Real P/E of ETF's and Mutual Funds  are the published ones misleading
Real P/E of ETF's and Mutual Funds  are the published ones misleading
Surprised to read an article by Jason Zweig in the WSJ today. Titled 'when cheaper P/E ratios mean nothing'. He explains that ETF's and hence I assume mutual funds aggregate earnings in different ways. He takes three S&P 500 ETF's and explains that the S&P500 index has a PE of 23.56 and the three large SP500 ETF's all report different PE's. The article is on the WSJ site today under money beat or you can search through google.
He says that even though S&P reports 23.65 for june30th, state street (SPY) has 18.65. Black rock IVV at 21.69 and VG VOO at 21.5. I was really surprised that they all EXCLUDE companies with negative earnings and SPY uses analyst estimates of forward earnings. He explains that VG uses Factset to aggregate the earnings and Blackrock uses some other method hence there is still a difference between the two.
I am assuming that Mutual Funds have the same issue and given that mutual fund companies have an incentive to make their funds look attractive and there seems to be no SEC mandate around how to aggregate earnings, funds must be biasing their earnings estimates to appear attractive. Wondering if Bogleheads have thoughts on how one can get the REAL P/E (Current total price /12 mo GAAP trailing total GAAP earnings). I am quite surprised to see that funds can exclude loss making companies and that there is no standard method. This makes the P/E ratios for total market stock and international funds esp. misleading as they own a lot of companies that are not profitable so the P/E is most likely quite inaccurate and really much higher.
He says that even though S&P reports 23.65 for june30th, state street (SPY) has 18.65. Black rock IVV at 21.69 and VG VOO at 21.5. I was really surprised that they all EXCLUDE companies with negative earnings and SPY uses analyst estimates of forward earnings. He explains that VG uses Factset to aggregate the earnings and Blackrock uses some other method hence there is still a difference between the two.
I am assuming that Mutual Funds have the same issue and given that mutual fund companies have an incentive to make their funds look attractive and there seems to be no SEC mandate around how to aggregate earnings, funds must be biasing their earnings estimates to appear attractive. Wondering if Bogleheads have thoughts on how one can get the REAL P/E (Current total price /12 mo GAAP trailing total GAAP earnings). I am quite surprised to see that funds can exclude loss making companies and that there is no standard method. This makes the P/E ratios for total market stock and international funds esp. misleading as they own a lot of companies that are not profitable so the P/E is most likely quite inaccurate and really much higher.
Re: Real P/E of ETF's and Mutual Funds  are the published ones misleading
As I look deeper into this. It seems that the arithmetic mean may not be the best way to aggregate ratios. A weighted harmonic mean might be better.
I understand that Vanguard's approach using harmonic means  I am assuming weighted harmonic means but not sure if this is done for all funds 
from Wikipedia on harmonic means 
In finance
The weighted harmonic mean is the preferable method for averaging multiples, such as the price–earnings ratio (P/E), in which price is in the numerator. If these ratios are averaged using a weighted arithmetic mean (a common error), high data points are given greater weights than low data points. The weighted harmonic mean, on the other hand, gives equal weight to each data point.[6] The simple weighted arithmetic mean when applied to nonprice normalized ratios such as the P/E is biased upwards and cannot be numerically justified, since it is based on equalized earnings; just as vehicles speeds cannot be averaged for a roundtrip journey.[7]
For example, consider two firms, one with a market capitalization of $150 billion and earnings of $5 billion (P/E of 30) and one with a market capitalization of $1 billion and earnings of $1 million (P/E of 1000). Consider an index made of the two stocks, with 30% invested in the first and 70% invested in the second. We want to calculate the P/E ratio of this index.
Using the weighted arithmetic mean (incorrect):
P/E=0.3*30+0.7∗1000=709
Using the weighted harmonic mean (correct):
P/E=(0.3+0.7)/0.3/30+0.7/1000)=93.46
Thus, the correct P/E of 93.46 of this index can only be found using the weighted harmonic mean, while the weighted arithmetic mean will significantly overestimate it.
