Tracking How Far They Fell: April 2019 Edition

|
Includes: AXP, BA, BBY, BRK.B, CAT, COST, CSX, DE, FISV, GWW, IR, LUV, NKE, O, PAYX, PGR, ROST, RSP, SPGI, SPLV, TROW, UNP, VPU
by: Cory Cramer
Summary

In the first half of 2018, I warned investors about 27 cyclical stocks that could fall quite far in a bear market, and in 2019, we've added two more.

I suggested some alternative investments I thought would be more defensive and also suggested a rotational strategy that could increase one's shares of the target stocks at no extra cost.

So far, we've had 15 successful rotations with large free share gains.

This article will track the performance of the remaining 14 ideas through the month of April.

This month I'll categorize the stocks by their historical earnings cyclicality and see what we can learn.

Source

Introduction

Beginning in mid-January of 2018, I wrote a series of articles that examined how far some popular large-cap cyclical stocks might fall if we were to have a downturn within the next three years. While all the articles were generally bearish in nature and meant to be a warning to current investors that even the stocks of good companies could fall quite far during a bear market, I didn't stop there. In each article, I suggested alternative investment ideas for the cyclical stocks in question. The four most frequent alternatives I suggested were the Invesco S&P 500 Low-Volatility ETF (NYSEARCA:SPLV), the Vanguard Utilities ETF (NYSEARCA:VPU), Berkshire Hathaway (NYSE:BRK.A) (NYSE:BRK.B), and Invesco S&P 500 Equal Weight ETF (NYSEARCA:RSP), or some mix of them. I called this series of articles the "How far could they fall?" series. The goal of the articles was to warn investors of the potential downside these stocks had while also offering alternative investment ideas that current shareholders could rotate into while the prices of the target stocks were high. Then, after the target stocks had fallen significantly, rotate from the defensive positions back into the target stocks. The idea was that this process would prevent buy-and-hold investors from suffering big declines while also producing free share gains in the target stocks compared to a buy-and-hold strategy.

For example, if one rotated out of the target stock and into the defensive ETF while they were both priced at $100 per share, then during a bear market, the defensive ETF dropped to $80, and the target stock to $40 per share. At that point, you can rotate back into the target company stock and own twice the number of shares at no extra cost. Then, when the stock eventually recovers, you have doubled your wealth compared to what it would have been if you held the target stock through the entire period (minus taxes, of course).

In order for all this to work, one needs to 1) identify a quality company, 2) understand when it is overvalued, 3) get somewhat close at identifying the late stages of the business cycle, 4) correctly identify a more defensive alternative, 5) have the guts to rotate back into the stock when it looks like the world is ending near the bottom of the cycle, and 6) wait for the stock to recover.

The "How far could they fall?" series essentially takes investors through this process with a fairly wide swath of large-cap names. In total there were 28 stocks in 2018 that I both wrote a warning article about and also offered a defensive alternative I thought would be better at the time. There were a few stocks I wrote "How far could they fall?" articles on that I decided not to offer alternatives on like Ross Stores (ROST) and T. Rowe Price (TROW) that were too hard to predict at the time. I didn't track those. And there was one that I recognized should have been put in that category a few months later, W.W. Grainger (GWW). I noticed after I reread my original Grainger article that I explicitly said in the article Grainger's turnaround was too hard to predict, and I didn't even perform a risk/reward analysis on it because of that, but then I went ahead and offered an alternative idea anyway, which I shouldn't have done. I kept tracking Grainger's performance until the end of November 2018 when it was nearly perfectly even with my suggested alternative, and I decided to correct my mistake, call that idea a wash, and stop tracking it. So, going into 2019 we were tracking 27 large-cap cyclical ideas and how they were performing vs. my chosen defensive alternatives.

The deep correction that started in October and bottomed in December 2018 created an opportunity to rotate out of many of the defensive alternatives back into the target stocks. I wrote a fairly detailed description of those moves in my article "Tracking How Far They Fell: 2018's Rotational Winners", in which I went through all 14 of the 27 rotational ideas I'd been tracking that had been completed in 2018, and we added one more completed rotation with Caterpillar (CAT) in January 2019. So far, in 2019, I've added two more new stocks to track, Realty Income (O) and Ingersoll-Rand (IR). Here is the table, which shows the free share gains achieved from the rotational strategy for the 15 completed rotational trades so far:

Ticker Free Shares Gained Ticker Free Shares Gained
CMI 53% EMR 30%
STT 38% GD 50%
MGA 40% ROK 35%
MMM 32% ETN 28%
NOC 47% CCL 33%
FDX 41% PNC 23%
TIF 29% UTX 29%
CAT 37%

These are great gains, all achieved in a year or less. For those readers who aren't used to thinking in terms of free share gains, here were the average returns during the times when we were out of these target stocks and in the defensive alternatives: Defensive Alternatives: +3.42%, S&P 500: -6.22%, and Target Stocks: -24.80%. So, for a group of 15 stocks, we had incredible outperformance while we were rotated out of them, and we even significantly outperformed the S&P 500 with this group over this time period.

