Performance versus trust: what really wins the allocation

# Performance versus trust: what really wins the allocation

Bill Gross left PIMCO on 26 September 2014. Not a single bond in PIMCO Total Return traded differently that morning. Investors pulled something like $23bn out of the fund that month, the largest monthly redemption a US mutual fund had recorded at the time, and the outflows ran for years afterwards, taking the fund from a 2013 peak near $293bn down to a small fraction of it. The three-year numbers on the screen had barely moved. The reason to hold the fund had.

That is the trade the allocation decision actually turns on. A buyer weighs three things that are only loosely correlated: the track record, the credibility of the people producing it, and the philosophy the firm has told the market it follows. When performance is good, the three agree and nobody has to arbitrate. When performance turns, they pull apart, and the marketer finds out which one was really holding the money.

Why identical numbers produce opposite decisions

Performance is the price of entry. Clearing the threshold (roughly top-half over the relevant period, top-quartile for anything sold as high conviction) gets a fund into the comparison set that the rating and buy-list machinery described in the gatekeepers lesson feeds. Inside that set, the numbers stop discriminating, because everyone in it has acceptable numbers.

What decides it among lookalikes:

• Whether the buyer can say in one sentence why this fund earns its return, and say it again next year without changing the sentence.
• Whether the return arrived the way the manager said it would.
• Whether the person or process producing it is still there, and will still be there in three years.
• Whether reporting, service and ownership are boring enough to be invisible.

Flows (net money in minus net money out) respond to that fourth item far more slowly than they respond to the third. A firm can survive weak reporting for a decade. It rarely survives losing the name on the door.

The trust stack in asset management

Each layer has to hold before money moves, and they fail in a particular order.

Layer 1: Fiduciary safety

The payoff is asymmetric for the buyer. A pension trustee who picks a top-decile manager gets a line in the minutes. A trustee who picks a fund that later blows up faces questions from a board, a sponsor, and possibly a regulator. So the first thing sold is the absence of downside surprise: clean audits, stable ownership, no key-person cliff, no strategy drift.

This is why alpha (return above a benchmark) is a second-order question in the room. It gets asked after the quieter one: will this embarrass me?

Layer 2: A philosophy specific enough to be broken

Fundsmith states its strategy in three lines: buy good companies, don't overpay, do nothing. Terry Smith repeats it in the annual letter and at the annual meeting, names his own mistakes there, and has attacked companies he holds when he thinks management has wandered off (his 2022 broadside at Unilever for worrying about the purpose of mayonnaise). The story is falsifiable, and that is the point. It also costs him something real: he cannot buy a cheap bank or an oil major without contradicting his own pitch, and when quality-growth lagged across 2022 and 2023 he had no permitted place to hide. Fundsmith Equity has come well down from a peak close to £30bn.

Compare the manager whose philosophy is "high conviction, bottom-up, quality at a reasonable price." Nothing there can be broken because nothing there was ever specific. When performance turns, that manager has no defence to hand the buyer, only a sympathetic tone. The vague pitch feels safe to write and is worthless in the meeting where the money is decided.

The test is simple and most firms fail it: have you told clients, in advance and in writing, the market conditions in which you expect to underperform? Baillie Gifford had. When Scottish Mortgage lost close to half its value in 2022 and firm-wide assets fell by roughly a third from the 2021 peak, the partners published the same growth argument rather than quietly buying value stocks to flatter the tracking error. Money still left. But the story stayed intact, which is why buyers can still describe the firm accurately today.

Layer 3: Manager credibility and key-person risk

Track record is attached to people, and people are a single point of failure. PIMCO's 2014 lesson is that a firm can own the best-known fixed income brand in the world and still watch the asset base halve when one investor leaves, because a large share of the trust was in him rather than in the institution.

The structural answers are unglamorous and slow. Baillie Gifford's partnership, with no external shareholders and team-managed strategies, spreads the credibility across a group. PIMCO's own rebuild ran the same way: the growth that followed came largely through PIMCO Income, fronted by a team rather than a single star. A marketer who lets a firm's entire narrative concentrate on one face is building an asset that walks out of the building.