I understand that Vanguard's approach using harmonic means  I am assuming weighted harmonic means but not sure if this is done for all funds 
from Wikipedia on harmonic means 
In finance
The weighted harmonic mean is the preferable method for averaging multiples, such as the price–earnings ratio (P/E), in which price is in the numerator. If these ratios are averaged using a weighted arithmetic mean (a common error), high data points are given greater weights than low data points. The weighted harmonic mean, on the other hand, gives equal weight to each data point.[6] The simple weighted arithmetic mean when applied to nonprice normalized ratios such as the P/E is biased upwards and cannot be numerically justified, since it is based on equalized earnings; just as vehicles speeds cannot be averaged for a roundtrip journey.[7]
For example, consider two firms, one with a market capitalization of $150 billion and earnings of $5 billion (P/E of 30) and one with a market capitalization of $1 billion and earnings of $1 million (P/E of 1000). Consider an index made of the two stocks, with 30% invested in the first and 70% invested in the second. We want to calculate the P/E ratio of this index.
Using the weighted arithmetic mean (incorrect):
P/E=0.3*30+0.7∗1000=709
Using the weighted harmonic mean (correct):
P/E=(0.3+0.7)/0.3/30+0.7/1000)=93.46
Thus, the correct P/E of 93.46 of this index can only be found using the weighted harmonic mean, while the weighted arithmetic mean will significantly overestimate it.
Re: Real P/E of ETF's and Mutual Funds  are the published ones misleading
A couple of thoughts here.Osp62 wrote:I am assuming that Mutual Funds have the same issue and given that mutual fund companies have an incentive to make their funds look attractive and there seems to be no SEC mandate around how to aggregate earnings, funds must be biasing their earnings estimates to appear attractive. Wondering if Bogleheads have thoughts on how one can get the REAL P/E (Current total price /12 mo GAAP trailing total GAAP earnings). I am quite surprised to see that funds can exclude loss making companies and that there is no standard method. This makes the P/E ratios for total market stock and international funds esp. misleading as they own a lot of companies that are not profitable so the P/E is most likely quite inaccurate and really much higher.
Mutual funds have some discretion on how to calculate P/E earnings. However once they have picked a method they have to keep using that method.
Are you sure you want to included negative earnings? You have the "Big Bath" problem, when companies clean house they tend to dump all of there losses at once. Most of those losses are paper losses, not real. i.e., aggressively writing down good will and other intangible assets. So you get these big lumpy things that don't mean much.
Lastly, I don't think you want historical earnings for 2 reasons. First, they are on average 8 months old. Second, inspite of all of the faults of forecasted earnings, historically they have had more predictive power than historical earnings.
So things are not as clean as you might think. If you go to S&P's web site I think they post both forward and trailing P.E ratios daily.
Re: Real P/E of ETF's and Mutual Funds  are the published ones misleading
Alex686. I understand the issues with forecasted and TTM earnings. The real issue I have is how the P/E ratios of the constituent companies are aggregated. Is it a weighted arithmetic mean or is it a weighted geometric mean or is it something else. The harmonic mean almost always gives a lower number but I had assumed that they had just added up all the market caps in the numerator (weighted by the weight in the index) and then divided by the total GAAP earnings. It is clear to me that this will give a much higher number for the mutual fund PE ratio but this seems like the mos intuitive method  how much you are paying for the total earnings  all weighted by the weight in the index. Wondering if you know where to find this simple weighted arithmetic mean P/E ( both historical and foreword looking) for the total stock index., tot international and sp500. Thx.
Re: Real P/E of ETF's and Mutual Funds  are the published ones misleading
Any company can be split up into multiple companies ("spinoffs") or multiple companies can be aggregated into one ("mergers").
Suppose ABC Corp makes $5M. XYZ Corp loses $1B. If ABC buys XYZ, the combined earnings of the new company are $4M, not $5M.
In a capweighted index, the PE should essentially be the sum of market cap / the sum of earnings.
So, in the case of this two company index:
There are more quirks in calculating an actual index that the math folks can be sure to jump in an correct me on ( ), but this is essentially the calculation.
That is how the various index providers do it which is why I like to rely on their information whenever possible.
Suppose ABC Corp makes $5M. XYZ Corp loses $1B. If ABC buys XYZ, the combined earnings of the new company are $4M, not $5M.
In a capweighted index, the PE should essentially be the sum of market cap / the sum of earnings.
So, in the case of this two company index:
Your PE should be (150 + 1) / (5 +.001) = 30.2.two firms, one with a market capitalization of $150 billion and earnings of $5 billion (P/E of 30) and one with a market capitalization of $1 billion and earnings of $1 million (P/E of 1000)
There are more quirks in calculating an actual index that the math folks can be sure to jump in an correct me on ( ), but this is essentially the calculation.
That is how the various index providers do it which is why I like to rely on their information whenever possible.
