Over the years, I have observed that investors have a much more difficult time assessing when to sell a stock than they have assessing when to buy a stock. At least partially because of this dynamic, I have dedicated a lot of my time on Seeking Alpha writing articles that suggest when stocks should be sold. Over the past year-and-a-half, I have written bearish articles on over 50 S&P 500 stocks. But I do occasionally buy stocks if they are trading at good values, and I purchased SYSCO Corporation (SYY) on 3/18/20 when it was trading at a reasonably good valuation. I explained the process I went through to identify Sysco as a good investment in my 5/15/20 article “Stocks I Bought On The Dip: Sysco“. This Wednesday, on 9/23/20, I took profits in the stock, and the closing price that day was $60.52 (I had a trailing stop in place so I sold at a slightly higher price of $62.51.). Here is how the investment’s performance compared to the S&P 500 over the same holding period using Wednesday’s closing price:
This turned out to be a very profitable investment, returning nearly 100% in about six months’ time and significantly outperforming the S&P 500 over the same time period.
In this article, I’m going to explain my process and reasoning for selling the stock.
It’s all about likely future returns
When I invest in less-cyclical stocks like Sysco, I focus almost exclusively on expected future returns and the risk that I may not get those returns. My approach assumes that future returns are driven by two primary factors: the cumulative earnings of the business over time and the market sentiment surrounding the stock. For the business returns, I simply imagine that I buy the whole business at its current market price and that I get to keep all of the cumulative earnings for the next 10 years, then I convert those cumulative earnings to a CAGR percentage and use that CAGR % to value the stock/business. For the market sentiment returns, I assume that the market will at some point over the next decade revert to the mean P/E ratio of the last full economic cycle, and I convert the change in price if the mean reversion were to occur to a CAGR percentage also. Then, I add those two CAGR percentages together to obtain an estimate for the ‘Full-Cycle’ return over the next 10 years if we were to experience a similar cycle as the last one.
The next stage of the process is to determine what expected 10-year CAGR % I am willing to buy, sell, or hold, the stock in question. A lot of this comes down to how each investor thinks about risk and what alternative investments are available at the time. If an investor thought that a 10-year Treasury represented an equal risk to stocks (or to a particular stock), then the 10-year Treasury yield is currently +0.677%. So, if one thought Sysco stock had equal risk to the 10-year Treasury and it was the only alternative investment available, then perhaps Sysco would become a ‘sell’ if the expected 10-year returns fell below a 0.677% CAGR expectation.
Personally, I think using the 10-Year Treasury yield threshold as a ‘sell threshold’ for a stock like Sysco is too low for at least two important reasons. The first is that Sysco is far riskier than a 10-year Treasury. As we have seen with COVID, risks are hard to know in advance (and the farther we project returns into the future, the more difficult making accurate estimates becomes). Businesses can suffer from a variety of unforeseeable dangers. This makes their stocks inherently riskier under most conditions.
The second important reason we might demand a higher return than that of the 10-year is that there may be other investments an investor can make that have a guaranteed return that is higher than the 10-year. For a great many investors, paying down debt will provide better returns than owning expensive stocks that have low expected forward returns. One survey from bankrate.com this year showed the average mortgage rate is around 4.41%. One of the lessons Warren Buffett revisited in Berkshire Hathaway’s (BRK.A) (BRK.B) annual meeting this year was that if one carried high levels of expensive debt (in his example someone who was carrying credit card debt and paying 18% interest) wondered what sort of returns Berkshire or Buffett could get them. I’m paraphrasing, but Warren pointed out that the best return they would get would be paying off their credit card debt.
The case I’m making here is a similar, yet less-extreme example. If someone has a mortgage with an interest rate of 4% or higher, then it doesn’t make much sense to hold onto a stock that has a 10-year expected return of less than 4% because the return of paying down their mortgage is guaranteed, while the stock carries much more risk, and could, in fact, lose money over that time period. So, knowing that a great number of investors (including myself) would be better off paying down debt rather than owning a stock that has expected long-term returns of less than 4%, I have chosen to use a 4% expected 10-year CAGR as my initial threshold for selling a stock, and I have chosen to use a 12% expected CAGR as my initial threshold for buying a stock under the assumption that the long-term average return of the market is 8% or so, and I can build in a margin of safety by requiring higher return expectations.
