This week, I posted the following on X:
Investing is mostly about the coexistence of the never-sell philosophy with the acknowledgment that very few companies deserve to be forever in a portfolio
While the post itself did not require much thought, it sparked an internal discussion about the concept of long-term investing. My objective in this article is to share my thoughts on the topic and outline several actions that I am implementing in my investment process. I understand that other long-term investors may criticize or disagree with parts of this article, but I don’t believe that what I am about to share here contradicts a long-term investment strategy. Hopefully, you understand why after reading it.
A natural segue into the article is to discuss the concept of never sell. The “never sell” philosophy stems from the concept of compounding. Although there’s no way to verify that Albert Einstein said that “compound interest is the 8th wonder of the world”, the person who attributed this quote to Mr. Einstein was onto something. Explaining compounding is as easy as saying that Warren Buffett has amassed over 65% of his net worth after celebrating his 80th birthday. As the snowball gets larger, it collects much more snow with every roll.
The interesting thing about compounding is that, despite everyone understanding how it works, only the truly patient, long-term investors will ever enjoy it, and waiting is hard in a dopamine-driven world. Compounding is a mathematically straightforward concept, but a challenging one to experience. Most humans have trouble envisioning exponential curves, and that’s precisely what compounding is all about.
So, what does compounding have to do with the concept of never-sell? The latter tends to be a necessary philosophy to enjoy the former. If compounding can only be experienced over more extended time frames, it goes without saying that an investor would likely need to hold a position for a very long time to experience it. An investor should consider compounding the value of their portfolio, rather than focusing on the individual companies within it. However, constantly buying and selling companies is an inefficient way to compound value over the long term, primarily due to taxes (I crunched some numbers in this article). This means that the thought of compounding the value of a portfolio begins with the idea of compounding the value of its individual constituents.
Let’s put some numbers around this. If you are looking to achieve a 100 bagger, you’d need to hold a position that compounds at a 15% CAGR for over 30 years. While aiming to hold great companies for the very long term is definitely a good starting point and probably good advice, I would say that thinking about it in black-and-white terms is most likely a mistake. A long-term investor should, in principle, consider holding a position indefinitely when they include it in their portfolio (to maximize the potential for tax-free compounding), but this should be an aspiration, not an objective. Things rarely turn out to be as we expect them to in such a dynamic world, so having never-sell as an objective can lead to subpar outcomes.
This ultimately means that an investor should always remain flexible to change their mind and be ruthless to sell companies that no longer comply with their criteria. This is hard primarily because most investors are not great at admitting they’ve made a mistake or in throwing hours of work to the trash (I don’t believe this is necessarily the case due to pattern recognition, but that’s a topic for another article). Investors should be aware of the following when thinking about selling:
Making mistakes is part of the game
Most companies don’t deserve to be in a portfolio forever
Let’s briefly talk about #1. Very few investors have achieved hit rates above 50%, which means most of us must be prepared to make one mistake for every two decisions (at least). The good news here is that, due to the asymmetric nature of stock returns (capped to the downside but uncapped to the upside), one can still do well while making quite a few mistakes. Later on, I’ll explain how one can do even better by being ruthless about limiting the magnitude of those mistakes.
The second point is probably less straightforward. If an investor acquires shares in a given company and these shares perform well for a while, the likelihood that they will continue to outperform over the long term is slim. There are very few companies that have compounded for decades within a base of thousands of publicly-traded corporations. Let’s try to look at this with some numbers. According to Henrik Bessembinder, approximately 4% of all publicly traded stocks have been responsible for nearly all the value creation in the stock market since 1926. While this might sound like “just” a number (and there are criticisms of Bessembinder’s work), I believe it carries some interesting implications.
