that shareholders could possibly bear. These legal rules reduce the potential number of proposed resolutions that can be put to shareholders, which in turn limits the scope that management has to manipulate the agenda in order to procure a resolution contrary to a majority’s wishes. The curve describing the value of externalities, E (in Figure 2), as a function of the required level of consent is likely to be flatter, which justifies a lower consent requirement.

This conclusion is reinforced when it is understood that the law in this area was formulated at a time when the fiduciary principle was largely contractible.[47] That is, it was possible to waive the conflict rule by a provision in the articles of association.[48] Empirical evidence suggests these provisions were very common.[49] Thus, it took only a special resolution, with a requirement of a three-quarters majority, to include such a provision in the articles — notwithstanding that such a provision presented greater risks of moral hazard, since it had of necessity to be voted on without reference to any of the details of future transactions to which it might apply.[50] It seems logical that the law would require a lower level of consent when managers put forward the full details of a specific transaction. This provided an incentive to managers of corporations lacking the term to relax the conflict rule to seek the approval of a bare majority for the transaction rather than to put forward a special resolution for the generic provision for the approval of a special majority.

Before discussing the issues associated with the initiation of collective action, I should say a word about the duty of care. As is well known, the requirements of this duty were historically highly undemanding, and there are very few instances of affirmative findings of negligence in the 19th and early 20th centuries.[51] The use of low standards of care made it unnecessary for managers to seek ex ante shareholder authorisation to engage in particular businesses or transactions. This minimised the scope for collective action by shareholders in the field of management of the corporation, and thus reduced consequent delays and impoverished choice (owing to apathy and limited information). Any other rule on the duty of care would have the indirect consequence of attenuating the discretion of management vis-à-vis the shareholders. The optimal rule in this respect is therefore one that delegates managerial decisions to managers only, and which holds them to a low standard of care.

(b) Rights to Initiate Collective Action

The final and considerably complex question to be considered in this part is the process associated with initiating collective action in relation to self-dealing and moral hazard. It is clearly the case that a director has the right to requisition a vote in relation to his or her self-dealing — putting such a transaction before the shareholders is the proper course for him or her to take. When the transaction is inchoate (as where ex ante approval is sought),[52] the director must determine the price that he or she should offer in the transaction. This therefore has the quality of a take-it-or-leave-it offer, because shareholders have limited opportunities to counter-offer, and, in any event, face collective action problems in doing so effectively.

Does the shareholder have an independent right to initiate the review of such a transaction by the general meeting after the transaction occurs, or to propose the restitution of benefits obtained by the director thereunder? This question is not answered very clearly in the law.[53] The shareholder, faced with a fiduciary breach, can choose between two possible courses of action — he or she could attempt to litigate the transaction, seeking its judicial review, or he or she could attempt to put the transaction before the general meeting. We know that until the recent enactment of the derivative suit provisions in the then Corporations Law,[54] the rule in Foss[55] took a very limiting approach to the entitlement of individual shareholders to seek judicial review of such transactions. Although Foss is now substantially obsolete, it is worth understanding the doctrinal logic and economic effects that justified the decision. These not only conform to the economic logic of constitutions, but also highlight the problems that arise in the new law. In the remainder of this part, I discuss the relation between Foss and the shareholders’ right to initiate review. I then use a game-theoretic illustration to justify the approach of the old law, and briefly compare it with the new law.

The following passage of Foss outlines the basis for limiting the entitlement of a shareholder to litigate in relation to an act or transaction contrary to the law:

That is, there is no entitlement to litigate the transaction as of right. This premise remained until the enactment in the then Corporations Law of the statutory derivative action. Instead, the validity of the transaction or voidable act depends on the will of the majority — on collective action. This is, of course, subject to well-known exceptions,[57] each of which has clear ‘constitutional’ justifications:

In the absence of these exceptions, the individual’s substantive entitlement is subject to majoritarian disposition. However, there is no suggestion in Foss, or the authorities that followed it, that the individual cannot invoke the processes that lead to a constitutionally valid majority decision. This is presumably a personal right, albeit procedural in nature, which the shareholder would be entitled to enforce by declaration or injunction. The only bar to this would be if the right were taken away by the constitution[62] or otherwise regulated by statute.

There are also strong consequential justifications for the traditional legal position in relation to overreaching. It lies close to the optimal legal regime, one which would afford both the directors and the shareholders the right, but not the obligation, to seek a decision regarding ratification by the shareholders at the same time as limiting the right of individual shareholders to litigate these matters of their own motion. Such a regime is likely to reduce litigation costs while deterring transactions that make shareholders worse off. To understand these justifications, it is useful to provide an example illustrative of the strategic incentives.

A director may choose whether or not to enter a transaction, which will yield a private benefit, B, to him or her (B > 0). The pay-off to the corporation — that is, to its shareholders collectively — may be positive or negative. Let us assume that an individual shareholder’s share of that pay-off is C. It would be efficient to encourage transactions where (B > 0, C > 0) and to discourage all others. The director may choose not to enter the transaction, to enter the transaction and seek ratification, or to enter the transaction and not seek approval. In the latter event, a shareholder also has three choices — do nothing, table a resolution to seek restitution of the benefits of the transaction, or bring a derivative suit. Let the probability of a majority approving the transaction when the director seeks ratification be represented by the variable p. Let the probability of a majority granting restitution at the instance of the shareholder be represented by the term (1 – q), where p ≥ q. The intuition of this inequality is that, by taking the initiative to seek ratification, the director is more likely to get ratification, being in better control of the information disseminated to shareholders. Where the shareholder litigates, assume that the court accepts jurisdiction under some exception to Foss with probability j; if it accepts jurisdiction, it applies the strict fiduciary duty and orders that the transaction be rescinded. Assume further that, where the director seeks, but does not get, ratification, or where the shareholder seeks restitution or litigates, the director suffers a reputational ‘penalty’ (as a decrease to the value of his or her human capital), represented by r. No penalty occurs where the director seeks and gets ratification, or where the shareholder does nothing.

The game structure is depicted in Figure 4, and the variables used in this example are summarised in Table 2. Pay-offs and actions are in the order director, shareholder.

B	Private benefit of the interested transaction to the director.
C	Pay-off from the interested transaction to an individual shareholder.
j	Probability that the court will permit the derivative suit to continue.
L	The cost to the shareholder to litigate.
p	Probability that the shareholders will ratify the transaction at the instance of the director.
q	Probability that the shareholders will not resolve to require the director to make restitution at the instance of the shareholder.
r	The reputational sanction the director bears if the shareholders fail to vote to ratify, or the shareholder seeks restitution or litigates.

To determine the equilibria likely to hold, we have to identify the move that the shareholder would make if confronted by an unratified transaction, in order to determine what action the director will take. The shareholder will choose to do nothing under various circumstances — where C > 0 (that is, the transaction is beneficial), or where both q = 1 and the value of j is below a certain point, specifically j < L/C.[63] Where q < 1, the shareholder is better off seeking restitution than doing nothing since (1 – q)C < C for all C < 0. However, the shareholder will litigate (take individual action) rather than seek restitution (proposing collective action) where 1 – q – j < L/C.[64] These results indicate that litigation becomes increasingly worthwhile the lower the fraction litigation costs are of the gains from litigating, and the lower the likelihood of the shareholders agreeing to restitution.

We then must determine how the shareholder’s move will affect the director’s decisions to enter the transaction and to choose to seek ratification. If the transaction makes shareholders better off (C > 0), the director is better off transacting without ratification. What of transactions that make shareholders worse off? We must distinguish between four different cases. First, the director will transact without seeking ratification in those cases in which the shareholder will do nothing. The director is better off not taking the chance of bearing the reputational sanction (equal to (1 – p)r) which occurs where ratification is sought but not obtained.

Second, where the shareholder is better off seeking restitution than doing nothing, the director is better off seeking ratification.[65] This is for two reasons, both by hypothesis but intuitively reasonable. One flows from the assumption that p ≥ q, since by taking the initiative to seek ratification the director is in better control of the information disseminated to shareholders. The other reason flows from the assumption that the director always suffers the reputational sanction if the shareholder seeks restitution, but, if the director first seeks ratification, his or her reputation is only damaged if he or she fails to get it (and, if p ≥ q, it is, by definition, more likely that the director will get ratification than that the shareholder will fail to get restitution). This seems reasonable, because a director in respect of whom prima facie evidence of an undisclosed interest is produced is more likely to suffer a tarnished reputation even if he or she succeeds in blocking the restitution resolution.

Third, the director will seek ratification where the shareholder’s best response to the director’s not seeking ratification is to litigate. The intuition here is simple. Remember that, for litigation to be the shareholder’s best response to non-ratification, it must produce a better outcome than seeking restitution. Since we found in the previous paragraph that the director always seeks ratification when the shareholder is better off seeking restitution, the director has no incentive to deviate when the shareholder is willing to litigate.[66] Finally, there are circumstances in which the director, knowing that the shareholder will seek restitution if the transaction is not disclosed, will be better off not transacting at all. This will occur where pB < (1 – p)r.

We can therefore see what happens if the shareholder is regarded as lacking the right to initiate collective action to seek restitution from the director in breach. Our concern should be that there may be parameters in which a shareholder will not litigate where he or she would have been willing to seek restitution from the general meeting. The director is better able to ‘get away’ with an unratified transaction. If litigation does not occur and the seeking of restitution is impossible, the best response by the director is to undertake the transaction without seeking ratification. The parameters in which {Director Transacts without Ratification/Shareholder Does Nothing} is an equilibrium that would not occur if collective action were permitted are q < 1, j < L/C — in other words, where litigation costs are high relative to the probability of jurisdiction being accepted. We can understand from this a basic doctrinal point that confirms my interpretation of Foss — for courts to apply that rule consistently (ie, j → 0), the shareholders must have substantial liberty to initiate collective action. If we relax the implicit assumption that ratification is not costless to the corporation, this conclusion remains valid provided only that the cost of convening the general meeting is less than L.[67]

Compare the old law with more recent legislative amendments, such as the related parties provisions and the derivative suit legislation.[68] The old law was premised on a symmetrical right for directors and shareholders to seek a decision of the majority of shareholders. The related parties provisions mandate the director’s action. Compared with the old law, this has several disadvantages. It may impose deadweight costs in requiring ratification to be sought when both directors and shareholders would rationally prefer not to approach the general meeting (where C > 0), and it may impose deadweight litigation costs in determining whether the related parties provisions have actually been satisfied.[69] The derivative suit litigation seeks to substitute litigation for collective action by increasing j and providing for costs indemnities to diminish L. The problems here are obvious — total litigation costs are likely to be higher, and greater weight is placed on courts to make determinations about the justifiability of particular transactions in place of shareholder voting. The old law only really needed supplementing to create more onerous disclosure obligations with respect to transactions involving fiduciary breaches. Only in very large corporations in which shareholder choices are most likely to be impoverished by apathy is there a theoretical justification for a more litigious model, and that justification is not well supported by empirical evidence.[70]

To conclude this subsection, my point is that, although there are no special reasons to require shareholder consent to a fiduciary breach substantially to exceed a bare majority, there are strong reasons to permit shareholders to raise in general meeting the question of overreaching and fiduciary breaches. The law should not disable shareholders’ rights to raise these objections or impose excessive transaction costs on the complaining shareholder (for instance, by requiring a threshold interest that the shareholder must meet).

