Someone emailed me this question:
“My other big initiative lately is collecting a list of investment mistakes (both my own and investors I admire) and implementing them into my research process. So if you ever find the time, I’d be really curious to hear how you implement learnings into your investment process and whether you use a checklist or set template?”
This is a topic where I think I differ a bit from a lot of other investors. Especially since the publication of Atul Gawande’s “The Checklist Manifesto,” I feel like there has been a lot of focus on learning from your own – and other people’s – mistakes. That’s a good goal. Like a lot of things in investing there’s this divide between the theoretical and the practical. How do you make a checklist practical? How do you know what matters and what doesn’t matter – and keep the list simple, easy to follow? How do you make it a habit?
The other problem is the risk of learning too much from your mistakes. Ask writers to quote from bad reviews they’ve gotten – and they can always do it. Ask writers to quote from good reviews they’ve gotten – chances are they don’t remember. It’s the same in investing. You’re likely to remember your mistakes a bit too well. Let’s say – just as a hypothetical here to illustrate a point – that you’re the kind of person who feels a loss twice as intensely as you feel the same gain. That’s going to really screw with your memory. The losses are going to “pop” in your memory. The gains aren’t.
I’ve made plenty of mistakes. I bought Barnes & Noble (BKS, Financial), Weight Watchers (WTW, Financial), and Town Sports (CLUB, Financial). Those were mistakes. I even lost meaningful amounts in Town Sports (a position I sold) and Weight Watchers (a position I haven’t eliminated yet). Weight Watchers in particular was a controversial pick. A lot of people told me it was a bad idea to buy it. They were right. It was heavily shorted. I could use that information to be more careful – guarding against arrogance and all that – in picking stocks that a lot of other people were betting against, telling me I was wrong to invest in, etc.
Here’s the problem with that. When Quan and I were writing a newsletter, we looked back at our past picks. We included stocks we were going to pick – but on which we (purely for scheduling reasons) never got a chance to publish an issue. We looked at data on those stocks from the time we were considering buying them. We weren’t using hindsight as much as investors normally do. We did a full catalog of all the stocks we’d considered investing in and then we collected the same data on those stocks so we had a truly comparable breakdown in an Excel sheet.
Guess what we found? Short interest wasn’t a reliable indicator of anything. Yes, some of the most heavily shorted stocks we looked at ending up performing badly. But some of our best-performing picks were also heavily shorted. If you plotted short interest against outcome – you didn’t get any sort of pattern. Also, if you tried to look for other similarities between poor-performing stocks – there were several symptoms our poor performers were more likely to share than short interest. The conventional wisdom would be that heavily shorted, controversial, etc., stocks are dangerous. We should be careful picking them. Some mistakes – like Weight Watchers – had these problems.
But other data we collected like Z-Score, F-Score and fixed costs (things like debt and capitalized leases / EBITDAR) were better indicators of possible bad outcomes than short interest was. In fact, companies that were financially sound, had improving ratios versus last year, etc., and yet were heavily shorted actually formed a really good group of possible picks. Short interest, level of controversy, etc., wasn’t actually a good indicator of risk at all. The only reason short interest seemed like a good indicator is because people naturally shorted financially risky stocks. It makes sense for them to do that, but they sometimes also shorted financially sound stocks.
For example, PetSmart (PEM, Financial) had a fair amount of short interest. If you looked at Weight Watchers’ financial ratios without knowing the short interest, you could guess it was probably heavily shorted. If you didn’t know which business you were looking at and just saw the stats on PetSmart, you’d never guess it would be shorted at all, and yet it was. In fact, those were the stocks that were often the best picks. Stocks that the numbers told you were fine but were being shorted for some reason.
I should mention, this pattern isn’t unique to the stocks that we considered. There’s been research into short interest as an indicator of future poor performance for stocks – and the record is mixed. Momentum can sometimes be a good indicator of some short-term poor performance lying ahead for a stock. And, in that way, short interest would also indicate some pretty negative sentiment. But it’s just not a good indicator of risk.
Something else we looked at was “beta.” It would make sense that stocks that move around a lot are riskier. We should – perhaps – avoid these kinds of stocks. Likewise, there have long been suggestions that some value stocks outperform because they are risky and that beta represents a form of risk of the price bouncing around a lot and making investors feel sick – that could help explain why some stocks are so cheap.
About five years ago, I took a good look at net-nets and which ones worked and which ones didn’t and why they might be cheap. I wrote a net-net newsletter for GuruFocus for a little while. One thing that concerned me was the pattern I saw from readers who emailed me telling me which net-nets they were buying and which ones they weren’t. We picked a net-net a month. I had to pick a stock every month – so I had to pick some net-nets I liked less than others. The pattern I saw from readers is that they consistently bought the net-nets I liked least and they consistently avoided the net-nets I liked most. What were they doing?
They were avoiding boring stocks. Now, I’m not saying that boring stocks always outperform exciting stocks. The first issue Quan and I did for Singular Diligence was on a stock called John Wiley (JW.A, Financial). It’s a boring stock. And it didn’t outperform the Standard & Poor's 500 from the time we picked it to today. Likewise, I bought George Risk (RSKIA, Financial) about six years ago. It was a boring net-net at the time. That stock also didn’t outperform the S&P 500 from 2010 to 2016. If you knew the market was going to perform as well as it did from 2010 to 2016, it would make plenty of sense to avoid boring stocks like John Wiley and George Risk. This is what most investors I talk to do. They don’t like boring stocks, they don’t like illiquid stocks, they don’t like stocks they hadn’t previously heard about (can’t get much information on, etc.), and they don’t like stocks that they have to talk to their broker about (stocks in another country, currency, etc.). For this reason, it seems exciting, liquid, well-known, U.S. stocks should be more expensive than other kinds of stocks.
