Same Question, Opposite Outcomes: Oakmark's Public Record of a -19% Sell and a +140% Hold

Judgments on holdings that ultimately split into a 19% decline and a 140% gain — the record of a sell and a hold that Oakmark made public.

On September 30, 2026, Harris | Oakmark, the manager of the Oakmark International Fund (OAKIX), released its third-quarter letter answering client questions. It was the first time the Q&A format — running for two years in the U.S. home-office CIO letter — had been applied to the international equity fund, and one passage in it stands out. In the same fund, at roughly the same time, two holdings faced the same question — "Should we keep holding this stock?" — and met diametrically opposite endings. Canadian enterprise software company OpenText was exited; German pharmaceutical and agrochemical company Bayer is still held.

Let's start with the story of OpenText, the position that was exited. When OpenText developed differently from expectations, the management team handed coverage to a new analyst. This is a practice described in last year's letter: when an analyst feels they have "lost their voice" on a stock, the team decides that a fresh perspective is needed. The new analyst's review surfaced three concerns: product duration, underinvestment, and the possibility that a major margin reset would be required to regain competitiveness. The team concluded that "a seemingly cheap software company was in fact carrying a structural competitive problem larger than we had anticipated," and closed the position out of concern that the problem would press on margins and limit intrinsic value growth.

And then the record continues. After the exit, OpenText moved roughly -19% in local-currency terms. Worldline (-50%) and WPP (-5%), which were also sold during the same review process, are mentioned alongside — though the team itself added the caveat that short-term price moves don't prove a decision right or wrong. Even so, the fact that it avoided that decline was a plus for the fund's performance.

Bayer was the judgment in the opposite direction. Having gone through last year's formal bear-case challenge process, the team revisited the key risks — litigation risk, business quality, and execution — and, with the stock trading near €20 at the time, judged that "management still has much to prove, but the intrinsic value and the potential for operational improvement remain attractive," and stayed in. Since then, litigation risk has decreased, the pharma business has improved, and the crop business has shown signs of self-help. The stock now trades near €48. On a simple calculation, that is 2.4x, a 140% gain (a comparison of price levels, not a realized return).

What's interesting is not that these two decisions took completely different approaches, but that they share the same process. When fundamentals deviate from expectations, Oakmark follows a procedure that asks again: "What changed? How does that affect our intrinsic value estimate? Has the range of potential outcomes widened?" If needed, it swaps in a new analyst (OpenText) or runs a formal bear-case process. The "Devil's Advocate (DA)" process that Bayer passed through is exactly that: a full bear-case evaluation → analyst rebuttal → formal team-meeting debate → vote. An analyst can trigger it themselves, or it can be triggered mandatorily when fundamentals deviate significantly from expectations. Edenred, which passed through the same process, moved +13% in local-currency terms after the hold decision. To sum up: the process is the same, but depending on how the facts changed, the conclusion can be an exit or a hold.

Of course, there is a limitation: this is a record in which the manager shows its own judgments in the most flattering light. The -19% and the 140% are hindsight results, and the exact timing and prices of the exit and the hold were not disclosed. Even so, I think the very fact that the two cases are placed side by side is the most valuable part of this letter — you can read the manager's intent to show, on its own record, that the sell and the hold were products not of luck but of a repeatable process that re-verifies the thesis whenever the facts change.

Let me add a few more numbers. As of September 30, 2026, the Oakmark International Fund held 59 stocks, with a forward P/E of 13.2x (comparisons: MSCI World ex USA 14.8x, the value version of the same index 12.4x), ROE of 14.5% (14.9%, 12.7%), and one-year consensus EPS growth of 20.0% (9.3%, 5.6%). Versus the value index, it trades at a 15% discount on FCF while carrying higher EPS growth and ROE. Emerging-markets exposure is 8.7%. These numbers read like a report card for a "cheap and good" style, but today's protagonist is not this spread itself — it is the process that has driven these judgments.

Conclusion

Oakmark's latest letter is, in truth, closer to a record of how judgments get reworked than a boast of "how well we did." If Bayer is the confident hold, perhaps that decision came not from a prophecy that the stock would rise 2.4x, but from formally re-asking the open questions — litigation, business quality, execution — through a structured process and waiting for the answers. Conversely, the OpenText exit is not a record of failure but a record of a process working: one that separates the "seemingly cheap" from the "structurally cheap." Same stocks, same question, different conclusions — I believe the difference was made not by affection for a name or by how long it had been owned, but by the existence of a process that reopens the judgment every time the facts change. Reading the letter that way, I think, is what makes it stay with you longest.

• Disclaimer: This post is informational content summarizing filings, analysis, and the author's views; it is not a buy or sell recommendation for any specific stock.

References

• Harris \| Oakmark, "3Q 2026 International Equity Market Commentary — A conversation with shareholders," September 30, 2026, Tony Coniaris — https://oakmark.com/wp-content/uploads/sites/3/2026/10/A-conversation-with-shareholdersInternational-market-commentary3Q26_vF.pdf