Buffett Harvests Tax Losses While Letting Winners Run And So Should You

Summary

• Selling losing stocks and letting winners ride is a successful investment strategy.
• Selling losers in November and evaluating them after 30 days can be a disciplined approach.
• Selling losers is easy and can help offset capital gains and earned income on IRS forms.

Are you trapped by your past success? My suspicion is that many successful investors are. My own basic principle for selling stocks goes back to an adage which I first heard in the 1950s: sell your losers and let your winners ride. There are two principles behind this adage. Principle number one is that you save on taxes, an outcome which doesn't need to be explained. Principle number two involves admitting that so far your purchase decision has been wrong. That's W R O N G wrong. Don't bother with making excuses. Just admit it before a small mistake becomes a large mistake. Most investors would do well to follow these two principles with persistent discipline.

I am always on the lookout for stocks in my portfolio which have dipped into the red. Sometimes I sell immediately but I always pick a day around Thanksgiving and look over my portfolio to see if any tranche of any holding is under water. If there is such, I sell anything in my portfolio which is in the red. It's a discipline I have never regretted. I tell myself that I can always buy the stock back thirty-one days later if I decide after a month that my original premise was correct after all. Buying back within 30 days turns the two transactions into a "wash sale" in which the tax loss is disallowed. The only time I don't sell right away is when I bought the stock so recently that any little dip is probably just noise. Thirty days is about right for clearing my mind and deciding whether I want to own the stock after all. It doesn't bother me if the stock has bounced back up a point or two. That's the price of following a disciplined approach.

For an individual investor harvesting tax losses serves in two ways. The first is taking the $3000 income reduction on page one of your 1040 (an advantage that corporations don't share). It also reduces your state and local taxes by reducing taxable income by that same amount. The second is taking advantage of the opportunity to maintain a tax loss bank. The numbers for both short and long term losses are calculated on the final page of IRS Form 1040 Schedule D where they are added to past short and long term losses. My tax loss bank from the 2008 Mortgage-Backed Securities crash, mostly from selling financials, lasted me for years. I repurchased most of them in 31 days or replaced them with similar stocks.

There hasn't been a year since 2008 that I have not been able to take that $3000 deduction on page 1 of my 1040. The most powerful use of a tax loss bank is the ability to sell a position which is entirely in the black without paying any capital gains taxes or letting the fear of taxes influence your decision. A good example was dumping Alibaba (BABA) and other Chinese positions at a significant profit in 2020 and 2021 as I decided that I was done with China for this lifetime. Luckily, thanks to my tax loss bank I was able to exit China forever with major capital gains without worrying about taxes. That's the major advantage that derives from tax loss harvesting.

I could end this article here and tell myself I have covered the major topics, but there are a few subtleties worth discussing. The first is the way tax loss harvesting shapes your portfolio. Tax loss harvesting tends to leave you with a portfolio that is entirely in the black - deeply in the black. For the most part that is a good thing. The second is the approach of Warren Buffett.

The Shape of Buffett's Berkshire Portfolio

Selling losers tilts the odds in your favor and reduces the risk that you will lose in a major way. And here's a surprising side effect: after a few years you are left with a portfolio in which everything or almost everything is in the black. Not only are your stocks in the black but as time passes they get deeper and deeper in the black. You have pruned the stuff that didn't work and what you are left is a portfolio of successful businesses with a cost basis that keeps getting lower and lower compared to the market value of your stocks. You learn that if you own conservative quality companies - low debt, good return on equity, consistent CAPEX - you are very likely to compound your money at a good clip.

The one catch is that more and more and more of your portfolio consists of embedded capital gains. If you look regularly at the table of the largest Berkshire Hathaway (BRK.A)(BRK.B) stock holdings, usually found somewhere in Buffett's Annual Shareholder Letter which will come out this coming Saturday February 24 (but note that the table was missing in the 2022 Annual Letter), you will get an interesting picture which yields important pieces of information. The table below is a slightly reduced version which does not include the percentage of the company owned. All prices are as of December 31, 2021, and the list is in alphabetical order:

With a little attention to the last two numbers on each line you will get an education on how the large positions in a portfolio come into being. The first thing needed is a period of time. The second thing needed is a company which at some point in its history had a significant amount of growth. The two major stars on the list are Coca-Cola (KO) and American Express (AXP) both of which have been in the Berkshire portfolio for more than two decades and both of which consist of about 5% cost (tax basis) and 95% embedded capital gains. In this February 14 John Vincent piece on Berkshire's portfolio they are listed among positions under the heading KEPT STEADY (no buying or selling) along with Vincent's comment that Buffett had once described his holding period for the two of them as "permanently."

