Halls of Shrewd'm / US Policy❤
No. of Recommendations: 33
Sort of following on from the bearish slant to Mr Klarman's recent comments--
I don't claim to know what the market will do, nor the few gigantic firms that dominate it.
But you don't have to be the sort of person who tries to time the market to be interested in valuation levels.
If you care about what your portfolio returns will be in future, you should care about how much you're paying for what you're getting.
It's quite reasonable to consider valuation levels when you estimate what it would take to arrive at a decent forward return.
If you care about the forward returns of the broad US market, or the potentially over-optimistic valuations of the giants within it, here's one simple back of the envelope exercise which I think is worthwhile considering.
I assume that nothing experiences super-high growth rates for more than a decade (it does happen, but not predictably).
Since high valuation multiples come from high anticipated growth, I assume that terminal multiples for anything will be in the teens after a few years when middle age sets in.
So, I pencil in a multiple of 19 time cyclically adjusted earnings, tops.
My usual investment horizon for predictions is 5-10 years, being the longest any person can reasonably have an opinion about the prospects of a business.
Whatever good thing you think is going to happen, most of it had better happen by then, since the rest is peering through a glass very darkly : )
The middle of that range is 7.5 years into the future.
If you combine those assumptions with the current price and current earning power, you can estimate the rate of earnings growth that will be required to hit any given desired rate of return from the stock.
A representative "monkey with a dartboard" hurdle rate is inflation + 6.5%/year, what you'd historically have managed with a random US stock bought in a random year.
So I wouldn't invest in something that doesn't give me a better-than-even chance of achieving that result.
The top 7 popular firms in the US market by market cap are currently trading at a collective weighted-average earnings multiple of 36.96.
That's the aggregate market cap divided by the aggregate earnings. (AAPL MSFT GOOG AMZN NVDA TSLA META)
For earnings, I'm using the average of trailing and forward numbers--a plausible estimate of what's going on right now.
If they were all merged and those 7 together were one stock treading at $100 per share, it would have current-run-rate earnings per share of about $2.71.
Here's a possible future table, with the multiples gradually falling at a slowing rate, but earnings rising at a constant rate.
Real Real
Year P/E EPS Price
1 36.96 $2.71 $100
2 31.9 3.22 103
3 28.0 3.83 107
4 24.9 4.55 113
5 22.5 5.41 122
6 20.6 6.43 133
7 19.2 7.64 147
8 18.0 9.09 164
9 17.1 10.81 185
10 16.5 12.85 211
The average multiple in years 5-10 is 19.0, my chosen upper bound for the medium/long term.
The average price in years 5-10 is $160.20, giving a rate of return of inflation + 6.49%/year for ~7.5 years, my "monkey with a dartboard" rate.
In order for that table to work out, earnings for the group as a whole have to rise at a rate of inflation + 18.9%/year.
In a word, yikes!
I'm not saying it won't happen. I can't see the future.
But that does seem like a very ambitious hurdle for the minimum earnings growth needed to get a historically average real return.
Any slower growth rate, still with a sane terminal multiple (which is all the more likely with a slower growth rate), will give a lower or possibly negative return.
This exercise is of course going to be wrong as a forecast of the returns for those firms, but I think it's still an important one to try out.
You can mess with the assumptions to taste, but whatever rate of return you anticipate has to be consistent with the assumptions you're making.
Those 7 firms make up about 27.5% of SPY these days, so their outlook matters to anyone holding the broad cap-weight US market.
Jim
No. of Recommendations: 5
Thanks Jim! That is the dose of reality, put into numbers, that I always ask for from those with blind faith in the index being superior to BRK going foreward.
No. of Recommendations: 1
Thanks Jim. I own googl. and Amzn (combined 6pct of portfolio). Sold my Meta on Friday for a healthy profit and have for the first time buying RSP which you brought to my attention in previous posts. Berkshire is already 64pct of my portfolio so not keen to add to it unless it gets ridiculously cheap again
No. of Recommendations: 0
I own googl. and Amzn (combined 6pct of portfolio). Sold my Meta on Friday for a healthy profit and have for the first time buying RSP which you brought to my attention in previous posts. Berkshire is already 64pct of my portfolio...
Now that is a diversification strategy.
64% is good but 65% is too much!!
Do you bbelieve BRK will outperform the index surely or not?
No. of Recommendations: 0
With Berkshire, perhaps it slightly beats the index or matches it but i dont see it underperforming however i suppose it depends on your time horizon.
No. of Recommendations: 9
For earnings, I'm using the average of trailing and forward numbers--a plausible estimate of what's going on right now.
The problem with your analysis is that P/E ratios have proven to be near worthless when valuing software-driven tech companies. You can't conclude any of them are overvalued by looking at P/E ratios alone and projecting them into the future. That's because you don't know what new multi-million or multi-billion dollar revenue streams they are in the process of unlocking by leveraging their current technology and customer footprint.
Unless you were in the board room, you couldn't have known in advance that Amazon would expand from selling books online to selling just about any item that can be shipped to your door through a vast 3rd party seller network, thereby more than 10x'ing their revenue in just a few years. You couldn't have known that they would create a multi-billion dollar subscription service called Prime that would then compete with Hollywood movies. You would have been skeptical when they first announced that they would offer a cloud service. No way you could project that service has grown to a $60 billion revenue run rate and expected to hit $100 billion in a few years.
Similarly, you couldn't have known what Microsoft's AI strategy was 10 years ago. You knew that Office 365 was doing well, and Azure was getting started. You now know that MSFT is the primary investor in OpenAI. Generative AI and AI in general is expected to have a $7T impact on global GDP. Yes, that $7 trillion, with a T.
Similarly, you can kind of speculate that Tesla will get into the Robo Taxi business and so will Uber. What will be the impact on world GDP when people have less of a need to own their own cars? Did you know in advance that Tesla would create a new $5 billion revenue stream just by opening up its charging stations for other car companies that they compete with?
Trying to be an accountant using a P/E ratio as a measuring stick in the face of disruptive forces like these is useless.
And you don't have to take my word for it. Here's a snippet from William O'Neill, the founder of Investor's Business Daily:
"Decades of market research finds that winning stocks tend to have P-E ratios that value investors consider too expensive ' even at the start of their giant price runs.
As long as a stock has superior fundamentals, institutional support and other traits of market winners, the valuation doesn't really matter. Google parent Alphabet (GOOGL) famously reached a P-E ratio in the 50s and 60s while its share price went from 115 in September 2004 to 475 in January 2006.
The folly of P-E ratios was discovered when IBD founder and Chairman William O'Neil conducted research on what exactly caused stocks to explode in price. Strong earnings growth, especially in the two most recent quarters, was a key factor shared by these winners. But the P-E was usually well beyond acceptable levels when stocks began their big price advances. In fact, P-Es get even higher as the advance gains strength."
No. of Recommendations: 6
Similarly, you couldn't have known what Microsoft's AI strategy was 10 years ago. You knew that Office 365 was doing well, and Azure was getting started.
Well I did pretty well when I bought MSFT (and Leaps) at ~10 earnings after the GFC. At that point no heroic assumptions about the future were necessary to conclude that MSFT was a good investment. Today I don't know, perhaps they'll win big with AI or perhaps not. I certainly wouldn't base my investment decision on some rosy projections.
No. of Recommendations: 36
For earnings, I'm using the average of trailing and forward numbers--a plausible estimate of what's going on right now.
...
The problem with your analysis is that P/E ratios have proven to be near worthless when valuing software-driven tech companies.
Um, nope, not buying it. It's not different this time.
Recall that I am talking about a collection of firms with a collective market cap over $11 trillion, not a single software startup.
But ignore that.
The only reason for a firm to have a low earnings yield now is because people expect (and deserve) lots of earnings later. It makes sense.
It can't be forever postponed, because people will eventually stop believing the forever-delayed hyped future.
Eventually, the profits have to arrive, or the price will fall along with the strength of belief.
Low earnings now aren't a problem; low distant-future earnings relative to price now is a big problem.
If even the purchaser doesn't believe in biggish future profits, but rather relies on a future optimist to buy from him at a high multiple because of the assumed high profits in an even-more distant future, the purchaser today is (statistically) very likely to be disappointed.
The key insight is that the future eventually arrives. No firm trades at nosebleed valuations forever. The valuations will fall to earth with 100% certainty; the only question is the time frame.
Especially if they already account for a visible fraction of all national business activity.
(one simple proof: the value ultimately comes from the future earnings, which have to come from the future sales.
Neither sales nor net margins can grow forever faster than the economy upon which they depend. Though most firms hit the growth rate wall very much sooner than that)
I have no doubt that some few of today's high flyers of various sizes will still be trading at high valuations of future earnings more than ten years from now.
However, I am willing to assert with confidence that nobody knows which ones those are with sufficient certainty to be sensible betting cold hard cash on it.
Everything is easy with hindsight. Very little is easy in advance.
It was blindingly obvious that Microsoft was a great buy at 76 times earnings 23 years ago, but...wait...no it wasn't.
Despite the fact that the business did extremely well, it was trading at single digit multiples of much higher earnings within a decade.
That was an overshoot, but a descent into the teens was completely predictable.
And you don't have to take my word for it. Here's a snippet from William O'Neill, the founder of Investor's Business Daily:
"Decades of market research finds that winning stocks tend to have P-E ratios that value investors consider too expensive ' even at the start of their giant price runs.
This is a fine statement, but 100% backwards looking.
It's like saying that every single lottery winner was a purchaser of lottery tickets.
True, but without predictive power: lottery tickets will lose you money on average.
The average firm that did very well had a high multiple at the outset? Fair enough. (though not always true)
But prospectively, what is the average long run rate of return of firms purchased at very high multiples? It's negative.
Or perhaps merely very low, depending on the precise test set-up.
Trying to be an accountant using a P/E ratio as a measuring stick in the face of disruptive forces like these is useless.
Again, no.
I'm not measuring a firm based on its price to earnings ratio, but on its price to FUTURE earnings ratio--many years out. There's quite a difference.