Stocks we are still tracking

I've learned a lot by tracking the performance of these ideas over the past year. One of the key lessons is that there are many different types of cycles that stocks can go through and it helps tremendously if one can understand the cycles more clearly. One of the most important ways to categorize any particular stock's cycles is by looking at their historical earnings cyclicality.

When I first started writing the "How far could they fall?," I focused almost entirely on price cyclicality because earnings for highly cyclical stocks don't tend to be reliable indicators of medium-term stock price movements. But the mistake that I made was not checking first to see if earnings were the primary driver of the stock price for some of the stocks I examined. For some of the stocks, earnings (along with market sentiment) did end up explaining the historical price movements better than the price movements alone.

This revealed itself pretty clearly during the -20% market correction we saw in Q4 of 2018. My strategy worked well during this correction for stocks with cyclical earnings, but since we have not yet experienced a recession, those with less-cyclical earnings continue to perform relatively well, and most of them we are still tracking.

Recently, I have developed five ways to categorize the historical earnings cyclicality of stocks so that I can better classify the cyclicality of any given stock. For stocks whose earnings are less cyclical, I have developed a new, longer-term way to analyze them over the course of a full business cycle, and for stocks whose earnings are more cyclical, I use the same method I used during the "How far could they fall?" articles. What I'm going to do in this tracking article is retroactively classify the 12 stocks from 2018 that we are still tracking, and see which of the five categories they fall into.

The first category I call "secular growth." This category describes earnings that continue to rise every year even during economic recessions. The next three categories are "low", "moderate", and "deep." "Low" is usually for businesses which have earnings that have a history of declining in the single digits percentage-wise during downturns but not much farther than that. "Deep" I consider earnings that fall more than -50%, and "moderate" somewhere in between low and deep. And last but not least are businesses whose earnings go negative during cyclical downturns, but recover soon after that, which I call "highly cyclical."

In this article, I'll share the category of earnings cyclicality that I think each stock belongs in and see if there is anything we might learn.

Stocks from 2018

Boeing

Chart Data by YCharts

I first wrote about Boeing (NYSE:BA) on 1/17/18 and warned investors about its potential downside during a recession. In that article, I suggested that SPLV was a good defensive alternative for investors who wanted to stay long stocks but mitigate Boeing's dangers. SPLV as of the end of April 2019 has outperformed both BA and SPY with far less volatility along the way. I have since followed up with more detailed articles about the price I would look to buy Boeing, though its lackluster response to these 737 Max crashes should worry investors more than it apparently has. I've mostly remained quiet on this issue in an attempt to not jump to conclusions or pile on Boeing in the face of a tragedy. I will likely wait until the price falls at least -40% before digging into the implications of the crashes more since I wouldn't buy Boeing stock until the price was lower anyway. At that point, it will be worth my time to gauge the damage of these crashes.

From 2001 to 2003, Boeing's earnings fell -72%m and from 2007 to 2009, its earnings fell -65%. This puts it in the "deep" category when it comes to the cyclicality of its earnings. For stocks in this category, I think it is appropriate to use historical price cyclicality to estimate the potential drop in price during a downturn, so I think my original analysis still holds for Boeing's stock and it's wise to be out of the stock and in a more defensive position right now.

Deere

Chart Data by YCharts

I first warned about Deere (NYSE:DE) on 1/24/18 and suggested VPU as a good alternative. That turned out to be a very good suggestion and VPU has dramatically outperformed both the index and Deere since then. Deere came close to being low enough relative to VPU to rotate back into last October but I didn't catch it. Deere's earnings fell -69% in the 2001 recession, -41% in the 2008 recession, and -47% from 2013 to 2016. I consider Deere a "deep" cyclical stock. And I think we will get another opportunity to rotate in and gain considerable shares.

Union Pacific

Chart Data by YCharts

I first warned about Union Pacific (NYSE:UNP) on 3/6/18, and I mentioned all three of my most popular defensive ideas in the article so I've been tracking it against an average of all three, SPLV, VPU, and Berkshire Hathaway. This is interesting because we can get a view of how all three are performing relative to each other from the same date. Berkshire has been lagging (as we'll see in many other comparisons in a moment), but I have every confidence that during the next recession, it will dramatically outperform more cyclical and overvalued stocks. But the question is was I correct in treating UNP as cyclical and/or overvalued?