Think for a moment about a few implications. For one thing KO now has a yield on cost around 60% while AXP is just under 50%%. Buffett once mentioned casually that he probably should have sold Coke in 1998 when it was one of the mundane stocks which followed the dot.coms to an absurd height. It didn't return to its price at the 1998 top until 2016. The reason he won't sell now is less a love story than a matter of numbers. Remember that both KO and AXP have a market cap consisting of 95% cap gains meaning that, at the 21% corporate tax rate, each dollar sold reduces Berkshire's capital in KO by about 20%. Selling his entire position in Coke would produce a $4.7 billion dollar gift to the IRS. That's a number Buffett would surely find intolerable. Corporations don't crush the S&P 500 (SPY) at a ratio of two to one by making gifts like that to the IRS. The actual long term performer on the list (now overpriced with a 34 P/E) is Moody's (MCO) which now consists of about 97.5 per cent capital gains and thus yields about 140% on cost.

There's another manipulation of these numbers that merits some thought. Let's stay with the 2021 table as the same principle would apply to data from 2022 and 2023. As stated above, selling the Coca-Cola position would reduce Buffett's total capital by about $4.7 billion. Looking at it a slightly different way, it would create a breakeven point 20% below the current stock price of $60 per share, or $48 dollars per share. A smaller decline in the stock price would mean the sale wasn't justified as a decline to something below $48 would be required for the elimination of KO stock to be a positive decision. Then there's the dividend of about 3.25%. To generate the same annual return after taxes reduced total capital by 20% would require a yield of 4.06% with the same safety and promise of future increases as Coke. A similar hurdle exists for American Express, Moody's, and Charter Communications while Apple may be headed in that direction counting both dividends and share buybacks as shareholder returns.

The Berkshire Hathaway portfolio in the aggregate looks like the table above. The reason is that the table contains the bulk of the Berkshire portfolio market cap. The only loser in the table is Verizon (VZ), a sure goner at the time it appeared in the table. (It has now been sold.) The remaining stocks in the table seem unlikely to be sold any time soon although I suspect that Buffett is not any more pleased with BAC than I am (the board was displeased enough to cut Brian Moynihan's pay probably because BAC was caught flatfooted with long duration bonds as the Fed raised rates). Unlike me, Buffett has small tranches of his Bank of America which he could sell at a loss and trim his position a little. His recent trim of 10 million Apple (AAPL) shares, for instance, was probably around break-even. His Apple position now consists of more than 80% capital gains and Buffett most likely plans to hold on to it and watch it become the next Coca-Cola rather than reward the IRS for doing nothing.

My Own Portfolio Differs From Berkshire's In Nuance

Once you get into the habit of doing it, selling your losers is easy. There's really no way to lose in a major way. One psychological effect comes from the fact that after a few years you are left with a portfolio in which everything, or almost everything, is in the black. It derives from a single principle: selling losers. It doesn't mean that the positions have been up every year. It just means that no tranche of any holding is selling for less than I paid for it when Thanksgiving comes around. It's a rule. I'm a fanatic about it. A few years ago I went so far as to use the Specific Identification Method to sell every dividend reinvestment share that was trading in the red. My tax accountant thought I had lost my mind, but there were several positions with dividend reinvestment shares in the red and it added up. It's a rare year that everything I own is up - that's an intentional effect from serious diversifying - but in the aggregate the total is up over any period longer than a year or two. I'm happy to settle for that.

Investing is just what the word says it is. It is taking an ownership position in something with predictable growth at a good price. There are no gimmicks which improve upon this simple approach. Sometimes your initial view turns out to be mistaken and in those cases you sell and take the loss, which can be used not only to offset any selling of positions with embedded capital gains but also to offset up to $3000 of earned income on page one of your IRS form. It's one of the great no brainers.

In general I am quicker to pull the trigger on losers than Buffett. That's the "nuance" mentioned in the subhead. There are lots of reasons to sell a stock and lots of reasons that the time is right. I suspect Buffett sees it that way. One reason is that he is more concerned with dividends than I am. He had to watch and think for a while to be sure he wanted to get out of Verizon and HP Inc. (HPQ) (neither of which I would have bought in the first place, to be honest, again because of my lower level of interest in dividends).

The same was not true of Paytm (PAYTM) and StoneCo (STNE), which were a Todd Combs purchase for $300 million each. It was just a matter of trying to be patient while two you young companies revealed their value. I assumed Todd Combs had some special insight which I had missed beyond the obvious fact that they were based in India and Brazil, two very large markets which did not at the time of the purchase have advanced payment vehicles. Combs threw in the towel in Q4 2023 as recorded under STAKE DISPOSALS in the earlier link to the February 14 John Vincent article. By that time both PAYTM and STNE had suffered a sustained period of miserable performance. My policy of selling everything in red pretty much immediately would have saved some regrets. They were losers. They would have been goners by Thanksgiving.

Red ink is all it takes to warrant a sell. Reasons to sell a winner are more complicated.

