What We Are Trying to Do
Davis Rea exists to compound our clients' capital durably over decades while avoiding the kinds of losses from which recovery is difficult or impossible. We are not trying to maximize return in any single year. We are not trying to track a benchmark. We are trying to own businesses we want to own for many years, and to avoid owning businesses whose long-term prospects rest on assumptions we do not believe will hold.
Most of what we do follows from that statement. The long-term compounding goal and the capital preservation goal usually pull in the same direction — owning a great business at a reasonable price tends to produce both compounding and protection. When the two goals pull in opposite directions, however, we choose preservation. A year of underperformance against an index is a recoverable outcome. A permanent impairment of capital in a business whose long-term cash flows turn out to be lower than the market assumed is not. We can recover from a year of being out of step with the market. We cannot recover from owning a business that quietly stopped being worth what we paid for it.
This priority is not unique to us. It is, in some form, what every careful investor would say if asked. What makes it operationally meaningful at Davis Rea is the discipline with which we act on it, including in moments when the cost of doing so is visible and uncomfortable. The rest of this piece is an attempt to describe what that looks like in practice.
How We Work
We use AI tools as part of our analytical process — to pressure-test arguments, to structure complex frameworks, to surface counterarguments we might otherwise miss, and to draft material that we then edit substantially. We believe AI is a meaningful productivity tool for the kind of work we do, and we expect to continue using it in ways that improve the quality of our research.
We want to be explicit about something that matters more than the tool itself. The judgment, the philosophical framework, the instinct built up over many years of investment practice, and the responsibility for every conclusion we reach and every position we hold remain entirely human and entirely ours. Current AI systems are remarkable in certain kinds of analytical work. They also lack forms of memory and instinct that humans develop through lived experience — the recognition that a particular configuration of market behaviour resembles a prior moment in ways that matter, the awareness of a specific regulatory history that shapes how a sector responds to stress, the pattern recognition that comes from having been wrong in specific ways before and having learned from it. These are not capabilities that AI can replicate today, and they are precisely the capabilities that make our work valuable to our clients.
There is something deeper still that we want to name plainly. The work we do is not only analytical. It is relational. It involves a genuinely human instinct to protect the people whose capital we manage — an empathy for what is at stake for each client and family, a sense of responsibility that no tool can carry on our behalf. Part of this shows up in the most practical question an advisor can ask, and one that is rarely answered well: not just what risks a client can mathematically afford to take, but what risks they are willing to tolerate. Capacity for risk is a number on a page. Tolerance for risk is a person — their history, their temperament, their family circumstances, the conversations we have had over many years about what they are really trying to accomplish with the wealth they have built. AI does not weigh tolerance. It cannot. The relationship between an advisor and a client is precisely the place where the human element of this work lives, and that is the place from which the most important portfolio decisions are made.
This distinction matters for two reasons. The first is straightforward transparency: our clients should know how we work. The second is that the relationship between tool and judgment is itself a useful lens for thinking about the businesses we want to own. Many companies are using AI to lower their costs and improve their output. The companies that will compound durably are those that use AI as a tool that enhances human judgment, not necessarily those that try to replace human judgment with AI. The same logic applies to investment management. Davis Rea's clients are not buying AI-generated analysis. They are buying judgment that AI tools help us sharpen — and they are buying a relationship with people who care, in a deeply human way, about the outcomes their capital produces.
Why Our Portfolio Looks Different
Our portfolio does not look like a typical institutional equity portfolio, and that has been a deliberate choice anchored in the priority we have just described. We have written about the specific positions elsewhere and will not catalogue them here. What we want to acknowledge directly is that the philosophy that produces our portfolio also produces consequences that are visible in our relative performance, and that those consequences cut in both directions.
The most visible expression of our different positioning has been our absence from the semiconductor manufacturers throughout the AI cycle. That absence has been costly. The semiconductor names have produced returns over the past several years that have meaningfully exceeded the broad market, and our relative performance has reflected that. We want to acknowledge this plainly rather than buried in a footnote. The reasoning behind the decision —that semiconductor manufacturing is a deeply cyclical business whose long-term economics depend on capital allocation discipline that the industry has rarely demonstrated in practice —was sound when we applied it. It has also been the wrong call for the cycle to date. Both things can be true.
Less visibly, the same investment philosophy kept us out of the broad enterprise software category. That decision has not been costly — it has been beneficial. The category has materially underperformed the broader market over the past several years, and many of the highest-profile franchises have suffered substantial drawdowns from their 2021 highs, with several still trading well below those levels today. Our absence from this category has saved our clients from losses that would otherwise have flowed through their portfolios. The full picture of our non-correlation with the index is therefore considerably more favourable than the conventional performance discussion typically acknowledges. We have missed one source of return. We have avoided a much larger source of loss. On a risk-adjusted basis and viewed through the lens of capital preservation that anchors our philosophy, we believe this trade-off has served our clients well.