The second-order consequence is the one boards underestimate. Outflows are not just lost revenue; in concentrated, small-cap or credit strategies, redemptions force selling into the same market that caused the underperformance, which degrades the numbers further and triggers the next round of redemptions. A trust shock becomes a performance shock within two quarters. This is why the communication work has to start before the flows do.

🎬 [VIDEO: "How Investment Consultants Shape Institutional Investing" - youtube.com - a clear overview of the gatekeeper role consultants play in directing pension and endowment assets]

What this means for the marketer's job

1. Own a position narrow enough to be defensible

PIMCO means fixed income. Fundsmith means one strategy, run one way, with no product extension to blur it. For a boutique the position is smaller but the logic is identical: be the obvious answer to one question rather than a plausible answer to twelve.

2. Turn performance into attributionattributionA framework for assigning credit to the touchpoints that contributed to a conversion, so you can measure which channels and interactions actually drive results.View full definition →

Every competitor has a fact sheet. The edge is explaining why the return happened and why it should repeat, which means publishing the answer to what you got wrong as well. Surviving a drawdown (a peak-to-trough decline) with the process unchanged is a stronger sales asset than a great year, because it is evidence rather than a claim.

3. Write the defence the buyer will read aloud

Whoever recommended you has to justify keeping you when the number is ugly. Give them the sentence: what the strategy owns, what it deliberately does not own, and which market this was always going to be hard in. If your material only works when performance is good, it is advertising, not support.

4. Communicate hardest when the news is worst

Flows leave fastest when investors feel surprised, not when they lose money. A pre-committed cadence (same letter, same author, same explanation of the same philosophy, in a bad quarter as in a good one) is the cheapest retention tool in the sector and the one most firms abandon exactly when it counts.

For industry-level flow and fee data, the fact books from the Investment Company Institute, the US fund industry's own trade association, are a solid free reference, with the usual caveat that they represent the industry they publish for.

A concrete scenario

A quality-growth manager is three years into underperformance. Assets are down 25 percent, and the sales team wants two changes: add a few cheap cyclicals to close the gap on the benchmark, and soften the language in the factsheet.

Both look like marketing decisions. Both destroy the only asset still working. The moment the portfolio contains positions the philosophy cannot explain, the buyer's one-sentence defence stops being true, and the next review becomes a redemption. Meanwhile the record is no longer evidence of anything, because it now mixes two different strategies.

The defensible move is the harder one: hold the process, publish exactly why the last three years happened, and accept a smaller, more patient book. Firms that drift under flow pressure end up with neither the record nor the story.

The tension: performance still matters

Trust buys time. It does not buy immunity. A fund lagging its benchmark by two points a year is roughly ten points behind after five, and no relationship survives that arithmetic; the narrative simply changes from "this is what we said would happen" to "this is what we are."

The passive case is the useful counter-example. Where the product is the index, there is no manager credibility to defend and no philosophy to break, so the allocation goes almost entirely on cost, tracking difference and operational reliability. Trust there is about the firm's plumbing. Fee compression across the industry has pushed active managers into the opposite position: with no cost story available, the only durable pitch is a credible reason to expect the process to repeat.

Performance gets you considered. Trust gets you allocated. Consistency gets you kept.

Where marketers get it wrong

• Selling last quarter's number, which attracts money that leaves at the same speed and sets an expectation you cannot meet twice.
• Building the entire brand on one manager's face, then discovering what PIMCO discovered in 2014.
• Going quiet in a drawdown. Silence reads as concealment, which costs more than the loss did.
• Widening the philosophy to fit whatever is working, which converts a track record into an uninterpretable blend.

Key takeaways

• Between funds with similar returns, the allocation goes to the one whose story the buyer can repeat and defend without editing it.
• A philosophy that cannot be broken cannot be trusted. State in advance when you expect to struggle, and accept the constraint that puts on the portfolio.
• Credibility concentrated in one person is a liability on the balance sheet, not an asset.
• Redemptions feed on themselves through forced selling, so retention work has to be running before performance turns.
• Trust buys patience; only results refill the account. Five years of lagging breaks any narrative ever written.