Re: Real P/E of ETF's and Mutual Funds  are the published ones misleading
Keep in mind that the S&P 500 index does not include companies that don't have positive earnings, companies that report negative earnings get cycled out of the index, although it isn't instant so there are a few stragglers. In this case it might make sense to remove companies with negative earnings since they won't be there long anyways.
For a total stock market index, removing companies with negative earnings from the calculation seems misleading.
For a total stock market index, removing companies with negative earnings from the calculation seems misleading.
Re: Real P/E of ETF's and Mutual Funds  are the published ones misleading
It includes a good number of companies with negative earnings, particularly in a recession. It is the case that they need to have positive earnings to get added in the first place. And they are very slow to move companies out of the 500 if they shrink. Actually, a very reasonable "smallcap" strategy would be to own, say, the smallest 50 companies in the S&P 500.Keep in mind that the S&P 500 index does not include companies that don't have positive earnings, companies that report negative earnings get cycled out of the index
Re: Real P/E of ETF's and Mutual Funds  are the published ones misleading
Do you have data on this? What is a "good number of companies?" Curious because it's a large portion of my portfolio specifically to exclude negative earnings.stlutz wrote:It includes a good number of companies with negative earnings, particularly in a recession. It is the case that they need to have positive earnings to get added in the first place. And they are very slow to move companies out of the 500 if they shrink. Actually, a very reasonable "smallcap" strategy would be to own, say, the smallest 50 companies in the S&P 500.Keep in mind that the S&P 500 index does not include companies that don't have positive earnings, companies that report negative earnings get cycled out of the index
Re: Real P/E of ETF's and Mutual Funds  are the published ones misleading
I agree with you but the whole point of this post is that all companies do NOT aggregate the PE this way. Even VG supposedly uses the Weighted Harmonic mean which will always give a smaller number than simple weighted arithmetic mean. See the posts I have above yours explaining my point.stlutz wrote:Any company can be split up into multiple companies ("spinoffs") or multiple companies can be aggregated into one ("mergers").
Suppose ABC Corp makes $5M. XYZ Corp loses $1B. If ABC buys XYZ, the combined earnings of the new company are $4M, not $5M.
In a capweighted index, the PE should essentially be the sum of market cap / the sum of earnings.
So, in the case of this two company index:
Your PE should be (150 + 1) / (5 +.001) = 30.2.two firms, one with a market capitalization of $150 billion and earnings of $5 billion (P/E of 30) and one with a market capitalization of $1 billion and earnings of $1 million (P/E of 1000)
There are more quirks in calculating an actual index that the math folks can be sure to jump in an correct me on ( ), but this is essentially the calculation.
That is how the various index providers do it which is why I like to rely on their information whenever possible.
Re: Real P/E of ETF's and Mutual Funds  are the published ones misleading
This is going to be a tough hunt. You might get lucky with the S&P 500. Lots of different people do things with that one. Almost nobody mucks around with the CRSP indexes. I would start with the respective index providers web site. I know S&P has multiple fields for the P/E ratio calculated in different ways. Some of those fields are restricted to data subscribers so I don't know if that would help.Osp62 wrote:Wondering if you know where to find this simple weighted arithmetic mean P/E ( both historical and foreword looking) for the total stock index., tot international and sp500. Thx.
Re: Real P/E of ETF's and Mutual Funds  are the published ones misleading
Are you sure about this? What is your source? If I recall correctly, negative earnings might prevent a company from inclusion to the S&P 500. Most of the companies in the index have had periods of negative earnings last a few years.kosomoto wrote:Keep in mind that the S&P 500 index does not include companies that don't have positive earnings, companies that report negative earnings get cycled out of the index, although it isn't instant so there are a few stragglers. In this case it might make sense to remove companies with negative earnings since they won't be there long anyways.
Re: Real P/E of ETF's and Mutual Funds  are the published ones misleading
They will be removed if there are "ongoing conditions" in violation of the admission criteria. Not sure how long they need to be ongoing before they get removed.alex_686 wrote:Are you sure about this? What is your source? If I recall correctly, negative earnings might prevent a company from inclusion to the S&P 500. Most of the companies in the index have had periods of negative earnings last a few years.kosomoto wrote:Keep in mind that the S&P 500 index does not include companies that don't have positive earnings, companies that report negative earnings get cycled out of the index, although it isn't instant so there are a few stragglers. In this case it might make sense to remove companies with negative earnings since they won't be there long anyways.
https://us.spindices.com/documents/meth ... ndices.pdf
Re: Real P/E of ETF's and Mutual Funds  are the published ones misleading
I don't know of any companies that have been removed for this reason. IIRC, there are companies that have had negative earnings for over 5 years and still remain on the list.kosomoto wrote:They will be removed if there are "ongoing conditions" in violation of the admission criteria. Not sure how long they need to be ongoing before they get removed.
 privatefarmer
 Posts: 327
 Joined: Mon Sep 08, 2014 2:45 pm
Re: Real P/E of ETF's and Mutual Funds  are the published ones misleading
just own the total market and you don't have to worry about this nonsense.