I have put lots of money into as-yet unprofitable businesses...but never into businesses that I didn't see a probable path to solid future earnings that look good relative to the price paid at the start.
I'll pay a bird in hand for two in the bush. But those willing to pay three or four birds in hand for two in the bush will do badly on average.
But it's good to see that the tech bubble mentality isn't dead : )
In order for some folks to beat the market, it's necessary for there to be other investors who underperform it.
More often than not, they are the overoptimistic ones who are willing to overpay for a rosy future because we old-fashioned folks just don't get it: we keep stubbornly insisting that a business eventually has to be worth what we pay for it.
Bottom line, going into an investment and expecting the ability to sell 7+ years down the road at a very high multiple of future earnings is a recipe for disappointment.
If the earnings don't come by then, those future optimistic buyers are usually very thin on the ground.
There isn't any need to require current earnings from a business, but it's usually best not to overpay for what you yourself believe their future income will be.
If you pay today over 15 times what the earnings turn out to be a decade later, you'll have a poor return with very high probability.
Jim
No. of Recommendations: 4
If you pay today over 15 times what the earnings turn out to be a decade later, you'll have a poor return with very high probability.
That might be absolutely correct, but I think wasn't the point of mechinv. Instead it was the impossibility to estimate for an individual(!) Tech company what a decade later those earnings might be, based on "average of trailing and forward numbers - - - a plausible estimate of what's going on right now.". That this technique in the world of Tech simply does not work to "measuring a firm based.... on its price to FUTURE earnings ratio--many years out". That it does not work because
you don't know what new multi-million or multi-billion dollar revenue streams they are in the process of unlocking by leveraging their current technology and customer footprint.
That for estimating earnings in X years for Tech other means are required, like (he didn't say that, I do) the "Deep Dive" of tech insiders at Saul's, or whatever. Or simply nothing might work as the future of Tech companies simply is too unpredictable, even for experts.
All that of course does not apply to Tech companies as a group, only to an individual one.
No. of Recommendations: 2
Or simply nothing might work as the future of Tech companies simply is too unpredictable, even for experts
How about you create a basket of tech companies, no way of knowing which will succeed, don't want to cap-weight them, again, can't predict which will go BK and which will flourish. If only there was an ETF that would do that for you???
Rgds,
HH/Sean
No. of Recommendations: 2
I think that's opposite to what both have to say, mechinv and Jim. Jim's original argument was that exactly such a basket is super-expensive:
The top 7 popular firms in the US market by market cap are currently trading at a collective weighted-average earnings multiple of 36.96. That's the aggregate market cap divided by the aggregate earnings. (AAPL MSFT GOOG AMZN NVDA TSLA META)
So even if one or even two of them are 5x winner in 5-10 years,what would be the gain (or loss) if the other 6 then did revert to "normal" valuation levels?
No. of Recommendations: 5
If you pay today over 15 times what the earnings turn out to be a decade later, you'll have a poor return with very high probability.
...
That might be absolutely correct, but I think wasn't the point of mechinv. Instead it was the impossibility to estimate for an individual(!) Tech company what a decade later those earnings might be, based on "average of trailing and forward numbers - - - a plausible estimate of what's going on right now.". That this technique in the world of Tech simply does not work to "measuring a firm based.... on its price to FUTURE earnings ratio--many years out".
That's certainly a reasonable point for a young firm whose profit-making years haven't really started, or whose business model hasn't settled down yet.
But I don't think that can be applied to the big 7 I mentioned, all of which are over 20 years old.
Four are mature and very profitable (AAPL MSFT GOOG META).
The three others are also into their profit making years (TSLA, NVDA, AMZN), but it seems that much more profit growth is anticipated based on current valuation levels.
For firms well into their intended operating mode with a real business model and real net margins, there is no particular problem using current earnings as the baseline for estimating an earnings growth rate.
Only Tesla is still a bit early in their profit making history, and only Amazon has a recent net profit margin under 13% (due to the nature of the retail business, not their corporate youth)
Certainly as a group there is no problem of a near-zero profit baseline causing mathematical difficulty talking about earnings growth rates meaningfully.
They each made many billions of dollars of profit in the last year, and collectively made $216 billion.
It's only because these firms ARE collectively mature and into their profit making years that I talked about an earnings growth rate, which I agree does involve having a reasonable and positive starting level.
The more general rule is simpler: no matter how young and unprofitable a firm is, the market price has to be in proportion to the buyer's anticipated distant future profits.
If not, then even many years of upcoming explosive growth in revenue and profitability still amounts to simple overvaluation.
If the future profits are paltry, to get a profit you will need the stock to be trading at a very high multiple of earnings in the distant future.
It doesn't pay to rely on finding such a buyer when the time comes, especially as about the only thing you know about that future firm is that its earnings trajectory has disappointed.
Jim
No. of Recommendations: 0
How about you create a basket of tech companies, no way of knowing which will succeed, don't want to cap-weight them, again, can't predict which will go BK and which will flourish. If only there was an ETF that would do that for you???
Well, QQQE ain't that. It continuously sells winners and buys losers. Over the long term, hard to see how it can beat "buy all components once, in equal measure, and hold forever".
No. of Recommendations: 15
Well, QQQE ain't that. It continuously sells winners and buys losers. Over the long term, hard to see how it can beat "buy all components once, in equal measure, and hold forever".
Bear in mind that different time frames have different properties.
They are selling winners and buying losers only on the time scale of quarterly price fluctuation, which is (as I have measured) a modestly profitable activity on average.
Most short term squiggles are transient, though certainly not all of them.
But over the long haul, the index membership gradually evolves.
It is based on sorting by market cap, but not too frequently, and with a "fuzz" around the limit to avoid buy-high-sell-low at the boundary.
(membership requires entering the biggest 100, but you don't get tossed out till dropping below #105, so it's pretty sticky)
New successful firms are added, very long run winners remain members, and firms not keeping up get tossed out. So at a longer time scale it's the reverse effect, keeping winners and cutting losers.
The thing that interests me is how the economics work.
In any rational world, the set of "the 100 biggest companies that aren't financials and happen to be listed on our exchange" does not make sense as a recipe for higher than average earnings growth over time.
Some would be great, but you'd also expect your fair share of duds, netting out to ... average. Nature abhors a vacuum, and markets abhor a get rich quick scheme.
Yet the trend of average earnings within the Nasdaq 100 set has (barring dips in recessions) risen at a rate that has been remarkably consistent, and shockingly higher than for the S&P 500.
Average earnings (which is what the value of QQQE tracks) have been rising around inflation plus 6-8% for decades now.
That dwarfs the equivalent figure for the S&P 500.
Even a substantial slowdown (which any rational person should expect) would leave the Nasdaq 100 set as the clear leader.
Jim
No. of Recommendations: 18
Um, nope, not buying it. It's not different this time.I agree that it's
not different this time. The fact that P/E ratios, by themselves, are useless for valuing high-growth companies was known back in the 1ate 1970s. That was when
Peter Lynch took over the reins at Fidelity Magellan and delivered an astounding
29.1% annualized return for his investors from 1977 through 1990. A $10,000 investment in the Magellan Fund would have grown to $280,000 during the 13 years Lynch ran the fund. As he explains in his classic book
One Up on Wall Street (which I recommend you read), Lynch decided long ago that P/E ratios were not useful for deciding when to buy growth stocks. He used the PEG ratio ratio instead, but his most important criteria was to "buy what you know".
So you are correct that it's
not different this time. P/E ratios were useless for valuing growth stocks in the 1970's and they are
still useless for valuing growth stocks 50 years later.
If you pay today over 15 times what the earnings turn out to be a decade later, you'll have a poor return with very high probability. ... The valuations will fall to earth with 100% certainty; the only question is the time frame. 100% certainty? Really? What would you say if I told you that in the 26 years that Amazon has been publicly traded that it has
never traded at a multiple lower than 20 times earnings? The same is true of Dow member Salesforce.com whose
lowest P/E ever was 48. $10,000 invested at Amazon's IPO would be worth over
$16 million today. And $10,000 invested in Salesforce's IPO nineteen years ago would be worth $570K today. Even during the Great Recession of 2008, you had to pay 96 times earnings to get a share of Salesforce. So what? You have to pay up for the type of revenue growth that Salesforce delivered. If you're curious, that company under Mark Benioff grew from $5M to $100M in revenues in only 2 years from 2001-2003, and then went on to generate $1 billion in revenue in only 6 years. I can give you many other examples where you could retire rich from these market-dominating growth stocks without their P/E ratios
ever "falling to earth".
Now, this is the Berkshire board, and I think that Buffett is running a fine company. If you're an investor in BRK.[A or B] I support that. Good for you. What I object to is the constant assertions over the years that the S&P 500 was overvalued relative to Berkshire. Those assertions were just plain wrong. Let's look at the annual return numbers over the past 5, 10 and 15 years. Easy to do using PortfolioVisualizer.com.
Annualized return
5 yrs 10 yrs 15 yrs
------------------------------------
S&P 500 12.3% 12.8% 10.9%
BRK.B 12.8% 11.8% 10.1%
If index funds were "overvalued" relative to Berkshire 10 years and 15 years ago, why has the index beaten BRK.B over those time periods? Yes, the S&P 500's P/E ratio back in 2013 was much higher than Berkshire's but this just shows that trailing P/Es are faulty measuring sticks for gauging forward returns.
If you're in your wealth-building years, and especially if you're a Millenial or Gen Z, ignore these assertions about the market being overvalued. These claims have been made many times ad nauseum over the years, but they've mostly turned out the be wrong, as the above figures show. The best way to reach your retirement goals is to save 10% of your income into a Vanguard-type index fund in your 401K. Put that investment plan on autopilot, focus on your career and have some fun with your hobbies. Your youth is a very precious time in your life.