For these reasons, I thought beta would make sense as information to keep track of in Excel when considering which stocks to buy. But the more I kept track of beta, the less useful it seemed to be. Here’s why. Take net-nets. It’s true that lower beta net-nets were the ones I liked better. This had nothing to do with their beta though. You could just as easily come up with a list of “low risk” net-nets by testing for the number of years of profits in their past and by testing for the level of their liabilities.
In fact, two numbers that seemed to make a lot more sense to me were simply the number of years of positive EPS in the company’s history, the Z-Score, etc. So, if a net-net had a profit in 10 of the last 10 years and a Z-Score of 10 it was safer than another net-net that had reported a profit in six of the last 10 years and had a Z-Score of 4. These measures have nothing to do with beta, short interest, etc. These are pure measures of the business as apart from the stock. And they seemed to work at least as well – better in fact – in predicting how risky a stock was. This is important, because if you asked most people for an explanation about what makes a stock risky, they’d talk to you about things like beta, short interest, etc.
In fact –Â I’ve mentioned this before – but simply looking at the number of consecutive years of profitability is a very easy way to gauge the safety of a business. And yet it’s something I almost never hear talked about. For example, which airline stock is the safest one to buy? There are plenty of ways you could try to figure that one out. I’d say the answer is Southwest Airlines (LUV, Financial) because it has the best long-term record of profitability. It also has a good credit rating, stronger financial position, etc. If you just looked at the last 10 to 20 years for every stock you considered buying and eliminated those that weren’t profitable in every year – you’d eliminate most of the potential mistakes you could make.
Not all of them. When I bought the stock, Weight Watchers hadn’t had a loss in a long time. Its financial position was weak. And – here’s the critical thing as far as Quan and I were concerned – its customer retention rate was abysmally low. Those are the two factors that created the risk in that stock. It had a lot of debt. And it depended a lot on attracting new customers each year.
Those are good lessons to learn. Those are valid things to look out for. I don’t think short interest is a valid thing for me to look out for. Why not? Couldn’t it be a good indicator of risk? Even if it has a lot of false positives – it wouldn’t hurt too much to avoid stocks with high short interest, right?
Wrong. See, the problem with eliminating either high short interest or high-beta stocks from consideration, is that these are likely to be some of the best values. Stocks that aren’t shorted, aren’t controversial and have more stable share prices are less likely to be mispriced for long enough that I can find them, research them and buy them before their share prices rise again.
I’ll give you an example of learning too much from your mistakes. The obvious mistake a lot of people would say I keep making is buying into a business that is exposed to some disruption in the industry, societal shift, etc. Buying a “buggy whip” business.
Quan and I considered that possibility. We looked at our best performing candidates for further research and our worst performing candidates for further research. Both the best and worst performing stocks fell in the high risk of “societal shift” category. I’ve picked both good and bad stocks that had a high risk of societal shift. Bad performers include stocks like Barnes & Noble, Weight Watchers and Town Sports. Good performers include such stocks as Greggs (ASX:GRG, Financial), PetSmart and Babcock & Wilcox (BW, Financial). Hindsight is a problem in separating these cases. When Quan and I looked at Greggs, there was plenty of talk about how the reason for that stock’s decline in same-store sales was the type of food (unhealthy) it was selling. This kind of thing happens all the time.
A great example is Luxottica (LUX, Financial). This is from so long ago that no one will remember it, but decades ago there was a rise in contact lens use. There was a lot written about the risks that contacts posed for Luxottica’s business of selling eyeglass frames. No one writes about this anymore. That’s not because contact lens use has declined. It’s just because the company has grown its earnings for so long that the threat from contacts became old news. So people stopped talking about it.
This is what happens in all these cases. If a company reports improving earnings, has a rising stock price for a couple of years, people stop writing about the risk of some cultural change and technological obsolescence. Unless you go back into the news archives (as Quan and I always did when researching a stock), you wouldn’t even know about all the risks analysts, investors and business reporters saw on the horizon for stocks that now appear to have had completely smooth historical growth trajectories.
It’s important to learn from your mistakes. But you have to learn in a way that allows you to make a process improvement. I have lost money investing in stocks that were harmed by change. I have also made money betting on stocks that other people thought would be harmed by societal change – and then they weren’t harmed, and the stock price rebounded quickly. There is no way for me to make a process improvement. If I avoided stocks where the general perception was that the company would soon face societal change – I’d miss out on both losers and winners. This kind of knowledge isn’t helpful. It might make me feel a little better to implement the change. I wouldn’t make the same mistake in the future as I did in the past, but I wouldn’t actually be improving my process.
There’s no reason to believe the change would net out to a benefit for me across a series of say 10 such bets, and that’s all that matters. I’d be avoiding both some future losers and some future winners. It would just net out to be change for change’s sake. I don’t have any evidence it would be a process improvement. It would just be something I could change and might or might not help. There’s no reason to make a change like that and, frankly, that’s the kind of change I see most investors make. They learn from their mistakes. They learn not to do what led to one or two bad outcomes.
It’s not an actual process improvement – because it teaches them to miss out on winners as well as losers in the future. An actual process improvement would have to be something that helped you pick fewer losers while causing little or no reduction in the frequency with which you picked winners. I know that implementing a “societal change” screen would – in my case – cost me a lot of big winners.
It’s the kind of lesson I try not to learn.
Disclosure: Long Babcock & Wilcox, Weight Watchers, George Risk.
Start a free seven-day trial of Premium Membership to GuruFocus.
