Building Something That Lasts

The recommendation was for a mortgage fund.

I'd been at AAA for a few weeks — long enough to understand the organization, not long enough to have rebuilt anything. The portfolio was managed by an OCIO: a capable, credentialed firm that ran institutional portfolios for dozens of clients. The recommendation they brought us was a perfectly reasonable yield-oriented strategy. For a typical institutional investor, it would have been defensible.

AAA insures homes.

A mortgage fund whose downside scenario involved widespread foreclosures — mass defaults, property values under severe stress, precisely the economic conditions that generate our largest insurance claims — was a fund whose worst-case outcome was also our worst-case outcome. The manager understood mortgages. They had not thought carefully about their client.

I saw it immediately. The OCIO hadn't — or hadn't considered it worth raising.

That wasn't the OCIO's failure as a firm. It was a structural limitation of the model. They served dozens of clients. Their research was built to scale across those clients. Understanding one client's specific business at the level required to catch a correlation like that is not scalable. It requires being inside the institution, understanding the mission, knowing what the bad scenario actually looks like for this organization. The OCIO couldn't provide that. Their model didn't allow for it.

We moved on from the OCIO. That decision led to the blank canvas.

What the blank canvas actually contains

Most people who haven't done it picture building an investment program as primarily a portfolio construction exercise — pick the right asset classes, find the right managers, set the right targets. That part is real, but it's not the hard part.

No investment policy statement worth keeping. No reporting infrastructure. No manager relationships that belong to the institution rather than to the previous advisor. No institutional memory of how past decisions were made or whether the reasoning still applies. No investment committee with a clear sense of its own role.

The hard part is building the governance. An investment policy that genuinely reflects this organization's mission, risk tolerance, and liquidity constraints — not a template from another institution's document, but something that actually describes this one. Board reporting that tells the truth about what's happening and why, rather than presenting performance in the most flattering possible frame. Standards against which every future decision will be evaluated — explicit enough that they mean something when challenged.

None of that was in place when I arrived. I built it from scratch, with no team and no institutional memory to draw on.

First principles

The discipline I carried from TRS — every allocation has to earn its place, position by position, continuously — doesn't change with scale. What changes is the application.

At TRS, that discipline operated across a $10 billion external program within a $100 billion fund, with a full team, a mature governance structure, and a culture built by people who understood institutional investing deeply. At AAA, it had to operate with one other person, within a culture I hadn't designed and couldn't transform overnight, at a fraction of the scale.

The question I started with wasn't: what should a typical investment program look like? The question was: what does this organization actually need from its investment program? The answers are different, and confusing the two questions is exactly how you end up recommending a mortgage fund to a home insurer.

AAA needed stability and liquidity appropriate to an insurance company. It needed a governance structure the board could genuinely understand and own — not a rubber stamp on recommendations from people they couldn't evaluate. It needed managers selected for edge relevant to this specific mandate, not track records that made sense for other clients. And it needed a reporting framework that told the truth about risk, not just performance.

Starting from first principles meant writing those requirements down before choosing any strategy or manager. The investment policy drove the asset allocation — not the reverse. The governance was built before the portfolio.

Operating within constraints you didn't design

I want to be honest about the environment. AAA is not an organization where a new hire walks in and redesigns everything. The culture is deliberate and durable — mostly a strength, occasionally a constraint. I couldn't hire freely. I couldn't build the team I might have designed on paper. For most of this period, it has been two of us.

That constraint forced a discipline I didn't expect to be valuable but has proven to be: building the process so that it doesn't depend on having the right team.

If the process only works when specific people are in specific roles, you haven't built a process — you've built a personnel dependency. The documentation, the standards, the governance framework, the reporting: all of it has to be rigorous enough that a qualified person can pick it up and run it. That's a higher bar than most investment offices meet. Most programs are held together by the judgment of one or two key individuals, and the board doesn't find that out until one of them leaves.

The discipline I learned at TRS — the SLOG, the documented accountability, the pre-specified exit conditions described in Parts I and II — is specifically designed to prevent exactly that. The rigor lives in the process, not the person. I tried to build that at AAA with the tools I actually had.

What five years produced

Five consecutive years of benchmark outperformance. A governance framework the board understands and actively uses. A manager roster built on edge analysis rather than track records or consultant shortlists. An investment policy that reflects the actual mission of the organization.

I won't cite specific performance numbers here — that's not my place, and it's not the point. What I'll say is this: the discipline works at $1 billion as directly as it did at $10 billion. The question of whether the methodology travels from a large pension to a smaller program was real when I started. I know the answer now.

What traveled wasn't the headcount or the budget or the institutional prestige. It was the process. The standard.

Most institutions assume you need large-institution resources to build a large-institution investment program. You don't. You need a standard — and the discipline to hold it regardless of market cycle, staff turnover, or the path of least resistance.

If you're running a program at a smaller institution — an insurance company, a corporate investment office, a foundation, a mid-sized endowment — the resources of a large institution are genuinely not what you need. You need clarity about what your organization actually requires, standards for how decisions get made and defended, and the discipline to hold both consistently. That's what endures. Not the managers. Not the specific allocations. The standard you build and keep.

Perhaps what I'm most proud of isn't the returns or the fee savings. It's that the program doesn't depend on me. The process is documented. If I walked out tomorrow, it would continue to perform — because we built it to last, not built it around a person. That's the difference between a program and a personality. And it's the thing most boards never think to ask about until the person walks out the door.

In Discipline IV, we'll look at the most common question I get asked about the future of this work — and why the honest answer is more specific, and more useful, than either side of the AI debate usually admits.