When I bought Sysco back in March, it had a 10-year expected CAGR of over 12%. When I wrote about it in late May, it had a +11.95% expected 10-year CAGR and so I listed the stock as a ‘hold’ in that article. Since May, we have been able to better assess the effects of COVID on Sysco’s business and how this recession is affecting them much worst than the 2008/9 recession I based my previous estimates on. (In that May article I estimated my sale price for Sysco would be $78.50, but I have since adjusted that number lower.) In this article, I will calculate the current expected 10-year CAGR and explain why I decided to sell the stock after running these updated numbers.
Mid-Recession Full-Cycle Analysis
During the middle of recessions is always a tricky time to analyze a stock. And the unique nature of COVID and the rapid aid from the Federal Reserve and the federal government, has made the analysis even more difficult during this recession. Sysco is an extreme example of this difficultly.
In the FAST Graph above, I want to draw attention to the long-term earnings trend, which is represented by the dark green shaded area in the graph, with the annual numbers and percentage of change listed along the bottom. I have annotated the graph with a series of red circles (and one green circle for the year we got corporate tax cuts). As a general rule, what my “Full-Cycle Analysis” does, is to go back to the last recession and try to learn from that recession how a stock might behave during the next recession. As we see in the FAST Graph, Sysco’s earnings only fell -2% during the 2009 Great Recession, which is a very small decline given the magnitude of that recession. However, earnings were also very slow to grow post-recession and even included some additional modestly negative earnings years in 2012 and 2014. It really wasn’t until 2015 when earnings began to grow again. So, on one hand, we have the positive news that earnings didn’t fall very much, but on the other, there wasn’t a quick return to growth after the recession. When I calculate my cyclically adjusted earnings growth numbers, I take all of this into account.
An important feature that the graph shows is that the current recession, because of the unique effects of COVID, was much worse for Sysco than the last recession was. Earnings fell a whopping -43% this year, and are expected to continue falling another -6% next year. It’s important to keep in mind that when I bought the stock in mid-March, while I knew things would be difficult, it was impossible to really know how the numbers would turn out or how fast a potential recovery would be. I based my assumptions on the last cycle which Sysco weathered better (but I also included a margin of safety precisely because sometimes downcycles are different, just like we see here). So, now that we have more data that could potentially change my projections, it’s important to take that data into account as best we can. And that’s what I plan to do here by running a new “Full-Cycle Analysis” using the updated earnings and estimates from analysts.
I have shortened the timeframe of this graph so that it now ranges from 2011 through 2022, using analysts’ estimates for earnings in 2021 and 2022. The first thing I’ll point out is the average P/E ratio has been 19.86 during this time-frame. You’ll notice that the current P/E is over 30 on the FAST Graph. I typically like to use peak earnings and the current price when calculating a P/E ratio at the start of a recession. But in this case, since we are mid-recession, what I am going to do is use 2022 earnings estimates of $3.06 as my ‘E’ for my current P/E ratio. This will measure what the P/E will be after the expected earnings recovery in 2022 using the current price. The reason I do this is that when earnings are depressed during a recession, using a P/E with depressed earnings can send the wrong signal to investors if earnings recover within a year or two. If you use recession earnings, the P/E can make a cheap stock look expensive (as it does now with a P/E of 30). So, using $3.06 for the expected earnings combined with the current price, I get a current PE ratio of 20.04. If that were to revert to the average PE ratio of 19.86 this cycle over the course of 10 years, it would produce a 10-year Mean Reversion CAGR of -0.16%. So, assuming earnings go as analysts expect them to, if you hold the stock for a couple of years, the market is valuing the stock pretty close to what it has done historically, and essentially it’s ignoring this year’s and next year’s poor earnings. I think this is a reasonable assumption to make given what we know right now.
We previously examined what would happen if market sentiment reverted to the mean. This is entirely determined by the mood of the market and is quite often disconnected, or only loosely connected, to the performance of the actual business. In this section, we will examine the actual earnings of the business. The goal here is simple: We want to know how much money we would earn (expressed in the form of a CAGR %) over the course of 10 years if we bought the business at today’s prices and kept all of the earnings for ourselves.
There are two main components of this: the first is the earnings yield, and the second is the rate at which the earnings can be expected to grow. Let’s start with the earnings yield. If we use 2022 earnings estimates I get a current earnings yield of about +4.96%. The way I like to think about this is, if I bought the company’s whole business right now for $100, I would earn $4.96 per year on my investment if earnings remained the same for the next 10 years.
