CEO Turnover at Dual-Class Firms

Dual-class structures remain a hotly contested topic in corporate governance.  Over the past two decades, the use of dual-class structures has expanded significantly, particularly among venture capital (VC) backed tech companies. Critics contend that high-vote shares increase agency costs and entrench underperforming executives. Proponents argue that dual-class structures shield founders from short-term market pressures, preserving their ability to pursue innovation and long-term growth. As a form of compromise, the current policy debate largely focuses on sunset provisions that would convert high-vote shares to single-vote shares after a set period or upon certain triggers. But regardless of their stance on dual-class structures’ merits, most participants in the debate have assumed that dual-class structures entrench founder-CEOs.

In our working paper, “CEO Turnover at Dual-Class Firms,” we analyze CEO turnover at U.S. VC-backed firms that went public between 2002 and 2020. Our findings reveal that CEOs at dual-class firms generally remain in their roles longer post-IPO compared to those at a matched set of single-class firms. However, once we account for the higher rate of M&A sales of single-class firms, this gap in CEO tenure largely disappears. More specifically, despite dual-class CEOs holding, on average, three times more voting power post-IPO than single-class CEOs, their risk of being replaced internally – where the board replaces the CEO or the CEO resigns – is not significantly different. Further, contrary to expectations, we find no evidence that a larger gap between voting rights and economic ownership reduces turnover risk. Both single- and dual-class firms are more likely to replace their CEO, whether voluntarily or involuntarily, following poor shareholder returns, and this performance sensitivity persists regardless of the CEO’s voting power—suggesting that market accountability mechanisms remain active even in dual-class structures.

Research Design and Findings

Our sample consists of 79 U.S. VC-backed dual-class IPOs from 2002 to 2020, in which CEOs hold high-vote shares and public investors hold low-vote shares. We employ 1-to-1 propensity score matching to identify a set of single-class IPOs that are similar on observable measures. The dataset encompasses detailed CEO attributes—including age, being a founder, voting power, equity ownership, and the wedge between voting and cash flow rights—as well as firm-level characteristics such as time-based sunset provisions, accounting data, and stock returns. We categorize CEO turnover events into external (M&A, bankruptcy), internal (performance-related or voluntary departures), and other causes (illness, death). To provide context for these transitions, we examine press releases and news coverage, while also tracking post-turnover leadership roles retained by departing CEOs within their firms. Given that CEOs in our sample have different lengths of observable history and many remain in office at the end of our study period, we employ time-to-event analysis (survival analysis), a statistical method commonly used in epidemiology to handle censored observations.

At first glance, the results appear to support the entrenchment narrative: the median CEO tenure at dual-class firms is 6.6 years after an IPO, compared to 4.3 years at single-class firms. However, this difference largely reflects the significantly higher rates of M&A sales of single-class firms. When we employ competing risks analysis to separate merger-related turnover from internal replacement, we find no significant difference in internal replacement rates between dual-class and single-class firms. This finding challenges the conventional wisdom that dual-class structures protect CEOs from internal governance pressure to step down.

Contrary to the common assumption that CEOs with greater voting power are less likely to be replaced, our analysis reveals a more nuanced pattern. At dual-class firms, CEO turnover probability shows no systematic relationship with voting power levels, despite CEOs often holding majority voting control. In contrast, at single-class firms, CEOs with lower voting power face higher turnover risk. Specifically, across the full sample, CEOs holding less than 5% of votes have an annual turnover probability of 21.4%, compared to 9.2% for those with greater voting power. Once a CEO holds more than 10% of votes, additional voting power—even reaching levels typically associated with hard control—appears to have minimal impact on turnover probability.

Moreover, we find no clear relationship between turnover risk and the size of the wedge between voting rights and cash flow rights in dual-class firms. The annual turnover probability for CEOs with wedges below 5% (9.4%) is similar to that of CEOs with larger wedges (11.0%). Even wedges exceeding 20% are not associated with lower turnover probabilities, suggesting the relationship between voting-ownership divergence and CEO retention is more complex than commonly assumed.

Finally, sunset provisions—which convert dual-class shares to single-class shares after a fixed period—appear less consequential than commonly believed. CEO turnovers in our sample of dual-class firms commonly occur well before any sunset triggers, turnover does not spike in anticipation of a pending sunset date, and median CEO tenure remains similar whether or not the firm adopts a time-based sunset. We note, however, that these findings should be interpreted with caution as sunset provisions have only recently become widespread and relatively few have triggered to date. More time may be needed to fully assess the impact of sunset provisions on CEO turnover.

If CEO turnover is not primarily driven by voting rights, voting-ownership wedges, or sunset provisions, what factors do correlate with CEO departures? We find that economic performance is associated with CEO turnover through multiple channels. Both dual-class and single-class CEOs face higher turnover rates when shareholder returns decline, with CEOs in the bottom performance quartile experiencing roughly twice the replacement rate of those in the top quartile. Additionally, changes in financial constraints, measured by year-over-year increases in short-term liabilities relative to liquid assets, correlate with higher turnover rates in both dual-class and single-class firms. This latter finding aligns with research suggesting that financial constraints can serve as an external disciplining mechanism on management.

Nonetheless, our analysis requires several important qualifications. First, CEO turnover alone does not fully capture founder control. Our data show that founders with substantial voting rights often maintain influence after stepping down as CEO, frequently through board chairmanship or other leadership positions. Specifically, while former CEOs in dual-class and single-class firms are equally likely to retain leadership positions initially, we find that over time, dual-class CEOs are more likely to maintain officer roles and less likely to relinquish all leadership responsibilities. We also observe that ‘boomerang CEOs’—where founders return for a second term—while infrequent, occur more often in dual-class firms. Finally, we emphasize that our study of CEO turnover patterns does not address whether shareholders benefit from dual-class structures overall. Even if dual-class structures do not entrench CEOs, they may create other governance challenges, such as hindering an effective market for corporate control, limiting shareholder activism, enabling CEOs to negotiate lucrative golden parachutes, or otherwise complicating leadership transitions.