Looking at UNP's earnings cyclicality, its earnings fell -33% from 2002 to 2004, -18% during the 2008 recession, and -12% during the oil bust in 2015-16. This evidence suggests that UNP's earnings are actually "moderately" cyclical. For that reason, the sort of historical price cyclicality analysis that I performed on the stock last year wasn't the best choice. Partially because of what I have learned from these tracking articles, in the early part of this year, I developed a better method of analyzing moderately cyclical stocks and I'll probably publish an update soon on UNP using that method of analysis and see what it turns up. Fortunately, my suggested alternatives, on the whole, haven't performed too badly. Despite my mistake, the alternatives have on average returned +16.34% compared to SPY's +10.28%. Additionally, we saw that during the correction in December that UNP did drop low enough to match the average of these investments, so there is a good chance we will be able to rotate back into UNP and break even in the future even with my error.

Costco

Chart Data by YCharts

I warned about Costco (NASDAQ:COST) on 3/8/18, and it has performed very well since then compared to Berkshire, which I suggested as an alternative. Additionally, Berkshire has underperformed the index. However, it's worth noting how much better Berkshire performed during the correction in December where it was outperforming the index and nearly performing the same as Costco. This is probably a good preview of what to expect during a recession. Costco's earnings only declined -4% during the 2001 recession and -14% during the 2008 recession. I would classify its earnings cyclicality as "low". However, it does go through sentiment cycles, and right now sentiment is very high for the stock, so I still think during a recession, Berkshire performs at least well enough to get back to even with Costco, which I'll probably write a follow-up article on soon.

S&P Global

Chart Data by YCharts

I warned about S&P Global (SPGI) on 3/19/18. It is currently outperforming Berkshire and the index, but you'll notice that during the 2018 correction that was not the case. While I still feel confident that Berkshire will outperform during a recession, SPGI's earnings didn't fall during the 2001 recession and only fell about -21% during the 2008 recession. Most of SPGI's historical price decline can be attributed to sentiment cycles. Currently, the sentiment is quite high for SPGI, so I think we'll still have an opportunity to eventually gain shares, but since SPGI only has moderate earnings cyclicality, it would have been better to use a 10-year, full-cycle approach rather than a historical price cyclicality approach. I'll put it on my list for a follow-up using that approach.

CSX Corp

Chart Data by YCharts

I warned about CSX (NASDAQ:CSX) on 3/22/18 and it is the only stock so far out of the 30 or so I covered that has broken out above its three-year projected price appreciation estimates. Over the past 20 years, CSX has seen earnings declines of -41%, -16%, -18%, and -9%, which puts it in the moderately cyclical category. I'll eventually do a follow-up on this one using the 10-year full-cycle analysis, and we were almost back to equal with Berkshire during the correction. I think we're just going to try to get back to even with this one.

Southwest Airlines

Chart Data by YCharts

I first warned about Southwest (NYSE:LUV) on 4/4/18, and most of the time since then, it has underperformed, but not quite enough to rotate back into the stock from the defensive position in Berkshire. From 2000 to 2002, Southwest's earnings fell -70%, from 2000 to 2009, they fell -73%, and from 2010 to 2011, they fell -42%. I consider Southwest's earnings to fall into the "deep" cyclical category. So, using historical price cyclicality was a good way to estimate future drawdowns, and I think that eventually we'll get a good enough spread between Berkshire and Southwest to gain significant free shares in the future.

Paychex

Chart Data by YCharts

I warned about Paychex (NASDAQ:PAYX) on 4/13/2018 and suggested SPLV as a good alternative. SPLV is doing fine, but Paychex has been doing much better. While I do think sentiment is too high for Paychex right now, earnings only fell -9% in 1999 and -13% in 2008/9, which puts it in the "low" earnings cyclicality category. So, I don't think it was optimal for me to historical price cyclicality to estimate the danger to Paychex and it would have been better to use the 10-year full-cycle analysis. Like a few other stocks we're still tracking, sentiment is quite high for Paychex right now, though, so I do still expect it to trade low enough to at least trade back in and break even during a recession or economic slowdown.

American Express

Chart Data by YCharts

I warned about American Express (AXP) on 5/2/18 and broke a bit from tradition by suggesting a combination of Fiserv (FISV) and VPU was a good defensive alternative mix. Over time, even though Fiserv has performed very well in this case and the FISV/VPU combo has returned +22.29%, over time I have come to the conclusion that for this particular rotational strategy, it is best to use ETFs instead of individual stocks for defensive alternatives. As for AXP's earnings cyclicality, earnings fell -53% in 2001 and -52% in 2008/9, so I consider AXP a "deep" cyclical stock with regard to earnings and it was appropriate to use the approach I did last April. Eventually, we will likely get a good chance to achieve some free shares in AXP during a downturn.

