Vanguard's proxy-voting choice reaches 500,000 participants as small plans lag on auto-enrollment

Half a million retirement investors have now elected to direct how the shares in their accounts get voted, through a proxy-voting choice program that Vanguard says has passed that participant mark and expanded to nearly 80 plan sponsors holding $120 billion in assets and more than one million participants, according to the firm's 2026 voting report. Put those two figures side by side—the program reaches a population north of a million people, and fewer than half of them have taken it—and you see a feature that is still being installed, not one that has become routine.

The same report shows automatic enrollment at 26% of small plans against 61% of large ones, making large sponsors more than twice as likely to run the design. Where small plans do enroll workers automatically, participation reaches 79%, still short of the 94% large plans record. Defaults do most of the work of getting people into a plan, but not all of it, and the plans that decline to run them are the ones where every enrollee has to be recruited by hand.

Defaults and participant choices do not spread the same way: a default is a switch the sponsor flips at the plan level—one decision, made once, reshaping the experience for every participant thereafter—while proxy voting choice is a switch each participant flips individually, so its reach is bounded by how many people bother to flip it. The first scales through sponsors; the second scales through participants, and only after those participants are already in the plan. That asymmetry is what makes the two Vanguard data points sit so unevenly beside each other: one tracks sponsors who took an action, the other tracks individuals who opted in.

$120 billion spread across nearly 80 sponsors averages out to roughly $1.5 billion per plan, and that is not the profile of a sponsor sitting at the 26% mark. Vanguard's voting report does not break the proxy program out by plan size, so the concentration is inference rather than disclosure, but the average is difficult to read any other way. The participants who have been handed a voice in corporate proxy voting are, by and large, participants in large plans, while the participants still waiting on the more basic intervention—a nudge into the plan at all—are in the smaller ones.

The line between a small plan and a large one is not spelled out in the reporting, and the threshold matters because plan size decides how much per-participant infrastructure a sponsor can absorb. A feature that lets individuals direct the voting of shares held in their accounts has to be built once and then administered continuously, with the cost spread over however many participants the plan carries; at tens of thousands of participants, that cost is a rounding error, and at a few hundred it can be the difference between offering the feature and not offering it. Vanguard's figures do not state that mechanism, but they are consistent with it.