Re: Real P/E of ETF's and Mutual Funds  are the published ones misleading
This P/E trickery is much worse in the Nasdaq where many stocks have negative earnings. See below on "how to turn a PE of 90 into 22 in three easy steps", using QQQ as an example. There are real companies with 200x and 100x PE ratios, so this is not an academic exercise, but these basically get ignored in the calculations. They also discuss the S&P500 index.
http://horizonkinetics.com/wpcontent/u ... _FINAL.pdf
Horizon Kinetics is pretty good on these issues.
http://horizonkinetics.com/wpcontent/u ... _FINAL.pdf
Horizon Kinetics is pretty good on these issues.
Re: Real P/E of ETF's and Mutual Funds  are the published ones misleading
Thanks Tanelorn. This horizon kinetics paper was quite good.Tanelorn wrote:This P/E trickery is much worse in the Nasdaq where many stocks have negative earnings. See below on "how to turn a PE of 90 into 22 in three easy steps", using QQQ as an example. There are real companies with 200x and 100x PE ratios, so this is not an academic exercise, but these basically get ignored in the calculations. They also discuss the S&P500 index.
http://horizonkinetics.com/wpcontent/u ... _FINAL.pdf
Horizon Kinetics is pretty good on these issues.

 Posts: 868
 Joined: Sun Jul 09, 2017 2:53 pm
Re: Real P/E of ETF's and Mutual Funds  are the published ones misleading
I think there are two issues here which are separate.Tanelorn wrote:This P/E trickery is much worse in the Nasdaq where many stocks have negative earnings. See below on "how to turn a PE of 90 into 22 in three easy steps", using QQQ as an example. There are real companies with 200x and 100x PE ratios, so this is not an academic exercise, but these basically get ignored in the calculations. They also discuss the S&P500 index.
http://horizonkinetics.com/wpcontent/u ... _FINAL.pdf
Horizon Kinetics is pretty good on these issues.
1) Should you use the weighted harmonic mean
2) Should you exclude companies with negative earnings.
The harmonic mean (unlike the arithmetic mean of P/Es) handles the negative and 0 earning companies well. There are good reasons to exclude companies with small/no earnings from the arithmetic mean but no particular reason to exclude from weighted harmonic.
Re: Real P/E of ETF's and Mutual Funds  are the published ones misleading
A simpler solution is to just divide the amount invested in the fund by the total GAAP earnings. This will give you the PE ratio of the fund. You don't need to take individual PE ratios and then remove outliers and then aggregate them.jbolden1517 wrote:I think there are two issues here which are separate.Tanelorn wrote:This P/E trickery is much worse in the Nasdaq where many stocks have negative earnings. See below on "how to turn a PE of 90 into 22 in three easy steps", using QQQ as an example. There are real companies with 200x and 100x PE ratios, so this is not an academic exercise, but these basically get ignored in the calculations. They also discuss the S&P500 index.
http://horizonkinetics.com/wpcontent/u ... _FINAL.pdf
Horizon Kinetics is pretty good on these issues.
1) Should you use the weighted harmonic mean
2) Should you exclude companies with negative earnings.
The harmonic mean (unlike the arithmetic mean of P/Es) handles the negative and 0 earning companies well. There are good reasons to exclude companies with small/no earnings from the arithmetic mean but no particular reason to exclude from weighted harmonic.

 Posts: 868
 Joined: Sun Jul 09, 2017 2:53 pm
Re: Real P/E of ETF's and Mutual Funds  are the published ones misleading
What you are suggesting is the weighted harmonic mean of the P/Es.Osp62 wrote:A simpler solution is to just divide the amount invested in the fund by the total GAAP earnings. This will give you the PE ratio of the fund. You don't need to take individual PE ratios and then remove outliers and then aggregate them.jbolden1517 wrote:
The harmonic mean (unlike the arithmetic mean of P/Es) handles the negative and 0 earning companies well. There are good reasons to exclude companies with small/no earnings from the arithmetic mean but no particular reason to exclude from weighted harmonic.