3 Ordinary Management

I have already demonstrated the basic logic behind the idea that shareholders should have little or no right to intervene in ordinary management.[71] The above analysis of the duty of care shows how a rule which extends to shareholders a substantial role in making investment decisions would produce delays and bad decisions. If shareholders were permitted to intervene in management issues, they might use that power to hold up contracts. Shareholders might also use the power to propose investments that yield them private benefits, without the constraint of a fiduciary duty. Diminishing shareholders’ entitlement to intervene in management also prevents shareholders from making credible threats to hold up managers by denying them access to assets in which they have made investments in match.[72] The only threat that the shareholders can make is to sack the director, which can only be credibly threatened in a more limited range of situations in which the director is not maximising the value of the firm. The rule is incentive-compatible in much the same way as a characteristic norm of internal labour markets, which posits that an employee can be sacked, but cannot have his or her wages reduced.[73]

Of all areas, the risk of cycling in decisions by shareholders is greatest in the case of ordinary management decisions. In these situations, the directors are likely to manipulate the agenda to force shareholders to agree to resolutions that make the managers, but not the shareholders, better off — for instance, by offering them as alternatives to highly unattractive ‘straw proposals’ that the shareholders will reject. A related problem is the risk of inefficient discontinuance or continuance of certain projects. Consider a case in which there are three possible options: (A) to discontinue a project; (B) to continue it without change; and (C) to continue it with alterations. Assume that the company has 11 shareholders and that their preferences are distributed in the manner described in Table 3:

	1st preference	2nd preference	3rd preference
Shareholder 1	A	B	C
Shareholder 2	A	C	B
Shareholder 3	A	B	C
Shareholder 4	A	B	C
Shareholder 5	A	B	C
Shareholder 6	B	C	A
Shareholder 7	B	A	C
Shareholder 8	C	A	B
Shareholder 9	C	A	B
Shareholder 10	C	B	A
Shareholder 11	C	B	A

In this distribution of preferences, the results depend entirely on how shareholders are asked to vote. There is no option which commands a majority of first preferences, but A commands more first preferences than B or C, and commands fewer third preferences than any other. If, however, the directors structure the vote so that shareholders are first asked whether the project should be continued or discontinued, a majority will vote to continue (since B and C combine more votes than A alone). If the shareholders are then asked whether or not changes should be made, the decision will be made to make no changes, since shareholders 1 and 3–7 will form a majority of six to five. Thus, B beats C, which both beat A, even though A is, in fact, a Condorcet winner (that is, paired against either individually, A would win).[74] A voting rule in which Condorcet winners do not prevail, as is the case here, is said to fail the Condorcet criterion, which is a basic standard of fairness in the design of voting rules. These possibilities are clearly very likely in the area of management resolutions.

The limits on the capacity of shareholders to dictate to the board of directors also incidentally accord with a recent economic analysis of the board of directors. Margaret Blair and Lynn Stout argue that the principal purpose a board serves is as a ‘mediating hierarchy’: a body able to adjust the rewards and pay-offs to the different constituencies contracting with the firm.[75] They argue that this is a compromise necessary to avoid the shirking that would occur if rewards were divided up ex ante,[76] and the destructive transaction costs arising from quasi-rent-seeking if the constituencies were permitted to divide up returns amongst themselves ex post.[77] Thus, they argue that it is necessary for the shareholders to be incapable of dictating to the directors when they are serving in this function. Limiting the capacity of shareholders to dictate to directors reinforces this result.

4 Constitutional Change

The next subject I wish to consider is the matter of constitutional change. In Anglo-Australian law, this topic largely encompasses the law associated with the alteration of articles of association.[78] The provisions included in the articles of association, although often heavily standardised, are typically fundamental to the rights of shareholders inter se, the powers of managers, and the relation between shareholders and managers. These provisions are clearly very important, but there may be a range of situations in which it is appropriate for them to be changed. Changing legal rules may require some degree of modification or exclusion. The ownership of the company may change, as, for instance, may occur where a company goes public. Because these provisions generally establish the parameters for the participation in governance functions by shareholders, officers and managers as a whole, the desirability of majoritarian collective action as the means for making such changes is clear. Leaving the matter to individual agreement would impose excessive transaction costs, which may thwart desirable changes.

On the other hand, majoritarian processes are not without their risks. The capacity for majoritarian change diminishes the credibility of the contractual commitments given in the original constitution. This is particularly true if the processes of collective choice are tainted by ‘rational apathy’, such that management has the capacity to effect suboptimal changes.[79] The other obvious risk is the capacity of the majority to expropriate the minority.

How will rational shareholders draft a ‘meta-constitutional’ rule — that is, a constitutional rule for the changing of constitutions? It is clear that a rule of unanimity is unlikely because of the high incidence of decision-making costs it would occasion. The use of a majority requirement substantially greater than a bare majority is, however, likely to have substantial appeal. A supermajority makes it less likely that the cycling tendencies of voting will result in changes contrary to the will of a majority — only amendments that are Condorcet winners should normally be approved. This reinforces the content of the ex ante contract and reduces the likelihood of midstream opportunism that imposes externalities on some or all shareholders.

The law reinforces these changes in a number of ways. First, it requires a supermajority of three-quarters to change the articles, since a special resolution is necessary to do so.[80] Second, by regulating the procedure for passing a special resolution, the law effectively minimises the capacity of managers to manipulate the voting agenda. They must, for example, specify the text of the resolution for specific approval or rejection to be passed in substantially that form. This minimises the ways in which resolutions can be put to shareholders, and the cycling that is possible when a third option is introduced. Third, prior to the enactment of the Company Law Review Act 1998 (Cth), the law enabled the parties to create their own immutable constitutional provisions by including particular articles in the memorandum.[81] It is no longer clear that this can be done.[82] Fourth, the law on members’ rights, the fraud on the power tests which apply to the alteration of articles, and other provisions in the law constrain the capacity of the majority to impose externalities on the minority.

We may also regard related transformative events in which the structure of rights in the corporation are substantially or even irreversibly changed as being closely analogous to constitutional change. One example is the law on voluntary winding-up.[83] Like the law on the alteration of articles, it also imposes a supermajority requirement,[84] again as an attempt to permit winding-up only when that option is at least a Condorcet winner vis-à-vis other proposals for the company. Clearly, winding-up can impose substantial externalities on various constituencies of the firm, since it may require the sale of assets to which other investments are co-specific, thus dissipating the quasi-rents of the firm. The capacity of shareholders to use the voluntary winding-up procedure to do that is limited to the extent that debts and other forms of claims must be paid out in full before the shareholders are entitled to the assets.[85]

5 Other Issues and Conclusion

There are other changes for which the law requires a special resolution. Some of these relate to the distribution of funds of the company in a selective manner.[86] The risk of these mechanisms being used to expropriate the capital of the company uncontroversially justifies the supermajority requirements of a special resolution. This works in a two-sided way, since it limits both insiders attempting to expropriate assets and outsiders attempting to greenmail the corporation.

From the above sections, we can see that the law has structured the scope of, and procedure for, collective action by shareholders in the modern corporation in a range of very different ways. These differences are not merely historical incidents or neutral mutations, but fit closely with the economic logic of constitutions, amongst other things. We have seen that the significance of external effects corresponds to the degree of consent required for collective action, so that requirements are the lowest in the election of directors, and highest in constitutional change. The risk of cycling and hold up considerations explains why the law is reluctant to place substantial discretionary powers in the hands of shareholders. The law in relation to moral hazard and overreaching is one of the more surprising areas, since it takes a quite liberal approach to collective contracting with the director. The risks that this creates are reduced by other principles, such as a requirement of full information and the absence of coercion, which constrain the number of possible resolutions that can be put to shareholders. One of the most important arguments that I have developed in this section is that shareholders should be recognised as having a broad right to raise issues of overreaching; the presence of that right is important to the capacity to discipline overreaching. This is also implicit in the rule in Foss.[87] Part III examines aspects of some of the issues connected to the relation between the board and the general meeting in the 20th century.

