Dual class capital structures have spread exponentially in recent years across much of the corporate world, as has previously been reported on this blog. Dual class listed companies today account for around $4 trillion of US total stock market value including 9 percent of the S&P 100, and the dual class stock (“DCS”) phenomenon has become one of the most significant issues in global corporate governance today.
A dual class capital structure is one that gives certain inside investors voting rights disproportionate to their economic interest in a company. This enables them to prevail over less privileged outside investors on any important shareholder vote, including the important question of who sits on the company’s board of directors. The standard dual class capital structure is for multiple-vote shares to carry 10 votes each, with other shares carrying only one vote each. However, in the 2019 IPOs of the ride-hailing firm Lyft and social media giant Pinterest (and, somewhat ominously, in the initial filings for WeWork’s recently aborted IPO attempt), the super-voting shares’ votes outweighed those attached to common stock by a 20:1 ratio. A small group of U.S.-listed companies including Alphabet, Snap, and Under Armour have recently gone even further and adopted triple class stock structures, which include a third class of non-voting stock.
Dual (and triple) class stock structures are controversial because they enable an investor holding multiple-vote shares to enjoy effective voting control over the company and its board, while holding only a relatively small economic interest in the future success of the company’s business. The general policy debate on DCS hangs on the conflict between two equally important considerations. On the one hand, there is the need to protect the freedom of founders and other trusted corporate leaders to implement their long-term entrepreneurial vision, unimpeded by the destabilizing short-term profit demands of hedge funds and other aggressive investors. On the other hand, there is the need to protect outside (i.e. non-controlling) minority investors from having their wealth destroyed by unaccountable and irremovable corporate controllers.
The general tolerance shown towards DCS by U.S. regulators puts the United States in stark contrast to numerous other jurisdictions including Australia, Belgium, Brazil, Germany, Italy, Spain and the United Kingdom. A common feature of all the above countries’ corporate governance frameworks is their adoption of a much more protectionist, if not altogether prohibitive, stance concerning the regulatory treatment of DCS in listed companies. For instance, in the United Kingdom, the London Stock Exchange’s enhanced listing regime effectively prohibits DCS and other discriminatory voting structures within premium-listed companies, a category that covers many of the UK’s biggest and most influential corporate names.
Interestingly, hostility towards DCS has recently intensified in the United States, especially within the country’s institutional investment community. For instance, the Investor Stewardship Group (ISG), which represents over 60 major institutional investors with combined assets of over $31 trillion in market value, has effectively advocated near-universal adherence by U.S.-listed companies to a one share/one vote norm in its influential Corporate Governance Principles.
Moreover, since 2017, S&P Dow Jones Indices has systematically excluded any new companies conducting IPOs with multiple-voting stock from its influential S&P 500 Index, which forms the benchmark for major index-linked funds’ multi-billion-dollar investment portfolios. In a similar vein, the influential proxy advisory firm Institutional Shareholder Services (ISS) has a general policy of recommending that its investor clients vote against re-electing the incumbent directors of any company that offers differential voting stock as part of an initial or midstream public offering.
By contrast, certain jurisdictions that have traditionally been averse to permitting DCS have recently come to recognize the potential benefits of taking a more permissive stance. One of the most intriguing examples is Singapore. In June 2018, the Singapore Exchange (SGX) took the landmark step of liberalizing its formerly preclusive listing requirements with respect to discriminatory shareholder voting structures. It consequently now permits issuers to deviate from the one share/one vote norm subject to specific regulatory restrictions.
In a similar vein, Hong Kong’s recently reformed (post-2017) listings regime has made limited allowance for DCS in the case of certain “innovative companies,” which essentially denotes firms involved in the production of new technologies or other innovations where significant R&D outlays are entailed. Indeed, the Chinese e-commerce giant Alibaba’s recent $13 billion secondary listing on the Hong Kong Stock Exchange represents a significant endorsement of the country’s newly liberalized listings framework in otherwise turbulent socio-political times.
