Financial statements record transactions. They capture purchases, sales, write-downs, and realized gains. What they do not capture is the cost of the decision that was never made. The foregone opportunity leaves no entry. The position never initiated produces no loss. Over long holding periods, these unrecorded events can determine more of an investor's compounded outcome than the recorded ones — not because the losses were small, but because the foregone gains were large.
I. The Problem: Invisible Errors
The standard framework for evaluating investment mistakes focuses on errors of commission: positions that declined, theses that proved incorrect, capital deployed at prices that proved too high. These errors are visible, appear in performance attribution, and invite scrutiny.
Errors of omission produce no such signal. When a high-quality opportunity is passed over, or sized far below what the evidence warrants, the cost accrues as a gap between actual and achievable returns. It does not appear as a loss. It generates no reporting line.
This asymmetry in visibility creates a systematic incentive problem. Investors evaluated primarily on realized losses will rationally tilt toward inaction. The career risk of a visible commission error exceeds the career risk of an invisible omission error. The predictable result is persistent under-action: too little capital committed, too infrequently, to the opportunities that were most clearly understood.
II. The Engine: Forces That Produce Under-Action
Four distinct forces reinforce this pattern.
First, measurement blindness. Accounting systems record what happened. They have no mechanism for recording what would have happened under a better decision. Absent measurement, there is no feedback, and the behavior that produced the suboptimal outcome is not identified or corrected.
Second, friction. Every reallocation involves transactional, emotional, and reputational costs. Selling one position to fund a better one requires acknowledging that the original is no longer the best available use of capital. Revising a prior view carries social and institutional costs. These costs are real but finite. The compounding cost of not acting is also real — and is not finite. Treating friction as a reason rather than an obstacle conflates the two.
Third, entry price anchoring. Investors frequently evaluate current positions relative to what was paid rather than what is expected going forward. The price paid is a sunk cost. It has no bearing on future returns, current intrinsic value, or whether a better alternative exists. Decisions anchored to cost rather than forward expected value will systematically misprice the hold-sell question.
Fourth, the asymmetric salience of commission versus omission risk. A loss from a position taken is vivid and concrete. The cost of a position not taken is probabilistic and diffuse. Behavioral research documents that vivid outcomes are weighted more heavily than abstract ones in decision-making, even when expected values are equivalent. This is not a correctable cognitive error so much as a stable feature of judgment under uncertainty — one that requires structural compensation rather than simple awareness.
III. Inversion: How Investors Compound the Problem
Inversion is useful here. Consider what behaviors would reliably produce a poor long-term allocation record.
Requiring certainty before acting. No investment of consequence is certain. An investor who waits for certainty will find that by the time the uncertainty resolves, the price has adjusted. The relevant discipline is not certainty but probability: defining acceptable downside parameters and acting when the evidence crosses a threshold — not when it eliminates doubt.
Permanent pilot sizing. A position initiated at minimal size and never scaled, regardless of how the thesis develops, captures little of the value identified. If a business proves more durable than expected, or the downside narrows materially, a position held at 1% when the evidence supports 5% is a structural underperformance relative to the investor's own thesis.
Friction as justification for inaction. Transactional and psychological costs are real. They are also one-time. The return differential between a position held at suboptimal size and one appropriately sized compounds annually. An investor who declines to incur a one-time cost to prevent a permanent annual drag has made a poor arithmetic decision.
Evaluating decisions by reputational defensibility rather than expected value. An action that produces a loss is harder to defend than an inaction that produces an equivalent opportunity cost. Where incentive structures punish the former more than the latter, rational agents will under-act. Recognizing the incentive distortion does not eliminate it, but failing to recognize it guarantees its effect.
IV. Discipline: Principles for Structural Correction
Measure decisions against counterfactuals. Every significant allocation decision should be evaluated relative to the best available alternative within the investor's domain of competence at the time. If the foregone alternative was superior, inaction was an active decision with a measurable cost — not a neutral non-event. Reviewing it as active creates accountability that reviewing it as passive does not.
Scale position size to evidence quality. As a thesis strengthens — as unit economics improve, as competitive position clarifies, as the downside range narrows — position size should respond proportionally. A sizing process that does not respond to evidence is not a process. It is a fixed allocation. The two are different in their long-term consequences.
Schedule reallocation reviews and document the reasoning. Episodic, ad hoc reallocation decisions are more susceptible to friction and inertia than scheduled ones. A regular review process distributes the decision cost, reduces its perceived weight, and creates a written record of why capital remained allocated as it was. That record introduces accountability for inaction.
Re-underwrite independent of entry price. The anchor for any current allocation decision is the forward expected value under current information. Entry price is a sunk cost. An investor who finds they cannot evaluate a position's merits without reference to what was paid has introduced an irrelevant variable into the analysis.
Audit near-misses. Reviewing opportunities that were passed over — particularly those that subsequently performed well — is uncomfortable and therefore rarely done. It is also the most direct way to identify the specific friction points and recurring reasoning patterns that produce omission errors. A pattern visible in five near-miss reviews is a correctable process problem. A pattern that has never been reviewed is not correctable.
V. Worldly Wisdom
Over long horizons, the composition of an investment portfolio is shaped less by what went wrong than by what the investor declined to do when the evidence was clear. Financial statements record transactions; they do not include entries for foregone alternatives. This measurement gap, compounded by institutional incentives that penalize visible errors more than invisible ones, produces a persistent structural bias toward under-action.
The allocator who accounts for this will treat inaction with the same analytical rigor applied to loss. They will ask not only whether a position costs them capital, but whether the decision not to size a position adequately costs them more. The relevant standard is not whether a decision avoided a loss. It is whether it was the best available use of capital, relative to identifiable alternatives, at the time it was made.
When an opportunity lies clearly within one's domain of competenc,e and the expected value is asymmetric, the larger structural risk is typically inaction, not error. That asymmetry does not resolve itself through awareness. It requires a process designed to counteract it.
Closing Reflection
Most investors have a clear record of their commission errors. Few have an equivalent record of their omission errors. The two are not symmetric in their effect on long-term outcomes, and the accounting system reinforces the asymmetry by recording only one of them.
The question is not what losses might have been avoided. It is what the portfolio would look like if the investor had acted each time the evidence was sufficient and the opportunity was within competence. That counterfactual is not recorded anywhere. It is nonetheless real.
If this memo was useful, consider forwarding it to one thoughtful investor.