III THE MOMENTUM OF 20TH CENTURY LAW

As is well known, in the 19th century, the general meeting was regarded as the supreme corporate organ.[88] Its resolutions bound the directors in the manner of a principal’s instructions binding an agent. With the increased use of the unincorporated joint stock company in the 18th and early 19th centuries, those establishing the structure of firms had to make choices about the content of the contracts between the shareholders and the directors. The most common provision was the use of a term in the deed of settlement conferring managerial power on a board of directors or analogous body. This was eventually replicated by the suppletive rules of early companies legislation, and the articles of companies incorporated thereunder.[89]

Against this background, cases arose in the 20th century which required courts to interpret these provisions of the articles. The context of these cases usually involved disagreement between a majority of the directors and a majority of the shareholders. The authorities in general reach results that are both logical and acceptable, but their reasoning is sometimes confusing and distorting. I discuss three leading English cases which have been followed or emulated in Australia.[90]

Perhaps the most commonly cited decision is Automatic Self-Cleansing Filter Syndicate Co Ltd v Cuninghame.[91] In that case, a shareholder, who had the support of a majority shareholder bloc, sought to require the directors to give effect to a general meeting resolution, supported by a bare majority, to sell the assets of the company. The articles of that company provided that powers of management were vested in the directors, but subject to ‘regulations, not being inconsistent with these presents, as may from time to time be made by extraordinary resolution’.[92] The trial judge refused orders sought by the shareholder, and was upheld on appeal. In his judgment, Collins MR made the point that, by the articles, the directors were empowered by all shareholders to manage the company and to dispose of its assets.[93] It was therefore not possible for a majority of shareholders to withdraw that ‘mandate’. Collins MR also agreed with the trial judge, Warrington J, that to permit a bare majority to dictate to the board would undercut another article, which exposed the director to dismissal only under a special resolution.[94]

This is a sensible judgment which reaches the correct result. The constitutional provisions are designed to mitigate the externalities or hold-up costs that a bare majority can collectively impose, either on the shareholders or on other constituencies. The other point that could be made to support this finding is the usual requirement for a special resolution or a voluntary winding-up; the sale of the company’s undertaking has very similar overtones. It is a matter of constitutional significance. I have already argued that such an event is sufficiently important and ‘one-off’ to require a substantial majority — such events must be Condorcet winners, at the least, in order to be supported. It is of course true that agency costs may be increased by minimising the capacity of shareholders to direct the sale of the assets of the firm. The assets may have ceased to return more than the cost of capital, and their sale to another firm valuing them more highly may have been justified. However, this argument is not a sufficient justification, by virtue of the fact that this is a known risk when constitutional provisions imposing supermajority requirements for collective action are used. The constitutional provisions are presumably justified on other grounds, such as minimising the incidence of externalities or hold-up.

The one criticism that can be made of the approach of the Court of Appeal is that it is silent on the circumstances in which a majority is entitled to enforce its will over the board. The suggestion of the Court is that its judgment is purely a matter of construction — the will of the general meeting is decisive only when there is scope explicitly afforded to it in the articles. Such a passive view of interpretation has attractions,[95] but it has rarely been an unalloyed principle in corporate law.[96] This theme becomes apparent in subsequent authorities.

In another much-cited case, John Shaw & Sons (Salford) Ltd v Shaw,[97] a company included the standard provision vesting managerial power and all powers not explicitly conferred on the general meeting in the board of directors. The facts involved the bringing of suit by the company against two of its directors (who constituted a minority of the board). The general meeting resolved that those proceedings be discontinued; the defendants had voted with the other shareholders. On that basis, the defendants sought orders that the proceedings be dismissed. The Court found against the defendants. This is again clearly the correct result at two levels. The enforcement of debts is a fairly straightforward managerial judgment, and the use of majority power to vitiate a debt seems to expropriate rights from the minority. Since the debts did not appear to arise from constitutional provisions, the case against permitting them to be varied by collective action seems very strong. The judgment, however, emphasises that it is a result reached by construction. Greer LJ makes the comment that:

The dictum was unnecessary because all the judge needed to conclude was that this exercise of power was within the grant of the article — there was no occasion to determine the outer limits of the article or to comment that voting and altering the articles are collectively exhaustive of the shareholders’ entitlements. Also, to say that the shareholders cannot usurp the powers vested in the directors begs the question as to whether the shareholders have powers that are an implicit consequence of the use or application of power referred to the board by the articles, just as the law imposes limits on the use by directors of their powers in certain circumstances.[99]

A third case, Marshall’s Valve Gear Co Ltd v Manning, Wardle & Co Ltd,[100] decided after Cuninghame but before John Shaw, shows the distortion that such an approach to the law engenders. In that case, a corporation purchased a patent. The patent’s previous owner became the managing director and had a controlling interest in the company (albeit falling short of a supermajority sufficient for a special resolution). The dispute arose from the acquisition by the other directors of the rights to a patent over an invention competing with the patent acquired by the company. Those directors voted in favour of a board resolution not to oppose the grant of this latter patent nor to bring proceedings for the infringement of the earlier patent. The managing director brought an action to establish a patent infringement on behalf of the company.

On the facts of this case, there is a clear conflict of interest on the part of the other directors, which the trial judge, Neville J, recognised.[101] After considering Cuninghame, he held that there was no ground to order the removal of the writ. His judgment is a complex one. He noted that, according to 19th century law, a majority of shareholders was entitled to determine whether an action should proceed.[102] He distinguished the result in Cuninghame on the dubious basis that the article vested the power of management in the board subject to regulations ‘as may from time to time be made by extraordinary resolution’.[103] Thus, Neville J said:

In other words, the case is based on a premise that shareholders do have a right to control the directors, but that the articles may cut down those rights as in Cuninghame. This principle has become increasingly doubted in later authorities.[105] The problems with this view cut in two directions — if it is right, the directors are subject to interference by shareholders in managerial functions; if it is wrong, the directors would have almost complete immunity from suit or collective action by the shareholders. The principle is thus too widely drawn, since it takes the view that the article creates either a generic indemnity from interference as in Cuninghame or John Shaw, or offers basically no protection unless there are explicit limitations on shareholders.

What these cases fail to do is to describe how the law on the board–general meeting relation meshes with the entitlement of an individual shareholder to activate a process which leads to a collective decision. Part II(B)(3) above analysed the latter subject in detail. I argued that, under the traditional law, the individual shareholder had an entitlement to propose a motion related to a breach of fiduciary duty by the director (despite the historical absence of a general right to litigate such a question). That entitlement could be taken away by the explicit language of a statute or the constitution. The question, then, is how the generic provisions of the constitution or the replaceable rules of the Corporations Act affect this entitlement. A provision such as s 198A(1), addressed specifically to managing ‘the business of a company’ is not apt to destroy the entitlement to raise a fiduciary issue. It is appropriate to draw a distinction between litigation arising in the general course of business, which is beyond the general meeting’s remit, and litigation associated with a breach of duty, as in Marshall’s Valve.[106]

The more generic referral of power in s 198A(2), regarding the exercise of all powers of the company, is a more complex provision. My conclusion is, however, unchanged. One could say that this provision could not be contemplated to apply to fiduciary breaches, since, if it did, that would imply that the board was competent to waive breaches of duty. Clearly, the board is not so competent. Alternatively, one could say that ratification is not a ‘power of the company’ in the language of the provision, though the issue of shares or acquisition of property may be but a right of a majority of shareholders granted for prudential reasons.

Thus, the decision in Marshall’s Valve is unfortunate. Neville J’s view of the article vesting managerial power in the board was too narrow since it allowed the shareholders a wider remit than was desirable. The appropriate conclusion was that the article deprived the shareholder of the right to intervene in matters of management but that this case fell outside the article. The shareholder holding a majority of shares would clearly prevail in the general meeting, and was entitled to litigate as though the general meeting had resolved accordingly.[107] The concern is that later cases might reject the validity of the result in Marshall’s Valve because of its dubious reasoning.

When we approach questions relating to the rights of shareholders in general meeting, it is appropriate to evaluate the nature of the resolution that the general meeting is considering. The matter cannot always be resolved only by reference to a provision in the constitution or to replaceable rules conferring residual power on the board. I take up this issue in Part IV, in which I study reform proposals.

IV REFORM PROPOSALS

In this part, my principal focus is the report by CSAC on the role of shareholders in the modern company,[108] together with certain related proposals. I pay particular attention to the requisitioning of meetings, the proposing of resolutions, and shareholder voting.

A Requisitioning Meetings

1 Some General Considerations

One of the implicit premises of the CSAC report is that requisitioning meetings and proposing resolutions are similar, and should be resolved with reference to similar factors.[109] CSAC’s proposed approach recognises that if meetings can be requisitioned or resolutions proposed by a small number of shareholders, there will be substantial costs imposed on the body of shareholders.[110] In a recent press release, the Minister for Financial Services and Regulation, Mr Joe Hockey, outlined a proposal to address this problem. In that press release, this proposal is described as stopping ‘political activists from hi-jacking company annual meetings’.[111]

The basic principle at stake — that a small number of shareholders can externalise costs by abusing the power to activate collective processes — is valid and familiar from the discussion of externalities in Part II. But it seems intuitively unlikely to be of the same order as between requisitioning a meeting and proposing a resolution at a meeting that would have been held anyway.[112] The marginal costs associated with the former are the cost of circulating information in relation to the proposal, and the time spent by shareholders and directors in considering the proposal and voting on it. The marginal costs associated with the latter appear to be greater. We should therefore ensure that requirements for proposals and meeting requisitions are not equated. Increasing the difference between the requirements as to resolutions and proposals encourages those who would consider requisitioning a meeting to propose instead a resolution at a meeting that will be held anyway, since the implicit cost is lower. There is, however, a more general objection to a focus that is exclusively on the costs imposed on the corporation by requisitioning a meeting or proposing a resolution. Such an approach ignores the marginal costs of collective action imposed on the shareholder. Increasing these costs decreases the ex ante incentive of a shareholder to take any action at all, which may offset the corporation’s costs while decreasing the capacity of shareholders to discipline managers.

Both the CSAC report[113] and the Minister’s proposal[114] seek to change the existing rule in relation to the requisitioning of a meeting. The law in this area has been the subject of recent change. Prior to 1998, the requirements for convening a meeting were the requisitioning of either (a) 100 shareholders holding shares with an average paid-up sum per shareholder of at least $200, or (b) shareholders entitled to at least 5% of the total voting rights in the company.[115] In July 1998, the test was changed, deleting the average paid-up capital requirement.[116] In 2000, regulations deleted the former test and substituted a requirement for the requisition of 5% of the shareholders by number, pending a response to the then anticipated Final Report by CSAC.[117] CSAC recommends the deletion of a test predicated only on shareholder numbers, and the retention of the 5% of voting capital test.[118]

Not to be outdone, the Minister has struck out in a different direction. His proposal retains the 5% of voting capital test, but adds an alternative numeric test[119] that would permit a meeting to be requisitioned by a body of shareholders greater than or equal to the number that is the square root of the current number of shareholders. The proposal is quite curious, when compared with the existing and previous rules. Figure 5 is a graph which depicts the proportion of shareholders required to requisition the meeting under the square root rule and the 100 shareholders test. Figure 6 is a ‘close-up’ of the graph, which enables the reader to see in more detail the difference between the two rules at the upper end, beginning at the arbitrary value of 6400 shareholders:

Formally, the graphs in these figures chart two functions, one being 100n-1, the other n-1/2, where n is the number of shareholders in the corporation. Together with the 5% rule, they produce radically different results. It is simply not valid to assume, as the graph shows, that any of them is systematically more demanding than the others. Table 4 shows the relative magnitude of the different rules:[120]

Company size	Most demanding rule	Least demanding rule
0 – 400	100 shareholders	5% rule
401 – 2000	100 shareholders	Square root rule
2001 – 10 000	5% rule	Square root rule
10 001 –	5% rule	100 shareholders

It is therefore impossible to extol the capacity of any one of these rules to save costs without making assumptions about the size of the company. As can be seen from the chart, a 5% rule in the companies with a million or more shareholders would impose transaction costs of a vast magnitude on any shareholder considering requisitioning a meeting. The costs of communicating with so many thousands of shareholders would be enormous. In his press release, the Minister refers to the considerable costs associated with staging a meeting requisitioned by shareholders, of the order of a million dollars.[121] It is easy to see an analogous sum being imposed on the shareholder who must persuade what could be many more than fifty thousand other shareholders.