Why An Investor Might Sell A Winner

There's no firm principle about selling winners. There are many prompts which might force you to think about it, but in the end it's a judgment call. How strong does the case need to be? I should say that I usually take no more than three or four actions of any kind in a year, often fewer, and the majority are sales. Your portfolio does best when you are doing nothing, just reading anything that might imply an important change and watching from time to time as the stocks you own continue to compound money. SA Quant rating shifts have helped a bit when I'm on the fence. Recessions and bear markets are generally not things that call for selling anything. If you made good choices, the stocks that go down will come back up and go to new highs. One of a kind business problems also don't require selling unless there is a reasonable probability that the problem is long term.

Here are some reasons that might justify selling:

• Something has clearly gone wrong with the company in a sustained and major way. My current leading candidate is Johnson & Johnson (JNJ). Ten years ago I wrote this article comparing JNJ with Parker-Hannifin (PH) saying that PH had some brief vulnerability to recessions which JNJ did not, but JNJ tended to have regular self-created recessions with some ridiculous recall followed by class action suits which should have cost top management their jobs. This problem has continued, alas. Despite the JNJ-manufactured single-company recessions I am up about 150% in the stock - less than half what I am up in Parker-Hannifin - which nevertheless means that 60% of my JNJ position consists of embedded cap gains (taxed at the max possible rate, higher for individuals than corporations). I should have swapped JNJ for Parker, which is my biggest long term winner (though currently expensive). As to JNJ I have stood on the fence for too long on this problem and may soon pull the trigger and use my tax loss bank.
• Something has gone wrong with the industry. Buffett has said that he doesn't like banks despite a huge past holding of Wells Fargo (WFC) and a current huge position in Bank of America. Early in 2023, as poorly thought out strategies were catching up with a few regional banks, Buffett kept BAC which was making the same mistake with some additional wrinkles but sold Bank of New York Mellon (BK), U.S. Bancorp (USB), and Marsh & McLennan (MMC). BK was a slow grower and USB had weak results and poor stock price action in recent years. MMC was doing fine along with most of the insurance business but had become quite expensive. How about the Citi (C) position? It's a sick puppy with a good and fairly safe dividend and the hope would be that halfway decent management might make it get a little better. The first corrective moves by a messed-up company are often so easy and obvious that riding them from awful to merely bad can be profitable. Buffett built his tremendous success with the Buffett Partnership (1956-1969) by using this principle. Not my game, though.
• Recessions, bear markets. Never mind them. If you own good companies they will likely recover. If you maintain a good amount of cash you can buy some new businesses which are leaders as we emerge from recessions and bear markets.
• Your portfolio of stocks is too overweight. You are too overweight in a particular holding. As I get older I am increasingly skeptical of this rationale for selling. If a holding is very large because of long term success you made a good decision in the beginning. Because I write frequently about Berkshire Hathaway I hear many stories from commenters whose Berk position has grown outsized in a big way. None are sorry. Very few ever bothered to trim. All just realized at some point that they were onto something really fine. Until something very fundamental changes, why worry about being overweight. What would you replace it with?

Most other things that would imperil good companies would be likely to make market problems the least of your worries. Buffett has said and written in Shareholder Letters that the only thing that could really harm Berkshire Hathaway would be a nuclear event. He seemed worried enough at the 2020 Annual Meeting that he might now add a more destructive global pandemic. The trouble is, you just can't plan for them.

No Hedging Your Bet With Call Writing

Until a few years ago I used to write calls against perfectly good holdings - winners selling in the black. I would buy the calls back a week or two before expiration. Being greedy and overconfident I wrote them a strike or two in the money. This strategy worked brilliantly until it didn't. At one point I had written calls against an already successful position - Parker-Hannifin as it happened - which has a way of running up too far when things are going especially well. With several weeks until expiration I woke up one morning to discover that my calls had been assigned. Supposedly "assigned" options are chosen by a lottery of some kind, although that may be an urban legend. This had the awful effect of losing what eventually became the best stock in all family accounts as well as paying capital gains taxes on the position, the thing I was most determined to avoid. Luckily, the PH position in my portfolio was relatively small while my wife's portfolio had a much larger position. Fate needed to punish me alone and remind me of it regularly every time I notice her current value in Parker (which is now about 85% embedded capital gains). I should have bought my position back but I couldn't get over my disgust.

That was the last time I used an option position to hedge or try to add a small amount to returns. There is one rule I follow: selling all stocks in the red by early December is mandatory. I give myself no leeway. There's no sitting on the fence by writing a call. When a position is under water at that point I was wrong to buy it, at least about the timing. The market has sent me a message to sell. That's my suggestion for readers. If you feel you have to sell a winner, try to have a tax loss bank in place to offset the capital gains. Good investing!