Geoff Gannon February 23, 2020

How Buffett Holds: The Incredible Importance of the “Contrasting Trajectories” of Long-Term Winners and Losers

In my first article about Warren Buffett’s annual letter to Berkshire Hathaway shareholders, I talked about how difficult it would’ve been to hold Berkshire stock during all the years when it rose so much so fast. That’s one underappreciated part of how successful Berkshire has been as an investment. Buy and hold often sounds like a simple strategy to follow. Berkshire has returned 20% a year compounded over more than 50 years. It would’ve been easy to hold the stock if it rose 20% a year every year. But, sometimes it got far ahead of itself – jumping 100% or more in price in a single year. And then, other times, the stock price lagged the intrinsic value gain for several years in a row. But, the long-term trend in Berkshire Hathaway’s results was one of compounding at about a ten percentage point a year advantage over the S&P 500. Today, I want to talk about a different underappreciated aspect of Berkshire’s compounding. Yesterday, we talked about how uneven the compounding was over time. Today, we’re going to talk about how uneven the compounding is in terms of sources.

Berkshire’s results are fueled a lot by its insurance operations. As an insurer, Berkshire tends to turn an underwriting profit. This gives Buffett a cost free form of money called float. Berkshire uses some of this float to invest in stocks and to buy entire businesses. The company uses its retained earnings to finance the rest of these two portfolios – one made up of private businesses 100% owned and the other made up of minority stakes in publicly traded companies. The underappreciated part of what Buffett does that we’ll be talking about today is exactly how he buys these businesses and these stocks. How he buys businesses is pretty normal. Most acquisitions done by big U.S. companies are done the same way. You buy basically 100% of a company using a combination of debt you raise, cash you have on hand, and maybe (Berkshire does this only occasionally) shares of your own stock. You make the purchase at one single point in time. Sometimes you might draft some kind of earn-out agreement where the former owners (who often stay on to manage the business you now own) make more money if the business they sold to you hits certain targets over the first few years you own it. But, that’s a very small part of the overall purchase price. We can simplify this by assuming you pay “x dollars” for the stock upfront in cash and then hold the business forever. That’s how most acquisitions work.

What a lot of investors don’t appreciate is that Buffett runs his stock portfolio a lot like he runs his stable of 100% owned businesses. Buffett buys and holds his shares in a company in a totally different way from almost any other investors out there. And this has some pretty big implications when it comes to just how the compounding of his stock portfolio works. Basically, Buffett isn’t joking when he talks about how he buys his stocks the way he’d buy entire businesses. That’s not just a metaphor for him. It’s literally how he makes many of his purchases. For example, we can see in Berkshire’s 1977 letter (this is the oldest letter on Berkshire’s site, if we went back further we’d see the same info) that the company’s Washington Post investment had a cost of $11 million. Thirty years later – in 2007 – Berkshire’s cost basis in the Washington Post was still $11 million. During this time, the market value of the Washington Post shares Berkshire owned rose from $33 million to $1.4 billion. However, the cost basis stayed the same. What does this tell us? Berkshire didn’t buy more Washington Post stock during this time. It also didn’t sell out of the position, trade around the position, etc. Buffett just acquired a minority position in the Washington Post in the 1970s and then kept that same position – without adding or subtracting from it – for more than 30 years. That’s not how traders think. It’s not how most investors think. It’s not even how most long-term investors or even investors who say they are “buy and hold investors” think. Most people I know who refer to themselves as buy and hold investors mean they stay in a stock for a long time in some form. But, they might add to that stock each month as they save more money in their retirement account. They might trim back the position to get it more in line with the original percentage of their portfolio they devoted to the stock. In other words, they might “rebalance”.

Buffett doesn’t rebalance. He doesn’t rebalance the 100% businesses he owns. And he doesn’t rebalance stocks in his portfolio. He buys a lot of Coke stock basically once. Buffett has sometimes added a bit more to stocks like Coca-Cola, Wells Fargo, etc. over the years. But, if you look at Berkshire’s history of buying stocks – Buffett is remarkably one-off in both his entry into a stock and his eventual exit (if he ever exits). The other investment managers at Berkshire might work differently. But, the biggest difference between Warren Buffett and most people reading this article is the way he both buys and sells. Buffett doesn’t – at least not at Berkshire, we don’t have as detailed info on his partnership years – tend to buy a stock slowly over a very, very long period of time. He definitely doesn’t rebalance his stock portfolio by trimming back winners and buying more of losers. If a stock starts out as 10% of his portfolio and outperforms his other holdings – it becomes 30% of his portfolio. And he lets that happen. But, just as importantly – if Buffett puts 10% of his portfolio into a stock and it turns out to be a real loser, that loser just vanishes from his top holdings table without him ever needing to sell it.

