No. of Recommendations: 33
You say: before taking into account future prospects (which is of paramount importance), the big 7 tech do seem to be very richly valued compared to the rest of the crowd: over twice the price for each dollar of current earnings.
"richly valued ..." measured on "... current earnings". Does that matter at all? Or does it only distract from what counts if trying to compare?Allow me to indulge in a bad metaphor : )
Say I tell you I weigh 220 pounds. Knowing nothing else, would you conclude I'm fat?
Probably so, statistically speaking. But maybe I'm super tall, or a body builder, or have gigantism, or some of my hard parts have been replaced with solid gold.
So, you can't be sure.
The lessons from the metaphor:
(1) The most basic obvious metric gives you a probability, but not anywhere near a certainty, of the obvious conclusion.
(2) If you want to have any confidence that the obvious conclusion is not correct, you have to find an alternative explanation sufficient to explain the discrepancy.
Now, imagine I told you that the average weight in my town is 220 pounds.
Is our population having a weight problem?
Yes, you can be pretty sure. It's not just one person, so the "exceptional" argument is unlikely to be true for all of them.
If it's a super tiny town full of tall male bodybuilders
maybe it's OK, but the odds are getting pretty remote.
Thus our third rule of generalizations from obvious metrics:
(3) It's impossible for the entire population to be exceptional.
So, back to investing:
A high current P/E is not proof that a company is overvalued, not by a long shot.
It absolutely can't be relied on, but it correlates positively. So to conclude that a high-P/E company is NOT overvalued requires you to find the evidence of why it's exceptional.
That's not always hard: Very high earnings expected with high confidence a few years from now? That will do fine.
Only the aggregate future earnings matter, not the current earning rate.
But, axiomatically, not all firms can be above average.
If the same metric is seen in a big collection of companies, the "it might be an exception" reasoning becomes tougher.
At the extreme, we know that the sales of all companies in aggregate can't grow faster than the economy that they're embedded in, and that their earnings can't grow faster than their sales over time.
The bigger the group of companies you're talking about, the more sure you can be that the obvious conclusion is likely the correct one.
The seven tech companies account for 11% of all revenues of the non-bank firms in the S&P 500, and it covers a huge variety of operating units...that's starting to look like a statistical sample.
It's a very very skewed sample, but it's starting to be too big in aggregate to be a huge outlier.
Now, others may disagree, but I observe that it is vanishingly unlikely for a richly valued firm well into its profit-making years still to be trading at over 20 times average earnings a decade later, so let's assume that won't happen.
It's doubly certain for a large sample of business units. We can't predict each one trivially, but a bigger collection is more predictable.
So, we assume that the Big Tech Seven are trading at a multiple of (at most) 20x many years down the road.
The math is then easy: if their total aggregate earnings grow at (say) inflation + 14%/year in the next 10 years and then you sell at a multiple of 20 times those cyclically adjusted future earnings, you'll still get a below-average return. (6.37%/year)
It then becomes a matter of how much confidence you have that they will have aggregate earnings growth of inflation + 15%/year or more. Some might, but all of them taken together?
Given their size, and the fraction of the economy they represent, and the diversity of their businesses despite all being called "tech", I am personally dubious.
=====================================
Regarding your specific examples of price versus future earnings for certain companies---
I don't want to quibble about any specific one. The whole point of my bad metaphor is that you can have high confidence about a conclusion about a set, but almost no confidence about any specific member of the set.
I would add that my own rule of thumb is always to value a firm based on the "pretty darned sure" future.
Imagine seven possible outcomes, from "everything comes up roses for years to come", down to "headquarters is hit by an asteroid".
I generally plan on the second-worst scenario. Worst case is always a wipe-out so you'd never invest, so look a bit higher than that.
Lots and lots of companies will look good based on analyst estimates of revenue and sales growth rates, or even based on your own conservative "most likely" middle scenario.
But very very few firms look good based on the "pretty darned sure" outcome, because few firms are predictable enough that outcome 2 of 7 isn't a way worse than outcome 4 of 7.
So: I think the exercise you have done is the right one, but you have to stomp on every assumption with iron clad boots to crush all conceivable optimism out of the forecasts.
Once you stamp out the optimism, the sort order of attractiveness of investments shifts a lot, away from the ones with the best central expected return towards the ones with the best "things went badly and I still did well" return.
But, with sufficient pessimism in the forecasts, I think the ratio of price to future real earnings (say, average 5-10 years out) is still the best thing to look at for almost all equities.
Some specific comments, just because I'm never at a loss for an opinion:
Yes, BABA looks very good on the numbers, but there are other things at play. If the CCP decides they should no longer make money, or that foreign shareholders shouldn't participate in that, it simply won't happen.
More than 30% of the value of a share has been taken away without compensation two or three times already, so it's not exactly a fanciful concern.
So, in my estimation, the second-worst-of-seven outcome is a huge loss, and I will no longer risk any money on them.
Rule #1 is called that for a reason.
Yes, I too think Carmax looks pretty good. It's currently one of my largest non-Berkshire positions.
An excellent example of a firm where the current P/E isn't telling you much.
A good start is to estimate how much they'd be making in owner earnings per share this year, if this year were neither unusually good nor unusually bad.
Net profit margins would probably be in the 3.5%-4.5% range. Profits per share would therefore be running over $6.50, not around $3.
I think your 12% revenue growth rate might be a bit rosy for my conservative tastes...I would rather be a pessimist now and pleasantly surprised later than the opposite.
With real revenue growth rate per share of 7%/year and net margins averaging 3.5%, one can easily imagine 5-10 year returns in the low double digits from today's price.
An exit multiple of only 15 on real sales/share growth of 6%/year and average net margins of 3% would still get you a (slightly) above average return over 5-10 years: would that count as second-worst-of-seven?
Maybe not quite, but I do have a fondness for the firm.
I like the way they sail through recessions with very few bumps, and, because I like to trade, I like the way the stock price goes up and down nonsensically on a regular basis.
Re Alphabet, I agree that the outlook isn't a slam dunk high return now, as the price has moved up from $90 last winter to $130 now.
My calculation approach summarized here
https://www.shrewdm.com/MB?pid=687699769But, making up for that, in my assessment the "second worst" business outcome is still wildly profitable and successful. Reliability of the outcome counts for a lot.
In short, that's what determines prudent position sizing, which in turn is what is the biggest determinant of your absolute profit on a position.
I find Berkshire extraordinarily predictable (so far!), so it's my biggest position, and has been for many years. So it is what has made me the most money.
Jim