2 The Square Root Rule Considered

Because the square root rule is in fact the least demanding for a wide range of companies — those with more than 400 and fewer than 10 000 shareholders — its appeal must be predicated on the presumption that the abuse of the requisitioning privilege is in fact most likely in larger corporations. There may be a number of reasons why this might be the case. It is arguable that genuine cases of ‘hijacking’ arise most often in very large corporations. That is because the purpose of the resolution is not really concerned with the effectiveness of collective action. The requisitioners know they are highly unlikely to prevail in any resolution. However, they may be relying instead on the value of achieving substantial public exposure for particular causes and aspirations. There are many empirical studies documenting a ‘size effect’ in the context of the political sensitivity of corporations.[122] If that is so, the incentive to requisition is expected to be greatest in the largest companies. Such an incentive would explain why the square root rule only becomes more demanding when the number of shareholders exceeds 10 000. But, if that is conceded, it is worth noting that the advantages of the square root rule, vis-à-vis the 100 shareholders rule, only really start to cut in for the very largest companies (those in the top 100 or so). For example, a company with 25 000 shareholders can be requisitioned to meet with the consent of 159 shareholders — not radically more than the 100 figure. The proportional comparisons are between 0.4% and 0.64%. To take an example that fits these numbers, consider that, at 30 December 1999, John Fairfax Holdings, in the top 50 companies by market capitalisation, had 26 067 shareholders.[123]

If the square root rule’s protection is marginal for all but the largest companies, one must still wonder why the Minister would be willing to diminish the protection the current rules afford in even smaller companies. It may well be the case that such a company is less likely to be followed by politically-minded shareholders than larger companies; nonetheless, other forms of opportunism are also possible. Quite small factions in small listed companies that are approaching financial distress could use the meeting as a credible threat if their collateral aims are not fulfilled by the company. For instance, those shareholders could requisition the meeting to call attention to the company’s plight — which is credible if they themselves are creditors or allied to creditors. Just 40 shareholders could call a meeting of a company with 1600 shareholders — half the number required by the 5% rule, and even less than the 100 number. The adverse publicity could sink the company.

3 CSAC’s Case

This logic seems to suggest that the Minister is wrong, and that CSAC is right. CSAC favours only the 5% of voting shares rule, and no numeric test. In effect, such a test means that a meeting could only be requisitioned with the imprimatur of an institutional (or non-institutional controlling[124]) shareholder. In practice, this may be difficult to get. The relations between the institutions and the board of directors seem to be governed by norms.[125] From experience, we know that those social norms discourage active dissent in the absence of highly inappropriate behaviour, perhaps in return for other favours such as access to management, insider information, and so on. If these practices apply to the requisitioning of meetings, a 5% rule would often require acquiescence of a very large proportion of the non-institutional shareholders, all of whom have weak incentives to be informed and to take action. In these circumstances, a rule that explicitly forbade the requisitioning of meetings may not operate very differently for many companies.

The justifications for the 5% of voting shares rule that CSAC offers are not particularly compelling. It is no doubt true to say that shareholders who cannot muster 5% of the votes are unlikely to pass the resolution.[126] However, since we know that remarkably few votes ever go against management, the case for allowing a meeting to be requisitioned in the first place to consider a hostile proposal is equally weak.[127] Thus, the justification for a 5% rule as opposed to other requirements is weak. To say that a 5% rule effects the necessary balance between majority and minority interests begs a question of how significant the capacity to requisition a meeting is to the protection of minority shareholdings.[128] It would be both more humble and more accurate to say that a 5% rule is arbitrary, but is not obviously capricious. That point is amplified by the final justification, that a similar rule is adopted elsewhere.[129]

4 Two Options: The Counter-Requisition and the Default Rule

(a) The Counter-Requisition Procedure

My conclusions so far have been critical of both proposals. The Minister’s proposal does not take a consistent approach to the reduction of the externalities that the requisitioning shareholders can impose on the company, as defined across a wide distribution of company sizes. The CSAC approach takes a more consistent approach to decreasing the cost of externalities, but fails to consider the corresponding decrease in the costs of consent. Is there an alternative rule which is supportable on better grounds than either of these others? I consider two proposals. In the first, one might be willing to accept a numeric test as a supplement to a proportion-of-voting-capital test. Whichever of these tests is used by shareholders to requisition the meeting, one might then impose a second-round test which enables another, larger, subset of the shareholders (for which I would favour a proportion-of-voting-capital test[130]) to pre-empt the requisition by a petition to dismiss the requisition.[131] In view of the fact that there is a norm counselling voting with management,[132] it would be appropriate to impose a high figure, say 25%. Management need only marshal the support of a relatively small number of institutions, if the proposal is indeed unmeritorious.[133] The legislation would provide further that, if management marshalled the requisite opposition, the shareholders requisitioning the meeting could seek the leave of the court to proceed with the requisition under stated grounds. The only ground on which leave should be granted is where the resolution to be put at the meeting is effectively time-critical, so that there is a real risk that deferring it until the annual general meeting would render the subject nugatory, or that effective relief would be impossible. Again, this encourages shareholders to pursue a strategy which imposes lower marginal costs, and only permits relief where deferral is not a satisfactory option. Certain issues arising in a takeover and other issues involving the transformation of rights and entitlements might be examples. If the court does not grant relief, the shareholder still has an option to seek an injunction under s 1324 or a derivative action under s 236. One of the advantages of this rule is that it requires management to persuade block-holders to vote with it, as opposed to management being able to rely simply on a norm not to rock the boat.[134] In more meritorious cases, management may be unable to garner that support, leading to greater embarrassment for it.

We may compare CSAC’s and my own proposals by examining the net costs of collective action that they impose and how these might change. I do this by comparing the effects of the various processes of requisitioning a meeting in terms of a utilitarian social welfare function, U, which aggregates the respective costs of the corporation and the proposing shareholders, and the effects on the value of the corporation. U is made up of several components. First, there are the costs shareholders incur in order to comply with the rules on requisitioning meetings. Call these costs S. Second, there are the costs of convening the meeting, the disclosure to shareholders, and so on. Call these costs M. Third, under my second proposal, there are the costs incurred in mustering opposition to the requisition.[135] Call them O. Fourth, there is the net value implications of the resolution that the meeting has been requisitioned to consider, which may be called A, which should be multiplied by r, which is the probability of the resolution being passed. The fifth variable that needs to be introduced is the probability of the meeting actually being requisitioned, called t, which depends on the capacity of the shareholders to muster the numbers, or the directors to muster opposition. Thus, we may define U as follows:

The first term, t(r.A – M) is similar to the externality (positive or negative) imposed by requisitioning the meeting. The second term, S + O, is the cost of consent.

A	The advantage to the corporation from the resolution succeeding.
M	The cost to hold the meeting.
S	The cost to the shareholders of complying with the legislative requirements on requisitions.
O	The costs incurred by management in mustering opposition to the requisition.
U	Level of aggregate social welfare from requisitioning a meeting.
r	The probability of the shareholder’s proposed resolution succeeding.
t	The probability of the meeting being successfully requisitioned.

Obviously, the optimal rule is the one that maximises U. To consider how the different legal rules might affect U, we need to know how these variables behave. Figure 7 graphs how these costs might conceivably behave, modelled as functions of the variable A. M is modelled as being invariate with respect to A, that is, ∂M/∂A = 0. Naturally, M depends on the requirements of the law for shareholder support. S is modelled as an increasing function of A (∂–S/∂A > 0), on the basis that the decrease in shareholders’ wealth from requisitioning a meeting falls the greater the advantage from requisitioning the meeting. I have treated O as being a more complex function, which is essentially invariate with respect to negative values of A, on the basis that managers will find blockholders who are willing to dismiss negative requisitions at low cost. At positive values, I have treated the function as being curvilinear, on the basis that larger holders take a substantial degree of persuading as to why they should depart their passivity norms where A is relatively low. Beyond a certain point, shareholders will be unwilling to participate in blocking the requisition. The value r.A is also substantially invariant at its negative values, but is an increasing function of A where A > 0, suggesting that r rises from 0 to approach 1 at high values of A. Therefore, the function t(r.A − M) will run from around the value of S at low levels of A, and will run broadly parallel with r.A, becoming positive only at some positive value of A.

CSAC’s proposal, in effect, would move S down to S', shown in Figure 8. However, we know from equation (1) that ∂U/∂S < 0 — U falls as S rises. CSAC’s proposal can only increase U in the hope that there will be a compensating increase in the curve t(r.A − M) through a decrease in the value of t — the probability that the meeting is in fact requisitioned. Now, it seems logical that heftier requisitioning requirements will decrease t. However, CSAC must rely on the fact that this decrease in t only occurs for negative values of A. If t falls for positive values of A as well, then the rate of change of the function t(r.A − M) with respect to A must necessarily be lower, so it follows that U must also be lower. That means in substance that shareholders would find it harder to requisition advantageous meetings. To display my reasoning graphically, CSAC is hoping that the curve will look like t(r.A − M)', the dotted curve in Figure 8, which lies above t(r.A − M), rather than t(r.A − M)'', which crosses the original function. This seems a very difficult balance to achieve.

