May 18, 2009
– Comments (27)
Wow, S&P trading at a P/E of 122 now????
Look at these charts on Big Picture.
See here for a chart of the P/B ratio for the S&P 500 index and have at the paragraph following that chart:
More importantly, the book value of the S&P 500 is overstated relative to its lows in the early 1980s, as neither technology or biotechnology R&D spending (the biotechnology industry was not born until the late 1980s) is capitalized and treated as an asset on the balance sheet. For example, Amgen and Genentech – two of the better performing stocks on the S&P 500 over the last 18 months, have P/B ratios of 2.42 and 5.49, respectively. Even Microsoft, a mature cash cow in the software industry, has P/B ratio of 3.94. Assuming (conservatively) that 15% of the S&P 500 are made up of such companies that did not exist in their current forms in the early 1980s, and assuming that the book value of such companies are understated by about 50%, the S&P 500's book value for 2008 is closer to around $638 a share. On a R&D-adjusted basis, the P/B ratio for the S&P 500 is approximately 1.07, or its lowest level since the beginning of the greatest bull market in history in late 1982. In addition, analysts are still projecting the S&P 500's earnings to be in the range of $35 to $50 a share this year, suggesting that the S&P 500's book value will continue to grow this year. Based on the price-to-book ratio of the S&P 500, the core earnings power of the S&P 500's components, and the range of liquidity schemes implemented by the Federal Reserve and Bank of England, there is no doubt that stocks now present the greatest buying opportunity of our generation.
Links to more charts (and to the abovementioned chart) can be found at the bottom of this article ...
I think the contrast between dwot's p/e chart and portefeuille's p/b chart shows the length the financials are going to, to lie about their financial position. Lots of toxic assets are being rated as caviar, padding book value. The truth is though, you invest money to make money, so earnings are important.
The government is running out of money to pump into the financials, and they are hoping that if they committ enough financial fraud, the private sector will pick up the slack and buy these paper tigers.
Pay no attention to the man behind the curtain!
the p/b of my portfolio, even on average doubled off its lows, is about 0.5
what if some of whats on banks balance sheets is worth what they say, or, in the fullness of time more? ditto insurance, ditto BDCs... ???
For a chart with data until April 2009 see this chart (pdf). The weight of the financial sector in the S&P 500 index is probably around 15% (see here). 15% of 910 is ca. 136. The aggregate book value of the S&P 500 is around 580. The P/B for the financial sector is around 1. If you think it is overstated, say by a factor of 2, then you would need to subtract ca. 136/2 = 68 from the aggregate book value. So instead of 910/580 = ca. 1.57 you would get 910/512 = 1.78. These numbers are smaller if you follow the argument that companies that have invested heavily in R&D should have higher book values to make the numbers comparable to the early eighties (see comment #1 above).
More importantly, the book value of the S&P 500 is overstated relative to its lows in the early 1980s ...
The author means "understated" as can be seen from what follows ...
Ok what am I missing? I look at Port's link and it has a little box in the top part of the S&P chart listing earnings. Since 2007:
Dec 2007 66.18
Dec 2008 14.88
Dec 2009 28. 51(E)
Dec 2010 35. 31(E)
If you think it is overstated, say by a factor of 2, ...
By that I mean:
If you think the book value for the financials is overstated, say by a factor of 2, ...
The truth is though, you invest money to make money, so earnings are important.
So Fort Knox was more or less worthless?
Well, earnings estimates are "difficult" and usually "all over the place" in "volatile times" ...
If you have (like I had) trouble finding that: It is linked to the article I mention at the end of comment #1 above (click on "S&P 500 Price/Earnings Ratio (Monthly)" there or just click here for the pdf-document with those numbers).
Okay, I withdraw that somewhat "polemic" statement.
Okay, I withdraw that somewhat "polemic" statement.
LOL, I got your point.
The fact that those 2 charts are so opposite of one another is amazing.
I'm assuming that for most financials, (especially AIG) that assets have a negative value at this point which can't be calculated.
Your point of the weight of the financials did not go unnoticed though. Taking the financials out of the picture entirely, would not effect p/b enough to make the difference.
Inflation can't be pointed at for the difference, because it's inflationary assets/inflationary dollars.
So why the 2 different stories painted by the charts? I have a feeling that if someone can answer that, they would have a much better picture of the market
The charts look so different because of the hugely negative earnings by financials. Take a stock with earnings of $10 and one with -$9.99 and turn them into an index. Stock A is worth $150 (15 P/E), stock B is worth $0. Index value is $150 because, like the S&P 500 it is market cap weighted (stock B is effectively not in the index). Earnings however are $0.01 since earnings are not weighted. P/E is 15,000; A little pricey wouldn't you say? Except it's not. It's a flawed methodology that makes the index look more expensive than it is. The book value in the P/B chart porte posted is market cap weighted.
I would pay very little attention to these "don't buy stocks now because of the S&P's P/E" arguments. Look at P/Es of individual stocks you want to buy.
I still like Value Line's P/E weekly stats. VL covers a fairly diverse universe of 1700 above mid cap companies.
