Why lean managers still need an institutional risk office.

The objection we hear most from a $150–500m manager considering an outsourced risk function is some version of: "I already know my book. I look at my Greeks every morning. Why do I need an office for something I'm already doing?" It's a fair question, and it deserves a specific answer rather than a generic pitch about "institutional-grade risk management."

Knowing your book and having a risk office are different products

A PM who watches gross, net, beta, and factor exposures daily is doing position monitoring. That's necessary and most good PMs do it well. A risk office does four things a PM monitoring their own book structurally cannot do, no matter how disciplined they are:

Independence of the marks. A PM's read on their own risk is filtered through the same conviction that generated the positions. This isn't a criticism — it's just true that the person who built the thesis is a biased instrument for measuring its risk. A risk office reports through a different line, with no P&L attribution tied to being right about the positions it's marking.

Aggregation across correlations the PM isn't pricing. Single-name risk is visible to a PM by construction — they know what they own. Cross-position correlation risk, especially correlation that's regime-dependent and low in calm markets, is exactly the kind of exposure that doesn't show up in a daily Greeks read until it's already realized. This is where 2018 Feb, 2020 March, and 2023 March all rhymed: idiosyncratic-looking books that were secretly one factor bet.

Standing scenario infrastructure. Ad hoc stress tests run by a PM under time pressure tend to test the scenarios the PM already has an intuition about. An institutional risk office runs a fixed library of historical and hypothetical scenarios on a schedule, regardless of what's topical, which is the only way to catch exposure to a scenario nobody's currently worried about.

A record that exists independent of memory. When an allocator asks "what was your VaR and factor exposure the week before the drawdown," a lean manager without a risk office is reconstructing that from trade blotters and memory. A risk office produces a contemporaneous, time-stamped record — which matters enormously in due diligence and, less happily, in post-mortems.

The economics of building this in-house at $150–500m

The counterargument is usually "fine, but I can hire one risk person." At sub-$500m AUM, the math rarely works:

• —A credible risk hire with the quant background to build multi-factor attribution, real scenario libraries, and liquidity-adjusted VaR runs $250–400k fully loaded before you've built any infrastructure. One person is also a single point of failure with no second reviewer — which partly defeats the independence argument above.
• —The infrastructure itself — factor models, historical scenario data, liquidity curves by instrument, a proper attribution stack — is a multi-year build for a single hire working alongside their day job of actually running the numbers daily.
• —One risk person reporting to the PM who pays them is not independent in the way an allocator doing ODD is looking for. The reporting line matters as much as the analysis.

An outsourced risk office gets a manager the infrastructure of a much larger institutional platform — the factor models, the scenario library, the second set of eyes — at a cost structure and speed that a single in-house hire can't match at that AUM, and with a reporting line an allocator can actually underwrite.

What "lean" managers get wrong about what the office is for

The framing we push back on is that a risk office exists to catch the PM doing something wrong. Mostly it doesn't. Its real value in a well-run lean shop is:

1. Turning tacit risk awareness into an explicit, comparable number. A PM who "feels" the book is getting risky is right more often than not — but "I feel like this is getting risky" isn't something an allocator, a prime broker, or the PM's own future self six months from now can act on with precision. A risk office turns that instinct into a number with a trend line. 2. Freeing the PM's attention. Every hour a good PM spends building scenario infrastructure or reconciling factor exposures across a multi-strategy book is an hour not spent generating alpha. The lean managers who benefit most from an outsourced office are usually the ones who were already doing risk work competently themselves and would rather not. 3. Making the next capital raise easier. Institutional allocators increasingly treat "no independent risk function" as a structural gap regardless of how good the PM's own risk instincts are. This isn't about distrust of the PM — it's that ODD frameworks are built around verifiable, independent process, and a PM's own diligence, however genuine, isn't independently verifiable by definition.

The actual question to ask

Not "do I need this," but: if my book had a correlated drawdown tomorrow that came from an exposure I wasn't watching, would there be a contemporaneous record showing whether that exposure was visible and flagged beforehand? For most lean managers without a standing risk office, the honest answer is no — not because they aren't careful, but because careful and independently documented are different things, and only one of them is the thing allocators, and eventually the manager themselves, actually need.