My proposal relies instead on the likely shape of the function O. For low values of A, it is clear that a procedure for blocking a requisitioned meeting can save costs compared with a more onerous requisition procedure if the costs of mustering opposition are lower than the increase in the value of S under the latter procedure. Thus, it would be preferable for meetings requisitioned for collateral purposes. This may actually be reinforced if those who would seek to convene a meeting for public exposure expect that they could be effectively vetoed, leaving them with the costs of their campaign but no forum — this possibility would be reflected in the move of the curve t(r.A − M) to t(r.A − M)'' as a result of the drop in the value of t for negative values of A.

What about positive values of A? How would my proposal impact here? To answer, we must first recognise that even ‘good’ reasons to requisition a meeting — those with a positive value of A — are not always enough if A is small enough, and certainly not in any case where t(r.A − M) is less than zero. It is therefore not so troubling that large shareholders may be willing to indulge managers up to a certain point by opposing the requisition. However, I hypothesise that the willingness to indulge managers diminishes rapidly beyond a certain point, because of the negative impact it has on the value of the corporation. As such, for the mid-range values of A, the curve O should have a steep negative slope, making it impossible for management to muster the necessary opposition. In this respect, management may be willing to ‘take on the chin’ the requisitioned meeting and whatever proposal is to be put there, rather than to have to offer excuses or future commitments to the larger shareholders. If this is true, then it indicates the superiority of an opposition procedure at the high values of A — the only ones where a meeting is desirable — since it reduces S, while confronting managers with such high costs of opposition that it is easier simply to acquiesce to shareholders. The danger zone is in the intermediate range, where management decides it is better off engaging in a costly mustering of opposition. With this relatively narrow exception, the opposition procedure seems to have strong advantages compared with a more onerous requisition procedure. To conclude, the opposition procedure need not assume that the fulfilling of an onerous requirement becomes easier the stronger the shareholder’s case; it only needs to assume that, beyond a certain point, the costs of opposing are too high to bother in the first place.

(b) A Contractible Rule

A contractible rule on requisitioning meetings will be anathema to some — it suggests that the protection of a rule might be diluted. But it is implicit in my analysis that the content of the optimal rule is highly unclear. The optimal rule depends on the size of the company, the distribution of the shareholdings, the nature of the issue, and so on. We can only tell for sure whether the rule actually serves any useful purpose by examining statistically the effect on the share price of requisitioning meetings at all. In the absence of such evidence, and the countervailing tendency to gravitate to arbitrary ‘focal point’[136] proportions, such as 5%, 10% or 20%, the case for contract is very strong. Should it be the case that investors value the right to requisition meetings, a company that excluded that right altogether would presumably trade at a discount when its shares are first offered to the public. The offerers of shares have an incentive to minimise that discount. If, on the other hand, the requisitioning provisions are used only for reasons that add nothing to the value of the firm, those opting out of the rule or imposing a requirement for a larger proportion of the voting shares in favour of the requisition, would be, at the least, no worse off.

Unlike the Delaware rule, which provides that a special meeting can only be called by the board in the absence of authorisation being given to the shareholders in the constitutional documents,[137] I would propose to retain a provision permitting a meeting to be requisitioned by a group of shareholders who:

(ii) twice the number that is the square root of the number of shareholders at the time the meeting is requisitioned.

However, that rule would be a replaceable rule for both public and proprietary companies. The difference is that the rule entitles firms to opt out, rather than permitting them to opt in. It would remain to be seen whether the existence of a default deterred firms from opting out of it, as some theorists hypothesise.[138]

B Proposing Shareholder Resolutions

I have already had much to say on the subject of shareholder resolutions. The gist of my analysis in Parts II and III made it apparent that I favour an approach in which preconditions to shareholder resolutions are graduated along subject matter grounds.

Where there is a case of director overreaching or an issue that directly affects the shareholders’ entitlements as members to vote, participate in meetings, or to receive capital and income,[139] the preconditions should be minimal. Their only possible purpose can be to exclude frivolous or vexatious proposals along these lines. However, if the costs associated with putting these proposals to an existing meeting are relatively low, then the best response to frivolous or vexatious proposals may be for them to be voted down, rather than for them not to be put at all.

At the opposite extreme, the case for allowing proposals in respect of ordinary management must be heavily attenuated, since the grounds for permitting shareholders to have any role in the management of the company are few. The simplest provision would be to prohibit the proposing of these resolutions rather than imposing limitations referable to a required degree of collective action. It is also quite likely that these are the matters that are more likely to be politicised, such as environmental issues, labour practices domestically and abroad, and relations with government. Retarding these proposals may dampen the incentive of politically motivated shareholders.

There are a range of issues that fall somewhere between these arguments. An example is executive compensation. The proposition that shareholders should makes cases in relation to the structure and amount of compensation — for example, arguing for a higher proportion of variable compensation — is important because of the centrality of these provisions in the alignment of incentives.[140] It is these resolutions where the value of a minimum support requirement is useful.

I have indicated already that the considerations relevant to the preconditions for the proposal of resolutions are different from those that are relevant to the requisitioning of meetings. Nonetheless, the formal provisions imposed by the law are not especially different. Under ss 249N–P, the 100 shareholders/5% voting capital tests are the preconditions for a shareholder right to require the company to distribute the text of a proposed resolution and a statement in support of it. CSAC favours the continuation of these provisions, with the added requirement of a minimum average interest of $1000 per shareholder in the case of the 100 shareholders test.[141] The report considers and rejects a range of different possible approaches.

The most important of these is that CSAC rejects the use of relevance controls. CSAC holds that they would generate complexity and uncertainty. In doing so, they unfortunately pay only brief attention to the experience with such controls in the United States.[142] A full analysis of the relevance controls used in American federal law is beyond the scope of this paper. However, they strike a balance which approximates the one that I have advocated. The chief exclusion is for ordinary business and management functions. That head of exclusion has been the subject of variable interpretation. From 1992 to 1998, the Securities and Exchange Commission (‘SEC’) was willing to issue ‘no action’ letters in relation to the non-inclusion in proxy materials of proposals with social policy concerns, such as ones dealing with anti-discrimination policy.[143] This is consistent with the normative case that I have put forward.[144] Since 1998, the SEC’s approach has become more case-specific.[145] Various other relevance controls are imposed, such as the exclusion of resolutions relating to operations that are ‘insignificant’ in terms of the corporate business.[146] There is no single set of legal principles capable of serving this function. The United States principles are based on implicit inclusions and explicit exclusions; the reverse is also possible, but would need to depend on the explicit inclusions being drafted broadly enough to mitigate the risk of under-inclusion.

To shed light on the preconditions that should be prescribed for shareholders to be able to table a resolution, we need to know something about how shareholders have made use of these provisions. Perhaps the most pertinent experience here is the activism of trade union pension funds in the United States. Stewart Schwab and Randall Thomas, in an extensive review of this subject, argue that workers behave more as model shareholders than proletarian subversives.[147] They can be at once activist and intelligent by combining their deep knowledge of the operational side of the corporation with the absence of social conditioning and normative influences that encourage passivity in the relations between other institutions and management. Thus, unions have proposed review of a range of aspects of the shareholder–management contract, such as aspects of compensation plans, anti-takeover mechanisms and so on. Unsurprisingly, this model conduct does not always occur. Sometimes, the proposal of resolutions is used as a strategic tactic in the development of collective bargaining. What is more significant, however, is that the other shareholders seem to show considerable discrimination when voting on union proposals. They have been capable of distinguishing the proposals with rent-seeking objectives from those that seek to increase wealth. They have provided some support for the latter, but very little for the former.[148] This suggests that abuse of the right to propose is capable of being disciplined by shareholders, which in turn suggests that the case for substantial preconditions is weaker than it first appears.

Despite this somewhat up-beat diagnosis, it is fair to say that financial economists have found very mixed results in respect of the value created by shareholder proposals in the United States. There have been many studies of both the attributes of firms in which proposals have been made, and the effect of these proposals.[149] Review of these studies is beyond the scope of this already overlong article. Subject to a few exceptions, they suggest that there is no general relation between proposals and firm performance, and some evidence of negative returns.[150] It may, therefore, be appropriate to impose some limitations on proposals.

Rather than regulating shareholder proposals only in terms of the collective support required before a resolution can be put, legislation could more usefully ration the right to propose on both proponent and subject matter lines, and impose ex post punishments for certain proposals.[151] The use of rationing is desirable in two senses. It reduces the number of resolutions and so economises on the cost of dealing with them; further, it reduces the risk of inconsistent or intransitive results from the proposal of multiple resolutions dealing with the same subject matter. The American approach utilises rationing in various respects. First, duplicative proposals need not be tabled.[152] Second, shareholders are limited to one proposal a year.[153] Third, proposals cannot be resubmitted where they are substantially the same as an earlier proposal that has failed in the previous five years or so.[154] Rationing shareholders to one proposal a year has been difficult to enforce because of the capacity of shareholders to work through other associates or shareholders they control. The SEC’s ‘alter-ego’ test has had but limited effectiveness.[155]

Ex post punishment would impose a cost on the frivolous or vexatious shareholder. This could take either ‘price’ or ‘quantity’ forms. A price form would require the shareholder to bear the costs of his or her proposal, such as by submitting a bond or security. This may prove difficult to administer, in part because of the ambiguity of quantifying the marginal cost occasioned by the proposal in view of the presence of fixed costs.[156] It was rejected by CSAC,[157] but has been supported in the United States by various scholars.[158] A quantity punishment is, however, easier to administer, since it would impose restrictions on the right of demonstrably vexatious proponents to make future proposals, vis-à-vis other shareholders. This could be done by imposing a ban on the shareholders’ capacity to make future proposals in relation to that company, or, alternatively, by imposing on that shareholder or that proposal a more onerous set of preconditions relating to approval. Both would presumably operate for a period of three to five years.[159] I am inclined to favour the more onerous precondition proposal over the ‘sin-bin’ approach. It would take the form that the shareholders whose proposed resolution failed to garner a minimum level of support — somewhere between, say, 20% and 25% — be obliged to obtain the imprimatur of 5–10% of the voting shares of the company for the proposal. This encourages shareholders to put those resolutions that they expect are likely to attract wider support. It also reflects the possibility that there are trends and developments that might substantially change the perception of a proposal from silly to visionary.

