We often describe our work as "underwriting," and the verb is chosen deliberately. Underwriting, in the original sense, means to accept a risk at a price, with full knowledge of the downside, and with a contractual obligation to honour the commitment regardless of subsequent sentiment. An investment thesis, held seriously, has the same structure.
The three questions
Every position we take is required, before committing capital, to pass three questions.
What would have to be true for this to work?
This is a question about the minimal sufficient conditions for the thesis. It is not a question about the most likely scenario; it is a question about the specific, falsifiable chain of causation that delivers the return. If the answer is vague — "the company is undervalued and the market will figure it out" — the position fails the question.
A good answer is structural. Regulatory approval by a specified date. A capital allocation decision by the new CEO. A change in the competitive set driven by a pending industry consolidation. Something observable, datable, and testable.
What would have to be true for this to fail permanently?
The distinction between temporary and permanent impairment is the single most important one in fundamental investing. A business can trade at half of our assessed intrinsic value for three years without the thesis being wrong; it can trade at ninety cents of our assessed intrinsic value for three months while being decisively wrong. Confusing the two is how thesis drift begins.
A permanent impairment has a character: the cash-generative capacity of the business, or the terms of its capital structure, or the legal or regulatory context in which it operates, has changed in a way that cannot be recovered. The test is not "is the stock down" but "has the thing we were buying changed."
At what price does the thesis cease to be interesting?
Every position has an exit price at which it no longer clears our hurdle, and we commit to that price in writing before taking the position. This is not a target; a target is a best-case. It is the point at which the margin of safety has collapsed, and the position must earn its keep on a fresh underwriting if it is to remain.
The discipline of writing this down in advance is significant. It defends against the human tendency to rationalise a richer multiple after the gain, which is how durable mispricings are quietly traded for lottery tickets.
What the questions protect against
These three questions do not guarantee a good outcome. Nothing does. What they guarantee is that a poor outcome is traceable to a specific, identifiable error: a mis-specified condition, a missed change in the underlying business, or a violated discipline. That traceability is, in our view, the difference between a process that compounds and one that does not.
A fund whose losses are indistinguishable from bad luck has no process. A fund whose losses are, in hindsight, attributable to specific, auditable errors — and whose errors recur less frequently over time — has one.