Eight Problems, One Screen: Reading the Frontier as a Macro Map
The sub-funds are not sectors. They are eight of the century's largest unsolved problems, held as a selection standard: energy, food, shelter, movement, materials, water, resilience and the technologies that measure them. Read through a factor lens rather than a sector lens, the interlocks — and the concentration risks — look very different.
A sector map answers the question "what industry is this company in?" It is an administrative classification, useful for reporting and close to useless for understanding what a portfolio is actually exposed to. Two companies in different sectors can share almost all of their risk; two in the same sector can share almost none.
The eight problems are a selection standard, not a sector scheme. They describe what a company is trying to solve, which is a statement about demand. The risk question — what makes these holdings move together — has to be answered separately, and it is answered with factors.
Why a factor lens rather than a sector lens
Factor analysis decomposes a return stream into systematic exposures — broad market, size, value, quality, momentum, and the macro sensitivities that matter here: interest rates, energy prices, industrial demand, regulatory direction.
Applied across the eight problems, the exercise is uncomfortable in a useful way. A grid-storage developer, a retrofit materials business and a water infrastructure concession occupy three different problems. All three are capital-intensive, long-duration, and priced substantially off the real rate. On a sector map they diversify. On a factor map they are, to a first approximation, the same trade.
Cross-problem diversification is a property of the narrative. It is not automatically a property of the risk.
The shared drivers we track
- Real rate duration. Long-dated infrastructure cash flows discounted at a real rate; the dominant shared exposure across at least five of the eight.
- Policy and subsidy regime. Correlated by construction, since a single legislative change can reprice several problems simultaneously.
- Industrial input costs. Steel, copper, lithium, cement — shared upstream exposure across energy, movement, shelter and materials.
- Permitting and grid access. An execution constraint that behaves like a factor, because it binds across projects at the same time.
Where correlation compresses
The important asymmetry: these exposures do not sit still. Under benign conditions, idiosyncratic drivers dominate and the eight problems genuinely behave differently. Under systemic stress — a sharp real-rate move, a policy reversal, a funding-market freeze — the shared factors dominate and cross-problem correlation compresses toward one.
This is the ordinary behaviour of correlated risk, not a peculiarity of impact portfolios. It matters more here because the narrative actively encourages the opposite assumption. A portfolio described as spanning eight distinct global problems sounds diversified. Under the conditions where diversification is actually required, it may not be.
Sizing against the driver, not the label
The practical consequence is that exposure has to be sized against shared factors rather than against problem categories. Holding eight positions across eight problems that all carry long-duration real-rate sensitivity is one concentrated position expressed eight ways.
The discipline is to measure exposure at the factor level first, then check that the problem-level allocation has not quietly double-counted the same macro driver. Where it has, the choice is explicit: accept the concentration and say so, or deliberately hold something whose driver is different — typically shorter-duration, less policy-sensitive, or with revenue indexed to a different input.
Reading the map honestly
None of this argues against organising a portfolio around problems. Problem-based selection is a statement about where durable demand is likely to come from, and it is a reasonable one: these are large, unsolved, and unlikely to become less pressing.
It argues against confusing that selection standard with a risk model. The selection standard says what to look for. The factor map says what you are actually holding. A portfolio needs both, and the second one should be allowed to contradict the first.
This note sets out the framework at a high level. We expect to publish a fuller factor decomposition of the eight problems, ideally with an independent specialist, rather than marking our own homework.