First, it portrays compounding to perfection and the asymmetric relationship between the upside and downside of stock returns. Good companies that have compounded for decades have grown significantly, and as a result, have played a substantial role in driving stock market returns. Consider this: if Big Tech doubles from here, they would be responsible for a significant portion of value creation going forward (at the index level). Even if a smaller company were to compound at 10 times the rate of Big Tech, it would be relatively immaterial to index returns due to its smaller size. This means that one can find market-beating returns outside of the largest companies, but that absolute value creation belongs to the largest corporations just due to past compounding.
It also suggests that finding long-term winners and therefore stocks that deserve to be held for decades is not an easy task. As optimistic investors, we are inclined to believe that our portfolio holdings fall within the 4% group of long-term winners. One can definitely increase the probability of success just by being selective in terms of what characteristics to look for in a company. For example, excluding commodities and airlines would likely increase that 4% by a significant margin. Still, it doesn’t change the fact that the probability that our portfolio will be made almost entirely of these long-term winners is very slim (if not zero).
Another way to look at probability is by examining Michael Mauboussin Base Rate Book. I think we can agree that stocks that have performed exceptionally well over extended periods will mostly have enjoyed strong earnings growth. According to Mauboussin, only 7.5% of the companies in his universe were able to sustain a 10-year net income growth CAGR above 20%:
If you need to justify the price of a given stock by projecting a net income CAGR of 20% over the next decade, be assured that, though possible, probability is against you. I would argue, though, that scaling a business today is much easier than it was in the era of physical assets, but this also means that competitors can also scale faster. The conclusion is relatively straightforward: finding long-term compounders is very tough.
Therefore, the fact that compounding requires a very long holding period to bear fruit, coupled with the difficulty of finding long-term compounders, can only mean one thing: investors are constantly navigating a very thin line between selling a potential winner and selling a potential underperformer. However, the laws of probability strongly suggest that with any sale, an investor will most likely be selling a future underperformer. If probability is on our side, why is it so hard to sell, and why is “never sell” so popular? Due to the asymmetric nature of stock returns. Selling a stock that belongs to that “select 4% group” is far more costly than selling an underperformer too late. To this, we must add that selling is always emotionally challenging.
I believe this has led many to think that “never sell” is the optimal option, as you’ll do well over the long term thanks to the asymmetry inherent to stock returns. My only pushback to this thought process would be that it assumes there is indeed a massive long-term winner in the portfolio, which is not necessarily the case.
As I have repeated many times, selling is arguably the most challenging part of investing. Not only because knowing when to sell is tough, but also because most companies are worth selling but selling one that doesn’t deserve to be sold can be very expensive. This, however, should not lead to a situation where selling is seen as taboo. As I see it, a strict never-sell strategy is more a reflection of laziness and ego than a sound investment strategy.
Even if we are great investors, chances are that we will have to sell many companies throughout our investing career, so better to be mentally prepared for it and face the reality. Here’s some data from arguably the most famous long-term investor of all time: over the last 5 years, Berkshire Hathaway has fully exited 5 positions. Is Warren Buffett a worse investor for exiting positions? No, he’s arguably a better investor for it. When Buffett is right, he’ll hold for decades, but he will have no remorse in selling an underperformer. Nobody said this “game” was easy!
When to sell
So, here comes the key question: when should we sell a position? While the objective of this article was to demystify never sell, I also thought it’d be worthwhile to share my thoughts on selling. My sales mostly fall into three buckets (#1 and #3 are somewhat related):
Not comfortable with the new thesis
There’s a better opportunity across the portfolio
I’ve made a mistake
I rarely sell for valuation reasons alone because I tend to invest in companies with positive optionality. Optionality is challenging to value, and since I don’t typically own companies trading at nosebleed valuations (although they may be expensive at times), I prefer to stay put during periods of relative overvaluation. I would have done better if I had adhered to what I've just outlined here. Not long ago, I sold a chunk of my Intuit position only to see it climb to ATHs after management announced a new enterprise offering that could potentially be comparable in size to the current SMB offering.