I have tried to capture the essential features of a recommended approach in the following provision:

(a) Any shareholder can put a proposal in relation to restitution, rescission, compensation, or other redress in respect of a breach of fiduciary duty by a current or past director, or the contravention of any other law protecting shareholders against overreaching or moral hazard. There would be no requirement for concurrence.

(b) Any shareholder can put a proposal in relation to the rights and entitlements of shareholders, including, without limitation, rights under the corporation’s constitution, rights to be notified of, attend, and deliberate at, meetings of shareholders, and rights of capital and income. There would be no requirement for concurrence.

(c) Subject to provisions (a) and (b), no shareholder can put a proposal that deals with the ordinary management of the corporation, or decisions made in relation to production, operation or investment.

(d) If a matter falls outside of provisions (a), (b) and (c), a proposal may be made by:

(i) a shareholder who has the right to cast more than 5% of the votes that might be cast at a general meeting; or

(ii) shareholders whose interests in total would permit them to cast more than 5% of the votes that might be cast at a general meeting or who number no less than one hundred.

(e) The right of a shareholder or group of shareholders to put a proposal under provisions (a), (b) and (d) is subject to the following limitations:

(i) No more than one resolution per year per shareholder or per group of shareholders (as the case may be) may be put under provision (d).

(ii) Where two or more proposals concerning substantially the same subject matter have been submitted in accordance with provisions (a), (b) and (d), the proposal that is submitted first in time will be put to shareholders unless the person or persons making the proposal agree to withdraw their proposal in favour of a later proposal.

(iii) Where a director puts a resolution that deals with substantially the same subject matter as a resolution proposed by shareholders at the same meeting, shareholders must be asked to vote for and against each proposal separately, and also to express their preference as between the two resolutions. That resolution is passed for which a majority of shareholders have voted, and which a majority prefers to the alternative resolution.

(iv) Where a shareholder, either individually or as part of a group of shareholders, has proposed a resolution which has been opposed by no less than 85% of the votes cast in respect of that resolution at any time in the preceding three years, the shareholder may not propose a further resolution under provisions (a), (b) or (d) unless he or she proposes it as part of a group of shareholders who are entitled to cast more than 10% of the votes that might be cast at a general meeting.

The only provisions that need explanation beyond that already found in the text are provisions (e)(ii) and (iv). The former provision uses a test which minimises the discretion of directors as to which proposal they put, but nonetheless enables a species of Coasean trade between different factions.[160] Provision (e)(iv) is markedly different from the United States approach since it does not give management the option to propose only its own resolution, and in so doing pre-empt a resolution of the shareholders that might be considerably more demanding. The problem, however, is the risk of a Condorcet paradox. For example, the shareholders’ resolution may be preferred to management’s resolution, which may itself be preferred to the status quo — but the status quo may prevail over the shareholders’ resolution.[161] The voting rules in provision (d) require shareholders to express their preferences ‘pair-wise’ in order to ensure that a resolution is only successful if it meets the Condorcet criterion — preferred to doing nothing and preferred to the alternative. It is true that such a rule makes it harder for a proposal to succeed, but the preference for the status quo, for which there was substantial support at an earlier time, is probably the lesser evil. To conclude, this approach attempts to focus more specifically on how different types of proposals impact on decision-making costs and externalities. It is premised on the notion that a single rule is less apt to minimise the costs of collective action.

C Aspects of Shareholder Voting

In relation to the voting rules used where directors are elected, CSAC proposes the adoption in the ASX Listing Rules of two principles.[162] These are an equal opportunity principle and a majority vote principle. The former requires that all candidates have an equal opportunity to be elected in situations in which the number of candidates exceed the number of vacancies.[163] The majority vote principle requires that, whether or not there are more candidates than vacancies, a person should be elected only if that person receives more votes for than against.[164]

CSAC is, in this respect, trying to prove the unprovable — the existence of a voting rule which always produces results that represent majority wishes fairly and consistently. To see this, let us consider the logic behind some of its specific comments. The report castigates as unfair the use of sequential voting systems in which elections occur for particular candidates one at a time.[165] CSAC makes the claim that this infringes their equal opportunity principle. This is, apparently, the case because those at the front of the queue could be elected before those at the end of it, notwithstanding that they may be in fact more popular than others. Hence, says CSAC, one should have simultaneous voting systems.

It is somewhat unfortunate that CSAC should make such a comment. It is undeniably true that such an outcome might occur. However, in order to base a set of normative recommendations changing the status quo on this possibility, one ought to explicate the particular decision-rules that voters use when casting their ballots, and the distribution of information across the voters. I develop an example where parties would rationally prefer sequential voting. In this example, there are five voters (or five factions of equal size), whom I number 1–5; three candidates, who are designated A–C; and one board position. Table 6 is a distribution of their preferences. Assume further that every shareholder prefers his or her least preferred candidate being elected to a result in which no director is returned.

	1st preference	2nd preference	3rd preference
Shareholder 1	A	B	C
Shareholder 2	A	B	C
Shareholder 3	B	A	C
Shareholder 4	B	A	C
Shareholder 5	C	B	A

Assume that the distribution of preferences is private information. Under a simultaneous voting system, assume that shareholders are asked to specify whether they vote for or against each candidate. The voting rules might specify that the shareholders may cast as many ‘for’ votes as they choose. Nonetheless, each shareholder may only vote for a single candidate for strategic reasons.[166] Now, under the majority vote principle and the simultaneous voting method that CSAC advocates, no director will be returned, since none musters more votes for than against. Thus, shareholders end up with their least preferred option, which is that no director is returned.

Compare a sequential voting process, in which the results for each director are declared one at a time. In this process, A, being first up, will not be returned, since a majority of the five voters (shareholders 3–5) do not prefer him or her. B, however, will be returned, since in these circumstances, shareholders 1 and 2, whose preferred candidate has been eliminated, prefer B to the other remaining candidate, C. B is in fact a Condorcet winner, since different majorities prefer him or her to both A and C. It therefore follows that there are parameters in which sequential voting may be demonstrably superior to simultaneous voting, and also to the criteria that CSAC proposes.

Of course, if we relax the assumptions relating to the distribution of information or offer parties the opportunity to log-roll or to buy votes by way of side-payments, CSAC’s method may be able to produce the correct results in some situations. Depending on the intensity of relative preferences, either shareholders 1 and 2 support B, or shareholders 3 and 4 support A. But of course, under those particular assumptions, sequential voting is no less justified. In the final analysis, all methods are capable of producing results that seem to be unjustifiable according to majority will. CSAC might as well be in the business of perfecting the measurement of phlogiston, or the production of the universal solvent, as that of finding voting rules free of cycling tendencies.

V CONCLUSIONS

It is, perhaps, a little surprising that, in reports on the role of collective action by shareholders and the means by which preferences are transformed into votes, there should be no reference at all to the literatures on social choice, the theory of collective action, or the economic analysis of constitutional choice. These literatures are, it is true, unfamiliar to lawyers compared with other bodies of theory. Some of the reasons for that are understandable. First, the literature is quite complex, even when compared with other economic theories used by legal scholars. Second, the literature sometimes lacks clear-cut normative implications given the presence of impossibility theorems and the like. Third, and despite the efforts to the contrary by scholars like Steve Bottomley,[167] corporations have rarely been seen as having sufficient analogies to states or governments to import into corporate law the analytical tools applied to them. Obviously, this bias will be most intense in conservative lawyer-economists, whose normative imperative to free corporations from prescriptive regulation is the converse of their desire to diminish and restrain governments.[168] It is necessary, however, to adapt the economic theory of voting and constitutions to corporations. This is because it is in the adoption of majoritarian principles that corporations differ so sharply from the contracts with which they are otherwise appositely analogised.

In this essay, I have argued that the scope for involvement by the shareholder in corporate governance is a complex issue — it depends on a number of factors other than the size of the general meeting and board of directors. I have argued that there are four key influences on the entitlement of shareholders to participate either individually or collectively in governance functions. These are the costs of decision-making, the significance of external effects occasioned by individual or collective action, the indeterminacies and cycling resulting from majoritarian processes, and the effects of opportunistic hold-up behaviour. I have shown how these might be expected to influence the proper shape of the doctrine, and how the law broadly conforms to these expected influences. In the analysis of the relation between the board and the general meeting, it was argued that courts have taken an unnecessarily and unusually literal construction of the provisions. This has foreclosed a more contingent view of the entitlement of the individual shareholder to propose, and the collective body of shareholders to consider, resolutions relating to overreaching or breach of fiduciary duty. This failure is associated with the misunderstanding of the rule in Foss[169] and its implications for these issues (prior to its repeal).

In reviewing recent proposals relating to requisitioning meetings, proposing resolutions and electing officers, I have been critical of the undiscriminating broad-brush treatment of these subjects. In relation to requisitioning meetings, policy-makers toss rules around without alluding to their properties relative to company size, and select amongst them according to criteria of adoption elsewhere, rather than the incidence of costs they impose. I have suggested that a more contractible law in this area might have appeal. In relation to proposing resolutions, I have argued in favour of differentiating the preconditions for the type of resolution, prohibitive restrictions on management issues, and using both rationing and ex post punishments to address the abuse of the right to propose resolutions. Finally, I have expressed agnosticism about whether the principles CSAC propose for the election of directors are in any way an improvement over the current law. I have shown how the methods CSAC criticises may be superior to those CSAC proposes.

Further research in relation to the effects on stock price of the proposal of resolutions and the requisitioning of meetings and the extent to which these are differentiated for the size, board structure, shareholding patterns, and the relevant issues in corporations, is clearly needed. Until that time, regulators and reformers would do well to understand the processes of social choice in corporations and the relation between individual entitlements and collective action, if ‘economic reform’, as a process, is to warrant either that adjective or that noun.

[*] BCom (Hons), LLB (Hons), MFM (Qld), PhD (Griffith); Associate Professor, Faculty of Law, Griffith University; Director, Business Ethics and Regulation Program, Key Centre for Ethics, Law, Justice and Governance. I wish to thank an anonymous referee for useful comments.