No. of Recommendations: 8
"What I object to is the constant assertions over the years that the S&P 500 was overvalued relative to Berkshire. Those assertions were just plain wrong. Let's look at the annual return numbers over the past 5, 10 and 15 years. Easy to do using PortfolioVisualizer.com. "
Great post. Dedicated Boards like this tend to become echo chambers since many of us are heavily invested in Berkshire. Your point on the S&P performance is spot on. Of course, the argument you may get back is you can pick any start and end point. 15 years is long enough.
As Charlie says, you want to understand the bear case better than those supporting it.
No. of Recommendations: 35
I agree that it's not different this time. The fact that P/E ratios, by themselves, are useless for valuing high-growth companies was known back in
the 1ate 1970s. That was when Peter Lynch took over the reins at Fidelity Magellan and delivered an astounding 29.1% annualized return for his
investors from 1977 through 1990. A $10,000 investment in the Magellan Fund would have grown to $280,000 during the 13 years Lynch ran the fund. As
he explains in his classic book One Up on Wall Street (which I recommend you read), Lynch decided long ago that P/E ratios were not useful for
deciding when to buy growth stocks. He used the PEG ratio ratio instead, but his most important criteria was to "buy what you know".
So you are correct that it's not different this time. P/E ratios were useless for valuing growth stocks in the 1970's and they are still useless for valuing growth stocks 50 years later.I am asserting that the ratio of FUTURE earnings to CURRENT price is important. More than anything else.
An "expensive" current P/E ratio for a specific firm isn't bad, especially if it is youngish or smallish, provided the future earnings eventually arrive and are plentiful enough to make it all worthwhile.
The 7 largest firms in the US do not fall into either of those categories, and even less so taken as a group, but the same thing is true: it's the future profits that matter, and how much you pay today to get them.
The main thing we
learn from a low or negative current P/E is to identify the firms that investors are optimistic about. They expect a rosy future so they aren't requiring current earnings--fair enough.
Future growth is expected. It might happen or might not, but it's clear that people are expecting it and paying extra for it.
And the most useful
observation of that group is that they are, on average, poor investments.
This confirms your assertion that low P/E is not useful to predict a good investment, but a high or undefined P/E is even worse as a predictor of a good investment.
You have to look at something else to find the exceptions among high P/E firms that will do well, because as as group, you'll do badly.
(Among low P/E firms you're somewhat less likely to do badly because the average is not as bad*)
Among those firms with high valuations a decade ago (low or negative current earnings yields) among the 1700 covered by Value Line,
there is almost no difference in return between the set with low earnings (P/E over 33, say) and those with negative earnings.
The first group returned inflation + 3.0%/year in the last decade, and the second group returned inflation + 2.4%/year.
(those are error-prone figures: more than half of the firms in that group could not be easily matched by ticker now,
generally because they were bought out or delisted or changed tickers, so I averaged only those that matched. Some of the matches will be wrong, too.
I suspect, but am unwilling to go to the work to prove, that the numbers would be even worse if we found the fates of the missing tickers and included them in the average)
The point here is that both of those returns are really terrible: on average across the group, a low earnings yield (high P/E) was a successful predictor of a poor outcome.
The S&P 500 returned inflation + 9.6%/year in the same ten years.
The reason is simple: On average over the centuries, the typical stock with a low earnings yield (high P/E, if there is a P/E) is overvalued.
Some definitely aren't, but the average one is.
They generally do have high growth, and high growth is definitely worth more.
But the typical situation is that people get excited and overpay for that bright future they anticipate.
A firm might be worth a lot more because it is going to grow a lot--nobody would dispute that--but people tend to pay 3 or 4 times as much for something that's ultimately worth only twice as much, and end up with poor returns.
In this particular case, we can even put a number on it: if we assume for the sake of argument that the return from the S&P 500 is what they "should" have got,
the optimism a decade ago cause people to pay 97% more than they should have, on average.
The numbers will vary depending on the time frame you look at, this is a pretty typical result. People usually overpay for growth prospects.
It isn't useful to find some exceptions and declare high valuation levels to be the sign of rapid growers or good investments.
Some are, but most aren't as good as their current prices assume.
The ones that turn out well are either (a) distant future profits turned out to be sufficient relative to the price originally paid, or (b) the investor got lucky enough to sell to a future optimist.
I do like investing in super-high-growth firms, with no profits or low profits. But I won't overpay for them.
If I'm putting $10 at risk, I want to see that achieving honest profits on that of $1/year 7-10 years down the road doesn't require any absurd assumptions.
The only other alternative is to rely on luck, which isn't much of a plan.
If you pay today over 15 times what the earnings turn out to be a decade later, you'll have a poor return with very high probability. ... The valuations will fall to earth with 100% certainty; the only question is the time frame.
...
100% certainty? Really? What would you say if I told you that in the 26 years that Amazon has been publicly traded that it has never traded at a multiple lower than 20 times earnings? I haven't checked the precise lowest figure for Amazon's multiples, but sure. What's your point?
They will trade at a multiple in the teens as a norm, as does every firm sooner or later. As mentioned, the only question is when. For them, unusually, it's "not yet".
The poor ten year outcome I mention is stated quite clearly as "very high probability"---Amazon is simply an outlier in that high valuation period is much longer than typical.
They aren't enough to change measurably the average outcome of stocks bought at high multiples of distant future profits, which is dismal.
As an aside:
Interestingly, in any year of your choice, forecasts for Amazon's profits ten years later were very much higher than what ultimately occurred.
At any individual date of high valuation, people were quite sensibly expecting a lot of profits within a reasonable time frame, much as I suggest is prudent.
So Amazon investors in 2000 weren't willing to wait 20 years: they expected lots of profits in 10 years.
They then morphed into Amazon 2005 investors willing to wait 10 years, then they into Amazon 2010 investors willing to wait 10 years, and so on.
That's not a problem for them--Amazon has built some formidable businesses and there is still an expectation of future profits--but it's an interesting phenomenon of investors rewriting their memories.
I presume that if you told the Amazon investors of 2000 (or whenever) what the total profits per share would be 10-20 years later (but not the share price) they mostly wouldn't have been buyers.
Annualized return
5 yrs 10 yrs 15 yrs
------------------------------------
S&P 500 12.3% 12.8% 10.9%
BRK.B 12.8% 11.8% 10.1%
If index funds were "overvalued" relative to Berkshire 10 years and 15 years ago, why has the index beaten BRK.B over those time periods? Well, since you ask, it's because of the endpoint effect of the date you chose (now).
In the specific interval you cite, the S&P 500 has become more expensive and Berkshire hasn't.
Berkshire's business results have been moderately superior, by an amount about the same as the amount by which the relative valuation levels changed.
Based on smooth trend earnings yield, the S&P is 32% more expensive than it was 10 years ago, and 55% more expensive than it was 15 years ago, and 32% more than 20 years ago.
That's based on smoothed recent profits, so it's on top of the upswing in profits from the tax cuts and recent net margin expansion.
By comparison Berkshire's valuation is relatively unchanged in the last 10 or 15 years based on price-to-peak-book or other methods.
Maybe the S&P 500 will continue to get ever more expensive. I personally wouldn't bet on it.
FWIW, a wild guess as to a plausible outlook for the S&P 500:
Starting from valuations similar to this on starting dates since 1995, if the future rhymes in terms of value growth and valuation levels and ongoing multiple expansion,
you'd expect a real total return from the S&P of around inflation + 1.9%/year in the next 7 years.
That's defined as the average of the annualized rates of return, across all possible holding periods starting now and ending in the stretch 4-10 years later.
Maybe more, maybe less, but higher or lower outcomes are probably a 50/50 shot.
Jim
*
Among the ~1700 firms covered by Value Line, last 37 years, total return CAGR among the ~250 firms with highest current P/E at purchase: 6.81%/year
Among the ~1700 firms covered by Value Line, last 37 years, total return CAGR among the ~250 firms with lowest current P/E at purchase: 14.55%/year
S&P 500 same period: 10.58%/year, between the two.
No. of Recommendations: 3
Speaking of big companies, I've explained why Microsoft shares gained almost 4% today over at the MSFT board here at
https://www.shrewdm.com/MB?pid=101761464 Those of you who are interested in chatting further on this topic, please reply at the MSFT board, not this one.
Thanks.
No. of Recommendations: 29
I know better than wading into these facts and figures discussions, and it's fine for anyone to always look at every detail of a company's numbers and such, so be it, but it just isn't necessary to get really good investing results. I will wade in just to speak up for the concept I've followed for 3+ decades of simply buying good to very good companies and holding them for the long term, rarely selling, never looking at a financial statement, or paying much attention to anything about the numbers. The long ago abandoned, but orignial concept of TMF was LTBH. They have provided an investor with capable companies over the years, and it wasn't a stretch to figure out that technology was the place to be, in good part, and it hasn't been too difficult to build a really good portfolio for the long haul using their picks. And I had to g r o w my money back in the day, not preserve capital. There wasn't much to preserve 20 years ago. And selling based on short term info, looking back, stand out as my biggest mistakes (like selling NVDA about 10 years ago).
I have mind boggling returns and I don't 'follow' the companies' fundamentals or anything else, other than sometimes hearing or reading about quarterly results. 1997 AMZN shares, cost basis four cents (not a typo), AAPL up 2,100%, NFLX up 4,200%, GOOG up 1000%, BKNG up 1,500%, UNH up 1,400% and a bunch of others like SBUX, WM, BRK-b, ATVI, CMG, COST with also truly stellar gains. I have about 50 stocks, don't follow any of them, and 90+% of them have done really well or well enough. The few losers don't matter. It has created a lot of wealth for me.
I'm glad I was around TMF in the earlier years. They had a simple message and they stuck to it. Not so much now.
conifer
No. of Recommendations: 8
Congratulations on picking so many successful companies so early and holding onto them.
And avoiding Nokia, Motorola, blackberry, IBM, Intel, Compaq, Cisco', all must buy at one time or the other in the last 3 decades.
No. of Recommendations: 1
Those were not really TMF picks. They had their losers, but it really wasn't that hard to figure out a good mix to build a portfolio, from the favorites of TMF at the time. Common sense.
conifer
No. of Recommendations: 10
They had their losers, but it really wasn't that hard to figure out a good mix to build a portfolio, from the favorites of TMF at the time.