The next step is to estimate the business’s earnings growth during this time period. I do that by figuring out at what rate earnings grew during the last cycle and applying that rate to the next 10 years. This involves calculating the EPS growth rate since 2011, taking into account each year’s EPS growth or decline, and then backing out any share buybacks that occurred over that time period (because reducing shares will increase the EPS due to fewer shares).
Sysco bought back about 13% worth of stock this cycle so I will back those out of my earnings growth estimates. When I go back to 2011 through estimated 2022 earnings, I get a cyclically adjusted earnings growth rate of +2.60%. So, basically, once we include the current downturn and we take into account buybacks, there hasn’t been much earnings growth at all. On one hand, this is a bit of a pessimistic estimate because it’s unlikely that we will experience a similar pandemic as we have with COVID over the course of the next decade. But on the other hand, we are unlikely to get another corporate tax cut, and we don’t really know if 2022 earnings are going to be better, it is possible they could be worse. So, while expecting no growth is probably a little bit pessimistic, looking at Sysco’s history, there have been many no growth years. Estimating a +2.60% earnings growth rate once earnings have mostly recovered in 2022 seems reasonable after taking everything into account.
Next, I’ll apply that growth rate to 2022 earnings looking forward 10 years in order to get a final 10-year CAGR estimate. The way I think about this is, if I bought Sysco’s whole business for $100, it would pay me back $4.96 the first year, and that amount would grow at +2.60% per year for 10 years. I want to know how much money I would have in total at the end of 10 years on my $100 investment, which I calculate to be about $157.22. When I plug that growth into a CAGR calculator, that translates to a +4.63% 10-year CAGR estimate for the expected business earnings returns.
10-Year, Full-Cycle CAGR Estimate
Potential future returns can come from two main places: market sentiment returns or business earnings returns. If we assume that market sentiment reverts to the mean from the last cycle over the next 10 years for Sysco, it will produce a -0.16% CAGR. If the earnings yield and growth are similar to the last cycle, the company should produce somewhere around a +4.63% 10-year CAGR. If we put the two together, we get an expected 10-year, full-cycle CAGR of +4.47%.
My Buy/Sell/Hold range for this category of stock is: above a 12% CAGR is a Buy, below a 4% expected CAGR is a Sell, and in between 4% and 12% is a Hold. Right now, Sysco is slightly above that 4% threshold so using the basic analysis this would narrowly be a ‘Hold’. However, I have some additional thresholds that I require as well as the basic thresholds. In the case of Sysco, the addition threshold involves the Business CAGR.
When I was buying stocks last March, including Sysco, I not only had a combined minimum expected 10-Year CAGR threshold of 12%, but I also had individual thresholds for the Business and Mean Reversion CAGRs of 9% and 1%, respectively. For most stocks that didn’t make a whole lot of difference, but for some, I had to wait for a lower price before they crossed that 9% Business CAGR threshold even though the 12% combined threshold had been crossed. Similarly, I have an individual ‘sell threshold’ for the Business CAGR of 5%. So, once a stock crosses that threshold to the downside it becomes a sell even if the combined 10-Year CAGR is still above its 4% sell threshold. Since the 4.63% Business CAGR is below the 5% threshold, Sysco is now a ‘sell’ even though the combined 10-Year CAGR is still slightly above 4%.
For many stocks, the way they are affected from one recession to another is fairly similar. But sometimes different recessions affect the same stock differently, as is the case with Sysco this time around. When that happens we have to use a lot more individual judgment when it comes to determining when we should sell. I think Sysco’s business will eventually recover, but I question the speed of that recovery, and it appears the market has already priced at least 80% of the recovery into the stock price. Most of the easy money has probably already been made in the stock today, and while long-term holders will likely do fine and generate market-type returns over the next decade, I want to find opportunities that can generate above-market returns. Since Sysco’s Business CAGR is expected to be less than 5% over the next decade, I have decided to sell and look for better opportunities.
If you have found my strategies interesting, useful, or profitable, consider supporting my continued research by joining the Cyclical Investor’s Club. It’s only $29/month, and it’s where I share my latest research and exclusive small-and-midcap ideas. Two-week trials are free.
Disclosure: I am/we are long BRK.B. I wrote this article myself, and it expresses my own opinions. I am not receiving compensation for it (other than from Seeking Alpha). I have no business relationship with any company whose stock is mentioned in this article.