Someone emailed me this question:
“Do you use checklists or research templates, and if so, what sorts of things are on them?”
For a couple years, I wrote a stock newsletter. The structure of that newsletter reflected the way I think about stocks. Basically, I just did my typical research process for a stock and published it. Each issue started with a data sheet. The two key parts of that data sheet were a very long-term historical record – as long-term as I could get – showing the income statement and balance sheet ratios, growth rates, etc. for the business. It wasn’t unusual for us to publish 20-25 years of data. That wasn’t for show. It wasn’t something we did because we though subscribers might like it. I always look at the company’s financial history going as far back as possible. I’m not interested in future projections. But, if you had fifty years of financial data on some company – I’d be interested in seeing all fifty years. In fact, one of the first things I do when looking at any stock is to read the most recent annual report (or 10-K) and to read the very oldest annual report (or 10-K) I can get my hands on. Companies that have been public for a long time have data on EDGAR (it’s a website run by the SEC) going back to about the mid-1990s. So, it’s possible to have about 20-25 years of financial data for most companies that have been public for a long time. There are a few companies – examples include Southwest Airlines (LUV) and Wal-Mart (WMT) – that archive even older financial data on their own websites. I believe Southwest Airlines has annual reports going back to the 1970s. The first thing I did when researching Southwest was to read the most recent annual report and to find the oldest annual report on the company’s website and read that one too. You can find historical financial data at sites like GuruFocus and QuickFS.net. You can also enter the data yourself using EDGAR. Like I said, EDGAR goes back to the mid-1990s. So, you should always also check the investor relations section of a company’s website. There may be annual reports going back before the oldest 10-K you can find on EDGAR. And, if you haven’t tried it before, I definitely recommend testing out the idea of reading the newest and oldest 10-K as the first part of your research into a company. It’s a habit I find very useful.
The second part of the data sheet was a peer comparison. Basically, we picked five peers for the company we were interested in. We then presented price ratios like enterprise value to sales, enterprise value to gross profit, enterprise value to EBITDA, and enterprise value to EBIT. The best investment ideas are often companies that are better than their peers in terms of quality but priced lower than their peers. Those ideas always get me excited. It’s also a good idea to collect peer data simply because one idea can lead to another. Many times, we moved from researching one stock in an industry to researching another related stock. We ended up writing reports on maybe five different banks. That process all started when we wrote about Frost (NYSE:CFR). Frost led us to banks in the same area like Prosperity (PB), which is a Texas bank. It also led us to look at banks that do the same sort of lending as Frost – energy lending – such as BOK Financial. And then researching BOK Financial and banks like that led us to look at other banks like Commerce (CBSH). The Kemper family runs Commerce. And a different branch of that same Kemper family runs UMB Financial (UMBF). And then studying certain aspects of Frost’s deposit base got us interested in Wells Fargo (WFC) – which we never wrote a report about. It also got us interested in Bank of Hawaii (BOH). And we did write a report about that bank. So, I’d say that research which started with Frost eventually led me to consider about six more banks. That was all due to peer comparisons. As it turns out, I like Frost as much or better than any U.S. bank. I only own one bank stock and it’s Frost. So, based on my actions, you’d have to say Frost is my favorite bank stock. Frost also happens to be the first bank stock I considered. But, I don’t think researching another half dozen or so bank stocks was a mistake at all. I liked those stocks more than non-financial stocks. Researching them went much faster because I had already researched Frost. So, I consider the selection of peers and the comparing of a stock I’m interested in to a group of about five peers to be a useful part of my investment process. That’s why I always include peer comparisons in the newsletter.
You could say there was a third part to the data sheet. We always included a graph showing the long-term gross profit margin, EBITDA margin, and EBIT margin. If you’ve read some of the stuff I’ve written in the past, you may know I have a bit of an obsession with the stability of long-term margins. I’m also very interested in potential long-term margin expansion and economies of scale. I care a lot about gross profitability. A company that has long had very good gross profits but was once a lot smaller or less well run at the corporate level may be able – through organic growth, through changes in management, or through horizontal mergers – to improve its net profit over time. However, a business that has long had poor gross profitability – and especially an industry that has long had poor gross profitability – is far less likely to turn things around. So, I am always interested in whether gross profitability has been high even if net profitability hasn’t been. And then I’m always interested in the amount of variation in the company’s margins. I prefer to invest in companies that have a long history of profitability and have especially stable margins compared to their industry and to other stocks.
