A large observational analysis of US public firms found that switching from a single-class to a dual-class share structure was associated with an immediate rise in company valuation and higher forward patent-citation output. Firms making the reverse switch, known as stock unification, showed lower valuation and declining patent-citation output relative to firms that kept dual-class shares. The findings describe associations, not proof that either governance change caused the outcomes.
The main comparison used stacked dynamic difference-in-differences event studies, which follow outcomes around a governance switch and compare switchers with firms that retained their original share structure. After industry, asset-size, data-availability and outlier exclusions, the panel contained 160,702 firm-years from 1971 through 2022: 9,950 dual-class firm-years across 1,151 firms and 150,752 single-class firm-years.
Valuation and innovation moved in different ways
For firms that recapitalized into a dual-class structure, Tobin’s q—a valuation measure used in the analysis—rose by 0.21 in the event year relative to firms that remained single-class. The estimate stayed elevated through year five. From year five to year 11, the difference fell by 0.366; the reported p-values were 0.018 for the event-year estimate and 0.004 for the later decline.
The reverse change showed a different valuation comparison. After stock unification, Tobin’s q was lower than among firms that continued with dual-class shares, although the gap moved closer to zero after year five. Those estimates were less precise because the continuing-dual-class control group was smaller.
Patent citations followed a separate event-time pattern. Forward patent-citation output increased after recapitalization compared with firms staying single-class and declined after unification compared with firms staying dual-class; the estimated magnitudes grew over event time. Most of the recapitalization increase came from the extensive margin, meaning firms beginning to generate cited patents. After unification, both extensive- and intensive-margin estimates were negative, although the intensive-margin estimate was statistically insignificant.
Market signals changed as firms aged
Announcement-return reactions differed by maturity. For a young-firm recapitalization, the estimated baseline cumulative abnormal return was 0.029. The age-over-12 coefficient was −0.035 for recapitalizations and 0.052 for unifications. The authors interpret these estimates as showing more favorable reactions to recapitalizations at younger firms and to unifications at mature firms, while noting that the comparison depends on whether announcing companies were otherwise comparable and on what investors already expected.
Valuation changes were concentrated among firms in industries with more specific capital and labor inputs. The unification-related patent decline was concentrated in the high-specificity group, while recapitalization’s patent pattern did not show the same concentration. Specificity was an input-based proxy rather than a direct measure of founder investment.
Investment behavior offered another maturity signal. In the low-sales-growth subsample, mature dual-class firms had significantly lower investment–q sensitivity than mature single-class firms, while young firms showed no such difference. Investment–q sensitivity captures how closely investment moves with Tobin’s q; the analysis treated it as an indirect proxy for private benefits. The mature comparison had a p-value of 0.011, and the mature-versus-young difference had a p-value of 0.014.
Voting premia—the price difference between superior- and inferior-voting shares—were also higher among older firms. Relative to the youngest age quartile, the third quartile had a 4.8-percentage-point higher premium, with a p-value of 0.054, while the fourth quartile had a 9.6-point higher premium, with a p-value of 0.025. The mean premium was 5.1%. Voting premia can also reflect control contests, liquidity and differences in investor beliefs.
The evidence leaves room for competing explanations
The comparison involving sunset provisions was more qualified. At unification, firms with sunset provisions showed no discernible valuation change, while the negative valuation pattern appeared among firms without sunsets. This was an observational subgroup comparison and does not establish that sunset provisions caused the difference.
Taken together, the authors interpret the cross-outcome evidence as more consistent with a treatment-effect reading of dual-class governance than with selection. Their proposed explanation is that dual-class structures may strengthen incentives for firm-specific investment and innovation early in a firm’s life, while private benefits and agency costs associated with continued control become more important as firms mature and erode valuation gains. The evidence bears on policy choices about governance structures and sunset provisions, but does not settle how that balance should be struck.
The study reported no evidence of differential pre-trends before either switch, supporting the parallel-trends assumption used in the event studies. That finding supports the comparison but does not by itself establish that a switch caused the outcome. Patent citations are a proxy for innovative output and founder effort, investment–q sensitivity is an indirect private-benefit measure, and voting premia may reflect other market and control factors.
Governance choices and the timing of switches were not randomly assigned, so the difference-in-differences interpretation depends on parallel trends and on the comparability of the control groups. The unification analysis was also less precise because relatively few firms remained dual-class as controls.
The document is an arXiv version 1 preprint dated 26 August 2026; the supplied metadata reports no journal publication or peer-review status. No funding source is reported in the supplied text, whose front matter says the authors’ views are not those of the Federal Reserve System and that all errors are their own. The authors say the dual-class datasets will be made available on their websites.
Paper data and sources
Original title: The Dynamic Trade-Off of Dual-Class Shares
Authors: Hyunseob Kim, Doron Levit, Roni Michaely
Journal/Repository: arXiv
Status: Preprint, not yet peer-reviewed
First online: 2026-08-26
DOI: Not available
Original paper · Full text