What Capital Discipline Owes You When the Macro Picture Will Not Resolve

For much of the past two years, nearly every investment committee discussion has opened with the same question: which way is the macro going to move. The question keeps getting asked, and the answer keeps failing to get clearer — interest-rate paths, demand structure, policy direction, three variables that once could be folded into a judgment with reasonable stability, are each harder to read than usual right now. This is not because our tools have gotten worse. It is because the variables themselves have entered a period that resists being read. Admitting that is harder than pretending otherwise, but the work that follows is only honest once it is admitted.

Facing this, an institution typically drifts toward one of two responses. One is to double down on macro research, try to force the uncertainty into a narrower band, and bet on it. The other, and the one we hold to, is to accept that some uncertainty cannot be compressed on any useful timeline, and shift the center of gravity from forecasting more accurately to ensuring a wrong call stays contained. The line between the two decides whether an institution keeps its judgment intact across an uncertain stretch, rather than being forced into an emotional decision at some turning point.

Forecast less, underwrite more

Forecast less, underwrite more can sound like a slogan, but it corresponds to a specific change in how work gets done. Forecasting works like this: form a judgment about where a macro variable is headed, then derive an investment's merit from it — if rates fall, this asset is worth this price. Underwriting works differently: instead of betting on where a variable goes, ask what outcome the investment produces under a range of plausible paths, and whether the worst of those outcomes still sits inside what can be absorbed. The first is only as sound as the forecast. The second is only as sound as the honesty of the downside math — and its failures are far easier to catch ahead of time.

The shift is easy to describe and harder to make, because it gives up a psychological comfort. Forecasting feels like control — as though a sufficiently refined model can tame uncertainty. Underwriting offers no such comfort. It requires a team to say plainly it does not know where rates will be next year, then face a plainer, less flattering question: if wrong, how wrong, and is that degree acceptable. Teams accustomed to showcasing analytical depth resist this instinctively, because underwriting-led conclusions sound less insightful. That is precisely the point — they depend on the honesty of the math, the one thing that can be relied on through a stretch when the macro will not resolve.

Price discipline: cheap, versus cheap enough to absorb being wrong

When the macro will not resolve, the role price discipline plays needs to be redefined. In a legible year, it mainly exists to avoid overpaying. In an illegible one, it has to carry an additional function — leaving room for being wrong. The two sound close, but the pricing looks different. The first only requires a fair valuation range against an asset's cash flows. The second requires an additional margin outside that range, and the only reason for it to exist is the acknowledgment that a judgment about the future could be mistaken, and the investment should still avoid an unacceptable loss under that mistake.

The question we habitually put to the committee is not whether a price is fair, but whether, if the macro turns out worse than expected over the following two years, buying today leaves us able to wait with composure, or forced into a decision at the worst possible moment. That is considerably harder to answer, because it requires imagining an unfavorable scenario honestly rather than dwelling on the base case. Most transactions we decline are not declined because the business is weak, but because the team could not establish that today's price still leaves room under that scenario.

Staged deployment: pushing back the moment a decision has to be final

Why a single deployment magnifies the cost of a macro misread

In a period when the macro will not resolve, completing a full capital deployment at a single moment compresses a judgment that could otherwise be diluted over time into a wager settling a view for years ahead. Staged deployment does not change the eventual total committed. It changes when a view has to be finalized — spreading that moment across several points in time, each carrying new information, each giving the team a chance to reconfirm or revise the judgment made at the last one.

For the mechanism to actually work, two conditions have to hold. First, each stage has to be large enough to represent genuine commitment, or it degenerates into a posture for delaying a decision — a stake too small to matter carries neither real downside exposure nor leverage to secure cooperation. Second, the interval between stages has to correspond to a real increment of information, not a calendar split — if nothing that could change the judgment has emerged between stages, the division is cosmetic and reduces no risk.

Concentration limits: discipline, not caution

The less reliable macro judgment is, the more a single position's size should be constrained — intuitive to state and routinely circumvented in practice, always for the same reason: this opportunity is too good to pass up an exception for. An exception is not always wrong. But if exceptions occur often enough that the limit becomes nominal, the constraint never really existed. We treat concentration limits as a vessel for discipline, not a concession against returns — their purpose is to hold a line against a misjudgment that has not happened yet, and fully could, at the exact moment conviction runs highest.

• A single-position cap: any completed investment's resulting share of the portfolio carries a hard limit, and an exception requires the full committee's written sign-off, not one deal lead's unilateral view that this time is different.
• An aggregate cap on shared macro exposure: several investments that look unrelated but share the same rate sensitivity or the same demand curve need to be counted against one combined limit, not calculated separately so each stays individually within bounds.
• A pacing cap on new commitments: even when each individual transaction is compliant, the aggregate pace of new capital commitments over a stretch of time should not outrun what the portfolio can absorb, to avoid too many unverified judgments sitting exposed at once near a macro turning point.

What doing nothing costs, and when it is right

Doing nothing carries a negative connotation in most institutional settings, as though it amounted to neglect. That deserves a finer distinction. Capital sitting idle simply because no one wants to take responsibility for a call is a real cost — idle capital carries genuine opportunity cost, one rarely put on the table because, unlike a failed investment, it has no specific moment that can be pinned on anyone. But a different kind of doing nothing is entirely different: having confirmed no opportunity on the table has a risk-return profile clear enough to warrant taking on, and choosing to wait. That is not neglect. It is evidence the system declines to force a bet when the information is not there.

The distinction is plain to apply: look for whether a written judgment sits behind the waiting — why not invest now, what conditions would trigger a reassessment, and whether that judgment is actively being revisited rather than left in a meeting record and never mentioned again. If the waiting is deliberate and carries a defined rhythm, it is discipline. If it exists only because no one wants to be first to speak, it is neglect, however carefully dressed up.

Discipline is not a substitute for foresight — it is the honest response to its limits

This position can be misread as passive — since the picture will not resolve, simply stop thinking about the macro and let the mechanisms run on their own. That is not what is meant. Studying the macro and forming a view has never stopped and should not stop. What changes is the weight judgment carries in the chain: it remains an important input, but no longer the sole basis for a transaction's size, pace and structure. The value of these three mechanisms lies precisely in needing no correct macro call to work — even when wrong, the loss stays within an absorbable range. That is the most honest response to the plain fact that the picture will not resolve.