TMF had several portfolios like Rule Maker and Rule Breaker. One of the portfolios (probably Rule Breakers) was a lot like Saul's is today - made up of companies the Gardners thought would be big winners in the Internet and WWW era. Among their popular picks in those days were IOMEGA & AOL.
You were unusually skilled in picking so many winners and avoiding the big losers. You had very uncommon sense - in the 3 sigma territory.
No. of Recommendations: 3
I think AOL did well for them, but not sure in the final end. They had gargantuan gains in it for a while, but they probably held too long. After all, that was their mantra, LTBH, and they had a lot of people watching. I remember Iomega, but not the details. They have certainly had losers, but they had their clear favorites in the early aughts through the financial crisis and that was a good time to be choosing from their picks. Like they say, 6 good out of 10 will do the trick.
I subscribed to their first 'real-money portfolio', MDP from '07 to'11 or so, and then SNova for a couple years. I consider Tom Engle (TMF1000) my investing mentor/hero, if you know who that is in the TMF world, and learning from him and following his advice was way more significant for me than TMF, although TMF was supplying the picks. From '08 to '10, the MDP boards were a very dynamic and interesting place, with TMF1000 an unofficial leader. Except for AMZN, all of my gains have come from the money invested at that time. I had cash in 2009 and because of the prior writings of TMF1000, I started buying. Patience is also very important, and that is in my DNA.
conifer
No. of Recommendations: 0
Nice post. We are few and far between conifer, but we do exist!
No. of Recommendations: 0
Surely though if you reviewed the numbers as well as thinking about the business / runway and concentrated funds in the top 5 or 10 returns would've been a lot better.
No. of Recommendations: 1
I side with mechiny on this subject. P/E's for high growth companies are deceptive. I highly recommend reading Where the Money Is by Adam Seessel. His discussion on the earnings power of Amazon (chapter 8) is an illustrative example on this topic.
No. of Recommendations: 25
I side with mechiny on this subject. P/E's for high growth companies are deceptive. I highly recommend reading Where the Money Is by Adam Seessel. His discussion on the earnings power of Amazon (chapter 8) is an illustrative example on this topic.
To be clear, the issue here is not the high P/E (low earnings yield) of high growth companies.
It's the ratio of distant future earnings to price today.
Would you consider that ratio misleading or not worth consideration?
A high current P/E (low current earnings yield) is not a problem at all, provided the earnings eventually arrive to justify the value of the firm.
After all, the whole idea of not worrying about current earnings from a growth company is that you'll get lots and lots of them later, which is fine.
But if the earnings don't eventually arrive, the only way to get a good investment outcome is if the market is still assigning a very high multiple in the distant future: it's still trading at a very low earnings yield.
That certainly does happen.
But it doesn't happen frequently, and the exceptions can not be identified in advance by mere mortals, which is why the average outcome from is so very poor if you overpay for future earning power.
Spotting today's optimists is easy, but spotting the things they will be very optimistic about a decade from now can not be done reliably.
It's interesting that so many people cite Amazon as the reason for not worrying about a company ever having any profits even far into the future.
It's by far the most famous example of an exception to the general rule of valuation multiples eventually falling to the teens, and so far Amazon investors have done very well indeed: good for them.
What's interesting is that, in any given year, the current Amazon investors generally DO expect substantial profits within a few years. Check out the consensus expectations written many years ago.
Time goes by, and the substantial earnings don't arrive. Past investors were flat-out wrong in their investment thesis.
But optimism holds, so today's investors now expect substantial earnings in a later year, leading to still-high multiples. This whole process then repeats.
So, this should not be cited as an example of the wisdom of not expecting far-distant profits. Amazon investors have never on average been like that: profits have always been expected quite soon.
It's an example of doing well even when you're wrong about your expectations for distant future profits, but then being saved by future optimists you couldn't have (and didn't) predict.
As usual, the current forecasts for Amazon's future earnings are quite substantial. A typical expectation for 2030 is $20 EPS. Today's price is only 6.6 times that future number, far below my usual threshold of 10x for a firm to be a likely good investment based on solid future earnings.
Broadly speaking, there are three possible outcomes for today's Amazon investors:
* Those expected earnings arrive. The current investors do well no matter what level of optimism Mr Market has about them 5-10 years from now.
* Those expected earnings do not arrive. Amazon investors rewrite their memories once again, and future investors bid it up to a high price expecting big profits to arrive in 2035 or 2040 instead.
Today's investors still do fine, provided those future optimists exist.
* The firm is trading at a relatively normal valuation level for a firm with their growth rates, and today's investors do poorly.
As a random example, maybe the company might be trading at 22 times net earnings averaging around $4, and today's investors lose 1/3 of their money.
I'm not saying which outcome is most likely. I don't have a strong opinion on the matter.
I'm merely pointing out that today's investors will do well only if (1) future earnings are solid relative to today's price, or (2) future multiples are still very optimistic after the years of earnings disappointment.
For any company other than Amazon, outcome (2) is very rare.
Which is why I keep suggesting that investment works best (even in profitless high growth firms) when choosing things that you're pretty sure will offer outcome (1).
That is, a low ratio of current P to distant future E.
Jim
No. of Recommendations: 14
After all, the whole idea of not worrying about current earnings from a growth company is that you'll get lots and lots of them later, which is fine.
But if the earnings don't eventually arrive, the only way to get a good investment outcome is if the market is still assigning a very high multiple in the distant future: it's still trading at a very low earnings yield.
That certainly does happen.
But it doesn't happen frequently, and the exceptions can not be identified in advance by mere mortals, which is why the average outcome from is so very poor if you overpay for future earning power.
Spotting today's optimists is easy, but spotting the things they will be very optimistic about a decade from now can not be done reliably.
From a growthy investor's point of view, the logical fallacy here is the requirement that predictions of future earnings be accurate with respect to individual stock choices. It is quite true that few if any investors have proven able to pick and choose in advance the relative handful of big winners out of the multitudes of aspirational growth companies at any given time.
As the old retail joke goes, they make it up on volume! Successful growth investors often buy scores of promising companies, hundreds in most cases, and let the winners, which neither they nor anyone else can name in advance, carry the portfolio.
It is also quite true that, in retrospect, examining the P/E ratios of these big winners at the time of purchase was not especially useful.
Instructive examples are The Motley Fool's Stock Advisor and Rule Breakers newsletters, which now have histories of 20 years or so. You can go to their website and see performance records for their many, many recommendations -- two per month is the longstanding policy. They helpfully offer a tab listing closed positions as well, but it goes back only a couple of years, so the full data set to check their performance results is not available on the current website. Perhaps the wayback machine would allow an industrious data miner to reconstruct it, but the amount of work this would require does leave skeptics an opening to doubt the published results. (You might need subscriptions to these letters to access the site; I bought mine years into the future sometime back after TMF moved on to more lucrative products and you could do it very cheaply.)
Unless they are a fraud -- always a non-zero possibility -- both portfolios, each including hundreds of stocks -- 187 active recs for SA at present; 142 for RB -- have beaten the market over their relatively long histories.
According to their website, SA, launched in March 2002, and RB, which debuted in September 2004, have both beaten the S&P 500 by 134% over their lifespans. The nearly identical results are not as surprising as they sound when you consider that David Gardner's gift for seeing around corners was the foundation of both letters. The outperformance is the result of a simple arithmetic fact: The most a poor choice can lose is 100%. The most a good choice can gain is many, many times that. So it doesn't take that many good choices to compensate for a boatload of poor ones.
For example, SA's April 15, 2005 rec of NVDA, at a split-adjusted price of $1.63, is up 28,537%, beating the index by 28,060%. Its Dec. 17, 2004 rec of NFLX is up 22,861%. The 2002 rec of AMZN is up a paltry 17,000%. The letters re-recommend stocks the authors especially like, and each re-rec counts as a new position. Of the top 10 SA picks of all time, five are Netflix.
Over at RB, where the gains are slightly less stratospheric having missed '02 and '03, Tesla is the biggest winner, at over 12,000%, followed by Mercadolibre, Intuitive Surgical, Salesforce, Chipotle and Shopify, all with four-figure percentage gains.
In short, like the QQQ of today, a few gigantic winners carried them to victory over the broad market, assuming you believe their math, of course.
As a subscriber for years, my objection to Stock Advisor and Rule Breakers was that few if any individual investors I know are willing to maintain portfolios comprising hundreds of stocks. That was the only way to replicate TMF's results. Most subscribers, I suspect, did what I did, picking and choosing among the many recs a dozen or two names that seemed most promising.
About that strategy, mungo is completely right. At least, he was in my case. But the volume strategy has worked for TMF, just as it worked for Peter Lynch. It works pretty much the way the minor leagues work in baseball. You sign every promising prospect you see and let the cream rise to the top.
No. of Recommendations: 0
<For example, SA's April 15, 2005 rec of NVDA, at a split-adjusted price of $1.63, is up 28,537%, beating the index by 28,060%. >
It's just pure luck or accidental. NVDA stock surged only after AI technology really took off, which happened after 2010s.
No. of Recommendations: 4
Ultimatespinach, thanks for the insightful response to Jim's point. It does appear that Stock Advisor and Rule Breakers have beaten the S&P 500 by a nice margin.
However, how many of TMF's other products (newsletters) have been shut down due to underperformance?
Personally, I briefly subscribed to the newsletter Hidden Gems many years ago, but I did not subscribe to the Stock Advisor or Rule Breakers.
Did you know in advance that one should subscribe to the Stock Advisor and/or Rule Breakers, and ignore the rest? If so, good for you. If not, does it prove Jim's point that it is difficult to identify winning bets (whether it is stocks or newsletters*) in advance?
*I am generalizing here and hope that that it is in line with Jim's views.
No. of Recommendations: 5
Did you know in advance that one should subscribe to the Stock Advisor and/or Rule Breakers, and ignore the rest?
