The Revolving Door Runs Backward: Institutional Timing and the Manager We Let Go Too Soon

Every allocator can recite the warning: past performance is not indicative of future results. Fewer can explain why their own institution's hiring record suggests they believe the opposite.

The evidence has accumulated quietly over two decades. Studies of public plan disclosures, consultant databases, and foundation grant records point to the same uncomfortable pattern — institutions tend to terminate managers after underperformance has peaked and hire replacements after outperformance has peaked. The decisions are not irrational in isolation. Each one can be defended with data, diligence, and a unanimous vote. Taken together, they trace a curve that reliably lags the manager cycle by two to four years, which in practice means the institution buys high and sells low while feeling, at every step, rigorous.

The Three-Year Window Problem

Most governing documents specify an evaluation window — commonly three years of underperformance triggers a formal review, and review frequently becomes termination. The window feels objective. It is anything but.

Three years is long enough for a style to fall decisively out of favor and short enough that mean reversion has not yet shown up in the numbers. A value manager fired in the depths of a growth regime, a small-cap specialist dismissed at the trough of a rate cycle: in both cases the committee's spreadsheet was accurate and its conclusion was wrong. The window does not measure skill. It measures the interaction of skill with a market environment the manager did not choose.

The deeper problem is that the evaluation window converts diligence into a momentum amplifier. By the time a three-year record is bad enough to trigger action, the entire institutional ecosystem — consultants, peer plans, board members who read the same trade press — is looking at the same numbers. Termination at that moment is not a contrarian act. It is a stampede with paperwork.

What the Allocation Data Actually Show

Look at the flow patterns in public pension reallocations following the 2008 crisis, or the flight from active equity in the mid-2010s, and the lag is visible in almost every cohort. Institutions that exited active domestic equity in 2014–2016 did so on the strength of a decade of underperformance data — much of it generated before the very fee compression and factor environment that would subsequently reward patient, concentrated stock-picking. Those that hired into passive did so at a point of maximum consensus.

None of this requires bad faith. It requires only that committees use the same inputs, weight recency heavily, and face asymmetric career risk: nobody loses a seat for firing a manager with a bad three-year record, regardless of what happens next. The asymmetry does the work that conviction used to do.

Counter-Cyclical Governance

If the problem is structural, the remedy must be structural too. A few mechanisms have shown promise in institutions willing to codify them:

Pre-committed decision frameworks. Before a manager is hired, the committee writes down — in advance — what would constitute a fireable offense versus a tolerable drawdown. Style drift, team departure, leverage, and compliance failures belong in the first bucket. Underperformance within a defined style mandate belongs, in most cases, in the second. When the termination criteria are set at hire, the committee cannot quietly rewrite them mid-drawdown.

A standing devil's advocate role. One committee member is assigned, on rotation, to argue the case for retention when the room leans toward exit — and for exit when the room leans toward entry. The role matters less for its conclusions than for forcing the dissent into the minutes before the vote, not after.

Decoupling review from termination. Committees should distinguish between a review, which is cheap and frequent, and an action, which is expensive and irreversible. Institutions that trigger reviews on schedule but require a supermajority plus a written counter-case for actual termination report fewer reflexive exits.

Contrarian capacity discipline. The hardest rule to follow: some of the best hires an institution will ever make will look worst at the moment of hiring. Committees that reserve a small allocation sleeve for exactly these decisions — funded deliberately, sized modestly — give themselves permission to act against the stampede without betting the endowment.

The Patience Premium

None of this argues for keeping bad managers. It argues for knowing, with evidence rather than instinct, which managers are bad — and recognizing that a three-year record cannot tell you. The institutions that compound best over a generation are rarely the ones with the fewest mistakes. They are the ones whose mistakes were made on their own terms, on a timeline they chose, rather than on a timeline imposed by the crowd.

The revolving door will keep turning. The question is whether the institution walks through it deliberately, or is pushed.