[2] See Companies and Securities Advisory Committee (‘CSAC’), Shareholder Participation in the Modern Listed Public Company: Final Report (2000); Joe Hockey, Minister for Financial Services and Regulation, New Rule Stops Hijack of Company Meetings, Press Release (18 December 2000).

[6] See generally Bernard Black, ‘Shareholder Passivity Reexamined’ (1990) 89 Michigan Law Review 520; Geof Stapledon, Institutional Shareholders and Corporate Governance (1996).

[7] Leslie Hannah, The Rise of the Corporate Economy (2nd ed, 1983); Brian Cheffins, Putting Britain on the Roe Map: The Emergence of the Berle-Means Corporation in the United Kingdom (2000) Social Science Research Network Electronic Library <http://papers.ssrn.com/sol3/papers.

[9] The idea of the corporation as a self-enforcing contract is developed in various works: Bernard Black and Reinier Kraakman, ‘A Self-Enforcing Model of Corporate Law’ (1996) 109 Harvard Law Review 1911; Edward Rock and Michael Wachter, ‘Islands of Conscious Power: Law, Norms and the Self-Governing Corporation’ (2001) 149 University of Pennsylvania Law Review 1619; Michael Whincop, An Economic and Jurisprudential Genealogy of Corporate Law (2001).

[10] Some elements of an analysis can be found in Jeffrey Gordon, ‘The Mandatory Structure of Corporate Law’ (1989) 89 Columbia Law Review 1549. Anglo-Australian scholarship has considered the issues associated with constitutional alteration: see Ian Ramsay (ed), Gambotto v WCP Ltd: Its Implications for Corporate Regulation (1996); Michael Whincop, ‘A Relational and Doctrinal Critique of Shareholders’ Special Contracts’ [1997] SydLawRw 18; (1997) 19 Sydney Law Review 314.

[11] James Buchanan and Gordon Tullock, The Calculus of Consent: Logical Foundations of Constitutional Democracy (1965).

[13] Mancur Olson, The Logic of Collective Action: Public Goods and the Theory of Groups (2nd ed, 1971).

[17] To take a corporate law example, any shareholder might be able to requisition a meeting.

[18] This point recognises the bargaining theory of the empty core. A bargaining game has an empty core where, for any existing set of agreements between the parties, there exists an alternative set of agreements that improves the pay-off for a party defecting from the former agreement and for a party who is excluded by it: Maxwell Stearns, Constitutional Process: A Social Choice Analysis of Supreme Court Decision Making (2000) 56.

[20] This is, of course, the basis of most of the economic analysis of corporations: see generally Frank Easterbrook and Daniel Fischel, The Economic Structure of Corporate Law (1991).

[22] See also Oliver Williamson, The Economic Institutions of Capitalism: Firms, Marketing, Relational Contracting (1985); Oliver Hart, Firms, Contracts, and Financial Structure (1995); Luigi Zingales, ‘Corporate Governance’ in Peter Newman (ed), The Palgrave Dictionary of Economics and the Law (1998) vol 1, 497. Cf Armen Alchian and Harold Demsetz, ‘Production, Information Costs and Economic Organization’ (1972) 62 American Economic Review 777.

[25] For accessible references, see Stearns, above n 18; Daniel Farber and Philip Frickey, Law and Public Choice: A Critical Introduction (1991) 38–62; Jerry Mashaw, Greed, Chaos, and Governance: Using Public Choice to Improve Public Law (1997) 12–15.

[26] One objection might be that shareholders will not have significantly variable preferences because they ascribe unanimously to value maximisation: Harry DeAngelo, ‘Competition and Unanimity’ (1981) 71 American Economic Review 18. For a discussion of the circumstances in which this premise ceases to sustain the objection, see Gordon, ‘Shareholder Initiative’, above n 8, 368–70.

[27] A Condorcet winner is an option that is preferred by majority to other options in pairwise competition: Stearns, above n 18, 45. A is a Condorcet winner compared with B and C if a majority prefers A to B, and A to C. In some choice contexts, no Condorcet winner exists. In the distribution of preferences in Table 1, there is no Condorcet winner. In these circumstances, the option of remaining with the status quo may be preferred since it was supported unanimously when preferences were originally expressed.

[28] See, eg, Benjamin Klein, Robert Crawford and Armen Alchian, ‘Vertical Integration, Appropriable Rents, and the Competitive Contracting Process’ (1978) 21 Journal of Law and Economics 297; Hart, above n 22; Williamson, above n 22.

[29] Rock and Wachter, ‘Law, Norms and the Self-Governing Corporation’, above n 9.

[30] Klein, Crawford and Alchian, above n 28; Hart, above n 22, ch 8; Williamson, above n 22.

[31] The principal difficulty is for those engaging collectively in hold-up behaviour to be able to commit to a basis for sharing gains.

[32] See generally Edward Rock and Michael Wachter, ‘Waiting for the Omelet to Set: Match-Specific Assets and Minority Oppression Close Corporations’ (1999) 24 Journal of Corporation Law 913.

[35] See, eg, David Yermack, ‘Higher Market Valuation of Companies with a Small Board of Directors’ (1996) 40 Journal of Financial Economics 185.

[36] I shall illustrate this in below Part IV(C) when analysing CSAC’s recommendations on the election of directors.

[38] See generally Henry Butler and Larry Ribstein, The Corporation and the Constitution (1995) 7–13; Andrei Shleifer and Robert Vishny, ‘A Survey of Corporate Governance’ (1997) 52 Journal of Finance 737.

[40] See, eg, Aberdeen Railway Co v Blaikie Bros (1854) 1 Macq 461, 473; [1843–60] All ER Rep 249 (Bagallay LJ); Parker v McKenna [1874] UKLawRpCh 130; (1874) LR 10 Ch App 96, 124 (Lord Cairns LC); Regal (Hastings) Ltd v Gulliver [1967] 2 AC 134, 144–5 (Lord Russell).

[44] See, eg, Provident International Corporation v International Leasing Corporation [1969] NSWR 424 (Helsham J).

[48] Commencing with the Companies Act 1929 (UK) 19 & 20 Geo 5, c 23, legislation imposed obligations on directors relying on these provisions to disclose their interests to the board. These provisions were anticipated by the Listing Rules of stock exchanges, which also required the director to abstain from voting. The earliest Australian provision of which I am aware is Sydney Stock Exchange, Official List Requirements, rule 4(q) (29 July 1912).

[49] Michael Whincop, An Empirical Investigation of the Terms of Corporate Charters and Influences on Term Standardization in a Laissez-Faire Environment (2000) Social Science Research Network Electronic Library <http://papers.ssrn.com/paper.taf?abstract_id=231048> at 6 July 2001 (copy on file with author).

[50] The courts arguably compensated for the insufficient state-contingency of such clauses by interpreting them in a literal and restrained manner: see Liquidators of the Imperial Mercantile Credit Association v Coleman [1832] EngR 517; (1873) 6 LRHL 189; Transvaal Lands Co v New Belgium (Transvaal) Land and Development Co [1914] 2 Ch 488; Re North Eastern Insurance Co Ltd [1919] 1 Ch 198; Victors Ltd (in liq) v Lingard [1927] 1 Ch 323.

[52] Cf the situation where the director engages in the transaction and then seeks ratification ex post: Miller v Miller (1995) 16 ACSR 73. In this situation, it may be that the ratification can subsequently be withdrawn by majority.

[53] Wallersteiner v Moir (No 2) [1975] 2 WLR 389 held that, where there has been a breach of duty, the shareholder is permitted to bring a suit. That outcome was probably appropriate since the breach of duty in that case involved misappropriation. However, creating this right in respect of a conflict of interest is more problematic: see generally Lord Wedderburn, ‘Control of Corporate Litigation’ (1976) 39 Modern Law Review 327.

[57] Edwards v Halliwell [1950] 2 All ER 1064, 1067 (Jenkins LJ); Wedderburn, ‘Shareholders’ Rights’, above n 46.

[58] York Corporation v Henry Leetham & Sons Ltd [1924] 1 Ch 557, 573 (Russell J).

[62] I examine whether the constitution takes this right away in below Part III.

[63] This is derived by rearranging the inequality (1 – j)C – L > C, which must hold for litigation to be better than doing nothing. The left-hand side of the inequality is the pay-off from litigating; the right-hand side is the pay-off from doing nothing. The rearranged inequality in the text should actually be –L/C, rather than L/C, but since we are interested in a situation in which C < 0, I have taken the liberty of cancelling the negative signs, for simplicity.

[64] This is easy to derive by rearranging the inequality qC > (1 – j)C – L. This is the inequality that must hold for seeking restitution to be a better option than litigating. The left-hand side of the inequality is the pay-off from seeking restitution; the right-hand side is the pay-off from litigating.

[65] Formally, this is because, as the text demonstrates, the inequality pB – (1 – p)r > qB – r always holds by hypothesis.

[66] To see this, recall that the director will be better off seeking ratification than not seeking ratification and facing litigation, where pB – (1 – p)r > (1 – j)B – r. From the analysis of the inequality in above n 65, we can see that this inequality will hold where q ≥ 1 – j. In the text above, I asserted that the shareholder will litigate, rather than seek restitution, where

1 – q – j < L/C. By rearrangement, we can see that q > 1 – j – L/C. Litigation only occurs where C < 0, so that it must be the case that 1 – j – L/C >1 – j. Because q > 1 – j – L/C for all C < 0, it must be the case that q > 1 – j. QED.

[67] Intuitively, one would expect that the cost of convening the meeting would rise with the number of shareholders, and that, if L did so, it would rise at a slower rate. If that is true, it offers some justification for invoking the interests of justice exception more often in larger companies. See text accompanying above n 60.

[69] The old fiduciary duty, in contrast, was strict and easy to interpret — but hard to litigate.

[73] Michael Wachter and Randall Wright, ‘The Economics of Internal Labor Markets’ (1990) 29 Industrial Relations 240, 243–4.

[74] See Stearns, above n 18, 45–6. In the pairing A versus B, A wins by a majority of seven to four (the majority are shareholders 1–5, 8 and 9). In the pairing A versus C, A wins by a majority of six to five (the majority are shareholders 1–5 and 7).

[76] This is self-evident — if team members have fixed shares of output, they will shirk.