I did not. Apparently like you, I began with Hidden Gems, a small-cap service. I subscribed to Global Gains, an international stock letter, and a truly god-awful attempt at small-cap value investing called Pay Dirt, which was shut down as soon as its performance stats became embarrassing. Global Gains recommended some of the stocks later implicated in the scandal around fraudulent Chinese companies listed on U.S. exchanges.
The difference between these and the original two is that David Gardner was not an advisor to any of the spinoffs. Tom, the businessman, was the brains behind this product expansion. If you were a dedicated growth investor, a Peter Lynch fan, you were more apt to gravitate toward Stock Advisor, where brothers Tom and David each made one rec per month, and Rule Breakers, which was all David and his team. This is where the tech and futurism were. Hidden Gems' top rec was an industrial company that made commercial cooking equipment.
If the conversation is about growth stock investing, and whether it's possible to consistently beat the market with it, I think SA and RB are objectively the appropriate TMF letters to examine. The proliferation was designed to appeal to other investing niches, which TMF wasn't nearly as good at. In truth, the only great stock picker TMF has produced is David Gardner, but in this business, one is all you need.
I do agree that survival bias has a major, misleading effect on TMF's self-reported returns in its advertising. You add all the dumpstered stuff to the current database and the reported numbers would not be what they are.
No. of Recommendations: 0
From a growthy investor's point of view, the logical fallacy here is the requirement that predictions of future earnings be accurate with respect to individual stock choices.
Not a logical fallacy if you apply it to the aggregate of your many individual stock choices.
The few winners among the many losers still need to meet predictions of future earnings, and those predictions also need to make up for the many losers.
I'm sure Jim could phrase it better than I did, though.
No. of Recommendations: 2
As far as market cap change Jim your incessant Apple over valuation posts over the years stand as the worst analysis I have encountered. Just plain facts. We all make mistakes, some obviously large ones.
No. of Recommendations: 35
As far as market cap change Jim your incessant Apple over valuation posts over the years stand as the worst analysis I have encountered. Just plain facts. We all make mistakes, some obviously large ones.
I think it's fair to say you're not reading very carefully what I've written about Apple : )
I am not bearish on Apple, and, as far as I can remember, never have been. Never.
I have sometimes expressed the opinion that I was steering clear because I wasn't too sure what the product mix would be in a decade,
and at other times (including now) I wasn't confident enough in the rich market price to count the entirety of that price as part of the conservative value of a share of Berkshire.
I am willing to allow that Apple might well be worth its current price, but I am merely unwilling to *assume* that this is true.
On the other hand, I've also been quite bullish when it was plainly cheap. It was trading at 12 times (smooth) earnings less than five years ago, and under 10 a couple of years before that.
My main heresy is that I just don't think that the value has risen any faster than the (smoothed) earnings trajectory. Gasp!
When the stock price is exuberant in the recent past, that stance sounds bearish, hence your comment.
When the stock price is in the dumps, my approach probably sounds exuberant.
For example, (smooth) earnings per share are around four times what they were nine years ago in 2014. Yet the share price is up by a factor of eight since then.
As a basic starting point, it seems reasonable to start with the notion that a share of Apple is worth around four times as much as it was then, not eight times as much.
Call me crazy.
When valuing Berkshire's stake in Apple, I simply don't care much about what the market price is. I treat is as I would any other business unit.
Estimate the current cyclically adjusted earning power, and apply a not-too-crazy multiple appropriate to the reliability and growth rate of the earnings trajectory.
Since the current market multiple for Apple is quite high, this might sound like a bearish comment to those not paying close attention.
I don't see it that way, as I'm not predicting that the stock price is going to fall.
I guess the strongest statement is that I wouldn't count on a particularly good return from Apple stock in the next few years. That outcome would require some unreliably bullish assumptions I'm not comfortable with.
About either implausibly high earnings growth rates, or far-future market multiples higher than one should be willing to count on.
Jim
No. of Recommendations: 5
I do agree that survival bias has a major, misleading effect on TMF's self-reported returns in its advertising. You add all the dumpstered stuff to the current database and the reported numbers would not be what they are.
I don't trust the motley fool, and I will tell you why. When I was just getting started with investing, I remember they had two portfolios. I think it was rule makers, and rule breakers. We had the 90s boom, and they very proudly showed how they were beating the market. Then we had the .com crash and they started trailing the market.
What did they do? They stopped publishing their results and claimed there was some problem with their data provider. This went on for a very long time.
That says a lot about their character.
No. of Recommendations: 16
When I was just getting started with investing, I remember they had two portfolios. I think it was rule makers, and rule breakers. We had the 90s boom, and they very proudly showed how they were beating the market. Then we had the .com crash and they started trailing the market.
What did they do? They stopped publishing their results and claimed there was some problem with their data provider. This went on for a very long time.Their slide in terms of performance and behaviour coincided with a big slide in their solvency.
Speaking from experience, having to make payroll will focus the mind tremendously, and sometimes trigger unfortunate decisions.
It is indeed sad that they have become the exact thing they mocked in the 1990s. Their emails are now astounding examples of the "shrill shill".
But I guess they have to make a living, and it is a fine demonstration how natural evolution in business models favours the most remunerative, not the most respectable.
Jim
As a totally unrelated aside, a bubble is a terrible thing to waste.
For those who miss it, I have an investment system very well suited to bubble times like that : )
Anybody remember super high growth / momentum investing?
Totally the wrong board, but maybe it has entertainment value
It can't get much simpler than this:
* Start with the ~1700 stocks in the Value Line database.
* Find the 40 stocks with the highest reported five-year rate of sales growth per share. (this field is updated only annually for each stock)
* Of those 40, buy equal dollar amounts of the 10 stocks which are currently furthest above their respective 52 week low prices. (measured one day before your trading)
* Hold those 10 stocks for a month, repeat.
Unsurprisingly, this likes exuberant markets, including bungee rebounds.
The surprising thing is that it does not that badly on average the rest of the time, too.
Final result 28.8%/year 1997 to date versus 9.0% for the S&P 500, at least in backtest.
On relatively sane risk metrics (size and frequency of rolling years having a shortfall below a reasonable return), this 10 stock portfolio is only a tad riskier than the S&P 500.
Mixing a little of this yeast into the "loaf" of your portfolio wouldn't be so crazy.
Come to the dark side!
Figures after trading costs, includes dividends, no provision for tax.
Year Screen S&P
1997 43.3 27.9
1998 86.1 34.5
1999 172.5 18.4
2000 19.3 -10.8
2001 7.3 -8.0
2002 -45.9 -18.9
2003 154.5 23.0
2004 32.0 8.9
2005 14.2 7.5
2006 28.1 13.8
2007 -2.4 1.9
2008 -29.9 -32.9
2009 144.0 25.5
2010 25.5 14.2
2011 -1.0 2.5
2012 -4.0 17.1
2013 106.8 27.6
2014 23.8 12.9
2015 9.9 1.9
2016 28.1 14.4
2017 36.8 21.8
2018 -8.1 -3.5
2019 12.4 29.9
2020 127.3 16.9
2021 36.5 30.5
2022 -10.9 -19.0
2023 18.4 17.5 (half year not annualized)
No. of Recommendations: 0
Hold those 10 stocks for a month, repeat.
Jim, a suggestion. As this is the Berkshire board not everybody here is a daytrader like you (exaggerated). Your suggestions might be even more interesting if one hasn't to get off ones ass ever month but only ever half year or even better, once a year only.
The question of course is whether you even check with those lazy rebalancing intervals, and if so, whether your screens still show interesting performances with those.
No. of Recommendations: 1
[TMF] had two portfolios. I think it was rule makers, and rule breakers. We had the 90s boom, and they very proudly showed how they were beating the market. Then we had the .com crash and they started trailing the market.
What did they do? They stopped publishing their results and claimed there was some problem with their data provider.
Actually, no. They closed them down saying that the portfolios did not work after the tech boom ended. I remember reading that at the time.
That's what all these stock-picking services do. And mutual fund families. Create a number of different portfolios and silently fold the ones that didn't succeed.
The Motley Fool ended up doing what they railed against then they started up.
No. of Recommendations: 19
The question of course is whether you even check with those lazy rebalancing intervals, and if so, whether your screens still show interesting performances with those. For this style of high growth/momentum investing, no, not really. It's not for the lazy investor.
The returns from this particular selection strategy traded annually drops to 15%/year. You're holding long past when the momentum figures were relevant.
Given the unreliability of backtests, that removes a lot of what you might call the "quant margin of safety".
But other, more sane/prudent/boring quant techniques work pretty well with reasonably long holds.
Not exactly "set it and forget it", but much further along the spectrum in that direction.
e.g, I like a strategy I created called LargeCapCash. If Berkshire didn't exist, it's how I'd manage most of my money.
This screen was designed with the express goal of giving safe long term returns similar to the S&P 500 (and even somewhat correlated), but with a decent chance of modestly higher long run returns.
It is not designed for the highest returns, or trying to make money even in bear markets, or any of the other things a quant screen might attempt.
Again, pretty simple, after a couple of housekeeping steps:
Start with the Value Line stocks with a "Timeliness" ranking (almost all of the 1700, but eliminating those with less than 5 years of data or undergoing buyout or bankruptcy)
Also, eliminate those not paying a dividend. I'm not fond of dividends, but for this particular screen it helps results materially.
Find the 30% of eligible stocks with the highest ROE (annual calculation, so not much change)
Of those, buy equal dollar amounts of the 40 stocks with the highest cash balance (cash minus long term debt). These will generally be very large cap firms.
Hold for a year. Repeat.
Since both the ROE and cash metrics tend to change pretty slowly, there isn't much trading. About 15 of the 40 stocks change each year, about 25 held over to the next year.
This would have beat the S&P 500 by 4.55%/year after trading costs in the last 34.5 years, including beating the S&P in about 80% of rolling years.
It has beat the S&P quite nicely since it was proposed.
Given that low turnover, it probably wouldn't even matter materially if you forgot to trade for several months, or even redid it every 2 or 3 years.