In recent years academic debate on DCS – especially in the United States – has begun to pivot less on the general merits of permitting as opposed to prohibiting dual class capital structures, and more on the specific middle-ground issues of, first, whether multiple-vote stock should be perpetual or terminate (or “sunset”) at some point and, second, the most appropriate means of determining when and how time-limited multiple-vote stock should sunset. While debate on these more granular questions is no less fervent, the continuing permissibility of dual class capital structures in at least some conditional format is now becoming an increasingly common premise of such discussions.
On the vexing question of optimal sunset design there would appear to be three broad schools of thought. The first, and increasingly most common, view is that sunset provisions should be time-conditioned, with the effect that any multiple-vote stock automatically sunsets after a pre-determined time period (e.g. seven, 10 or 15 years) unless affirmatively resolved otherwise by a majority of independent (i.e. non-multiple-vote) shareholders (the time-centered sunset model). The second common view is that sunset provisions should ideally be ownership-based, such that multiple-vote stock automatically converts (subject to offsetting independent shareholder resolution) to common stock upon the holder’s proportionate holding of cash flow rights dropping below a pre-determined minimum threshold (e.g. 10 percent, 15percent, or 20 percent) (the ownership-centered sunset model). And the third common view – as manifested notably in the case of the above Asian regulatory reforms – is that sunset provisions should be transfer-based, such that any multiple-vote stock automatically terminates (again subject to any countervailing independent shareholder vote) upon the death, incapacitation, or retirement (from the company’s board) of its original holder, or if otherwise transferred to someone other than that specified person (the transfer-centered sunset model).
In a recent paper Designing Dual Class Sunsets: The Case for a Transfer-Centered Approach, I argue in support of the third model both for private ordering at the individual firm l and as a blueprint for future SEC regulatory policy. The first key comparative advantage of the transfer-centered model of sunset design is its lesser degree of arbitrariness in comparison with the other two main models, and especially the time-centered one. Indeed, it is in this regard that the time-centered model’s key practical strength – its relative simplicity and uniformity – also becomes its main weakness by imposing a bright-line objective approach to what is, more often than not, a matter of nuanced and subjective business judgment.
Since a transfer-based sunset, by contrast, will only be triggered by the death or retirement (as a director) of the multiple-vote shareholder(s), or the sale of her equity stake, there is consequently no risk of a transfer-based trigger being activated during the period of an incumbent controller’s premiership. This insures against the risk of disturbance to the firm’s business trajectory. Moreover, since a transfer of corporate control (whether by succession or sale or purchase) inevitably entails a sudden change of trajectory for the firm’s business (whether strategically or at least culturally) in any event, it follows that conversion of the firm’s capital structure at this point will not in itself be a likely cause of organizational destabilization.
The transfer-centered model is not only less arbitrary in its application than the competitor models but also less inclined to elicit moral hazard and other perverse incentives on the part of corporate controllers. This is because a time-based sunset – whether prompted by regulation or private ordering – arguably creates an artificial incentive for an incumbent controller to dispose of her multiple-vote shareholding at some point within the specified pre-sunset period. By doing so, the controller will be able to recoup as much of her remaining control premium as possible before it dissipates on the triggering of the applicable sunset deadline. Meanwhile, the transfer-centered model has the additional advantage of remaining sensitive to the powerful non-financial incentives that incumbent controllers typically have to safeguard and promote firm value, even where their corporate control rights significantly outweigh their corresponding cash flow rights.
For the above reasons (and as explained more thoroughly in my recent paper) the SEC and main domestic market authorities should continue to resist calls from influential investor-related bodies for the introduction of mandatory time-based sunsets. For constructive inspiration, U.S. policymakers should instead look to Singapore and Hong Kong, where the transfer-based sunset model has garnered greater regulatory support.
This post comes to us from Professor Marc Moore at University College London, Faculty of Laws. It is based on his recent paper, “Designing Dual Class Sunsets: The Case for a Transfer-Centered Approach,” available here.