Best Buy

Chart Data by YCharts

I warned about Best Buy (NYSE:BBY) on 4/17/18. It's probably best described as volatile rather than cyclical, but earnings went totally negative in 2012 and have been somewhat unpredictable. I think that I was correct to have used the approach I did with such highly cyclical/volatile earnings, and we could have had massive share gains during the correction, but I was looking for a little bit more. Perhaps during the next downturn, we'll get another chance.

Progressive

Chart Data by YCharts

I warned about Progressive (NYSE:PGR) on 6/19/18. Progressive has experienced earnings declines over the past 20 years of -79%, -39%, -24%, -16%, so PGR probably falls in the moderate to deep categories, but the earnings declines don't always match up with the wider economy. My guess is that mother nature plays a role when it comes to insurance and earnings. But there will be periods of earnings volatility with Progressive, so I think using historical price movements is fairly reasonable, and we did see an opportunity to gain shares in the company during the 2018 correction. I think we'll eventually get that opportunity again.

Nike

Chart Data by YCharts

I warned about Nike (NYSE:NKE) on 8/14/19 and it has been reasonably resilient since then. Nike has only had one year of modest -4% earnings declines the past 20 years, which puts it in the "low" earnings cyclicality category, so I should have used a different approach to analyze the stock. However, like a few other stocks I'm tracking, Nike is trading at a very big premium to its historic multiple, so I think we'll end up doing okay with this idea eventually.

Alright, if we take a look at the performance of the stocks we are still waiting to rotate back into from 2018, the average performance of the 12 target stocks through the end of April was +21.09%, the defensive alternatives were +13.10%, and the S&P 500 +10.20%. I'm very pleased to see that not only were all the defensive ideas positive and performing much better than cash (which was the goal), but that they are also outperforming the index. And while it's true that current target stocks we are still tracking are up +21.01%, placed within the context of the market recently hitting all-time highs, that's not too bad. Additionally, it's important to remember that we've already had 15 successful rotations. When we include those in the mix, the target stocks have returned an average of -4.40%, SPY has returned +1.08%, and the defensive alternatives have returned +7.72%. So, overall, there really isn't much contest when we take the 27 stocks from 2018 in total even despite some of the minor missteps I've made along the way (including those I highlighted in this article). In the future, I expect the strategy to perform even better because of what I've learned writing this series and tracking the results because I'll be able to better categorize the stocks and select the best type of analysis for them.

Now let's add in the ideas from 2019.

2019 Stocks

Realty Income

Chart Data by YCharts

With Realty Income (O) being a REIT, it makes a little trickier to categorize it using earnings. This was actually more of a sentiment play from the beginning, and we're only looking for 20% share gains with the idea. I should probably develop a REIT-specific analysis in the future, but since there is so much REIT coverage on SA already, it's not something I have been in a rush to do. Currently, O is modestly underperforming my alternatives.

Chart Data by YCharts

I just wrote about Ingersoll-Rand (NYSE:IR) last week, so it has only had a couple of weeks to work. I suggested a SPY/SPLV mix as an alternative idea. IR has seen earnings declines of -45% and -53% in the past, so I consider its earnings "deeply" cyclical, and it was appropriate to use historic price cyclicality as a guide.

Currently, there are only two ideas from 2019 I'm tracking, and so far the target stocks have returned +5.25%, SPY +3.70% and the defensive alternatives +3.11%. If the market continues to make new highs this year, I'm likely to add more stocks to the 2019 list.

When taking all 29 ideas in their entirety, the total returns for the target stocks are -3.73%, SPY +1.26%, and the defensive alternatives +7.40%.

Conclusion

Overall, I think there is certainly enough evidence to consider using this rotational strategy in the late stages of the business cycle with stocks whose earnings are moderate to highly cyclical, or that are trading at higher than average P/E multiples. It's important to remember that I specifically chose the stocks of companies of reasonably high quality. These were all stocks that I might have wanted to purchase if the price was right. To be outperforming these relatively high-quality stocks by more than 10% after a year is very good. However, I think properly categorizing the earnings cyclicality of the target stocks, and then choosing the proper technique to analyze them based on what category they fall in will improve the results in the future. When I combine that approach with some of the other lessons I've learned about defensive investing and price cyclicality this past year, I expect even better performance of this strategy going forward. If you would like to learn more about the strategy, click the orange "follow" button at the top of the page in order to be notified when I publish new articles.

Disclosure: I am/we are long BRK.B, FDX, STT, TROW. 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.