True Edge

The meeting was in New York, sometime in the mid-2000s. Across the table was one of the largest fund managers in the world — a firm running tens of billions of dollars, a track record that had become almost obligatory reading for institutional allocators, the kind of operation whose quarterly letters circulated through investment offices whether or not you were an investor.

I asked: "Do you think your growing assets are affecting your performance?"

He nodded.

Then he said no.

I don't mean he paused and reconsidered. I mean the nod came first — reflexive, honest, his body answering before his professional self caught up — and then he reversed in the same breath. "No. No, I don't think so."

The performance data at the time supported my question. The fund had grown dramatically. The strategies that had produced its early returns were capacity-constrained in ways that mattered at scale. Returns had flattened. Correlation to broad market beta had risen. The math was right there.

He knew it. He said so with his face, before his words caught up with him.

That nod is what most due diligence misses.

In Discipline I of this series, I wrote about conviction: the word investors reach for when they haven't done the work. Manager selection has its own version of the same problem. Most of what passes for due diligence is conviction in a research costume — a track record, a brand, a peer group that includes this manager, and a meeting structured to confirm what you already wanted to conclude.

True Edge is different. It's specific. It's the structural advantage a manager has that explains why they produce returns other managers don't. It's the evidence — current evidence, not historical — that this advantage still exists. And it's the pre-specified conditions under which you would exit, established before you enter, not invented retroactively when things go wrong.

That is not how most allocation decisions are made.

A track record is not an edge

The most common mistake in manager evaluation is treating performance history as evidence of a replicable advantage rather than evidence that something happened.

A broken clock is right twice a day. A manager who was long risk assets during the 2010-2021 expansion looks, on a five-year lookback, like a skilled stock-picker. A long-short fund that ran a high net-long book during a bull market looks, on a track record, like a master of alpha generation. The track record doesn't distinguish between skill and exposure. Only a rigorous examination of what the manager actually did — position by position, period by period, market regime by market regime — can begin to separate them.

This is time-consuming. It requires more than a consultant's due diligence report. And it sometimes produces conclusions that conflict with the track record, which creates a specific institutional discomfort that most allocation processes are quietly designed to avoid.

The most common edge-killer

Growing assets under management is the single most reliable way to destroy a manager's edge, and the single thing managers are most reliably dishonest about.

This is not because they are bad people. It is structural. The incentives of a fund management business — management fees that scale with AUM, the reputational reward of a large AUM number, the natural desire to accommodate investors who want to add capital — all point in one direction. The honest edge analysis points in the other. The nod-and-denial I described above happens because both answers are simultaneously true: the manager knows the honest answer, and the business requires the other one.

Capacity constraints are strategy-specific, but the pattern is consistent. Small-cap equity strategies that worked at $500 million get diluted at $5 billion. Convertible arbitrage opportunities available to a nimble $1 billion book evaporate when you're running $20 billion and the market knows you need to buy. Credit strategies dependent on quick exits change character at scale.

The question isn't whether AUM affects performance. It does, in almost every strategy, at some scale. The question is at what scale, in what ways, and whether the manager will tell you honestly. That nod told me everything about the last part.

What they say versus what they do

Style drift rarely surfaces in a standard manager review.

Managers describe their strategy in terms that made sense when they wrote their offering documents. Then they adapt. The market changes, the specific opportunity that defined their early edge gets crowded or disappears, and they find something adjacent. This is not necessarily dishonest — managers are paid to find returns, and the world doesn't stop changing to accommodate their original thesis. But it means the portfolio you're holding today may not resemble the strategy you approved three years ago.

The fix is re-underwriting. Not reviewing — re-underwriting. Starting from current position-level data, current attribution, current factor exposures. Asking: if this manager walked in today with this track record and this current portfolio, would we invest? The relationship has no standing in that question. The prior approval has no standing. You're evaluating the strategy as it actually exists right now.

Sometimes the answer is yes. Sometimes the honest answer is no, and the only reason you're still holding is that you've been holding for three years and nothing has gone visibly wrong. That's not an investment thesis. That's a lapse in discipline wearing the costume of one.

If you can't specify the exit, you don't have a thesis

The most important question in manager selection isn't the entry question. It's the exit question.

If you cannot articulate — before you invest — the specific conditions under which you would exit this position, you don't have an investment thesis. You have a position. The difference matters enormously when performance disappoints, because without pre-specified exit conditions, every drawdown becomes a decision problem in real time: is this temporary? is this the end of the edge? should we add, hold, or exit?

Without a pre-specified answer, those decisions get made by whoever has the most conviction in the room. Which brings us back to the beginning.

Pre-specifying exit conditions before entry forces you to articulate the failure case before you've committed the capital. It makes the analysis more honest. It also makes it easier to exit when the time comes — because the decision has already been made.

The discipline of doing this is uncomfortable. Most investment committees don't want to articulate exit conditions at the time of investment because articulating the failure case feels like undermining confidence in the decision. That discomfort is exactly the signal. The process is working when it's uncomfortable to do it honestly.

In Discipline III, we'll look at what these disciplines look like in practice when you're building from scratch — not at a $100 billion pension with full institutional infrastructure, but at a smaller organization with one other person and a culture you didn't design.