Another key question worth pondering when considering valuation is…what is a nosebleed multiple? 45x? 40x? Where’s the limit? Much tougher to answer than it might seem at first glance. This said, selling for valuation reasons is somewhat included in the second scenario, as there might be a better opportunity across the portfolio that triggers the sale. I don’t consider this to be the same as selling solely for valuation reasons, as in this case, the money would stay still until we find a new opportunity.
Number 1 is interesting because even though many long-term investors argue that a thesis can’t change in a given quarter, it can. I do agree that such a case will be rare, but this doesn’t mean it’s impossible. Talking to a great investor this week made me realize that it might be a good idea to write down the initial investment thesis very clearly, along with KPIs, to understand if I am suffering from thesis creep. I believe that most theses are pretty likely to “shift” over long periods in such a dynamic world, but at least an investor should understand how it has moved and judge whether they remain comfortable with it.
Number 3 is arguably the hardest because we have to acknowledge that we have made a mistake, and this becomes a very emotional decision after having pulled the trigger. The problem with this being emotional is that we can let a mistake grow larger and larger. I’ve had many cases where I've seen a problem developing in a company and have given it sufficient time to escalate from a minor issue to a huge, irreversible one. It becomes increasingly emotional as the stock price drops, so it becomes incrementally harder to acknowledge there’s a flaw in the investment thesis.
I explained a few weeks ago that, to remove emotion from this process, I would sell any stock that is 20-25% below my cost basis. After speaking with the same investor I mentioned above, I have decided to change this requirement to selling it when it underperforms the benchmark by 20-25%. This should help me avoid selling through broad based drawdowns.
While this strategy might seem extremely short-term oriented (I know I would’ve thought this way a couple of years ago), I feel that it’s quite the opposite. First, it allows me to take away emotion from the selling decision. Second, I basically don’t lose anything by selling and later evaluating whether I should buy a larger position or simply leave it as is. What seems unacceptable is to see a position going against you so strongly and not take any action. In my opinion, there are two options when a stock drops considerably:
Selling
Buying more
Not doing anything can also work out fine over the long term, but more often than not, it will be a mistake of complacency that allows our mistakes to escalate from minor to significant issues. This is not about timing bottoms, this is about taking emotions out of the way.
Wrapping up
With this article, my goal was to demystify never sell. Due to the nature of investing, selling will inevitably become an integral part of a long-term investor’s career, even if the objective is to own a never-sell portfolio. Statistics and common sense suggest that most stocks are worthy of a sale at some point, so it's better to accept this fact rather than ignore it.
Hope you have a great weekend,
Leandro
Good write up, thanks. Considering the biggest mistakes great investors have made in their careerers, and the asymmetry in missed returns from a great compounder vs a mistake, erring on the side of holding (when a company has optionality and a decade+ runway ahead) is my default.
I've also this year started labeling a few positions I hold as "permanent" positions (my highest conviction ones which I've held usually for a few years). Whilst this label can change if the competitive dynamic changes for them, having a few labelled like this helps with my decision making process of what to trim for a new opportunity, or to hold and not trim when these "permanent" holdings get expensive-looking.
Similar to the Mungar "never sell CostCo" or Mohnish with Raysas (unless the father/son team changes)
completely agree with this approach. Selling purely based on valuation misses the bigger picture—especially when owning companies with strong optionality that’s hard to quantify. Like he said, valuation multiples alone don’t tell the full story, and defining a “nosebleed” multiple isn’t straightforward.
I also appreciate the emphasis on clearly documenting the investment thesis and tracking how it evolves over time. This helps avoid “thesis creep” and keeps the process rational rather than emotional.
Most importantly, shifting the sell rule to underperformance versus a benchmark (rather than just a fixed percentage loss) makes a lot of sense. It helps avoid knee-jerk selling during market downturns while still cutting losses on truly underperforming positions.
Overall, this disciplined and flexible mindset balancing patience with pragmatic risk management
is exactly how I approach investing too.