Re: Real P/E of ETF's and Mutual Funds  are the published ones misleading
Yes, it does seem that even the P/E ratio is in the eye of the beholder. You can find both forward P/E and historical P/E on Morningstar. It is useful to look at both though I look mostly at forward P/E based upon future estimated earnings. The market looks forward and so should I.
A fool and his money are good for business.
Re: Real P/E of ETF's and Mutual Funds  are the published ones misleading
Can you explain how dividing the invested amount by the total GAAP earnings for the shares the fund owns is same as weighted harmonic mean of P/E's. Any link to this info would also work.jbolden1517 wrote: ↑Sat Aug 12, 2017 7:23 amWhat you are suggesting is the weighted harmonic mean of the P/Es.Osp62 wrote:A simpler solution is to just divide the amount invested in the fund by the total GAAP earnings. This will give you the PE ratio of the fund. You don't need to take individual PE ratios and then remove outliers and then aggregate them.jbolden1517 wrote:
The harmonic mean (unlike the arithmetic mean of P/Es) handles the negative and 0 earning companies well. There are good reasons to exclude companies with small/no earnings from the arithmetic mean but no particular reason to exclude from weighted harmonic.
Re: Real P/E of ETF's and Mutual Funds  are the published ones misleading
There is really no reason to exclude any company, if you choose the alternative of calculating the PE for the fund (total invested amount/total GAAP earnings of all shares (of all companies) owned by the fund). This is the most intuitive option and shows how much GAAP earnings you are generating from all your invested capital. There is no reason to use the calculated PE of each constituent stock and then calculate an average of all the PE's. I think the funds are doing this to have an excuse to remove loss making companies and thus make the fund's PE look more attractive than the reality.jbolden1517 wrote: ↑Sat Aug 12, 2017 4:13 amI think there are two issues here which are separate.Tanelorn wrote:This P/E trickery is much worse in the Nasdaq where many stocks have negative earnings. See below on "how to turn a PE of 90 into 22 in three easy steps", using QQQ as an example. There are real companies with 200x and 100x PE ratios, so this is not an academic exercise, but these basically get ignored in the calculations. They also discuss the S&P500 index.
http://horizonkinetics.com/wpcontent/u ... _FINAL.pdf
Horizon Kinetics is pretty good on these issues.
1) Should you use the weighted harmonic mean
2) Should you exclude companies with negative earnings.
The harmonic mean (unlike the arithmetic mean of P/Es) handles the negative and 0 earning companies well. There are good reasons to exclude companies with small/no earnings from the arithmetic mean but no particular reason to exclude from weighted harmonic.

 Posts: 868
 Joined: Sun Jul 09, 2017 2:53 pm
Re: Real P/E of ETF's and Mutual Funds  are the published ones misleading
Sure. The reciprocal of P/E is E/P. Value of shares is price * (# of shares)Osp62 wrote: ↑Sat Aug 12, 2017 7:01 pmCan you explain how dividing the invested amount by the total GAAP earnings for the shares the fund owns is same as weighted harmonic mean of P/E's. Any link to this info would also work.jbolden1517 wrote: ↑Sat Aug 12, 2017 7:23 amWhat you are suggesting is the weighted harmonic mean of the P/Es.Osp62 wrote:A simpler solution is to just divide the amount invested in the fund by the total GAAP earnings. This will give you the PE ratio of the fund. You don't need to take individual PE ratios and then remove outliers and then aggregate them.jbolden1517 wrote:
The harmonic mean (unlike the arithmetic mean of P/Es) handles the negative and 0 earning companies well. There are good reasons to exclude companies with small/no earnings from the arithmetic mean but no particular reason to exclude from weighted harmonic.
Weighting is simply multiplying E/P by (P*(# shares)) = (earnings per share)*(# of share held)
The sum is just then the earnings for the fund.
http://corporate.morningstar.com/US/doc ... dology.pdf
Re: Real P/E of ETF's and Mutual Funds  are the published ones misleading
Yup, Jason Zweig describes a wellknown issue with PE computations for mutual funds, ETFs and entire markets. Please check this wiki page for a summary of findings following some research that was triggered by another (older) thread similar to yours.