Now, Buffett does sometimes sell losers. We’ve seen him do that. He will never sell a loser of a 100% controlled business. But, he will sell a loser as a stock. However, if you are adding more and more money over time to your investable savings – as Buffett is at Berkshire Hathaway – and you aren’t just automatically buying more of the stocks you already own to keep to some set portfolio weighting, your losers will become a smaller and smaller percentage of your portfolio. Imagine you have $100,000 in savings and add $10,000 in savings during the year. At the start of the year, you bought two 20% positions – we’ll ignore the other positions you bought for this illustration – where the stock with ticker “DBL” went on to go from a $20,000 market value to a $40,000 market value. The stock with ticker “HLV” fell 50% from a starting market value of $20,000 to $10,000. What relative importance will the returns in these two stocks have in year 2? Well, at the end of year 1 you’ve added $10,000 to your account and one stock had a gain of $20,000 and another stock had a loss of $10,000. We’ll assume your other 3 positions were a total wash. So, that leaves you with a portfolio at $120,000 at the start of year 2 ($100,000 + $20,000 + $10,000 = $130,000; $130,000 – $10,000 = $120,000). Now, the poorest performer you had was HLV which dropped from a market value of $20,000 to $10,000. And $10,000/$120,000 = 8%. Without selling a single share, you’ve cut your allocation to stock HLV from 20% to 8%. At the same time, your best performer was “DBL”. It increased in market value from $20,000 to $40,000. A lot of people would sell this stock now. But, you’re a buy and hold investor. So, you’ll keep it in year 2 when its new portfolio weighting is 33% ($40,000/$120,000). Last year’s biggest winner is now 4 times more important to your portfolio going forward than last year’s loser.

A lot of value investors might rebalance these. Buffett doesn’t. Clearly, he doesn’t in the case of 100% owned businesses. No one does that. That’s why any serial acquirer ends up heavily weighted to two kinds of acquisitions – past acquisitions that were very successful growers and recent acquisitions that were of a very big size – and very insignificantly weighted to one other kind of acquisitions (unsuccessful acquisitions from long ago). You’ll notice this in the long-term record of companies that acquire lots of other companies. If they did a deal a long time ago and that business unit never really grew much – it just gets dropped from discussion in presentations, annual letters, the 10-K, etc. It doesn’t get its own segment reporting. It doesn’t have to be sold. It just slowly gets put so far in the distance to be made a speck by the simple process of compounding happening at the other growing business units.

I was very impressed by Buffett’s discussion of this phenomenon in his 2019 letter. Now, he was talking about 100% owned businesses. But, the same force is at work – somewhat lessened by Buffett’s willingness to completely eliminate his positions in true losers when they are stocks, not 100% owned businesses – in Berkshire’s stock portfolio.

Here’s what Buffett said:

“…the fallout from many of my errors has been reduced by a characteristic shared by most businesses that disappoint: As the years pass, the ‘poor’ business tends to stagnate, thereupon entering a state in which its operations require an ever smaller percentage of Berkshire’s capital. Meanwhile, our ‘good’ businesses often grow and find opportunities for investing additional capital at attractive rates. Because of these contrasting trajectories, the assets employed at Berkshire’s winners gradually become an expanding portion of our total capital.”

This is only something that happens to buy and hold investor’s portfolios. It doesn’t happen if you actively trade around positions. It doesn’t happen if you rebalance. It happens because a buy and hold investor allows the “contrasting trajectories” of the underlying businesses in the portfolio to naturally shift the percentage of capital employed in one business into another.

Let me use a longer term example of what could happen with those two stocks – “DBL” and “HLV” – over a period as long as one decade. Let’s say you tend to be about a 10% a year type investor in the sense that the average business you pick out to own in your portfolio will tend to compound at 10% a year while you hold it. Okay. But, you’ll have some winners and some losers. There will be quite a lot of dispersion in your results around that central tendency of buying 10% a year compounders. However, unlike most investors – you won’t sell your stocks before 10 years is up. You also won’t sell your winners to buy more of your losers. Nor will you employ any kind of rebalancing. You’ll just buy however much you want of stocks “DBL” and “HLV” this year (2020) and keep holding them ten years from now (in 2030). In this case, I’ll imagine you aren’t saving more and adding that cash to your portfolio each year. If you did, that would water down the concentration in “DBL” but also water down “HLV” further than what I’m calculating here.