We say this not to make excuses, but to provide the context our clients need to evaluate our work fairly. An equity portfolio constructed to track a broad index inherits the index's exposures by default — both the exposures that have produced strong returns and the exposures that have produced losses. The investor who tracks the index is not making a choice about either. The investor who actively chooses not to own either category is making two specific decisions, and both should be understood for what they are.
The Kind of Risk We Worry About Most
We are deeply preoccupied with a category of risk that we believe conventional valuation methods underweight: the risk that a business is healthy and well-run, but the customers it serves are progressively weakened by competitive forces that have nothing to do with the business itself. The business looks fine on paper. Its competitive position is intact. Its operating metrics are healthy. And its long-term prospects are nonetheless quietly eroding, because the customers it depends on are themselves losing ground to faster, leaner, more technologically capable competitors.
This risk has always existed. What has changed is the speed at which it can happen. In prior decades, an established business could rely on the stability of its customer base for many years, because competitive landscapes changed slowly. The Sears and the Polaroids and the Kodaks of prior eras took decades to fade. The era we are now entering is one in which AI-native competitors can enter business-to-business markets and reach material scale within years rather than decades. The cost structure of a competitor built from the ground up around AI is so structurally lower than the cost structure of the incumbents in many industries that the competitive landscape can shift faster than the stock prices and income statements of the affected businesses currently reflect.
The portfolio implication of taking this risk seriously is straightforward. We have a strong preference for businesses whose moats are anchored in physical scarcity that cannot be replicated by an algorithm — pipelines, rail networks, real assets, telecommunications infrastructure, distribution and delivery networks. We have a similar preference for businesses whose customers sit firmly at the top of their respective competitive distributions — the largest enterprises, the highest-tier consumers, the institutions whose own competitive positions are durable. We have less interest in businesses whose customer bases sit in the middle of their respective industries, where the AI-native disruption thesis is most acute. The decision in 2024 to avoid the broad enterprise software category was, in significant part, an expression of this preference.
The Discipline of Being Wrong in the Short-Term
Long-term ownership, practiced honestly, requires being willing to look different from the index for extended periods. It requires being willing to bear the visible cost (the opportunity cost) of being out of step with categories that are currently producing strong returns, in exchange for the less visible benefit of avoiding categories that will produce permanent impairment. It requires being willing to be wrong in the short-term while standing behind the reasoning that produced the decision.
We want to be clear about what we are and are not saying. We have been wrong in the short-term on the semiconductor manufacturers. The names have outperformed and we have not owned them, and that has resulted in an opportunity cost for our clients. We are not making excuses for that outcome. What we are saying is that the reasoning behind the decision remains solid, that the same reasoning is what has kept our clients out of the enterprise software category that has produced material losses, and that our investment philosophy is built around the conviction that sound reasoning is what compounds capital over decades, even when it produces uncomfortable outcomes in particular years. We owe our clients honesty about both halves of that statement.
This kind of honesty is rarer in our industry than it should be. The temptation, when a long-held view has been costly for a sustained period, is to capitulate and chase the returns one has missed. We believe capitulation in moments like this is precisely the wrong response, because it risks converting the temporary cost of being wrong in the short-term into a permanent impairment of the discipline that protects capital over the long-term. The investor who sells what they should be holding and buys what they should be avoiding, at the moment when the cost of doing the right thing feels most acute, is the investor who risks converting recoverable underperformance into permanent loss. We do not intend to be that investor on behalf of our clients.
A Closing Word
The opportunity in front of us is unusually large, and the uncertainty is unusually high. We hold our views with conviction and with humility — conviction because the analytical work supports them, humility because the world will continue to reveal more of itself than any framework can fully anticipate. We will continue to revise both our framework and our positions as the evidence requires, and we will continue to communicate with our clients honestly about both what we know and what we are still working to understand.
Long-term ownership, done well, is not glamorous. It is the patient accumulation of small advantages — owning the right businesses, avoiding the wrong ones, being willing to look different from the crowd when the crowd is taking risks we are not willing to take, and being willing to acknowledge plainly when the discipline has carried a visible cost. These advantages compound over decades into outcomes that protect and grow our clients' capital in ways that more frenetic approaches cannot. We are grateful for our clients' continued trust as we do that work on their behalf.
For clients who would like a more detailed analytical treatment of how this philosophy is currently being applied across our portfolios — including the specific framework we use to think about the risk described above and how it applies to individual positions we hold — our companion piece, Why We Own What We Own — And What We've Avoided, develops these ideas more fully. Please feel free to reach out to us if you would like a copy or would like to discuss these ideas further.
John O'Connell, Davis Rea. Motovun, Croatia. May 20, 2026
Sincerely,
John O'Connell
Chairman, CEO & Portfolio Manager
