Here's a statement of the obvious: The opinions expressed here are those of the participants, not those of the Mutual Fund Observer. We cannot vouch for the accuracy or appropriateness of any of it, though we do encourage civility and good humor.
Support MFO
Donate through PayPal
Q: what does it actually MEAN when I see a neg. P/E?
Aside from various ways of handling outliers as Yogi described, there are a couple of different ways of calculating the average P/E. One is by taking a simple (weighted) arithmetic average of the P/Es. Another way, IMHO more meaningful, is to take the (weighted) harmonic average of the P/Es.
Effective November 30, 2005, we [Morningstar] will ... use a harmonic weighted average, rather than an arithmetic weighted average. The harmonic method prevents outliers from skewing the result...
Computing the harmonic weighted average is equivalent to adding up all the earnings of each holding weighted by the size of the holding (E), then dividing it into the total value of the portfolio (P) to get P/E.
For example, suppose a fund holds shares of two companies: $10 of company A with a P/E of 20, and $15 of company B with a P/E of 15. Company A gets weighted 0.4, company B gets weighted 0.6.
The total earnings from the first holding are $10 x earnings/dollar = $10 x1/20 = $0.50 Total earnings from the second holding are: $15 x 1/15 = $1. Total earnings of portfolio = $1.50 Total value of portfolio = $25 P/E = $25/$1.50 = 16.67
Weighted harmonic average = 1 / (0.4 x 1/20 + 0.6 x 1/15) = 1 / (.02 + .04) = 1/.06 = 16.67
OTOH, weighted simple average = 0.4 x 20 + 0.6 x 15 = 8 + 9 = 17.
Yahoo Finance is clueless. It reports VFIAX's P/E as 0.05, while reporting the P/E of a different share class VOO of the same fund as 22.19.
Despite its faults, I'm so familiar with the layout at M* that I still use it consistently----- though not by itself to make any final decisions about whether to buy or sell anything. So, I've become accustomed to their use of the "weighted harmonic" beast. Surely, they are not the only ones doing it. I see stocks listed in the neighborhood of 13-14, and more richly valued ones at 23 or so. That's "the page I'm on" with it.
And my ENTIRE portfolio, which I maintain at M*, tells me the whole thing holds a P/E of 14.14. And it tells me it's a 25% lower value than the average for the full SP500. THANK you all for answering my question.
Averages are often misused in the financial industry. Almost everything is averaged arithmetically (i.e. simple average) when other averages such as geometric averages (for TR, etc), harmonic ratios (for P/E, P/B ratios, etc) may be more appropriate; in some cases, no averaging method is applicable but simple averages are used anyway (SD, Sharpe Ratio, etc).
BTW, the M* document @msf linked also says (my bolds), "Morningstar generates this figure in-house, based on the most-recent portfolio holdings submitted by the fund and stock statistics gleaned from our internal U.S. equities databases. (Our U.S. equities department receives prices from ComStock, a division of Interactive Data Corporation. , and gathers earnings information from a company's most-recent annual and quarterly income statements.) Negative P/Es are not used, and any PE greater than 60 is capped at 60 in the calculation of the average." https://awgmain.morningstar.com/webhelp/glossary_definitions/mutual_fund/mfglossary_Price_Earnings_Ratio.html
I think that Stock Rover (SR) also uses harmonic averages for ratios (P/E, P/B, etc), but then goes ahead and uses simple average for SDs - I cannot find a link now. SR was also making some other errors in handling cash, m-mkt funds, yearend CG distributions, and after several back-and-forth communications to get to the right person/team, SR fixed them. https://ybbpersonalfinance.proboards.com/thread/318/stock-rover-sr
As for contacting organizations for errors or data, the worst are Yahoo Finance and Vanguard - either they don't respond, or they don't care and send form letter responses. The best in this respect are Fidelity, Portfolio Visualizer (PV), StockCharts, StockRover (SR).
I've become accustomed to their use of the "weighted harmonic" beast. Surely, they are not the only ones doing it.
FWIW, a SeekingAlpha piece asserts:
Most institutions like Morningstar calculate a fund's P/E by taking the harmonic average of the underlying equities and filtering out those with negative earnings.
Comments
If earnings are near 0, P/E can be very large.
For funds, the P/E calculation differs by data source. Some treat negative P/E as 0 and also cap high P/Es. Others may use P/E as they are.
For BRUFX P/E,
Yahoo Finance 0.06.
M* 13.86
Computing the harmonic weighted average is equivalent to adding up all the earnings of each holding weighted by the size of the holding (E), then dividing it into the total value of the portfolio (P) to get P/E.
For example, suppose a fund holds shares of two companies: $10 of company A with a P/E of 20, and $15 of company B with a P/E of 15. Company A gets weighted 0.4, company B gets weighted 0.6.
The total earnings from the first holding are $10 x earnings/dollar = $10 x1/20 = $0.50
Total earnings from the second holding are: $15 x 1/15 = $1.
Total earnings of portfolio = $1.50
Total value of portfolio = $25
P/E = $25/$1.50 = 16.67
Weighted harmonic average = 1 / (0.4 x 1/20 + 0.6 x 1/15) = 1 / (.02 + .04) = 1/.06 = 16.67
OTOH, weighted simple average = 0.4 x 20 + 0.6 x 15 = 8 + 9 = 17.
Yahoo Finance is clueless. It reports VFIAX's P/E as 0.05, while reporting the P/E of a different share class VOO of the same fund as 22.19.
And my ENTIRE portfolio, which I maintain at M*, tells me the whole thing holds a P/E of 14.14. And it tells me it's a 25% lower value than the average for the full SP500. THANK you all for answering my question.
Averages are often misused in the financial industry. Almost everything is averaged arithmetically (i.e. simple average) when other averages such as geometric averages (for TR, etc), harmonic ratios (for P/E, P/B ratios, etc) may be more appropriate; in some cases, no averaging method is applicable but simple averages are used anyway (SD, Sharpe Ratio, etc).
BTW, the M* document @msf linked also says (my bolds),
"Morningstar generates this figure in-house, based on the most-recent portfolio holdings submitted by the fund and stock statistics gleaned from our internal U.S. equities databases. (Our U.S. equities department receives prices from ComStock, a division of Interactive Data Corporation. , and gathers earnings information from a company's most-recent annual and quarterly income statements.) Negative P/Es are not used, and any PE greater than 60 is capped at 60 in the calculation of the average." https://awgmain.morningstar.com/webhelp/glossary_definitions/mutual_fund/mfglossary_Price_Earnings_Ratio.html
I think that Stock Rover (SR) also uses harmonic averages for ratios (P/E, P/B, etc), but then goes ahead and uses simple average for SDs - I cannot find a link now. SR was also making some other errors in handling cash, m-mkt funds, yearend CG distributions, and after several back-and-forth communications to get to the right person/team, SR fixed them. https://ybbpersonalfinance.proboards.com/thread/318/stock-rover-sr
As for contacting organizations for errors or data, the worst are Yahoo Finance and Vanguard - either they don't respond, or they don't care and send form letter responses. The best in this respect are Fidelity, Portfolio Visualizer (PV), StockCharts, StockRover (SR).
FWIW, a SeekingAlpha piece asserts: https://seekingalpha.com/article/4074061-your-mutual-funds-p-e-is-likely-wrong
(free subscription required)