Most recent statement of 5/15 is 14.7 current, 26 weeks ago of 12.1, market low (10-09-02) 14.1, Market high (7-13-07) of 19.7.
Off topic I know, but this is an important question I want to draw attention to.
Question for the community:
I red thumb a stock. Stock rises 2%, S&P500 rises 2%, shouldn't my score decrease by 4 points because I underperformed the index by 4%? Under the current system my score would remain unchanged in this instance. What do you all think?
My line of thinking:
A big part of the Fool approach is to invest long term in an index fund. Well, following that line of thinking, the benchmark is someone who is long the S&P 500. Well, scoring works great when I green thumb a stock. Stock rises 5%, S&P 500 rises 3%, then horray I beat the index/benchmark by 2%, so my score goes up 2 points. However, it doesn't work that way when I short. I red thumb a stock. Stock rises 2%, S&P500 rises 2%, then oh no I underperformed the index/benchmark by 4% (I lost 2% trying to short the stock while index rose 2% in the meantime). In this case, my score should have gone down 4 points, but instead under the current system it would remain unchanged. My score shouldn't remain unchanged because because when I'm red thumbing the stock, my benchmark shouldn't suddenly change to become shorting the index. The person long the index--aka my benchmark--isn't going to suddenly short the index when I short a stock. It makes no sense. Again, the benchmark is someone who is long the index. When while I short a stock, the benchmark should still be someone who is long the index. That's would make sense, and that would also be in agreement with the Fool approach. Do you follow? Similar error with one or two other combinations, but I wanted to use only one example to avoid confusion.
Stock A is worth $150 (15 P/E), stock B is worth $0. Index value is $150 because, like the S&P 500 it is market cap weighted (stock B is effectively not in the index). Earnings however are $0.01 since earnings are not weighted.
It took me some 2 seconds but then I realised (before clicking that link) that that must be that flawed Jeremy Siegel argument.
Please read the comment section to that article. They are weighted by definition (or construction if you will).
Please have a look at comment #53 here. I should really write a note on that because Siegel has apparently done more damage by publishing that flawed article than I would have imagined at that time. I just wanted to "save him from embarrassment" and did not think about the readers being fed this nonsense.
If you replace stock by company and stock price my market capitalisation and multiply the numbers by a million (to make it more realistic) you can leave the rest of your argument unchanged (just take out that Siegel nonsense). So you end up with:
Take a company with earnings of $10M and one with -$9.99M and turn them into an index. Company A is worth $150M (15 P/E), company B is worth $0. The aggregate market capitalisation of the index is $150M (stock B is effectively not in the index). Earnings however are $0.01M P/E is 15,000;...
(end of new version)
For those who cannot wait and have not figured it out or did not care to think about it:
Comment #3 of comment section to that article for example reads:
With all due respect to the good professor, I think he has it wrong on this one: Adding the earnings “dollar for dollar” is perfectly appropriate in that they are, by definition, already weighted according to the size of the enterprise. What would be wrong would be adding the individual per share earnings to come up with an aggregate figure for the index. Using dollar earnings is analogous to measuring price changes as the change in the dollar value of the entire enterprise rather than price per share. This would yield the same result as weighting the per share figures by capitalization. A $10 billion loss should affect the index equally whether it comes from a small company or large.
He got the idea.
For some Siegel article bashing see the first comment to this article:
Oh man, is this embarrassing.
Siegel is just wrong. This is a freshman mistake. I'm very embarrassed for Wharton's finance department. I'm even a little concerned about Siegel's health...how did he make such an error? And what were the editors thinking?
2/25/2009 2:48 PM
or comments #3,4 to this article:
BenE · 11 weeks agoWow elementary math FAIL. The reason the percentage share price is multiplied by cap in the example he gives is to make it an absolute value instead of a relative percentage. If the earnings were stated in relative terms then it would make sense to multiply them by previous earnings to obtain the absolute value before adding them to the index. For example, if it was stated that earning were down 20% from last year for a company it would make sense to multiply by last years earnings to figure out the absolute decline. What he is doing makes absolutely no sense.
Namazu · 11 weeks agoHe's out of his effing mind--you're just too polite to say it. The P/E on the S&P500 should be the same for an index fund as for a hypothetical (rich) owner of 100% of all the components. Period. The earnings are ALREADY cap weighted. I thought the whole point of Siegel's nutty mutual funds was that cap weighting was bad. You can stare at the Pareto curve all you want, and it doesn't mean you can predict anything. Back-testing doesn't count. I'd love to see Nassim Taleb tear this guy a new one.
You can find a lot more on the internet! Well, that comes from not taking portefeuille's advice (joke) ...
