The mix explains the arithmetic. The two highest-volume governance topics were independent board chair proposals, which nearly tripled to 56 and averaged 25.9% support without a single approval, and special meeting threshold proposals, of which 28 went to a vote at roughly 38% average support with one pass [10][11]. Together that is 84 of the 164 governance proposals voted on, 51.2% of the category, and between them they produced one approval, a rate of 1.2% [18][19]. That is how average governance support climbed while the category's pass rate moved the other way, and it is the DragonGC authors' own reading of their data: the slate loaded up on topics that reliably draw a quarter to a third of the vote and reliably lose [1][13].
The set that did pass is short enough to itemise. All four board declassification proposals were approved, as were both blank-check preferred stock proposals, three of eight simple majority vote proposals, two written consent proposals, one special meeting threshold proposal and one on director removal with or without cause [12]. Those six lines sum to thirteen, which is every approval in the season [20]. Each concerns the machinery of corporate control rather than conduct or disclosure, and outside that list the memo's authors describe a 14a-8 proposal as close to a dead letter [14].
The objection from an investor relations desk is that 25.9% for separating the chair is not nothing, and that a minority of that size is how a campaign starts [10]. Fair, but this record does not test it. What three seasons show is a band rather than a trajectory: chair proposals tripled in volume without an approval while the governance pass rate fell from 28.0% to 18.6% to 7.9% [9][10]. Whether a quarter of the vote predicts anything about next year's slate is not a question these numbers answer, and it is better to say so than to promote a correlation to an early warning.
One channel in the same body of work does show engagement and outcome moving together. The authors' 2024, 2025 and 2026 reports on adverse say-on-pay votes found that companies receiving them responded with robust engagement programmes, and that nearly all recorded say-on-pay support above 80% the following season [15]. That is an association they observed rather than a controlled test, and it sits alongside a season in which executive compensation proposals filed under 14a-8 were part of the 170 that failed outright [4][21]. For a board weighing where the marginal engagement hour goes, the pay conversation already has a ballot line that every holder votes.
What the memo does not explain is why 293 fewer proposals reached a vote than two seasons ago [2][22]. Thinning investor appetite and a narrowing supply of admissible proposals draw the same line on a chart, and this data cannot separate them. So the reallocation available this quarter is a one-season judgement: defensible on a 3.9% approval rate and a thirteen-item pass list [3][20], and revisable on a single season of contrary evidence.
