Some who reads my articles asked me this question: Hi Geoff,
I'm doing some quick valuation exercises on some of the stocks I own based on the methods you detailed in your past posts (Microsoft and Exxon). Instead of using the current interest rate or AAA bond yield, I am using the Shiller P/E. My question is should I use the 14.88 (Geometric mean) that you used in your Exxon example or do I use the current Shiller P/E which is approximately 22. I noticed you compared the latest Buffett/Munger Bargains Newsletter pick which had a P/E of 19 to the current Shiller P/E and therefore drew a conclusion that it was attractive. Obviously the current Shiller P/E would make my valuations a bit higher.
All valuation is relative. The key is not the precision of your intrinsic value estimates. It is their usefulness.
So, it depends on what you are looking for.
The Buffett/Munger Bargains Newsletter picks a stock every month. So, it has to compare its pick to the market by default. You always value the stock you are considering buying against the next best alternative. In the case of the newsletter, the next best alternative is always another stock. I could've done a long analysis comparing the stock I picked for the Buffett/Munger Bargains Newsletter last month to every other possible pick – or at least the other top few candidates. Basically, the stock I picked was safer than the alternatives that were cheaper than it. And it had more long-term upside than the stocks that were as safe as it was (which were very, very few stocks). So that's why the Buffett Munger Bargains Newsletter uses current market valuations to justify its picks. Because it has to make a pick. For the newsletter, the next best alternative is always another stock. That’s because newsletters keep making stock picks regardless of the market environment.
But do you have to keep buying stocks regardless of the market environment?
That's up to you.
Your next best alternative could be holding cash.
However, since you are talking about doing valuation exercises on stocks you already own – I would suggest using the current (higher) Shiller P/E. You should sell stocks you already own only because you have a better idea to invest in or because they are clearly overvalued. If you decide the stocks you own are overvalued using the 14.88 Shiller P/E (the geometric mean – GuruFocus uses a different averaging method and gets 16.4) then you are really not saying those specific stocks you own are overvalued. You might as well just be saying the market is overvalued. I don't think – and a lot of reasonable people disagree with me on this one – that you should sell the shares of companies you like and are familiar with simply because they have risen along with stocks generally.
I have no problem with selling an overvalued stock to buy an undervalued stock. Or even selling a slightly undervalued stock to buy a hugely undervalued stock. But I don’t see why you would want to sell your shares in a company you know and like simply to hold cash.
So, my own feeling is that you should always err on the side of inactivity. The bar for buying should be set high. And the bar for selling should be set high. The bar for doing nothing should be very, very low.
All valuations are ultimately relative. Even the Shiller P/E approach is basically using long-term interest rates over 100+ years. It's just doing it indirectly. In the long-run interest rates (especially the longest dated – 30 year+ corporate bond rates) help determine the value of stocks by either offering attractive or unattractive returns and thus being an alternative to stocks.
It is this competition between investment options: stocks, bonds, land, commodities, cash, etc. that leads to talk about interest rate based valuations.
Of course, bonds themselves could be badly mispriced. So, you should never compare the expected return in stocks against the expected return in bonds alone. You want to compare every asset you value to a variety of assets. In fact, right now, most individual investors should start by comparing their expected return in stocks to their expected return in owning a home. Since buying a house right now is clearly a much safer form of long-term purchasing power preservation than buying a bond (which might have some speculative appeal if you expect deflation – but that’s about all bonds offer right now).
So, I think you can – for the purposes of doing an exercise – always value the stock both ways and consider the band between the roughly 15 times historically normal P/E and the current 22 times Shiller P/E to be the band of possible intrinsic values for the stocks you are looking at.
In other words, if their earning power is $1 a share, you should understand that means they are neither definitively worth $15 or $22 but rather they are worth something in the $15 to $22 range.
Where true value lies may end up having as much to do with interest rates as with the particular business you are analyzing.
As for your buying and selling decisions, I would suggest you use the lower Shiller P/E (the long-term average of 15) for your buy decisions and today's higher Shiller P/E of 22 for your sell decisions.
Don't use the same number for your buy and sell decisions. This is just an invitation to overprecision and overtrading. Be less precise. And more patient. When in doubt, stick with what you know. Be slow to sell good stocks you own. And be slow to buy new stocks you've never owned.
Personally, I do not focus on intrinsic values. I focus on the returns I think are possible in the stock and a conservative calculation of what I think the stock is clearly worth more than.
Ideally, I try to find a stock that I think offers safety in the sense that if I know it is worth more than 2 times book value, and I pay 1.3 times book value – I am getting the stock for less than it is worth (even though I don't know exactly what it is worth, I know I paid less than that mystery number). And then I want to see a reasonable way for me to earn an acceptable annual return on my investment over the next 10 years. I define and acceptable return as 10% a year.
These are my criteria for buying a stock:
1) Price clearly lower than value
2) Acceptable return over 10 years (10% a year)
3) Better than alternatives
I use 10% as my minimum acceptable expected return over a 10 year holding period. For me, even if interest rates are at 4%, I would prefer to hold cash and wait for an opportunity I think can return 10% than ever invest in something that only has single digit return potential.
Because I think someone – anyone really – who is diligent, disciplined, and practicing the principles of Ben Graham and Warren Buffett can find a way to make 10% whatever the market is doing. I don’t mean in any one year. But I do mean over any one decade. My portfolio is 50% in Japan right now. You can invest in micro caps, foreign countries, net-nets, special situations – whatever. There are ways to find 10% returns even in tough environments as long as you are a value investor working with small sums of money. So that’s my hurdle rate: 10% a year. If I don’t think an investment can deliver that, I’ll hold cash.
For alternatives, I look at the other stocks I might buy. But even more than that, I focus on the stocks I already own. Would adding to them be a better use of my cash? This can be hard because sometimes there is an uneven comparison in terms of safety and return. Sometimes a company I like less has a share price that is so low its huge potential to offer very good annual returns over say – 3 years – is hard to pass up.
But I try to stay focused on the investments where:
1. Present day undervaluation is clearest
2. Acceptable 10-year returns (10%+) are likeliest
3. Safety is highest
4. Business quality is best
Also, remember that valuation tools are general tools. Not specific tools. I've talked sometimes about companies like DNB, CEC, and BDMS. Their capital allocation policies alter my views of their intrinsic value. Because I think the likelihood of adequate returns in those stocks is increased by their policy of not over doing reinvestment in their own operations but instead buying back their stock, paying dividends, etc.
This is why Warren Buffett values IBM so highly.
So, using the kind of approach I used in the Microsoft and Exxon examples is just a starting point. It just values generic free cash flow. It does not value the uses to which management will put those free cash flows. If the use is poor, the stock's earning power should be valued at less than $1 for every $1 of retained free cash flow. If the use is good, the stock's earning power should be valued at more than $1 for every $1 of retained free cash flow.
Investing involves the valuing not just of capital but of the return on that capital and then the subsequent return on that return on the original capital – and so on. This is critical. And it alters the valuation of companies like IBM.
If IBM used its capital differently, I Warren Buffett wouldn’t have paid $171 a share for the stock.
Keep this in mind when actually applying your intrinsic value estimates. A company is more than its future cash flows – it is the uses to which those cash flows will be put.
Ask Geoff a Question about Intrinsic Value Estimates
Check out the Buffett/Munger Bargains Newsletter
Check out the Ben Graham Net-Net Newsletter
Some who reads my articles asked me this question: Hi Geoff,