Introduction

The research begins by highlighting the June 2026 initial public offering (IPO) of SpaceX, which was the largest on record at a valuation of $1.77 trillion. SpaceX utilised a dual-class share structure where Class B shares, held by Elon Musk and insiders, carried ten votes each compared to the single vote of Class A public shares. While critics of such structures argue for the “one share, one vote” principle, dual-class listings have become increasingly common. Critics warn that these structures concentrate power, making it difficult to remove chief executive officers (CEOs) and allowing controllers to pursue private benefits at the expense of outside investors. However, this paper seeks to assess what has been learned about the costs and benefits of these structures over the last 20 years.

Dual-class shares and governance: data

The authors assembled a database of United States IPOs from 2000 to 2025, matching governance information from BoardEx with financial data from Compustat and CRSP. The final “core” IPO sample consisted of 3,290 firms. The research used artificial intelligence to assist in collecting difficult-to-find information on founders. The data includes various firm classifications, such as operating companies, special-purpose acquisition companies (SPACs), and real estate investment trusts (REITs). Variables were constructed to measure board size, independence, gender diversity, and founder control, including whether the CEO at the time of the IPO was a founder.

Dual-class structures: what are the facts?

The study documents several “stylised facts”. Fact 1 states that the share of dual-class IPOs increased gradually from approximately 10% in 2000 to 35% in 2025. Fact 2 notes that dual-class firms are generally larger by market capitalisation than their single-class counterparts. While tech and non-tech firms had similar propensities for dual-class shares before 2014, Fact 4 highlights that from 2015 to 2025, tech firms became much more likely to choose this structure. By the end of the sample period, most tech IPOs were dual-class offerings.

The research also addresses venture capital (VC) backing. Contrary to the idea that VCs act as a counterforce to dual-class structures, most of the recent increase in dual-class IPOs occurred in VC-backed firms. Fact 10 confirms that dual-class firms are more likely to retain founders as CEOs or board members. However, Fact 13 finds that boards of single-class and dual-class firms are remarkably similar in size, independence, and diversity. Crucially, Fact 14 finds no clear evidence that CEOs of dual-class firms stay in office longer than those of single-class firms, suggesting that the entrenchment narrative may be overstated.

Unbundling ownership and control: theory

The report reviews recent theoretical literature that challenges the traditional view that deviations from “one share, one vote” are inherently inefficient. Under conditions of asymmetric information, where an incumbent has private knowledge of a firm’s prospects, unbundling control from cash flow rights can actually facilitate efficient transfers of control. By reducing the number of shares that must change hands to shift control, dual-class structures may limit a controller’s ability to extract informational rents. This theory aligns with the empirical observation that dual-class structures are most prevalent in technology firms, where uncertainty and information asymmetries are highest.

Dual-class shares and governance: empirical literature

The empirical review updates the authors’ 2008 survey, noting that while early studies found negative associations between dual-class status and firm value, recent work suggests a more nuanced life-cycle effect. Some research indicates that dual-class firms may trade at a premium during their early years, particularly when founder-led, with this premium declining as the firm matures. Other studies suggest that dual-class firms may provide higher-quality financial reporting and be attractive targets for shareholder activists, indicating that the separation of ownership and control need not weaken accountability.

Directions for further research and conclusions

The authors conclude that the conventional entrenchment explanation for dual-class shares lacks robust empirical support. They suggest that founders may prefer these structures to implement a specific long-term “vision”. Future research should investigate how dual-class structures relate to a CEO’s ability to implement vision, the impact on long-term investments in intangibles, and the overall social optimality of ceding power in an era where shareholder democracy is under threat. The paper calls for new governance models that incorporate concepts of managerial vision into the analysis of firm performance and structure.