This post has a rundown of the performance of the main variants in the first three years after its introduction
https://www.shrewdm.com/MB?pid=864956561Beating the S&P by 3-7%/year depending on the specific variant examined.
A slightly subtle note: the requirement for a high cash balance is an absolute number, not a percentage of the market cap.
This is why it almost always picks very large cap firms, which in turn is why its short term correlation to the S&P is quite high.
At the one-month level, correlation coefficient is 94.6%. At one year it is 93%.
The best-fit model for one year performance of the screen is 1.037 times the S&P return in that year, plus 4.3%.
Its performance relative to market tends to be a tad bigger during weak markets. (in this case, one tad = 1%/year rate)
Here are some recent picks, to get a feel for what it recommends.
MSFT TSM CSCO ACN COST NKE NVDA BHP INFY TJX
ROST EOG AB EXPD STM AMP BBY EQH LOGI SIG
HLI TER TXN SEIC KLIC MPWR WIRE PBF NTAP MLI
RHI CNS NSP DKS WSM LPLA LSTR UFPI BCC DDS
The average dividend yield of that set is 1.81%.
The average earnings yield is 6.66% (P/E of 15.0)
Average ROE (last full fiscal) 49.1%
This is only 40 stocks, but they all have some good properties, and in any case it's less than a third the single-stock risk of the S&P 500.
Increasing it to 60 stocks for more diversification doesn't make a whole lot of difference. Maybe 0.4%/year lower returns, at a guess. More typing.
Jim
No. of Recommendations: 0
Do you use Value line to screen for these companies?
No. of Recommendations: 0
nevermind, obviously do. cant delete the previous post
No. of Recommendations: 0
Actually, no. They closed them down saying that the portfolios did not work after the tech boom ended. I remember reading that at the time.
That's what all these stock-picking services do. And mutual fund families. Create a number of different portfolios and silently fold the ones that didn't succeed.
The Motley Fool ended up doing what they railed against then they started up.
Ray, perhaps they closed them down eventually, but for quite a while, they said, their data provider could not provide results, unless I am having a delusion. I even remember discussing this with my Father.
No. of Recommendations: 3
This is how I value high growth, high pace of innovation, entering new product categories companies ... by looking at the change in value, not the GAAP earnings. My January post from the Tesla board:
Jan 6, 2023
Add bookmark
#392,295
Now that 2022 is over we can look back and calculate a (CK)non-GAAP PE for the year. Non-GAAP in that I make adjustments for changes in the value of assets held instead of looking only at earnings. An example of that would be that if you owned a $1M home that went up 100K in value. Your GAAP earnings were zero but your (CK)non-GAAP earnings would be 100K. Giving your house a (CK) PE of 10. Another example would be to look at the GAAP earnings of the Mona Lisa. The earnings are accrued over decades but only recognized on the sale. So, a PE of infinity and then almost zero. So much of our earnings at Tesla are not reflected in the profit and loss statement I think it makes GAAP near useless in determining the value of our shares.
So, here we go. My unofficial official earnings report for 2022:
- 10B. cash earnings.
- 100B. value of FSD. (based on 1T when done. We went from 70% to 80% complete.)
- 10B. Semi "start" of production.
- 10B. CT prep for mass production.
- 1B. value of work done to get lithium refinery going.
- 10B. Bot progress. Mostly in the form of real world AI.
- 5B. factory buildouts and improvements.
- 10B. Tesla Energy nearing ramp.
- 10B. skunkworks projects.
166B total of value created but not recognized.
Puts the current "real" PE at somewhere between two and three. Around eight at our highest valuation.
Someone earlier asked what they should tell their son who worried that Tesla was trending towards being a big car company that didn't reward its investors. My answer would be to go through all the progress Tesla made during the year in setting the stage for future earnings. Earnings that will be returned to shareholders in dividends, share repurchases or share price appreciation. (See BRK.A for an example.)
We need to get out of the 12 month Wall Street price target earnings modeling rut. Current profits only matter in that they help fund future innovation ... and we're way past having to worry about issues like that.
Bought more TSLA today.
No. of Recommendations: 6
Do you use Value line to screen for these companies?
Yes and no.
I use the Value Line database, so I'm using their data.
But I don't use their screening tool.
I have their "Analyzer" product, which lets me download the weekly updates to their Windows program, and export all the data fields to a spreadsheet via a flat file.
Jim
No. of Recommendations: 5
So much of our earnings at Tesla are not reflected in the profit and loss statement I think it makes GAAP near useless in determining the value of our shares.
Yes, P/E could be a misleading measure, expected distant E or not when we are just talking about GAAP earnings. It doesn't factor in growth prospect, nor does it properly adjust for investments/expenses made for future growth.
Obsession with P/E not only leads to error of omission but even commission (remember LYLT?). It's much more important to be a business analyst than a financial analyst in investing. It's better to be approximately right than precisely wrong by seeing P/E in the wrong mindset.
No. of Recommendations: 3
Here are some recent picks, to get a feel for what it recommends.
MSFT TSM CSCO ACN COST NKE NVDA BHP INFY TJX
ROST EOG AB EXPD STM AMP BBY EQH LOGI SIG
HLI TER TXN SEIC KLIC MPWR WIRE PBF NTAP MLI
RHI CNS NSP DKS WSM LPLA LSTR UFPI BCC DDS
The average dividend yield of that set is 1.81%.
The average earnings yield is 6.66% (P/E of 15.0)
Average ROE (last full fiscal) 49.1%
I had a go using the free library version of Value Line. I match 31/40 tickers.
I think mainly the small caps don't show on my list.
For example, AB (3.7B market cap) disappears when I add 'Cash And Mkt Securities' to the screen.
AB looks like one to avoid in any case.
My list:
MSFT Microsoft Corporation
TSM Taiwan Semiconductor Manufacturing Co Ltd
CSCO Cisco Systems Inc
ACN Accenture Plc New
PRU Prudential Financial Inc
COST Costco Wholesale Corporation
NKE Nike Inc
NVDA NVIDIA Corp
BHP BHP Group Ltd. ADR
INFY Infosys Limited
ROST Ross Stores Inc
EOG EOG Resources Inc
EXPD Expeditors International of Washington Inc
STM STMicroelectronics
AMP Ameriprise Financial Inc
LOGI Logitech International SA
SIG Signet Jewelers Ltd
TER Teradyne Inc
TXN Texas Instruments Incorporated
KLIC Kulicke and Soffa Industries Inc
MPWR Monolithic Power Sys.
WIRE Encore Wire
BBY Best Buy Company
PBF PBF Energy
MLI Mueller Industries Inc
WFG West Fraser Timber
PAYX Paychex Inc
CNS Cohen and Steers Inc
DKS Dicks Sporting Goods Inc
WSM Williams Sonoma
LPLA LPL Financial Holdings Inc
LSTR Landstar System
UFPI UFP Industries
SHOO Steven Madden Ltd
DDS Dillards Inc
NVO Novo Nordisk
MGY Magnolia Oil & Gas
BKE Buckle Inc
GGG Graco Inc
RGR Sturm Ruger and Co
Average dividend yield 1.94%
Average ROE 41%
No. of Recommendations: 18
This is how I value high growth, high pace of innovation, entering new product categories companies ... by looking at the change in value, not the GAAP earnings. My January post from the Tesla board:
It's an interesting approach, for sure. I presume it works fine if the assumptions are good.
I think the only thing I would suggest is to remember that ultimately all of the things on your list exist for the sole purpose of creating greater profits down the road
for existing shareholders.
If a given line item is estimated to have increased the value of the firm by X% beyond what it was a year earlier, then it will have to have increased future owner earnings by X% in some fixed future year beyond what they would otherwise have been.
If the value-gain estimates you've done are all individually consistent with that conclusion, go for it.
As has been beaten to death, current P/E means very little.
But the future earnings mean everything.
Personally, I find it easier to work backwards from the current price: is it low enough to offer a decent return?
It's unwise to assume that any equity security can be sold at a multiple of more then 20 times cyclically adjusted owner earnings far in the future (a decade or more, say).
It can happen, but when it does it's luck, not prudent prescience.
And I'm not interested in anything that doesn't get me a return of (say) inflation + 6.5%/year with some confidence, since that's about the average return from the average US stock in the average year.
Putting those two together, you can figure out what it would take for any stock to be a good deal at its current price.
It's simply a matter of assessing whether it's sufficiently plausible that they'll make enough money to manage the required outcome at a far future multiple no more than 20.
For example, with Tesla stock at $265 today, it's simply a matter of estimating whether one can see a pretty plausible path to owner earnings of (say) $22 in 8 years, in today's money.
$265 * 1.065^8 / 20 = about $28. If you prefer 12 years out, it would required owner earnings of around $28 in today's money.
If your assessment of the business prospects in the next many years is that they can manage that profit or more with confidence, the stock is not overpriced.
Some current estimates for 2030 EPS at Tesla are about $11, which wouldn't do it. Other estimates are $100, which would definitely do it.
The nice thing about working backwards is that as soon as you have the answer to "am I confident the future earnings be more than $22" (or whatever the number is), you can stop worrying about the details too much.
Anything more is just icing on top, and it doesn't matter precisely how thick the icing is.
Jim
No. of Recommendations: 11
I had a go using the free library version of Value Line. I match 31/40 tickers.
I think mainly the small caps don't show on my list.
There shouldn't be many small caps: it requires a very big pile of cash to make the cut : )
It's more likely something small, like the fact that the list I posted was from a database snapshot from a few weeks ago.
Or smaller things, like ensuring that the stocks have all the required fields populated. (notably a Timeliness rank)
Even a small change will change which set of firms make the 30% cutoff.
The most important thing is to use the annual, not quarterly, data field for ROE. The quarterly one is too volatile because it's just the latest quarter, not the four quarters just ended.
It's better to be out of date than too random.
It's annoying that VL uses slightly different names for the same data field depending on which product you use.