After the data sheet, the newsletter was just a series of written sections. The first section was called the “overview”. This was mostly just a historical description of how the company got started and how it had developed. I’m always interested in learning what I can about the historical development of a company and an industry. So, the overview section was basically historical. Then there was a “durability” section. That section is different than moat. Usually, whenever I’ve eliminated a company from consideration – either as a newsletter pick or as a personal investment – I’ve done so because of some threat to the company’s or the industry’s durability. A good example is something like Q-Logic (NASDAQ:QLGC). I did plenty of research into this company, its product, and the industry. A lot of the research showed up really good signs. But, one thing I couldn’t be sure of is that the need for the company’s products would be as great in the future as it had been in the past. This wasn’t really a competitive position risk. It was a risk related to the wider issue of how storage area networks would be organized and used. I found a similar situation in Teradata (NYSE:TDC). There was plenty of scuttlebutt showing that Q-Logic and Teradata had strengths. But, there just wasn’t information I trusted giving me a good idea that things would be done the same way five years down the road as they are done today. I always do a “durability” check for a company. Sometimes I am aware of a risk to a company’s durability, but I don’t think it’s a big problem mathematically in terms of a discounted cash flow analysis. I never do a DCF. But, I’m aware of the principle when investing. Progressive (NYSE:PGR) presented a special problem in terms of durability. On the one hand, Progressive and GEICO both still have much lower shares of the overall auto insurance market than they have of the truly new business in the industry. If a company has say a 12% share of the auto insurance market but has a 25% share of the new business in this industry – we can pretty safely assume it will one day be capable of having twice the share of the market that it does today. So, you have a big tailwind for direct insurers like Progressive and GEICO. But, you also have the threat of driverless cars. And here we aren’t just worried about cars that don’t require a human driver at all. We also need to be aware of the possibility that technology will continually reduce accident frequency. In fact, that’s what had been happening for years. The situation has become less certain lately with some evidence that texting has increased accidents generally and fatal accidents specifically. Regardless, there are some kinds of accidents – like a driver falling asleep and hitting the car in front of them on the highway – that the use of computers can easily eliminate entirely without even being a true “driverless” car. So, I took a good look at how long it might take for driverless car technology to be adopted by a popular, mainstream auto manufacturer on one of its top-selling models. Then, I looked at the history of how long it has usually taken for a safety feature introduced by one company to become standard on all new cars. And then, finally, I considered how long it would take for half of the cars actually on the road – since most cars are pretty old – to be driverless cars. All of this was very rough and theoretical. It was an approximation. But, what it showed to me was that even if driverless car technology was ready now and eagerly embraced by consumers, most of the damage it would do to the auto insurance industry would come fifteen or more years from now. In a DCF type analysis, years 16 and on are a lot less important than you might think. I was also pretty sure that companies like Progressive and GEICO would have a much greater piece of the overall auto insurance pie when the industry was finally hit with this obsolescence risk. So, I assessed Progressive’s durability as being imperfect but adequate for a long-term investor. I didn’t feel I had to rule out Progressive just because driverless cars would reduce accident frequency in the medium term and perhaps obsolete the auto insurance industry entirely in the very long-term.
After the “durability” section, I did a “moat” section. I’m sure you’re familiar with Warren Buffett (Trades, Portfolio)’s definition of a company’s “moat”. So, I won’t bore you with a discussion of the “moat” section of the newsletter. Just know that moat is something I look at whenever I analyze a possible stock investment. There was a “quality” section. For me, a big part of a company’s quality is its unlevered return on net tangible assets. I always had a “capital allocation” section. This is a topic I focus on more than almost any other investor I know. I am often first attracted to a company because it has made smart acquisitions, paid a special dividend, pays a regular dividend, or has regularly bought back stock for years and years. In some cases, like with the selection of Omnicom (NYSE:OMC) for the newsletter – capital allocation was the primary reason for picking the stock. Omnicom is a high quality, adequate growth stock. It rarely trades at a low P/E ratio. If the company had more typical capital allocation, it would just be an okay stock. However, Omnicom has long focused on buying back more of its own stock than other ad agencies do. The business model of an ad agency is far superior to that of the average public company. Therefore, an ad agency that devotes as much of its free cash flow as possible to repurchasing its own shares can add a lot of value over time. About seven years ago, I bought Omnicom stock for myself. And then I picked it for the newsletter too. In both cases, I wouldn’t have considered Omnicom if not for its capital allocation.