[77] Each party has an incentive to incur negotiation costs in laying claim to a larger share of the output.

[89] Harold Ford, Robert Austin and Ian Ramsay, Ford’s Principles of Corporation Law (10th ed, 2001) 216–18.

[99] See, eg, Permanent Building Society (in liq) v Wheeler (1994) 11 WAR 187, 218 (Ipp J, Malcolm CJ and Seaman J concurring).

[105] John Shaw [1935] 2 KB 113; Scott v Scott [1943] 1 All ER 582; Clifton v Mount Morgan Ltd [1940] NSWStRp 5; (1940) 40 SR (NSW) 31. Cf G D Goldberg, ‘Article 80 of Table A of the Companies Act 1948’ (1970) 33 Modern Law Review 177; G R Sullivan, ‘The Relationship between the Board of Directors and the General Meeting in Limited Companies’ (1977) 93 Law Quarterly Review 569, 578. Goldberg and Sullivan argue that Marshall’s Valve was correctly decided and that the shareholders are entitled to interfere in companies. However, the basis of this argument lies in the existence of an article conferring power on the board ‘subject ... to any of these regulations, to the provisions of the Act and to such regulations ... as may be prescribed by the company in general meeting’: see Goldberg at 178 and Sullivan at 571. The latter provision, Goldberg (at 178, 183) and Sullivan (at 578) argue, preserves power in the general meeting. However, this latter reference does not appear in s 226A, nor in the model articles that preceded it: Corporations Law, sch 1, table A, reg 66.

[106] Gower recognised this point: L C B Gower, Modern Company Law (4th ed, 1979) 147–8.

[107] The appropriate course would usually be for a general meeting to be requisitioned to consider a proposal for restitution of those benefits, or for a suit alleging patent infringement to be brought. Since the managing director had the voting power to pass a resolution in this event, it was not necessary to conduct a general meeting in this particular case.

[109] There is some acknowledgement of this in the report, but the differences contemplated seem quite marginal: ibid 28–9.

[119] I use the phrase ‘numeric test’ to describe a test which determines the minimum number of shareholders required to requisition the meeting or propose a resolution by applying some function to the number of shareholders, or using a constant number of shareholders. The contrast is a test based on interests in voting capital.

[120] My analysis of the 5% rule treats it as a numeric test (ie, 5% of the number of shareholders), rather than 5% of capital or voting shares. In practice, the two will, of course, diverge. However, because larger institutional shareholders are traditionally hard to persuade to take a position against management, an analysis of the 5% rule as 5% of the number of shareholders will underestimate its demands in larger companies.

[122] See, eg, Armen Alchian and Reuben Kessel, ‘Competition, Monopoly and the Pursuit of Money’ in National Bureau Committee for Economic Research, Aspects of Labor Economics (1962) 157, 162.

[123] Huntleys’ Business Network Ltd, Huntleys’ Sharemarket Handbook: Top 350 Companies in Brief (2000) 202, 375.

[125] See generally Melvin Eisenberg, ‘Corporate Law and Corporate Norms’ (1999) 99 Columbia Law Review 1253; David Skeel Jr, ‘Shaming in Corporate Law’ (2001) 149 University of Pennsylvania Law Review 1811. Of course, it seems highly unlikely that a controlling non-institutional shareholder will ever be inclined to support minority shareholders in a resolution disfavoured by the board, since the board is bound to reflect the former shareholder’s wishes.

[127] See, eg, the findings in Geof Stapledon et al, Proxy Voting in Australia’s Largest Companies (2000).

[130] The reason for this is simply that a high proportion against the requisition makes the chance of the resolution succeeding almost impossible.

[131] In proposing this rule, I was influenced by Ian Ayres and Jack Balkin’s proposal for higher-order liability rules: Ian Ayres and J M Balkin, ‘Legal Entitlements as Auctions: Property Rules, Liability Rules, and Beyond’ (1996) 106 Yale Law Journal 703. They argue that, if A has a right to take B’s property for $x, B could conceivably have the right to take the entitlement back for $y, where y > x.

[133] I reserve my opinion on the question of whether it would be appropriate to exclude a substantial non-institutional shareholder from voting in favour of the counter-requisition. If such a shareholder could vote for the counter-requisition, no extraordinary meeting could ever be held without its consent. On the one hand, that shareholder may have incentives at odds with those of the other shareholders but in accord with those of the board. On the other, the shareholder is likely to be able to block any resolution at the requisitioned meeting, so ignoring his or her views in the first instance would be likely to occasion unnecessary deadweight costs. In the USA, shareholder proposals are usually made in relation to companies with low insider ownership: see, eg, John Bizjak and Christopher Marquette, ‘Are Shareholder Proposals All Bark and No Bite? Evidence from Shareholder Resolutions to Rescind Poison Pills’ (1998) 33 Journal of Financial and Quantitative Analysis 499.

[134] Stapledon et al, above n 127, 14 note that institutional shareholders sometimes abstain rather than vote for or against resolutions as a protest strategy. The distinction in the text may therefore be valid.

[135] If the meeting is in fact requisitioned, these costs may be incurred anyway in opposing the resolution.

[136] A focal point is a concept used in game theory. Where the nature of the strategic interaction between the players requires co-ordination (rather than involving competition), the parties may be able to achieve co-ordination by making apparently arbitrary choices which have some sort of historical, cognitive, intuitive, or psychological appeal, such as driving on the left-hand side of the road. These are called focal points. There are, however, no significant co-ordination issues in these cases.

[138] Michael Klausner, ‘Corporations, Corporate Law and Networks of Contracts’ (1995) 81 Virginia Law Review 757; Russell Korobkin, ‘The Status Quo Bias and Contract Default Rules’ (1998) 83 Cornell Law Review 608.

[139] To the extent that resolutions relating to voting or participatory or return entitlements affect the articles, they would have to be proposed as a special resolution, and be subject to procedural and supermajority requirements found in the definition of ‘special resolution’ in Corporations Act s 9.

[140] However, the US evidence on the value of these proposals is not positive: Jonathan Karpoff, Paul Malatesta and Ralph Walkling, ‘Corporate Governance and Shareholder Initiatives: Empirical Evidence’ (1996) 42 Journal of Financial Economics 365.

[143] It was influenced in this regard by the decision in New York City Employees’ Retirement System v Securities and Exchange Commission, 843 F Supp 858 (SDNY, 1994); rev’d [1995] USCA2 8; 45 F 3d 7 (2nd Cir, 1995), which upheld such an approach.

[144] Although it is to be hoped that this is implicit, I make this argument because of the risk of cycling and hold-up behaviour possible where parties are permitted to propose these resolutions. This is not because anti-discrimination policy is unimportant or undesirable. Shareholder resolutions are simply not an effective way of pursuing such aims.

[145] SEC, ‘Amendments to Rules on Shareholder Proposals’, Release No 34-40018, IC-23200, reported in CCH, Transfer Binder Federal Securities Law Reporter (at S7-25-97) 80,551.

[149] See, eg, Bernard Black, ‘Shareholder Activism and Corporate Governance in the United States’ in Peter Newman (ed), The New Palgrave Dictionary of Economics and the Law (1998) vol 3, 459; Roberta Romano, ‘Less Is More: Making Shareholder Activism a Valued Mechanism of Corporate Governance’ (International Center for Finance, Yale School of Management, 2000).

[150] Eg, Sunil Wahal, ‘Pension Fund Activism and Firm Performance’ (1996) 31 Journal of Financial and Quantitative Analysis 1; Karpoff, Malatesta and Walkling, above n 140.

[151] CSAC, above n 2, 33–5 considers the use of punishments in only a very minor respect. It does not favour them.

[154] 17 CFR s 240.14a-8(i)(12) (2000). The actual rule is more complex than I have indicated in the text. It operates with reference to proposals made in the last five years, but only permits them to be excluded if the last such proposal had been made in the preceding three years, and failed to achieve a certain level of support, which increases with the number of times proposed in the last five years — 3% if proposed once, 6% if proposed twice, and 10% if proposed three times.

[155] Schwab and Thomas, above n 147, 1047–9. The test was activated in a case where a union submitted several proposals, and then, learning of the one-proposal limitation, resubmitted them as proposals from the union and several of its members: Labor Policy Association, SEC No-Action Letter, Release No 34-93093 (20 November 1997) (‘Pacific Enterprises’).

[157] CSAC, above n 2, 13 rejects the use of a bond in the context of requisitioning meetings.

[158] See, eg, Romano, above n 149; George Dent Jr, ‘Response: Proxy Regulation in Search of a Purpose — A Reply to Professor Ryan’ (1989) 23 Georgia Law Review 815.

[159] This is analogous to American law: 17 CFR s 240.14a-8(i)(12) (2000), discussed in above n 154. The difference is that that ban is referable not to the proponent, but to the substance of the proposal.

[160] R H Coase, ‘The Problem of Social Cost’ (1960) 3 Journal of Law and Economics 1. Ie, the later proponents may cut a deal with the earlier proponents such that the earlier proponents agree to withdraw their resolution in return for a side-payment or other consideration.

[166] This may occur where there is a substantial difference between the marginal gain expected by each shareholder when electing their preferred candidate over their next preferred candidate. Under certain expectations regarding the change in probability of returning certain directors by voting in different ways, these shareholders will vote only for their preferred candidate. For example, shareholder 1 may value A’s election at $10, and B’s and C’s election at $3 and $2 respectively. Shareholder 1 estimates that the probability of no candidate being elected is 0.1, and estimates that each candidate has the same probability (0.3, being (1 – 0.1)/3) of being elected, should shareholder 1 vote only for his or her preferred candidate, A. That implies that the average number of votes for each candidate is 1.67, should each shareholder vote only for one candidate. Shareholder 1 knows that, if he or she were to cast a second ‘for’ vote for B, that would change B’s average expected vote to 2.67. In revising his or her estimate of the odds of the different candidates winning and of no candidate winning, shareholder 1 realises that, by voting for a second candidate, he or she may gain by reducing the probability of no candidate being elected. This may gain him or her up to $0.30 (being 0.1 × $3) should he or she vote for B. However, if the probability that A will be elected falls by more than 0.03, shareholder 1 will be worse off because of the reduced probability of securing the election of A, whose election he or she values at $10.

Melbourne University Law Review