In case it matters, the fields I use are (in my program) called
"Current Dividend"
"Timeliness Rank"
"Return on Shareholders Equity" [sic]
"Cash"
"Long Term Debt"
Jim
No. of Recommendations: 3
"It is indeed sad that they have become the exact thing they mocked in the 1990s" ...quoting Jim.
That's it in a nutshell, imo. I fall back on my common sense theme being important with using TMF in the past, and the changes afoot in the aughts were not hard to see, though maybe hard to accept. I became a vocal critic of the MDP service around 2010 when they strayed from what TMF had built their 'brand' on, LTBH. That simple LTBH advice (and an advertising-only business model) doesn't pay the bills with an ever growing staff and footprint.
Common sense again, it became clear that Tom Engle as TMF1000 on the boards was the only one in the room with logical, consistent and effective methods to invest money. It was pure gold, and most importantly, it works. Anyhow, for others to suggest it was pure luck to pick good stocks from the TMF picks and favorites from back in the day, for good future returns, is ridiculous to me. It was there for the taking, but there was some logical navigation required! Nowadays, I'm completely out of touch with what TMF is doing. I'm just so thankful for TMF1000 back then, like so many others.
conifer
No. of Recommendations: 1
My humble perspective'
+ Retired and now 70 years old. Married and Good Health.
+ Co. Pension, Soc. Securities and Traditional Investment Account cover living expenses
+ No short/long term debt;
+ Investments: 80% in IRA conservatively invested (e.g. Brkb, 2025 Trgt Date Fund, Private Co AG stock and Cash); 20% in Roth IRA filled with 40 Motley Fool stocks since 2918; higher risk stocks; 6 of which have 4-6 x'd yet still only up in total 2.7%; no complaints to date Re Motley Fool Service.
Have adopted philosophy of 10 possible outcomes from horror to life changing wealth and I'll be fine with either end of spectrum. I just enjoy the game.
Grateful Always to MF and you all for your inputs.
GLTA,
Silverlinin
No. of Recommendations: 2
For me the moment that marked the end of the good times at TMF was when Bill Mann left the Hidden Gems small caps newsletter and became the lead portfolio manager on their mutual fund product. There were a lot of winners from HG before he left, but I can't remember any from afterwards.
No. of Recommendations: 4
I do not have easy access to VL, have to go to a library 50 miles away. I can get figures for the Russell and S&P stocks from barchart{.}com. And can screen to discard all the ones that don't have a VL timelines rating.
My first trial did not have MSFT as a pick, which didn't seem right. Turns out that barchart lists "cash" as _only_ cash-on-hand, whereas the generally accepted figure is "Cash and Short Term Investments". There also is some discrepancy between various data sources, some list the financial figures from the lastest quarter and some from the preceeding quarter. From what I gather, VL sometimes is up to a year out-of-date.
Anyway....after getting all that resolved, the list I got 24 of the 40 the same picks as Jim mentioned in post 2931. Adrian had 31 same as Jim, I had 22 same as Adrian.
My 40 had average ROE 51.6%, P/E 20.9, yield 2.19%
Average market cap $116.3 billion. Because of MSFT being so large. Excluding MSFT was 54.8 billion.
Median mktcap 14.0 billion.
Probably close enough.
These picks:
EXPO CNS MED BKE MSFT SCHW CSCO ACN COST APO AMP MMC TJX ROST EXPD NKE
EOG MPWR MLI PAYX VVV V NTAP WIRE CALM RHI BCC SIG PAYC BBY WWE DDS NSP
LRCX LSTR IDCC MGY AOS PBF DKS
No. of Recommendations: 10
From what I gather, VL sometimes is up to a year out-of-date...Yes, some of their data fields are updated only from the last full year's report.
That includes "Return on Shareholders Equity" and "Sales Growth 5-Year".
They have a quarterly ROE figure, but it's only the ROE in the latest quarter, not for the trailing four quarters.
Trailing four quarters ROE would be better, but the annual figure seems to work fine, and leads to lower portfolio turnover too.
The picks I posted were from earlier in July, which would account for some differences in the results.
Using the current data, I think the picks would go like this:
* Number with any valid Timeliness rank 1-5: 1459 stocks
* Of those, also with a positive dividend: 965 stocks
* 30% of 965 = 289.5 stocks, let's say 290.
* Sorting the 965 stocks by ROE, the cutoff to get 290 stocks is between ROE 25.53 and 25.47
* For those 290 stocks, I calculate [Cash - Long Term Debt], and sort on that.
* The top 40 I get from that process, with the calculation fields shown:
Ticker Time Long-Term Return on Current Cash -
Symbol Rank Cash Debt Equity Dividend Debt
MSFT Yes 104757 47032 41.69 2.75 57725
TSM Yes 43697.2 27385.3 33.62 1.8 16311.9
CSCO Yes 19267 8416 35.43 1.56 10851
ACN Yes 7893.8 45.9 31.11 4.72 7847.9
COST Yes 11049 6484 28.31 4.08 4565
NKE Yes 12997 8920 39.56 1.36 4077
NVDA Yes 13296 9703 37.85 0.16 3593
BHP Yes 17236 13806 47.47 3.8 3430
INFY Yes 3185 0 29.85 0.42 3185
TJX Yes 6226.8 3354.8 54.68 1.33 2872
ROST Yes 4922.4 2452.3 42.42 1.39 2470.1
EOG Yes 5972 3795 32.6 3.4 2177
EXPD Yes 2034.1 0 43.64 1.38 2034.1
STM Yes 4518 2542 31.19 0.24 1976
AMP Yes 7097 5184 70.82 5.4 1913
EQH Yes 5803 4472 121.17 0.88 1331
LOGI Yes 1328.7 0 26.86 1.16 1328.7
SIG Yes 1166.8 147.4 29.97 0.92 1019.4
HLI Yes 942.8 0.5 30.32 2.22 942.3
TER Yes 894.4 0 29.18 0.48 894.4
TXN Yes 9067 8235 60.01 4.96 832
SEIC Yes 831.4 40 29.37 0.86 791.4
KLIC Yes 775.5 0 37.34 0.76 775.5
MPWR Yes 737.9 0 35.93 4 737.9
WIRE Yes 730.6 0 39.47 0.08 730.6
BBY Yes 1874 1160 57.13 3.68 714
PBF Yes 2203.6 1492.8 58.36 0.8 710.8
NTAP Yes 3070 2389 106.12 2.1 681
MLI Yes 678.9 1.2 36.75 1.3 677.7
WFG Yes 1162 499 25.92 1.2 663
RHI Yes 619 0 43.34 2.02 619
PAYX Yes 1222 798.2 44.58 3.7 423.8
CNS Yes 420.4 0 50.67 2.28 420.4
NSP Yes 765.9 369.4 220.67 2.28 396.5
DKS Yes 1924.4 1540.6 42.2 4 383.8
WSM Yes 367.3 0 66.3 3.6 367.3
LPLA Yes 3046.9 2717.4 39.01 1.2 329.5
LSTR Yes 393.5 67.2 48.56 1.2 326.3
UFPI Yes 595.4 275.2 27.01 1.05 320.2
BCC Yes 748.9 444.6 52.67 0.63 304.3
These picks are created from the data set released the morning of Monday July 31, with the download file name "2023-08-04"
(it's the data set for the week ending August 4, so they're always dated in the future)
I don't guarantee a perfect match, but to quote any aging programmer, "it should work".
Note, there is one "bug" in this process.
I believe it's the same issue in the calculation above as in the backtest, and (I think) also in the VL data set I have: it doesn't distinguish between long term debt field unpopulated, and long term debt of zero.
But the backtest works fine, so I'm good with it.
This is a "shotgut" approach to investing, not a "rifle" approach. A few little lead pellets will always go astray, but it's the overall result that matters.
Jim
No. of Recommendations: 0
I get 32/40 from Jim's updated list.
The first miss is TJX. For TJX the fields "Cash & Marketable Securities" and "% Return on Shareholders' Equity" are both blank in the VL data from my library.
Suspect it will be the same for the other 7 misses.
My screen picks Prudential Financial Inc PRU in the #5 spot. The others not on Jim's list are my bottom 7: SHOO DDS NVO MGY BKE GGG RGR.
Interesting exercise.
No. of Recommendations: 0
I don't get it.
Our figures for MSFT agree, except for timing. For netcash I get 41990 and you get 57725.
But I get EXPO as cash $1481740 and LT debt as $230420, giving netcash of 1251320. Which is more than 57725 (or 41990). Which is why I get EXPO as the 1st pick instead of MSFT.
Indeed, yahoo says EXPO cash is 148.17M and *TOTAL* debt is 28.79M. Other sites say LT debt is 13.34M. ROE is 29.32%, which is higher than the lowest you posted (25.92), so it should have made the ROE cut.
Ah, I see one thing that maybe explains it.
PAYX, Jim shows cash 1222 and LTdebt 798.2
Yahoo & WSJ & others show LTdebt (excl. Cap leases) 798.
But they show "Cash Only" 1,272
and "Cash & Short Term Investments" 1,645
Note that WSJ for MSFT shows "Cash Only" 34,704 and "Cash & Short Term Investments" 111,262.
So it looks like sometimes VL reports cash_only and sometimes reports Cash & Short Term Investments.
Probably doesn't make much of a difference overall, using this as a long-term hold screen. We get 24 of the 40 in common and the buys are equal weight; it's mostly just a matter of making the netcash cutoff. It's somewhat hard to equal weight 40 stocks, going with 50 makes it 2% per stock. (M1 Finance only takes whole numbers for the weight percentages.)
Still not quite ready to replace some of my BRK with this screen.
I am encouraged that this screen only picks one of the huge stocks that are the top holdings of SPY, QQQ, etc. instead of many of them. Of course, that means that it will probably underperform the S&P500 for the near future.
--------------
My screen picks Prudential Financial Inc PRU in the #5 spot.
PRU has quite low ROE, 6.8% (barchart says 14%), so it should have not made the ROE cut. Using the data from barchart the ROE cutoff is 25.82% at position #240, PRU is #486.