Let’s imagine “DBL” is an especially good compounder doing 20% a year for 10 years. A great pick by you. But, “HLV” is a laggard. It’s not a disaster. But, you misjudged it. The stock compounds at just 5% a year for the next 10 years. You have a 5-stock portfolio. You start with 20% of your portfolio in DBL, 20% in HLV, and 60% in 3 other stocks that are all truly “average” picks by your standards.

How much of your portfolio will be in DBL at the end of 10 years?

How much of your portfolio will be in HLV at the end of 10 years?

Well, $10,000 invested in DBL will become $62,000 in 10 years.

Meanwhile: $10,000 invested in HLV will become $16,000 in 10 years.

The rest of your portfolio will go from $60,000 to $156,000.

Your overall portfolio has grown from $100,000 in 2020 to $234,000 in 2030 (a 9% a year CAGR).

But, what I’m interested in is how much of your compounding in the 2030s will be driven by each of the stocks you bought in 2020. You started out with 5 equally weighted 20% positions. But, going into your second decade – your portfolio is now 26% in DBL and only 7% in HLV. DBL will have nearly 4 times the influence on your compounding in the 2030s as HLV will. And yet you didn’t make any active choice about either stock. You just let DBL outrun HLV by so much over those 10 years that it became a progressively bigger part of your portfolio.

If this same difference in compounding – 20% a year versus 5% a year – was sustained for a full 20 years, your winning position would be 14 times larger than your losing position. If you were running a fairly concentrated portfolio, adding more to your savings each year, and had even just a couple more big winners – that “losing” position would seem to simply vanish into unimportance in your brokerage statements. You wouldn’t have to sell a share. Your mistake would just gradually erode over time.

Is this a good strategy though?

Should you bet more and more on stocks with the highest CAGRs in your portfolio and less and less on the stocks with the lowest CAGRs in your portfolio?

I can’t answer that question.

Obviously, that is how fortunes are made. It may not be how jobs are kept though. I wouldn’t recommend it to the average mutual fund manager. But, no business or investor has grown their net worth to a tremendous size by rebalancing their winning businesses and losing businesses over time. It’s also kind of inefficient in other ways. But, even before worrying about taxes and fees and such – equal weighting your bets to start and then not adding to them or cutting them back would be the approach that seems closest to how super successful investors built their records over time.

It’s certainly the way Buffett did it. He tends to buy a lot of a stock till he stops. And then he holds that stock till he decides to eliminate it. He clearly does not target some sort of percentage allocation to various stocks. If Apple looks cheap to him, he’d buy it whether it was 10% of his total stock portfolio, 30%, or 50%. And, if it didn’t look cheap, he’d just keep holding it without rebalancing it with the other stuff he owns.

The section where Buffett discusses how the errors he made in buying 100% controlled businesses have been overwhelmed into percentage insignificance by the compounding of the successes he had is my favorite part of this year’s annual letter.

It’s definitely worth a reread. And it’s something I’ve been struggling with now that I run both a hedge fund and separately managed accounts. The fund structure allows you to buy like Buffett buys. And that certainly feels more natural to me. The separately managed accounts don’t really allow you to do that. And so you are really targeting percentages of the portfolio to allocate. That makes a lot of sense to a lot of investors. It seems to be the way most professionals think. But, it’s not the approach that seems most intuitive to me.

You can certainly buy the same kinds of stocks as Buffett does in any structure. You can buy what he’d buy and rebalance it. You can buy what he’d buy and trade around it. But, will you get similar results? Is Buffett’s success as an investor due to which stocks he picked? Is it due to how much he put into each of those stocks? Or, is it due to the fact he leaves both his winners and losers alone far more than the average investor does?

Probably some combination of all three. But, it’s usually only the first point – what to buy – that gets discussed. How he holds his stocks is rarely discussed.

Reading this year’s Berkshire shareholder letter gives us a chance to look at the way we hold the stocks we’ve bought and look for ways we can improve our “hold process”. Do we really want to be more like Buffett? Do we have the stomach for it? Or, do we need to constantly be tweaking our position sizes to feel like we’re earning our fees?

Become a Focused Compounding Premium Member

Get a Free Weekly Email from Geoff (enter your email in the sign-up box)