Say-on-Pay 2026: Averages Are Safe. Your Meeting Is Not an Average.

Aon's half-year review of 2026 say-on-pay outcomes, published on the Harvard Law School Forum on Corporate Governance, reads at first as reassurance: roughly 80% of Russell 3000 companies received 90%+ support through early June, up from 75% in the first half of 2025, with fewer low-support outcomes and failed votes than prior years.

The headline hides the mechanism. As the authors caution, strong headline results can mask increasing complexity. Opposition has not softened — it has concentrated. It now lands on specific, identifiable triggers: large one-time or "mega" equity awards and sign-on grants where the rationale or structure is not clearly articulated; perceived pay-performance misalignment; limited or unconvincing responsiveness to prior say-on-pay results; and non-standard structures lacking clear disclosure. Investors are separately scrutinising discretionary metric adjustments and executive security costs.

The proxy-advisor signal predicts less every season. Aon notes that ISS and Glass Lewis have extended their pay-for-performance evaluation windows toward five years — with, so far, "relatively muted" immediate impact. The structural story around that observation is larger: major managers are internalising voting entirely (JPMorgan's Proxy IQ; Wells Fargo's in-house build), Glass Lewis is retiring its single benchmark policy by 2027, and ISS is restructuring toward individualised voting research. A proxy-advisor recommendation no longer cascades predictably across a register. The same remuneration resolution now routinely draws a FOR from a passive index manager, an AGAINST from a European pension, and a conditional position from a stewardship-led super fund — each for reasons set out in that investor's own policy and, increasingly, its own published rationale. Anticipating "the ISS position" answers less and less of the question that matters: how will YOUR register vote?

Every trigger on Aon's list is screenable in advance — holder by holder. A one-off award, a discretionary uplift, an extended severance term, a thin disclosure: these are not abstractions, they are data points a board can test against each investor on its register before the notice of meeting is printed. That is the capability VINCIT QSV was built around, in three layers:

Layer one — the policy screen. Each register investor's written voting policy, held as verbatim extracts, cross-referenced against the company's actual framework features. If a holder's policy states it votes against uncapped sign-on awards, and the framework contains one, that flag exists months before the AGM.

Layer two — observed behaviour. Policies say what investors claim; votes show what they do. The platform holds over 2.1 million observed votes across six proxy seasons in Australia, the UK and Europe — so each policy flag carries the holder's real against-rate on that resolution type, with sample size and confidence attached. It also captures behavioural sequences: several major managers escalate remuneration dissent into director-election votes the following season at multiples of their baseline rate. A strike is rarely a one-year event.

Layer three — revealed reasoning. Where investors publish per-vote rationales, the platform attaches their own words, verbatim, to the vote. This is what makes Aon's most actionable trigger — "responsiveness to prior say-on-pay results" — genuinely manageable: responsiveness can only be demonstrated against what each holder actually said it wanted. When an investor has told the market, in writing, why it voted against your remuneration report, the response that matters is the one addressed to that stated concern.

The conclusion for boards is the one the averages obscure. Whether a company sails through or takes a strike is not determined by the Russell 3000 distribution — it is determined by which specific investors hold its register, what their policies say about the exact features in its framework, how those investors have actually voted on those features, and what they have said about why. All of that is knowable in advance, from the evidence rather than the averages — and the boards that treat it that way are the ones for whom the AGM stops being a surprise.

Source data: A. Patterson, R. Harrison & R. Kalb (Aon plc), "2026 Say-on-Pay Trends," Harvard Law School Forum on Corporate Governance, 16 July 2026.

Want to see how VINCIT QSV addresses these challenges?

Request a demo to see register-weighted governance intelligence in action.