No. of Recommendations: 8
I am encouraged that this screen only picks one of the huge stocks that are the top holdings of SPY, QQQ, etc. instead of many of them. Of course, that means that it will probably underperform the S&P500 for the near future.
I would definitely question that latter assumption : )
These things are hard to predict.
The very large caps have been the stars since around 2015 (other than 2022), but definitely the duds over the long run. Beats me what will happen next.
But if you want to lean towards the bigger firms for whatever reason, the "companion" variation of the screen simply skips the step that requires a dividend.
The long run returns are extremely similar. A hair lower in backtest, less than half a point difference, well within statistical noise.
The dividend test causes a lot of very large firms to be eliminated at that step, and makes it a much larger cap screen without it.
Other than one's preference for (and tax rate on) dividends, the no-dividend-required version tracks the S&P a bit more closely because of the greater gigacap tilt.
That can make it easier to live with. To outperform the index you have to diverge from it, which will obviously not always be positive in every time period.
It's emotionally hard (even when rational) when the short term divergence is much to the downside, so a method with closer tracking (beta closer to 1) will be easier to live with.
In terms of your mismatches on the picks, the "Cash" and "Long Term Debt" fields appear to be two more of the ones they update only annually.
So that may explain the fact that some of your picks match on that others don't: the update cycle.
Still not quite ready to replace some of my BRK with this screen...
I'm thinking about it.
I might sell some real estate soon, and I'm not sure whether I'll put the same proportion of the proceeds into BRK as I have now, so I'm thinking about moving money into quant like this.
Rationally it doesn't make a difference how big the portfolio is, it's the allocation percentage that matters, so I shouldn't be hesitant about putting more capital in and keeping the percentages the same.
But emotionally it would be a big "buy" order for BRK when it's a bit pricier than average, which is a tricky thing to do.
I kinda pity fund managers when they have big influx of capital - and even more when they get big redemption requests.
If I ever get back into fund management, it will be closed end!
Jim
No. of Recommendations: 0
PRU has quite low ROE, 6.8% (barchart says 14%), so it should have not made the ROE cut. Using the data from barchart the ROE cutoff is 25.82% at position #240, PRU is #486.
The library VL database has PRU "% Return On Shareholders Equity" = 35% and net cash of $7.5Bn. Can't trust the data.
Still not quite ready to replace some of my BRK with this screen.
Me either. I just know that without a valuation component, and trust in the data source, I just won't be able to stick it out when it inevitably lags for a few years.
No. of Recommendations: 4
The library VL database has PRU "% Return On Shareholders Equity" = 35% and net cash of $7.5Bn. Can't trust the data.
Their database certainly has errors, no doubt about it, but they aren't material very often.
I find that more often than not mismatches are a question of which statements were used, and what adjustments they make. (smoothing, skipping goodwill, whatever).
If (a big if) the mismatches are due to methodology rather than flat-out bugs, and the methodology is pretty consistent through the years, the historical results might be expected to keep some useful predictive power.
I would never put more than (say) 2% of a portfolio into a quant pick. And probably much less, these days.
As a rule of thumb, the allocation to a single carefully picked quant screen might be comparable in size to the allocation to a single carefully picked stock.
Jim
No. of Recommendations: 2
The library VL database has PRU "% Return On Shareholders Equity" = 35% and net cash of $7.5Bn. Can't trust the data.Indeed, can't trust it.
Goggling "pru roe" we see this: "During the past 13 years, Prudential Financial's highest ROE % was 15.71%. The lowest was -3.68%. And the median was 7.71%."
Various data sources say "Cash And Cash Equivalents" 14.6Bn (or 17.5Bn) and LTDebt 20.9Bn, thus a negative netcash. Both figures have been fairly stable in the last 5 quarters. And years.
See, for example,
https://www.wsj.com/market-data/quotes/PRU/financi...I don't know how VL could be that far off. Maybe a transcription error?, either yours or theirs.
Hopefully one bad pick out of 40 (or 50) won't hurt too bad.
without a valuation component, and trust in the data source, I just won't be able to stick it out when it inevitably lags for a few years.You mean like the Mechanical Investing screens such as YLDEARNYEAR? {ugh}
ROE, netcash, 5-year-sales-growth all seem to be valuation components.
At some point it becomes attractive to slow down on the frequent trading and move a chunk of money into something that doesn't trade very much, and just leave it sit.
FWIW (nothing!) the Jim-inspired 100-stock screen that I posted as "A new idea" has returned 2.33% vs SPY 0.88% since 7/3/23.
The (meaningless) 2 day return of my picks here is -0.40% vs SPY -1.42%.
No. of Recommendations: 9
Indeed, can't trust it.
Goggling "pru roe" we see this: "During the past 13 years, Prudential Financial's highest ROE % was 15.71%. The lowest was -3.68%. And the median was 7.71%."
I don't know how VL could be that far off. Maybe a transcription error?
It appears to be because of the wild swing in shareholders' equity at PRU.
At end 2021 net worth was $61.88 billion. At end 2021 it dropped to $16.25 billion, primarily because of a $41.1bn drop in accumulated comprehensive income.
If book drops by a factor of four for any company, ROE (naively calculated) rises by a factor of four.
So that part seems right: a normal level earnings would suddenly create a huge ROE. There is no inherent conflict between a high recent number and a 13 year average of 16% that you found elsewhere.
Separately, it looks like they are treating some part of the GAAP net income as exceptional and therefore to be excluded.
GAAP net was a loss in 2022 and (I believe) TTM, but VL seems to have only positive numbers.
So I'm not sure whether it's an adjustment they decided to do that makes some sense, or an error.
The 2022 net profit they show is fairly close to the 2021 figure, so it's not a random digit scrambling.
The equity drop figure seems right, and the income adjustment could conceivably make sense, so I wouldn't assume it's a flat out mistake without digging more deeply.
In any case, the ROE is going to give you an odd result without some sort of smoothing or adjustment to shareholders' equity.
This is a problem with a naive ROE calculation.
Ideally, you'd work through each firm that's a candidate and check it for sanity before putting money into it.
If book or income are really unusual going into the most recent ROE calculation, you'll get an odd number, especially when book is near zero.
(one of the oddities of this is that firms with persistently negative book but positive earnings should be shown as having an arbitrary but high ROE, like 100, not a negative one)
Without manual checking for sanity, the safest approach would be to use the lower of the TTM and five year average figures.
You'd skip some good firms that just had one bad year, but you'd eliminate the great majority of the ones inappropriately passing due to a quirk of ROE calculation.
The FT.com equity screener has a data field for five year average ROE. I wish VL, or one of my other databases, had that.
Jim
No. of Recommendations: 0
"I don't know how VL could be that far off. Maybe a transcription error?"
It appears to be because of the wild swing in shareholders' equity at PRU. ... it looks like they are treating some part of the GAAP net income as exceptional and therefore to be excluded.
There are ~800 stocks in the Russell 1000 & 2000 that have a VL timeliness rating. You cannot realistically do that sort of digging into all the potential candidate stocks.
I think you have to decide upon a decent data source and go with it.
I think I'll watch this screen for a while, run a paper portfolio for the next few month, also see how much the picks change from month to month.
No. of Recommendations: 0
Ah, I just remembered that Fidelity allows fractional share trades. Quantity to 3 decimal places (.001). Or enter dollars and it will convert to quantity. No need to fiddle with whole number percentages like at M1Finance.
No. of Recommendations: 3
There are ~800 stocks in the Russell 1000 & 2000 that have a VL timeliness rating. You cannot realistically do that sort of digging into all the potential candidate stocks.
I think you have to decide upon a decent data source and go with it.
Actually it's not a big deal at all to do sanity checks.
Say you want 40 stocks from this screen, or another.
Find the top 50 picks from the database you have.
Start at the top and do a quick sanity check on them starting from the top. Maybe a few minutes apiece.
(for example, for ROE, just pull up the VL report page and look at the row for equity, and the row for earnings--if both are trending reasonably well, there isn't a big recent anomaly)
If there are some stock picks that look fishy, skip them, and take the top 40 of those that are OK.
You won't catch some that were erroneously far down the list that should have been high, but that's not where the danger lies.
Skipping one good pick isn't a problem; buying a bad one is.
A lot of data fields aren't subject to anomalies as much as ROE is. All kinds of things can throw it off: assets booked at far from true value like real estate or dud goodwill. Or simply an unusually good or bad year for earnings.
Maddeningly, ROE is probably the best single data field one could use as a predictor of long run returns from a stock.
And oddly still works better than anything else even with all those problems.
For example, when I do a quant screen which has a momentum sort, I always check to see if the stock has just had a buyout offer.
Those will show good momentum looking backwards because of the bid premium, but will generally be flat going forwards.
You don't usually even have to read the news...it's one of the few things that really jumps out at you from a price chart.
All that being said, trying the paper portfolio is a great idea.
Performance in any short period might be good or might be bad--there is always variation.
But it is always very useful to do the picks, and look them up to get a feel for what sort of thing it tends to pick.
What industries, balance sheets, market caps, dividend styles.
I was looking at a screen yesterday, and it has an inordinate fondness for tobacco firms, not my favourite sector.
Jim
No. of Recommendations: 1
Another interesting screen Jim. From a screen logic perspective, would it make sense to eliminate financials (like you have in other screens) since banks and insurers are required to keep a chunk in cash?
Thanks!
-John
No. of Recommendations: 7
Another interesting screen Jim. From a screen logic perspective, would it make sense to eliminate financials (like you have in other screens) since banks and insurers are required to keep a chunk in cash?
That makes some sense, but it's not really necessary.
The typical bank has much less cash on its balance sheet than it has long term debt.
There are some outliers, but it's typically 1:2 ratio these days.
The really cash rich non-bank companies tend to crowd them out of the screen.
If you look at the recent picks I posted, there are no banks in the top 40, and only one investment bank.
Since they are such a minority of the picks it makes very little difference over time whether you filter them out or not.
In backtest, the difference was 0.14%/year...and the higher number was for the version allowing financials.
Jim