A big part of the Fool approach is to invest long term in an index fund. Well, following that line of thinking, the benchmark is someone who is long the S&P 500.Well, scoring works great when I green thumb a stock. Stock rises 5%, S&P 500 rises 3%, then horray I beat the index/benchmark by 2%, so my score goes up 2 points. However, it doesn't work that way when I short. I red thumb a stock. Stock rises 2%, S&P500 rises 2%, then oh no I underperformed the index/benchmark by 4% (I lost 2% trying to short the stock while index rose 2% in the meantime). In this case, my score should have gone down 4 points, but instead under the current system it would remain unchanged. My score shouldn't remain unchanged because because when I'm red thumbing the stock, my benchmark shouldn't suddenly change to become shorting the index. The person long the index--aka my benchmark--isn't going to suddenly short the index when I short a stock. It makes no sense. Again, the benchmark is someone who is long the index. When while I short a stock, the benchmark should still be someone who is long the index. That's would make sense, and that would also be in agreement with the Fool approach. Do you follow? Similar error with one or two other combinations, but I wanted to use only one example to avoid confusion.
I have answered to your question in your blog post (see comment #1):
Everything is fine.
The investment analogy to an "underperform" call on A would be going short A and long the benchmark ...
P/E and P/B are linked. If a company has $1 billion book value, but is expected to lose money over the next five years, it's P/B should reflect that, maybe residing under 1. The P/B on Fort Knox should be 1 because that's all you are ever going to get. With a deflationary environment, asset values are declining. I don't care what it was worth, if it cannot produce as much revenue, it is worth less now. Couple that with a punctured debt bubble that leaves the earnings outlook pretty bleak for years to come and you are left with a still overvalued market. Everything was inflated by the over-borrowing and zero savings rate including the prices paid for assets (book value). The book value of a lot of these companies is a mirage. Do you want to buy some company like the axle maker Dana with a P/B of 0.16? Good luck.
porte, you're misinterpreting my argument. My point is not that equally weighting earnings is mathematically incorrect. Far from it, it is the correct way to calculate a P/E. Rather, my point was that you can't use that as a valuation metric in the same way as you would for an individual company.
In the example, if you applied a 15 P/E to the index, you would get a value of effectively zero. But that is of course wrong because stock A is worth far more than that. I'm pointing this out in response to all the people saying "don't buy stocks, the P/Es are too high," then referencing the earnings of the index. When companies earn or lose money in amounts disproportionate to their market size, the P/E of the index can become very different from the P/E of the underlying components.
I believe that was the point Mr. Siegel was trying to make as well, but perhaps poorly communicated.
Siegel was just wrong. He did not understand that there was no need to "market weigh" anything artificially because that is done automatically. Just have a look at the comment section. About 1/3 of the commentators point out that mistake.
I wanted to end my comment #15 above with something like:
Apart from the Siegel thing I fully agree with what you say (if you replace the per stock values for price and earnings by "whole company" values). I somehow forgot about it when I finally got to the end (all this copy, paste, search thing keeps me somewhat confused at the end of my posts sometimes (have a look at comment #15 above, that somehow grew into some "monster-post" of its own ...)).
You are right. It is as easy as that. Nonpositive aggregate earnings are nothing scary.
This whole "what is the P/E ratio of that index" thing is somewhat annoying. If anyone really cares so much about P/E ratios then why does he not simply calculate it for every company, then he could do his "oh my god" the P/E should be higher/lower thing to come up with a "fair value" for that company and then he could add them all up and see what the "fair value" for the "combined market capitalisation" is. Then he could compare that with the current value and then finally he could do his "oh my god it's under-/overvalued" dance ...
Yeah, he may have just been wrong. I interpreted the piece as him trying to make a point about the indexes earnings not being useful for valuation of the index given the way they're calculated, but maybe he's just a fool. If that's the case, shame on the WSJ for publishing it.
The beginning and end of his article are okay (at least they are free of flaws).
I suggested to him that he should either correct that flawed part or leave it out but he apparently had this "idée fixe" that S&P was doing something wrong.
Maybe he did not take my advice that seriously because my email adress was from the physics department of my university ...
Wonderful discussion here...
I didn't have time to go an look at references and really study what is being said here ...
But, without looking too deeply, I tend to thing that book values are inflated and the ratio of price to book value increase. As assets get repriced I would tend to think the stocks will look more expensive.
As for earnings estimates, I have yet to be impressed with one when I have torn them apart to study them.
Bigpeach has an excellent point about checking the individual stocks.
more on P/E (including operating earnings estimates sectors) here
from that article:
A quote from Ian Cumming from the Leucadia (LUK) annual meeting I went to. “The science is in the “In”. The poetry is in the “Out”. The value buyer assesses the potential earnings power of the assets, but buys when there are no or at least very low earnings and the market not having the sense to do the same type of work is lost in the pricing and giving stocks away. But, when the earnings are at full potential and have recovered, the market believes that some new earnings trend has begun and priced the stock at “poetry levels”-“so beautiful” and it is this is where one should sell.
also see this from here:
Q22: How do you decide when to sell an asset? Cumming: The science is in the “in”. The poetry is in the “out”. Cumming does not use spreadsheet models to determine when to sell. Does have access to people who can use models and spreadsheets. They look at the future earnings stream and discount it back to determine a value The rate depends on how durable the company is and how much they love it. If the asset is selling above their estimation of value then they think about selling it
also have a look at this:
Operating Earnings v As Reported Earnings (chart 1 !)
The interesting part is the comment section of the FT article mentioned in the first sentence of the article linked to comment #25 above.
I guess ...
Rosenberg on S&P earnings