Bottomline:
 always compare PE quantities computed with the same methodology (i.e. from the same source)
 it is really hard to get historical information, except for the S&P 500 (thanks to Prof. Shiller)
 some sites do really weird stuff (did somebody mention Morningstar?)  Vanguard seems pretty reasonable
Re: Real P/E of ETF's and Mutual Funds  are the published ones misleading
Thank you very much. Greatly appreciate your response and the link to the detailed m* document.jbolden1517 wrote: ↑Sat Aug 12, 2017 7:10 pmSure. The reciprocal of P/E is E/P. Value of shares is price * (# of shares)Osp62 wrote: ↑Sat Aug 12, 2017 7:01 pmCan you explain how dividing the invested amount by the total GAAP earnings for the shares the fund owns is same as weighted harmonic mean of P/E's. Any link to this info would also work.jbolden1517 wrote: ↑Sat Aug 12, 2017 7:23 amWhat you are suggesting is the weighted harmonic mean of the P/Es.Osp62 wrote:A simpler solution is to just divide the amount invested in the fund by the total GAAP earnings. This will give you the PE ratio of the fund. You don't need to take individual PE ratios and then remove outliers and then aggregate them.jbolden1517 wrote:
The harmonic mean (unlike the arithmetic mean of P/Es) handles the negative and 0 earning companies well. There are good reasons to exclude companies with small/no earnings from the arithmetic mean but no particular reason to exclude from weighted harmonic.
Weighting is simply multiplying E/P by (P*(# shares)) = (earnings per share)*(# of share held)
The sum is just then the earnings for the fund.
http://corporate.morningstar.com/US/doc ... dology.pdf
Re: Real P/E of ETF's and Mutual Funds  are the published ones misleading
Hm. I was just rereading the WSJ article, and something is odd, the quote from Rich Powers (Vanguard) stating that negative earnings are eliminated from the Vanguard (FactSet) math.siamond wrote: ↑Sat Aug 12, 2017 7:20 pmYup, Jason Zweig describes a wellknown issue with PE computations for mutual funds, ETFs and entire markets. Please check this wiki page for a summary of findings following some research that was triggered by another (older) thread similar to yours.
Bottomline:
 always compare PE quantities computed with the same methodology (i.e. from the same source)
 it is really hard to get historical information, except for the S&P 500 (thanks to Prof. Shiller)
 some sites do really weird stuff (did somebody mention Morningstar?)  Vanguard seems pretty reasonable
Well, when I did my own research one year ago, I asked this precise question to Vanguard via the support group, got redirected to somebody who seemed to know what he/she was speaking of, and got this detailed response:
viewtopic.php?f=10&t=199804&p=3076893#p3076893
Here it was explicitly mentioned that negative earnings are fair game (which makes complete sense as the harmonic average formula makes sure that outliers do not have an outsized effect). Ok, I no longer know which is which now... I just sent an email to somebody I know at Vanguard to try to sort out which is which. Will share the findings.
Re: Real P/E of ETF's and Mutual Funds  are the published ones misleading
FWIW, There's an old WSJ article that also confirms the idea that Vanguard's methodology includes negative earnings
But as the article points out, accounting for the earnings/losses on a weighted basis of the shares weight in the portfolio produces different results than aggregating all the earnings and losses as if every stock together formed one giant company (the way index providers do). I think the way Vanguard does it (or the way we think they do it) weighting the impact by the weight of the stock to the portfolio makes sense. Billions of dollars in a accounting artifact on a single bank stock that makes up a small fraction of the portfolio shouldn't overwhelm the earnings of other companies in the portfolio.WSJ: The Inexact Business of Valuing ETFs wrote: http://archive.is/IcnyG
... Yet even the inclusion of unprofitable companies doesn't guarantee that ETF providers will come up with the same P/E for their funds as might be calculated on the underlying index.
For instance, indexfund giant Vanguard Group Inc. includes moneylosing companies in its calculations. The Malvern, Pa., fund firm says its SmallCap Growth Vipers ETF VBK 0.60%▲ has a P/E ratio of nearly 30. But the index that the ETF tracks  the Morgan Stanley Capital International Inc.'s US Small Cap Growth Index  has a P/E of nearly 47, as calculated by the indexprovider MSCI Barra.
Joseph Brennan, a Vanguard executive, says the company uses what mathematicians call the "harmonic mean" in calculating an average earnings figure for its funds' P/E ratios. He adds that there are various ways to calculate a P/E, and investors shouldn't be using one number to compare funds. "Buyers need to understand the statistics they are looking at," he says...
"To achieve satisfactory investment results is easier than most people realize; to achieve superior results is harder than it looks."  Benjamin Graham