A pension system can outsource almost anything. It can hire managers, pay for expensive advice, and delegate the administrative work of voting the shares it owns. But it cannot delegate its responsibility to know how those shares were voted.
In 2024, Missouri State Auditor Scott Fitzpatrick put that responsibility to the test. His report should be required reading for every state treasurer, pension trustee and investment committee in the country.
His office examined Missouri’s eight largest public retirement systems, which together hold more than $85.2 billion. More than $15.2 billion of that sits in managed accounts where outside investment managers hold the shares and cast the votes. That is where a great deal of public money lives, and almost nobody was checking what the votes added up to.
The analysis found that only two of the eight systems, MOSERS and PSRS/PEERS, had written guidance on how certain issues should be voted. The other six leaned on general language about acting in the best interest of participants. Only one system required a proxy voting advisor and specified what that advisor should weigh. Not one of the eight required anybody to review the votes cast on its behalf, and five of the eight did not regularly review the proxy reports they were already receiving.
General language feels responsible. It also tells a manager nothing about a contested proposal, and it gives a trustee nothing to check once the vote is gone.
But that wasn’t the worst of it. The audit examined seven ESG-related proxy votes. In five of them, managers for a single retirement system voted both for and against the same proposal. The same pension fund, on the same question, on the same ballot, canceled itself out.
Fitzpatrick said it plainly:
“When systems allow their asset managers to vote on both sides of an issue or prioritize advancing social and political causes above maximizing shareholder returns, they are squandering their members’ votes, and that is unacceptable. Going forward these systems need to improve their policies and do a better job of monitoring the proxy votes that are made to make absolutely certain they prioritize shareholder return.”
Fitzpatrick deserves real credit here. He had served on the MOSERS board as state treasurer during the audit period, so he recused himself from every portion of the review related to MOSERS, even though it cost him his strongest example. That is what an audit looks like when it is built to survive scrutiny.
He was just as careful about the limits of his own report. He never pretended the review was exhaustive. Seven votes sampled from a state with $85 billion invested is a smoke test.
If seven sampled votes turn up five contradictions, then the next step is to examine the full record, including every vote, from every asset manager, in every system, across a full proxy season. Missouri used a selected set of proposals by design, but analyzing all ballots, all managers, and all systems would mean split voting and policy divergence get measured completely.
A comprehensive audit can also look back across four or five years of proxy seasons, or longer. A single season gives you a snapshot, but multi-year coverage shows whether split voting, policy divergence and inconsistent delegation are recurring patterns or isolated anomalies.
That isn’t a criticism of Fitzpatrick’s report, but rather a testament to his work. His report sets in motion what should be a nationwide effort to uncover similar activity and root it out on site.
Fiduciaries across the country elsewhere should follow Fitzpatrick’s example and go further. Verify whether your votes match your policy. Then measure it over time and make sure your voting authority is doing what your fiduciary standard says it should.
The Missouri report got it right because it began with evidence. It did not assume that a policy statement guaranteed alignment, that an external manager’s reputation eliminated the need for review, or that delegation marked the end of fiduciary responsibility. It asked whether the institution could demonstrate control, and that is the standard every fiduciary should apply. You can delegate voting authority, but you cannot delegate accountability.
The Missouri audit only looked at a sample of seven votes from eight funds. With that very limited scope, the audit found five split votes between the funds. Just think what a comprehensive proxy voting audit would reveal.