You probably consider me a value investor, so it’s no surprise I had a section in the newsletter called “value”. I always do an appraisal of a stock. So, I don’t just determine it’s cheap. I actually look at the stock and come up with what I think it would be worth in normal times. Sometimes the difference in my appraisal and the market price is big. A good example of this is Frost. I still own the stock. That surprises some people because it probably rose 50% or so from where I wrote about it. However, I think Frost was trading at about $60 a share when I picked it for the newsletter and I put a “normal year” (meaning a year in which the Fed Funds Rate was at least 3%) appraisal on the stock of about $140 a share. It’s very rare for me to find a situation where the market value of the stock is less than 50% of my appraisal of the business. In the last few years, the only non-financial stock that leaps to mind is Hunter Douglas. I believe I appraised that stock at more than twice its stock price. Again, there was a cyclical reason for that. I do a “normal year” valuation. I don’t pay attention to what earnings will be this year or next year. So, my appraisal of Frost was based on a year in which the Fed Funds Rate had recovered to historically normal levels and my appraisal of Hunter Douglas – which is indirectly related to housing since it makes shades and blinds for windows – was based on a year in which the U.S. housing market had fully recovered.
I also do a section on “growth”. I do consider likely future growth at the companies I research. This isn’t a very important part of my research process. I have sometimes picked one stock over another because of better growth prospects. For example, I prefer Texas banks over banks in other states because I think Texas will grow faster than a state like Hawaii. One reason I own Frost and not Bank of Hawaii is that I feel certain Frost will grow its deposits faster than Bank of Hawaii will. Growth is also a factor when I consider investing in a chain like a retailer or a restaurant. I will pay more for a restaurant chain that has fewer present day locations compared to how many locations I think it can one day have in the country in which it operates.
I always include a “misjudgment” section. The misjudgment section is where I consider my potential biases, errors, etc. I’m not sure how useful it is. But, it’s something I always do. It’s different from the other parts of my investment process in the sense that it’s purely subjective – focused on me and my thinking – rather than objectively focused on the stock itself.
Finally, there was the “future” section of the newsletter. In many ways, I consider this the most important part of my investment process. I’ve said this many times. But, it’s still worth repeating. I don’t care what a business will report in earnings this year or next year. I also don’t pay any attention to what I think the market will value the business at. I don’t care about trying to predict what will change market sentiment. That’s too hard for me. I always frame the question this way. One, what will the business look like in five years? Two, what would an acquirer pay for such a business? That’s it. And that’s what I mean by future. For example, Kraft-Heinz just made a bid for Unilever. So, if I was analyzing Mondelez – which used to be part of Kraft – I’d ask what Mondelez would look like in 2022 and how 3G and Kraft-Heinz would value such a business. I really don’t think about how much Mondelez will earn in 2017, 2018, or 2019. That’s what a trader worries about. I also don’t look at what Mr. Market is valuing the business at now. I do collect data on peers. They’re public peers. So, I see how the market feels about the industry that way. However, I pay at least as much attention – and maybe more – to the multiples at which acquisitions have been done in the industry. I don’t pay more attention to what the market values a food company at today in terms of EV/EBITDA or P/E than I do to what one food company paid to acquire another food company bank in 2015 or 2012.
That’s my process. It is standardized in terms of my major concerns. And it’s fully standardized in terms of collecting and presenting historical data. It’s not really very structured beyond that. So, I guess you could say I have a checklist that reads: “durability”, “moat”, “quality”, “capital allocation”, “value”, “growth”, “misjudgment”, and “future”. I also always compare the company I’m interested in to publicly traded peers. And, most importantly, I always look at historical financials going as many years into the past as possible. I’d say I’m usually working from about 20 to 25 years of past financial data. That data is the bedrock of my process. It’s the quantitative part. Everything else is basically qualitative.
Ask Geoff a Question
Disclosure: Long CFR
Someone emailed me this question: